WORLD CLIMATE RESEARCH PROGRAMME (WCRP)

WMO/ICSU Joint Scientific Committee

REVISED INSTRUCTION MANUAL

ON RADIATION INSTRUMENTS
AND MEASUREMENTS

prepared by the Radiation Commission of IAMAP

l
Claus Fr6hlich & Julius London
Editors

WCRP PUBLICATIONS SERIES No. 7

WMO/TD - No. 149

OCTOBER 1986
 TABLE OF CONTENTS

3.8 SURFACE PROPERTIES 36

3.8.1 Albedo 36
3.8.2 Typical Albedo Values 38
3.8.2.1 Snow and Ice 38
3.8.2.2 Vegetated Surfaces 39
3.8.2.3 Mineral Surfaces 41
3.8.2.4 Sea Surface 42

3.8.3 Polarization by Surface Reflection 43

3.8.4 Emittance 43
3.8.4.1 Mineral Surfaces 43
3.8.4.2 Vegetated Surfaces 44

REFERENCES 45

4. RADIATION MEASUREMENTS 49

4.1 RADIATION DETECTORS 49

4~1.1 Thermal detectors 49
4.1.2 Quantum Detectors 51
4.1.3 Radiation Scales 51

4.2 TECHNIQUES FOR THE MEASUREMENTS OF TOTAL RADIATION 54

4.2.1 Pyrheliometers 54
4.2.1.1 Historical Reference Pyrheliometers 55
4.2.1.2 Self-Calibrating Pyrheliometers 56
4.2.1.3 Operational Pyrheliometers 58
4.2.1.4 Mounting of Pyrheliometers 61
4.2.1.5 Maintenance of Pyrheliometers 62
4.2.1.6 Calibration of Pyrheliometers 62

4.2.2 Pyranometers 62
4.2.2.1 Most Widely Used Pyranometers 63
4.2.2.2 Installation of Pyranometers 69
4.2.2.3 Maintenance of Pyranometers 69
4.2.2.4 Calibration of Pyranometers 70
4.2.2.5 Measurement of Diffuse Radiation 71
4.2.2.6 Measurement of Solar Radiation
on Inclined Surfaces 73
4.2.2.7 Special Pyranometers 73

4.2.3 Pyrradiometers and Pyrqeometers 74

4.2.3.1 Most Widely Used Pyrradiometers 74
4.2.3.2 Maintenance and Calibration
of Pyrradiometers 77

4.3 TECHNIQUES FOR THE MEASUREMENTS OF SPECTRAL RADIATION 77

4.3.1 Spectral Measurements with
Pyrheliometers and Pyranometers 78
4.3.2 Sunphotometers 79
4.3.3 Calibration of Sunphotometers 83
4.3.4 Measurement of Ultraviolet Radiation 84

4.4 TECHNIQUES FOR RESEARCH-ORIENTED MEASUREMENTS 85

4.4.1 Measurements of Surface Albedo 85
4.4.2 Measurements of Surface Emittance 86
4.4.3 Measurements of the Polarization of Sunlight 87

REFERENCES 88
 TABLE OF CONTENTS

5. DATA PROCESSING¥ VALIDATION, AND QUALITY-CONTROL 93

5.1 INTRODUCTION 93

5.2 QUALITY-CONTROL AND VALIDATION 93

5.2.1 Checks Using Independent Measuring Systems 93
5.2.2 Consistency Checks Using Observed
Horizontal-Surface Irradiances 93
5.2.2.1 Global Solar Radiation (Egt) 93
5.2.2.2 Diffuse Solar Radiation (~.) 94
5.2.2.3 Direct Solar Radiation (S) 95
5.2.2.4 Reflected Solar Radiation (Ert) 95
5.2.2.5 Downward Atmospheric Radiation (Elt) 96
5.2.2.6 Upward Terrestrial Radiation (Elt) 96
5.2.2.7 Net Radiation (E*) 96

5.2.3 Consistency Checks Using Calculated

Horizontal-Surface Irradiances 97
5.2.3.1 Global Solar Radiation (Egt) 97
5.2.3.2 Diffuse Solar Radiation (Edt) 97
5.2.3.3 Direct Solar Radiation (S) 97
5.2.3.4 Reflected Solar Radiation (Ert) 97
5.2.3.5 Downward Atmospheric Radiation (El t) 98
5.2.3.6 Upward Terrestrial Radiation (Elt) 98
5.2.3.7 Net Radiation (E*) 98

5.2.4 Consistency Checks Using Measured

Inclined-Surface Irradiances 98
5.2.5 Consistency Checks Using Calculated
Inclined-Surface Irradiances 98
5.2.5.1 Direct Solar Radiation on an
Inclined Surface (SI) 99
5.2.5.2 Diffuse Solar Radiation on an
Inclined Surface (Ed,l t) 100
5.2.5.3 Global Solar Radiation on an
Inclined Surface (Eg,l~) 100
5.2.5.4 Downward Atmospheric Radiation on an
Inclined Surface (E,,1~) 100
5.2.5.5 Net Radiation on an
Inclined Surface (E*I) 101

5.2.6 Quality-Control of Broad-Band Spectral Data 101

5.3 HISSING DATA 102

5.3.1 Interpolation 102
5.3.2 Calculations Using Variables Coincidently
Observed vith the Data Break 103

5.4 DATA ARCHIVE 104

5.5 SUMMARY STATISTICS AND GRAPHS 105

5.5.1 Distributional Characteristics 105
5.5.2 Frequency Distributions 107
5.5.3 Joint Frequency Distributions 107
5.5.4 Threshold Analyses 107

5.6 TEMPORAL REPRESENTATIVNESS 109

5.7 SPATIAL REPRESENTATIVENESS 110

5.8 CONCLUSIONS 114

REFERENCES 114
 TABLE OF CONTENTS

A.1. EXTRATERRESTRIAL SOLAR SPECTRUM 119

A.2. NOMENCLATURE OF RADIATION QUANTITIES 127
A.3. CRITERIA FOR THE CHOICE OF INSTRUMENTS 131

11..4. GLOSSARY OF SYMBOLS 133

11..5. CONSTANTS AND CONVERSION FACTORS 135

INDEX 137
 FOREWORD

In response to a request by the World Meteorological Organization, the

Radiation Commission of the International Association of Meteorology and
Atmospheric Physics (IAMAP) accepted the responsibility of revising and
making current the material contained in the Instruction Manual on Radiation
Instruments and Measurements published as a guide for the International Geo-
physical Year 1957/58. The present manual is the product of such revision. It
was agreed that this manual should provide, in addition to an up-to-date
description of different radiation instruments used in the present global
observing network, a broad physical rationale for the requirements for the
information derived from these instrumental observations. The first three
chapters of the manual address the radiation characteristics of the physical
(earth-atmosphere) system to be observed. The last two chapters deal with the
types of instrument available for the necessary measurements and how the data
derived from the observations can best be presented to the different users.

The manual is the result of a cooperative effort of a number of authors,

both members of the Radiation Commission arid other experts in the field of
theory and measurement of radiation in the atmosphere. Contributions to the
various chapters have been made by the following:

K.L. Coulson. Vacaville, California, USA (Chapter 3 & 4);

C. Frohlich. World Radiation Centre, PMOD, Davos, Switzerland

(Chapter 4);

J. E. Hay. University of British Columbia, Vancouver, Canada (Chap-

ter 5);

J. Hickey. Eppley Laboratory, Newport, Rhode Island, USA (Chap-

ter 4);

. A.A. Kmito and D.V. Lutsko. Main Geophysical Observatory, Lenin-

grad, USSR (Chapter 4);

J. London. University of Colorado, Boulder, Colorado, USA (Chap-

ter 1);

G. W. Pal tridge. CSIRO, Sidney, Australia (Chapter 3);

E. Raschke and J. Schmetz. UniversitiH Koln, Koln, Federal Republic

of Germany, (Chapter 2).

Acknowledgements are extended to U. BUhler, S. Gerber and S. Hanen of

the World Radiation Centre, Davos for the careful prepar.ation of the manu-
script and the World Meteorological Organization for financial support and
for the printing of the manual.

.Da vos, June 1985 C. Frohlich and J. London, Editors

1
 1. INTRODUCTION

Electromagnetic radiation in the ultraviolet, visible and near infrared

received from the sun, and infrared radiation emitted by the earth and its
atmosphere represent the ultimate energy sources and sinks affecting the
meteorology and climatology of the planet. However, in addition to providing
for the basic heating and cooling distribution that activates the circulation
of the atmosphere and oceans, radiative processes are responsible for the
existence and maintenance of the earth's biotic system, for providing an
energy source for domestic and industrial purposes, for activating the earth-
atmosphere hydrological cycle, and for providing for a general environment
suitable for human habitation.

The energy input to the earth-atmosphere system is a function of the

net-radiation budget at the upper boundary of the system; that is, the diffe-
rence between the average solar irradiance and the combined reflected solar
and outgoing long-wave radiation at the top of the atmosphere. The horizontal
variation of this difference constitutes the primary driving force for large-
scale atmospheric motions. Until recently, the components of the radiation
budget at the upper boundary were inferred only from application of radiative
transfer theory to different climatological models. The first such global
study was that discussed in the classic paper by Simpson (929), who used a
relatively unsophisticated radiative treatment and simplified climatology to
compute the seasonal variation of the geographical distribution of the radia-
tion budget at the top of the atmosphere. Later calculations were based on
numerical analysis of the detailed structure of the radiation spectrum and
more realistic, observed distributions of the relevant climatological varia-
bles, such as cloudiness, water-vapour concentration, etc. It is now possible
to measure directly the global distribution and time variation of the indivi-
dual components of the earth's radiation budget by use of suitable satellite
observational techniques (see, for instance, Step hens et al., 1981; Raschke
and Kondratyev, WCP-40) .

. Evaluation of the different components of the radiation budget at the

earth's surface indicates how the net radiative input to the earth-atmosphere
system is partitioned and used. These components have also been calculated in
various radiative/climatological models. Some of the components, notably, the
direct and diffuse solar radiation received at the earth's surface, have been
observed for over 80 years. It is, however, still not possible to specify the
radiative energy budget components at the lower boundary solely from satel-
lites with sufficient reliability, although this problem is currently under
serious study (see, for instance, World Climate Research Programme, 1984;
Raschke, 1985). At present, determination of these values still requires a
global network of properly designed instruments to supplement theoretical
calculations of the direct and diffuse downward solar radiances and the up-
ward and downward infrared radiances at the surface. The ground-based obser-
vations are also needed to provide validation measurements for satellite
techniques being developed to estimate the different contributions to the
surface radiation budget. This information is essential for an understanding
of the modes of energy exchange between the atmosphere and the earth's sur-
face. Although whole spectrum measurements are generally sufficient for ap-
plication to most atmospheric and engineering problems, in some cases, e.g.,
radiative effects on biological systems, the spectral distribution of ob-
served downwelling solar radiation, particularly in the UV, is required.

Before the advent of rocket and satellite observations it was only pos-
sible to infer the mean irradiance at the top of the atmosphere (the solar
constant) by extrapolation of measurements from ground-based, balloonborne or
aircraft mounted instruments. However, it is now feasible to observe directly
the total solar irradiance and its spectral distribution from platforms above

3
CHAPTER 1 INTRODUCTION

the earth's atmosphere. Observations reported during the past five years
indicate an average solar constant of about 1367 Wm- 2 with a standard devia-
tion of the observed values of about ±0.3 per cent (Willson, 1984). Approxi-
mately 95 per cent of this energy is in the visible and near-infrared spec-
tral range (0.325 tlm - 2.8 tlm). About 30 per cent of the solar irradiance
received on a horizontal surface at the top of the atmosphere is reflected
back to space, primarily by clouds, but also by the clear atmosphere and the
earth's surface. Slightly more than 20 per cent is absorbed by atmospheric
water vapour, carbon dioxide, ozone and by clouds. ·The remaining incoming
solar energy, just under 50 per cent, is received at the earth's surface and
utilized to evaporate large quantities of ocean-surface water and to heat the
land and ocean surfaces. A small but essential amount of this energy is
converted for use in sustenance and growth of the world's biological system.

The earth radiates in the infrared (4 tlm - 100 tlm), but most of this
upward radiation is absorbed by clouds, atmospheric water vapour and carbon
dioxide. As a result, less than 10 per cent of the earth's emitted radiation
directly escapes to space. The remaining outgoing radiative energy emanates
from cloud tops and atmospheric· gases. On an average, downward radiation,
chiefly from the base of clouds, and from water vapour and carbon dioxide in
the lower atmosphere almost balances the upward surface radiation. Details of
the radiation characteristics of the earth and its atmosphere are discussed
in Chapters 2 and 3 of this Manual.

It has long been realized that observation of radiation quantities that

affect basic atmospheric and biological processes is required for an adequate
understanding of how these processes evolve and are maintained. Recognition
of the importance of documentation of computed radiation fluxes through rou-
tine radiation measurements led to suggestions at the end of the 19th century
that an international network of standardized observations be established.
Al though many different radiation instruments were developed earlier in the
19th century, very few, with the possible exception of the Campbell - Stokes
sunshine recorder, could be adapted for general use. The first reliable in-
strument to be used as a basis for standard routine observations of direct
solar radiation was the electrical compensation pyrheliometer constructed by
K. Angstrom in 1893. It was agreed by the International Meteorological Com-
mittee in 1905 that the Angstrom pyrheliometer be the standard for solar
irradiance measurements made at all central observatories (Moller, 1972;
1980). This instrument, with some modifications from the original, is still
one of the instruments used today for solar radiation measurements. Through-
out the first half of the 20th century emphasis was placed on design and
improvement of new solar radiation instruments. This emphasis was principally
because of the obvious direct effect that such radiation has on agriculture
and the subsequent food supply for the world's population. More recently,
Concerns for dwindling energy supplies for domestic and industrial use has
increased the need for documented information of the geographic distribution
of total (sun and sky) solar radiation available at the surface. As a result
of these different needs, various types of pyrheliometers and pyranometer.s
were built for routine use in the global observing network (see, for in-
stance, Drummond, 1970; Goulson, 1975). Absolute radiometers were developed
starting in the late 1960's (e.g. Kendall and Berdahl, 1970) and are present-
ly in use in about one third of the radiation calibration centres (see the
discussion of absolute radiometers in Chapter 4).

Progress in the design of practical methods of implementing network

observations of infrared earth and sky radiances was much slower, probably as
a result of the inadequacy of the development ·of suitable filters and .detec-
tor techniques for operational field use. Very few regular infrared measure-
ments at the earth's surface were made prior to about 1940. The source of
almost all of the downward infrared radiation received at the earth's surface
is in the lower troposphere and is thus very sensitive to local atmospheric
conditions. Frequently, therefore, the observed data are not representative
of regions much beyond the restricted area of the measurements themselves.
Nevertheless., the need to validate and augment theoretical calculations of

4
CHAPTER 1 INTRODUCTION

the net infrared-radiation balance at the ground has led to some effort to-
wards improving the surface network of infrared observations. Such measure-
ments are also required for detection of long-term changes of upward- and
downward-directed radiative components at the earth's surface to provide
useful information of the mechanisms of climate change. Since there are no
methods other than ground-based observations to measure the downward infrared
radiation at the surf ace, it has been proposed by the Radiation Com mission
that the present, very limited network of net-atmospheric-radiation measure-
ments be extended. Detailed descriptions of radiation instruments, both solar
and terrestrial, that are presently operational, including principles of
measurements and techniques for their use, are described in Chapter 4.

It is essential, of cour.se, that the different instruments used in glo-

bal radiation network provide a compatible and consistent measurement set.
This requires periodic intercomparisons of the different instrumental obser-
vations and reduction of the data to a standard reference scale. Prior to
1956 two different radiation scales were used, each based on a different
pyrheliometer type. In 1956 a new standard scale, the International Pyrhelio-
metric Scale (IPS 1956), was adopted for use during the IGY and for sub-
sequent observations. It was later found, however, that corrections were
needed to IPS 1956. As a result the World Radiometric Reference (WRR) was
officially adopted in 1981 as the Standard Scale for all solar radiation
measurements. The history and rationale for the solution of this problem is
discussed in Chapter 4.

The generation of motions in different atmospheric layers is in response

to the distribution of radiative sources and sinks within the atmosphere.
These quantities are currently determined solely from radiative transfer
calculations based on the observed atmospheric structure or theoretical atmo-
spheric models. Solar and infrared research instruments, balloon borne or
mounted on aircraft, have frequently been used in large experiments as a
check on theoretical calculations of radiative heating and cooling rates
(e. g . , B0 MEX, GAT E, et c.) . Re g u 1a rope rat ion s 0 f s u chi n s t rum en t s nee d a
broad base of support and there are no current plans for routine network
measurements of radiative heating and cooling rates in the free air.

The first comprehensive handbook of radiation instrument,s and measure-

ments was prepared for the extended observational programmes planned for the
International Geophysical Year (CSAGI, 1958). Experience gained during the
IGY, and subsequently, indicated the value, and indeed the necessity, for the
expansion and maintenance of the global network of observed radiation quanti-
ties. This need became even more apparent during the formative years of addi-
tional international campaigns, such as the Global Atmospheric Research Pro-
gramme (GARP) and the World Climate Programme (WCP). A glohal effort to coor-
dinate ground-based radiation measurements with planned satellite measure-
ments would also be an essential part of the currently proposed International
Geosphere-Biosphere Program me <IGBP). In 1973, the Com mission for Instruments
and Methods of Observation (CIMO) recommended the establishment 'of a Working
Group on Radiation Measurement Systems that would, among othar things, assist
in the preparation of a new meteorological radiation manual (WMO, 1973),
During the past 25 years new instruments have been designed and older instru-
ments have been modified. The present manual is intended to make current the
relevant information contained in the original handbook.

It is important for clarity of description of radiation measurements

tha t a consistent and universally recognized terminology and system of units
be used. The general system adopted in this manual is based on the Interna-
tional System of Units (SI) with special application to radi,ation quantities
as discussed in the report by the Radiation Commission of- IAMAP edited by
Raschke (1978), Definitions of radiation quantities, symbols and units used
are given in the Appendix A.1,

5
CHAPTER 1 INTRODUCTION

REFERENCES

Coulson, K.L. (1975) Solar and Terrestrial Radiation, Academic Press, New
York.

CSAGI (1958) Radiation Instruments and Measurements, Part VI, IGY Instruction
Manual, p. 371. Pergamon Press; London, New York, Paris.

Drummond, A.J. (1970) A survey of the important developments in thermal ra-

diometry, A.J. Drummond, ed., Adv. in G'eopl1ys. 14, Academic Press.

Kendall, J .M., Sr. and C.M. Berdahl (1970) Two blackbody radiometers of high
accuracy. App.l. opt. 9: 1082.

Moller, F. (1972) Radiation in the At mosphere. In Meteorological Challenges:

A History, D.P. McIntyre, ed"., In£. Canada, Ottawa, p. 43.

Moller, F. (1980) A Brief History of the Radiation Commissions of the IMO and
of the IUGG, Radiation Com mission (lAMAP) p. 43.

Ra s ch k e , E. (e d) (1 978) Ter mi n 0 I og y and Un its of Ra d i a t i on Q u ant it i e san d

Measurements, Radiation Commission (lA MAP), Boulder, Colorado, p. 17.

Raschke, E. and K. Ya. Kondratyev: Radiation Budget of the Earth and Its
Atmosphere, WCP-40, World Climate Research Programme.

Raschke, E. (1985) On the Derivation of Radiation Budget Parameters at the

Surface from Satellite Measurements. Submitted to COSPAR 1984 for publi-
cation.

Simpson, G.C. (1929) The distribution of terrestrial radiation. Hem. Roy.

Het. Soc. 111: 53.

Stephens, G.L., C.G. Campbell and T.H. Vonder Haar (1981) Earth radiation
budgets measurements from satellites and their interpretation for cli-
mate modeling and studies. J. G'eopl1ys. Res. 86: 9739.

Willson, R.C. (1984) Measurements of Solar Total Irradiance and its Variabi-
lity. Space Sel. Rev. 38: 203.

World Meteorological Organization (1973) Commission for Instruments and

Methods of Observation, Abridged Final Report of the Sixth Session, 6~17
August 1973, Helsinki, WMO, No. 363, Geneva, Switzerland.

World Climate Research Programme (1984) Satellite Systems to Measure the

Earth's Radiation Budget Parameters and Climate Change Signals, WCP-70.

6
 2. RADIATION IN THE ATMOSPHERE

2.1 INTRODUCTION

The interaction of radiation with the atmosphere and the underlying sur-
face, such as land, water and ice, involves:

absorption of radiative energy causing mainly conversion into heat

-scattering and reflection of radiation causing angular and spatial redi-

stribution of the incident radiation

emission of thermal radiation causing cooling of the atmospheric layers.

These processes determine the transfer of radiative energy in the earth-

atmosphere system. They are influenced by the abundance of atmospheric gases
a nd the concen tra tion of aerosols and cloud particles. They will be de-
scribed in detail in Section 2.4. For all energy-related problems below 50 km
radiation can be separated in the spectral range of about 0.2 to 4.0 ~m
emitted by the sun from the long-wavelength range of thermal radiation. The
latter range extends, for atmospheric temperatures of 200 to 280 K, from 4 to
100 ~m.

2.2 ATMOSPHERIC COMPOSITION

2.2.1 Gases and Mean Structure

The most important gaseous atmospheric constituents of dry air are lis-
ted· in Table 2.1. Water vapour (H 2 0) has a molecular weight of 18.02 and is
present in variable amounts of 0-0.04 fraction of total molecules.

Table 2.1 Composition of the earth's atmosphere below 100 km (modified after
U.S. Standard Atmosphere, 1976).

Molecular Content
Constituent weight (fraction of total molecules)

Nitrogen (N 2 ) 28.01 0.7808 (75.51% by mass)

Oxygen (0 2 ) 32.00 0.2095 (23.14% by mass)
Argon (Ar) 39.95 0.0093 (1.28% by mass)
Carbon dioxide (C0 2 ) 44.01 340 parts per million (1983)
Neon (Ne) 20.18 18 parts per million
Helium (He) 4.00 5 parts per million
Krypton (Kr) 83.8 1 part per million
Hydrogen (H 2 ) 2.02 0.5 parts per million
Ozone (0 3 ) 48.00 0-12 parts per million (variable)

These gases are uniformly mixed, except as noted below, up to the meso-
pause (Le. to about 85 km), where photo-dissociation and gravitational dif-
fusive separation becomes increasingly important with increasing height. Car-
bon dioxide has a seasonal variation in the lower troposphere of a few parts

7
CHAPTER 2 RADIATION IN THE ATMOSPHERE

per million, and ozone, abundant mainly in the lower and middle stratosphere
(15 - 35 km) also has large seasonal and geographic variations. Atmospheric
water is found in the atmosphere in all three states: vapour, fluid and
solid.

Incident solar radiation is scattered by all these gases. Some, princi-

pally water vapour, carbon dioxide, bzone, oxygen and methane, absorb solar
radiation in the ultraviolet, visible and near infrared, and absorb and emit
thermal radiation in the infrared.

2.2.2 Aerosols

Soil particles, dust and sea salt form aerosols, which are radiatively
active components of the atmosphere. They are active in a direct way by:

absorption and scattering of incident radiation,

and an indirect way as:

condensation nuclei, which increase the cloud particle concentration and

thus the optical thickness of clouds.

The amount of the scattering and absorption of aerosols depends strongly

on their size and therefore on their size distribution. Two typical size
distributions are shown in Fig. 2.1. Depending on the origin of the par-
ticles, the maxima of number densities may lie at radii of about 0.001 to
1 J-lm. Further, hygroscopic aerosols, e.g. sea-salt crystals, change their
size with changing ambient water-vapour concentrations, with a consequent

8
10

6 MARITIME
...
'0
-10
AEROSOL MODE L

--"...
Z
10
4

2
CONTINENTAL /

-
'0
10 URIGINN
~

0 ~~, \
0
c 10
CD / / SEA SPRAY \
'0 -2 / PRODUCED \
... 10 I
CD
.0 I
/ \
E 10 -4 I
\
::::J
Z \
-6
,10 \
\
-8
10 -J -2 -1: 0 1
10 10 10 10 10
Radius <lAm)

Figure 2.1 Two size distributions of atmospheric aerosols (maritime) (from

Shet tIe and Fenn, 1976). Short-dashed line: sea spray; long-dashed line:
continental origin;, solid line: combination of the two.

8
CHAPTER 2 RADIATION IN THR ATMOSPHERE

change in their size distribution and hence their radiative properties. The
largest aerosol concentrations are found in the planetary boundary layer over
industrial areas and over the oceans. After strong volcanic eruptions a
different, much smaller, aerosol maximum may be found in the lower strato-
sphere (Fig. 2.2).

70
60 STANDARD AEROSOL MODEL
~ VERTICAL PROFILE S
50 ~

40 - ' HAZY
---CLEAR
30 "\
CHANGE \ ....... RAYLEIGH
IN \

-E
JiIi:
..,....
25
SCALE \
\
\
\
\ .
'
RAYLE IGH
SCATTERING

CD 20 BACKGROUND \ . . . VOLCANIC
"C STRATOSPHERIC I '.
.....-
-
~
".\I
15
et I
I w
I 0:
UJ
\
10 \
I
Cl..

,,
\ I./)
\ 0
Cl..
'. 0
5 \. 0:>-
~o:
\
'.',
'.
.....
oc{
0
Z
-::>
-4 -3 -2 -1 0 0
cD
10 10 10 10 10
Radius (pm)
Figure 2.2 Vertical profile of the extinction coefficients at 550 nm of
a tmospheric aerosol profiles (Shet tle and Fenn, 1976).

West
150 120 90 60 30 0

o
('I')

-
.c
: :l
o
(f)

Figure 2.3 Global, fractional cloud cover in January (National Academy of

Sciences, 1975).

9
CHAPTER 2 RADIATION IN THE ATMOSPHERE

2.2.3 Clouds
Clouds are found in all tropospheric layers and, occasionally, in the
lower stratosphere. Clouds contain water particles in one or two states:
liquid and/or solid. Although estimates of the global cloud distribution vary
somewhat, it is likely that the global cloud cover is around 60 per cent
(e.g. Berlyand et al., 1980). Fig. 2.3 (from National Academy of Sciences,
1975) shows the global distribution of fractional cloud cover for the month
of January.
The usual size of cloud droplets and particles ranges from 2 to 20 tim.
Larger raindrops may have diameters of up to 10 mm. The largest "particles"
are found in convective tropical and extra tropical cloud systems as hail-
stones with diameters of up to 10 cm. Typical size distributions are shown in
Fig. 2.4. Fig. 2.5 shows a vertical profile of the size range of droplets in
a low stratocumulus cloud. This figure also shows a profile of the liquid-
water content (LWC in g of liquid water per m- J air); on occasion the LWC may
reach even larger values. The content of solid water in dense cirrus may be
as high as 0.03 gm- J (Griffith and Cox, 1977). .

2
n(r) 10
1
10
0
10

-1
10

-2
10 0 2· 4 6 10 12 14
r(pm)
Figure 2.4 Size distribution of cloud droplets (Deirmendjian, 1969).

8.AUG.1978
Sc -3 8.Aug.197B_ 1000
LWC=0.4gm 3
Sc-1 cloud to~level
Sc-2 mid cloud leve I E800
Sc-3 cloud base level _ -a
Gi 600 ./
:c adiabatic LWC

OL.-=::::::.-::::::......_~~-..::::::-_-+---..-.. " OO.J--..........>-+_-4--+-o---l~--. .

o 2 4 6 8 10 12 14 16 18 20 o Q2 0.4 0.6 0.8 1.0
Radius (pm) LWC (gr m- 3 )

Figure .2. 5 Drop~size distribution <left) and liquid-water content (right)

measured in a maritime stratocumulus cloud.

Clouds primarily reflect and scatter incident solar radiation. They are
almost black for terrestrial thermal radiation and absorb practically all
incident IR radiation. The emission is thus determined mainly by the tempe-
rature at their surface.

10
CHAPTER 2 RADIATION IN THE ATMOSPHERE

2.3 BASIC TERMINOLOGY OF RADIATIVE ENERGY FLUXES

The description of atmospheric radiative transfer and of measured radia-

tion quantities should be based on a terminology which is coordinated in-
ternationally. Such a nomenclature system, recommended by the Radiation Com-
mission (IAMAP) and adopted by IUGG has been discussed by Raschke (1980) and
Cl MO Guide (1983). It is based on the International System of Units (SI) and
the recommendations of the International Commission of Illumination (CIE).
The main terms and some quantities important for meteorological field work
are sum marized below,

2.3.1 Radiometric Quantities

The terms and definitions of radiative quantities are given in Appendix

A 2, Table A 2.2 and A 2.4.

The directional dependence of radiation may easily be described in terms

of spherical coordinates (Fig. 2.6), where B is the zenith angle and t/> the
azimuthal angle with respect to a direction of reference. Fig. 2.6 also
illustrates the solid angle element dQ:

dQ = sinB dB dt/>. (2.1)

~=o

!p=

Figure 2.6 Spherical coordinate system

2.3.2 Properties of Media

The properties of emittance, absorptance, reflectance and transmittance

in a radiation field are given in the Appendix A 2, Table A 2.3. These can be
applied to layers of air, cloud sheets, surfaces, etc. The names of these
four quantities end with the suffix ..... ance .. replacing. the formerly used
suffix ..... ivity... The quantities relevant to volume elements are defined as
follows:

The extinction coefficient a., (Bouguer - Lambert law, also referred to as

Beer's law):

1 dL
a. = L --
ds
(m- 1 )
'
(2.2)

L is the radiance and ds an element of distance along the path. Similar

coefficients, a. and as, can be defined for absorption or scattering. For a

11
CHAPTER 2 RADIATION IN THE ATMOSPHERE

particle density N(m- 3 ), the single particle cross section (m 2 ) is given by:'

= .GE'A'S.
(2.3)
N

For a single scattering process within the volume element, the single scat-
tering albedo w is defined as the ratio of the part scattered to the total
extinction:

(2.4)

For a pure scatterer w == 1 and for a pure absorber w == O.

The "optical thickness", OE, between the points Xi and X2 is defined as:

X2
OE ::: JaE' ds (2.5)
Xi

The scattering by molecules, aerosols, and other particles as well as

the reflectance at surfaces causes strong angulat dependence of the radiation
fields in the atmosphere. Therefore, a terminology has been developed for the
angular description of radiation. It refers to light being scattered, re-
flected and transmitted and is summarized in Appendix A 2, Table A 2.1.

The scattering and absorption in gases, water droplets or at the ground

can be described by their dielectric properties along a path of light. That
is, by the complex index of refraction, which is defined as:

m ::: m.. + i· m\ (2.6)

where 12 ::: -1, and the subscripts rand i mean the real and imaginary
parts respectively. The reflection and refraction processes (including scat-
tering) are due to the real part m.. , while the imaginary part m\ stands
for absorption ("damping") of radiation. In general, the index of refraction
depends on the radiation wavelength.

2.4 INTERACTIONS BETWEEN RADIATION AND ATMOSPHERE

Although electromagnetic radiation does not, in general, penetrate into

soil, short wave radiation in the visible range (between about 0.45 and
0.7 pm) can penetrate to a depth of about 100 metres in water, at which point
it has been almost completely absorbed. It also penetrates partly into ice
and snow. In the absence of clouds these surfaces can be considerd as lower
boundary planes for infrared radiative transfer in the atmosphere.

2.4.1 Radiative Transfer Equation

Radiation is scattered and absorbed along its path through the atmo-
sphere - or, in more general terms - through any gaseous, fluid or solid
matter. Also, radiation may be added to a given beam, either due to scat-
tering into the direction of the beam from any other direction or due to
thermal emission within the volume element. Thus the transfer of a beam of
radiation through a volume element can be given' by a differential equation
which describes the energy budget along the path of the beam through this
volume element.

12
CHAPTER 2 RADIATION IN THE ATMOSPHERE

For simplicity, we assume an infinitesimally thin and horizontally homo-

geneous slab of optical thickness, do = 0E 'ds, and apply the coordinate
system given. in Fig. 2.6. Furthermore, j1 = cos8. Then, the equation for
radiative transfer may be written:

(2.7)

The first part of the right-hand side of Eq. 2.7 is the law of Bouguer -
Lambert as described above. The second part, J(j1,ijI), is called the source
function and describes the direct and scattered solar radiation and thermal
radiation emitted from and scattered into and out of this volume element.
Each contribution of the source function depends on the spectral range consi-
dered and on the properties of the atmosphere.

Consider an attenuating layer of large horizontal extent with the lower

and upper boundaries, in terms of optical thickness, at 0 = 00 and 01. The
formal solution for the upward radiance can then be written as:

dt
(2.8)
j1

A similar solution exists for the downward radiance. The first term in
Eq. 2.8 describes the radiance transmitted through the layer. The second
term describes the light scattered and radiation emitted from the layer in
the direction (j1, ijI) and integrated from 00 to 0 where t represents the dummy
integration variable. Many algorithms have been developed to solve this equa-
tion numerically for a great number of practical applications to realistic
atmospheric conditions. A summary of the algorithms and a comparison of
their performance has been published by the Radiation Com mission of IAMAP
(Lenoble, 1977 and Fouquart et aI., 1980). In terrestrial applications radia-
tion·from the sun dominates the source term at wavelengths less than about
4 j1m, whereas thermal radiation is dominant for A > 4 j1m.

2.4.2 Scattering and Reflection

Changes of the index of refraction within a medium cause changes in the

direction of the beam of radiation. If the beam impinges on a surface, part
of the radiation is reflected specularly or diffusely into the backward di-
rection and part of it penetrates into the medium. The reflection properties
of natural surfaces vary strongly with type, roughness, wavelength and angle
of incidence.

The term 'scattering' is. used to explain all processes occurring within
geometrical. dimensions of the order of, or smaller than, the wavelength of
the radiation involved. The spherical waves reflected in all directions to-
gether with the refracted part affect scattered light through subtle inter-
ference.

The angular distribution of single scattered radiation is described by

the "phase function" or "scattering indica trix" p , which is dimensionless
and defined by:

Cs • Lo
L(8,r) = 4nr 2 • p(8,r) , (2.9)

where L<8,r) is the scattered radiance at a large distance, r, from the

particle, and Lo the radiance incident on the particle. The term 'phase func-

13
CHAPTER 2 RADIATION IN THR ATMOSPHRRE

tion' describes the angular distribution of the scattered radiation. As the

cross section Cs is defined by the ratio of all radiation lost to the inci-
dent radia tion, the in tegral of p over 4n ha s to be nor malized to 4n in
order to satisfy its definition according to Hq. 2.9.

The scatterers in the atmosphere exist in a variety of dimensions. As-

suming spherical particles and, as the scattering properties depend on the
di!TIension (e.g. radius) of the particles, the overall scattering coefficient
has to be calculated by weighting the scattering cross section Cs by the
size disteribution n(r):

(J = fC s 'n(r)dr and N=fn(rJdr, (2.10)

with N as the total number of particles per unit volume. Similar formulas
have to be used for the phase function, the single scattering albedo, and
other optical properties.

If the wavelength is much larger than the dimensions of the scattering

particles, so-called Rayleigh-Scattering is valid. In that case, the scat-
tering coefficient as R is proportional to ).,- 4, and is given by:

32 1
aSR =3 N • (m-1)2. (2.11)

where m is the refractive index of medium (e.g. air), and l)p is the so-
called depolarization factor. For air l)p = 0.0279 (Young, 1980). The phase
function p(8 s ) is given by:

p (8 s ) - -4
3
. (1 + cos2 8s ) , (2.12)

where 8s is the scattering angle, Le. the angle between the direction of the
incident beam and the scattered ray (8 s = 0 for forward scattering, 8s = 180
for backward scattering). .

o
It) 0 = 1.50 + O.OOi

o
.
o·
('I)

o
C'I

o
X =604 om
o
o-1-- ....-:~~~;d::::::--_.__--,--~~--,--,..---.,~-__r___'-...__~

0.01 ·0.02 0.05 0.1 0.2 0.5 1.0 2.0 5.0 10.0
Radius (pm)

Figure 2.7 Scattering efficiency Q = Cs /nr 2 as function of particle radius

and for 5 different wavelengths calculated after Mie (1908).
CHAPTER 2 RADIATION IN THE ATMOSPHERE

Rayleigh scattering is valid as a first-order description of the visible

light field in extremely clear air or for the description of radar waves
within clouds.

N = 100cm'3

1.0

.6
...,.._.~~. / I
/
Lt ,\
. C3
.......
E
I

--...
CJ
10
2

''f-~ d'" ~.~

'C 3
...... 10
.4 c
\. 'C

\~
....>-
.2 III
c 10"
Gl

.. .1
..
'C

GI
~P3
\') .0
E 10 5
.06 L :::l
Z
.04
\ 10&
.02 ~
.01
H- N p4
_ 10.7
.4 .8 1 2 4 .01 .1 1 10
A (.-n) Radius (IJm)

Figure 2.8 Extinction coefficient aE a-s function of wa velength (left)

for different size distributions (right).
3
C 10 T"""-........~..................'-'-'"..............................................." - - - . ,

-.o...
() 2
10
-Sc-1
C ---Sc-3
....::I 550nm

,,

-2
10

-3
1 0 ....T"'T",.,-..,....,..,....",.....'T""I"........TT'"r..-r---y----i
2 6 10 30 50 70 90 110130150170175 180
Scattering angle
Figure 2.9 Phase functions for the drop-size distributions Sc.-1 and Sc-3 in
Fig. 2.5.

15
CHAPTER 2 RADIATION IN THE ATMOSPHERE

If the particle size becomes comparable to the wavelength, as and p may

be calculated from an explicit solution of the Maxwell equations (Mie, 1908).
The scattering cross section is no longer a simple function of the wavelength
(Fig. 2.7). Although some smoothing occurs if many particles of different
sizes are mixed together as still sh:ows a more or less complicated behaviour.
However, within a small wavelength range in the visible, as may be approxi-
mated by a power law with the exponent depending on' the particle-size dis-
tribution and type of the particles (Fig. 2.8). An example of the phase
function for particle mixtures, as described in Section 2.2, is given in
Fig. 2.9 and shows a strong forward peak and only weak backward Scattering.

As shown above, the scattering properties can be calculated using the

Mie theory from a priori knoWledge of, the index of refraction, the size dis-
tribution and the concentration, if the particleS are assumed to have a sphe-
rical shape. This may be, valid for cloud droplets, but not necessarily for
aerosol particles. However, several investigations, both theoretical and
experimental, have shown that the irregular particles may be quite well ap-
proximated by "equivalent spheres" (see e.g. Heintzenberg, 1977).

204.3 Polarization Effects

Up to now polarization effects have been neglected and only properties

of the radiance, Le. the sum of the squares of the polarized amplitudes,
have been described. Light has many of the properties of a wave oscillating
in a plane normal to its direction of propagation. In this case the oscilla-
tion is that of an electric (or magnetic) vector, and the state of polariza-
tion is determined by the character of the oscillation. If the vector de-
scribes a circle with no change of amplitude, the light is circularly po-
larized; i f i t describes an ellipse, involving a change of amplitude, the
polarization is elliptical; i f the motion is a single plane, then linear
polarization results~ In addition, there may be random variations of the
vector superimposed on its regular behaviour, in which case partial polariza-
tion is the result. Light in the atmospheric environment is in most cases
partially polarized, and since the circular and elliptic components are nor-
mally of very small magnitude, the usual state for the day Ught is partial
plane polarization. The human eye is almost completely insensitive to po-
larization, so we are not usually aware of the polarization which exists,
unless the eye is aided by some type of polarizer, such as the well-known
polarizing sun-glasses.

A partially plane polarized beam of light may be described completely by

its average radiance L, its degree of polarizationP, and the plane of pola-
rization X taken with respect 'to an arbitrary direction in the transverse
plane. Our usual method of obtaining information on the environment is by
use of the radiance alone, such as is sensed by the eye, photographic came-
ras, or some types of 'radiometers. It is being incre.asingly recognized,
however, that the fields of the degree and plane of polarization also carry
information, some of which is not available in the radiance measurements
alone. A realization of the possible utility. of this additional informaHon,
remote sensing and diagnosing of the radiative properties of the atmosphere,
has generated an increased interest in thepolatization of light in the
environment.

There is a comprehensive literature on the polarization of light in the

environment, and the various aspects of this phenomenon can only be touched
on here. Reviews of the subject,. with extensive'bibliographies, have been
given by Sekera (1957) and Coulson (1974). It is presently a veryacti.ve
area of research, and it is probable that interest will expand even further
as the potential use of polarization obse:r;:-vations in remote sensing applica-
tions becomes more generally recognized.

There are three principal processes in the natural environment that are
responsible for polarization. First, very strong polarization fields are

16
CHAPTER 2 RADIATION IN THE ATMOSPHERE

produced by scattering of sunlight by gaseous molecules in the atmosphere,

described as 'Rayleigh scattering'. In fact, scattering by a single molecule
will produce c;omplete polarization at certain angles, but in the real atmo-
sphere the numerous molecules cause successive scattering events among the
particles and prevent the light from being completely polarized. While the
ideal case of a pure Rayleigh atmosphere does not exist in nature, it is a
model which has received a great deal of attention and which may be approxi-
mated reasonably well by the very clearest cases of the actual atmosphere.
-,.
90 I
,/ \
\
\ 90
, I \
.)
,, /-,
\
80
, \
\
\
80 I \
\
\
-
~
70 \
\
70
,I \
\
z
Q 60 ~ 60 \
t-
,,- ....., \
,
« I z
N I 0 / \
ii 50 I
\ ~ 50 \
«
..A \ ~ I '
0 a: I ..........
\ \
::: 40
0
\ \
\ \ :}40
~ /' " \
W
w30
\
.... \ \ \
,,\.\.\ \ \ ~ 30 "' ,-..... \ '(ABO
\. \~0.25

,
;'
a: .... \ \ \ \. 0.50
I
w
'~0.75
Cl .
w 1&.1
0
20
~\
A-O ~\
\ \ \. \ ~ 20
1&.1
0
,\,\ \
\
10 0.25~ \ 10 '\
0.50.:=::--", \. '\
-.;:~
0.75 "
0 0

-10 -10

Figure 2.10 a) Degree of polarization as a function of angle in the vertical

plane through the sun for a Rayleigh atmosphere at a wavelength of 800
nm and at an altitude of 3400m for four values of surface albedo and two
zenith angles of the sun. BA and BR indicate the positions of the neu-
tral points of Babinet and Brewster. b) Same as a) but for 0.365 ,urn. In
addition AR indicates the position of the neutral point of Arago (from
Coulson et al., 1960).

The degree of polarization P (ratio of the linearly polarized intensity

to the total intensity) of skylight for a Rayleigh atmosphere varies with
wavelength, zenith angle of the sun, albedo of the underlying' surface, alti-
tude of the site, and zenith and azimuth angles of the position in the sky.
The distribution of P in the vertical plane through the sun is shown in
Fig. 2.10a for a wavelength of 0.80,um, a site altitude of 3400 m, two dif-
ferent solar zenith angles 8., and four values of the albedo A. These values
are calculated assuming a pure Rayleigh atmosphere and a Lambertian-type
albedo (isotropic angular distribution).

While most of these features correspond to reality, two would be modi-

fied somewhat by a more realistic model. First, the magnitude of P at the
maximum would be decreased by 3 to 4 per cent, if the anisotropy of the atmo-
spheric molecules were considered in the computations. Secondly, the albedo
effect would be slightly modified and the positions of the neutral points
would be a weak function of albedo, if the anisotropy and degree of polari-
zation of light reflected from real surfaces were taken into account in the
computations.

17
CHAPTER 2 RADIATION IN THE ATMOSPHERE

The maxima of P are considerably lower here than for 0.80 Mm, and the
negative portion of the polarization field is much more pronounced. The
neutral points are distinct in both cases, their angular distance from the
sun is greater, and the Brewster point of the previous diagram gives way to
the Arago point here. All of these features are explained by the greater
amount of multiple scattering which occurs at 0.365 Mm.

The second source of polarization in the atmosphere is due to numerous

types of aerosols, such as dust, haze, crystals, pollen, and even droplets of
water or sulphuric acid. In general, such particles produce much smaller
degrees of polarization than do Rayleigh particles. The polarization proper-
ties of these larger particles can also be calculated by the Mie theory. The
real atmosphere, even in the cleanest locations, always contains considerable
numbers of Mie particles, and the light which we see from the sunlit sky is
the result of scattering by the composite of Rayleigh- and Mie-type scatter-
ers.

The third type of a strong polarizing mechanism operating in the natural

environment is that of reflection from solid or liquid surfaces. still water
has the capability of producing complete polarization at one angle of reflec-
tion, and although this ideal case does not apply to a surface of rough water
or at other angles, light reflected from lakes or the sea does tend to· be
highly polarized. Somewhat less polarization is produced by reflection from
soils, sands~ rocks, and vegetation, but even here polarization values of 30
per cent to 40 per cent are not uncommon. Polarization effects are further
discussed in Sections 3.5.4 and 3.8.3.

2.4.4 Absorption and Thermal Emission

The molecules of several atmospheric gases possess electric or magnetic

dipole, or multipole, moments. Thus, they are able to absorb and emit ra-
diative energy at particular resonance wavelengths given by rotational, vi-
brational and electronic transitions. The basic theory of these processes
may be found in many texts (e.g. Goody, 1964; Paltridge and Platt, 1976).

The absorption per unit volume is given by the absorption coefficient aG

(m-i), which can be calculated for a single absorption line or for groups of
lines. However, in most applications the absorption is required for wave-
lengths in broad-band regions. For numerical calculations it is most con-
venient to use average transmittances .(},) as a function of the generalized
absorption coefficient C(;\) and the absorber amount w' .

• (A) =f (C(A),W') (2.12)

A very useful compilation of average transmittances for all relevant

atmospheric constituents can be found in Kneizys et al. (1980).

Kirchhoff's law states that a gas in local thermodynamic equilibrium

emits radiation with its ambient temperature T according to

(2.13)

where B(A,T) is the Planck functio'n.

(2.14)

G1 and Cz are the so-called Planck constants (C 1 = 3.7419,10- 16 Wm z ,

Cz = 1.4388,10- z m·K). In the case of atmospheDic radiation in the cloud-
·free at mosphere L., ~ of Eq. 2.13 is the only source term in the radiative
transfer Eq. 2.3 for thermal radiation.

18
CHAPTER 2 RADIATION IN THE ATMOSPHERE

The emission of thermal radiation from surfaces is described by the same

Eq. 2.8. a6(1I) is then replaced by the emittance E:(;\") of the surface. Most
soils, the oceans and vegetation emit at values between 0.95 and 0.99 in the
spectral range of thermal radiation (from 4 to 100 ,urn). Soils may have some
spectral features around 10,um due to reststrahlen-bands (see, also, Section
3.8.4).

REFERENCES

Berlyand, T.G., L.A. Stroknia and L.E. Gresnnikova (1980) Zonal Distribution
of Cloud Amount on the Earth. Het.lwd G.idr. 3: 15, Allerton Press.

CIMO Guide (1983) Chapter 9: Solar Radiation, Meteorological "Instruments

and Methods of Observation, 5th ed., WMO, Geneva.

Coulson, K.L., J. V. Da ve and Z. Sek era (1960) Tables Rela ted to Radia tion
Emerging from a Planetary Atmosphere with Rayleigh Scattering.' Univer-
sity of California Press, Berkeley and Los Angeles

Coulson, K.L. (1974) The polarization of light in the environment, pp. 444-
471, . in Gehrels, T. (Ed.) Planets stars and Nebulae Studied with Photo-
polarimetry, University of Arizona Press, Tucson, Arizona.

Deirmendjian, D. (1969) Electromagnetic scattering on spherical polydisper-

sions. American Elsevier Publishing Company, New York, p. 290.

Fouquart, Y., W.M. Irvine and J. Lenoble ed. (980) Standard Procedures to
Corn pute Atmospheric Radia ti ve Tra n sfer in a Sca t tering Atmosphere.
Vol.II. Radiation Corn mission, CIAMAP), Boulder, Colorado.

Goody, R.M. (1964) Atmospheric' Radia tion, Oxford Monographs on Meteorology,

Clarendon press, p. 436.

Griffith, K.T. and S.K. Cox (1977) Infrared radiative properties of tropical
cirrus clouds inferred from broadband measurements. Atmos. ScL Paper
No. 269, Colorado State University, Fort Collins, (NTIS PB 26853).

Heintzenberg, J. (1977) Particle size distributions from scattering data on

non spherical particles via Mie Theory, Proc. Rad. Symp. Garmisch Par-
tenkirchen, H.J .Bolle Ed., Science Press, p. 157.

Kneizys, F.X., E.P. Shettle, W.O. Gallery, J.H. Chetwynd, Jr., L.W. Abreu,
J.E.A. Selby, R.W. Fenn and R.A. McClatchey (1980) Atmospheric Transmit-
tance/Radiance: Computer Code LOWTRAN 5, AFGL - TR - 80 - 0067, Environ~
mental Research Papers, No. 697, Optical Physics Division, Air Force
Geophysics Laboratory, Hanscom, Massachusetts.

Lenoble, J. ed. (1977) Standard Procedures to Compute Atmospheric Radiative

Transfer in a Scattering Atmosphere; Vol.!. Radiation Commission,
CIAMAP), Boulder Colorado.

Mie, G. (1908) Bei tdige' zur Optik tri..iber Medien, speziell kolloidaler
Metalllosungen. Anna.len der Phys.ik 25: 377.

National Academy of Sciences (975) Understanding climatic change - A program

for action, Washington D.C.

Paltridge, G.W. and C.M.R. Platt (1976) Radiative Processes in Meteorology

and Climatology. Elsevier ScL Publ. Company, Amsterdam - Oxford - New
York.

19
CHAPTER 2 RADIATION IN THE ATMOSPHERE

Raschke, E. (198Q) Terminology and Units of Radiation Quantities and Measure-

ments, Radiation Commission (IAMAP), Boulder, Colorado, p.• 17.

Sekera, L. (1957) Polarization of skylight. £ncycJoped.ia of Physics 48:288,

Springer, New York.

Shettle, E.P. and R.W. Fenn (1976) Models of Atmospheric Aerosols and their
Optical Properties. In: AGARD Conference Proceedings No. 183. Opt.ical
Propagation in the Atmosphere. Presented at the Electromagnetic Wave
Propagation Panel Symposium, Lyngby, Denmark, 27-31 October 1975, AGARD-
CP -18 3 , available from U•S • Na t i on a I T e ch n ic a I Inform a t ion Se r vie,e
(No. AD-A028-615).

U. S. Standard Atm,osphere (1976) National Oceanic and Atmospheric Administra-

tion, National Aeronautics and Space Administration, United states Air
Force; Washington D.C. p. 227.

Young, A.T. (1980) On the Rayleigh-Scattering Optical Depth of the Atmo-

sphere •. J, App.l. Het. 20: 328.
 3. RADIATION QUANTITIES IN METEOROLOGY

3.1 INTRODUCTION

The radiant energy output of the sun is 3.94-10 26 W, of which 1.8 . 1017
W is intercepted by the disk of the earth. R6ughly 30 per cent of the inter-
cepted energy is scattered back to space, but the remainder is absorbed and
tends to heat the earth-atmosphere system. Since in fact the annual mean
temp~rature of the planet is fairly constant, the absorption of solar energy
must be bal~nced by an equal energy output. The planet has adopted a tempe-
rature and temperature distribution such that its average output of thermal-
infrared radiant-energy equals the solar input.

SPACE incoming Outgoing Radialion

Solar
Radiation Short-wave Long-wave

100 6 21 ,4 6 37 26

ATMOSPHERE

Net Emission by ~
Absorbed Water Vapour, CO 2 Emission
18
by Water by Clouds
Vapour, Dust, O.
Net Gain by Absorption
Absorbed 10 by Water Vapour, CO 2
by Clouds
Latent
Heat Flux
Net Surface Sensible'
Emission of H~at Flux
Long-wave Radiation

OCEAN,LAND 48 16 10. 22

Figure 3.1 Diagram of the annual mean, global average, vertical

energy flows in the atmosphere. The units are in terms of percentage of
mean solar input. (Modified from National Academy of' Sciences, 1975).

Fig. 3.1 gives some detail of what happens to the energy stream on its
way into and out of the earth-atmosphere system. Note that only 48 per cent
of the solar radiation impinging on the top of the atmosphere actually
reaches the bottom, and it is this 48 per cent which is the primary drive for
all processes on or involving the earth's surface. Most of the physical sur-
face processes (evaporation of water, surface heating and so on) involve the
total spectrum of solar radiation which, at the ground, covers the wavelength
band from 0.29 to about 2.5 ,urn. Biological processes, on the other hand,
often depend on the amount of solar radiation in specific wavelength regions
of the total spectrum. For instance, sunburning of skin, DNA breakdown, and
the degradation of paints are all primarily dependent on the shortest-wave-
length (ultra-violet) portion of solar radiation, and photosynthesis involves
mainly the visible part of the spectrum (see Fig. 3.2). The surface values of
solar radiation in different wavelength bands are complicated- by the fact
that absorption and scatter of solar radiation by the atmosphere itself is
highly wavelength-dependent.

21
CHAPTER 3 RADIATION QUANTITIES IN METEOROLOGY

;4
:', n
11
"
Cl) "
I,

0CO ""
C 0. I'
It

0 1\
:
Q. , \\
UiU) : \ 2
Cl) • I \;
~O I
I \
\
./ \
" \
Q)'lit
,
I

> . I I

_.
I
;;0 I I

...
I
(Q.. I I
1
, I
I
G)N \ ; I
\
, f I
o \ .' I
,
.' \ I
\ ' \
" \,
O~ __ ~..,)".. ......._~....-_ ...... ....... ..- ...........--~_
o 300 400 5.00 600 700
Wavelength (nm)
Figure 3.2 Relative "response spectra" to solar radiation for various bio-'
logical processes. Curve 1 = photosynthesis; curve 2 = erythema; curve 3
= product-ion of vitamin D; curve 4 = bactericidal effects.
The infrared radiative output from the surface and from objects on the
surface is also governed by the state of the atmosphere and is also spec-
trally dependent.

The most important variable quantities in the atmosphere which affect

the radiation balance of the surface are the cloud amount and type. Clouds
can refledt back to space upwards of 80 per cent of the solar radiation which
would otherwise reach· the ground, and can reduce effectively to zero the
infrared cooling of the surface.

The solar input generally decreases with increasing latitude, because

the solar beam impinges more directly on a unit horizontal surface at low
latitudes. In terms of the annual average, the solar input exceeds the in-
frared output in the equatorial regions, but is less than the infrared output
towards the poles. This implies that energy in various forms must betrans-
parted via the atmosphere and ocean in apoleward direction to maintain over-
all energy balance, and it is this process which can be regarded as the
fundamental driving mechanism of weather and climate.

3.2 SOLAR. CONSTANT

The flux density (solar irradiance) impinging on a unit area at the top
of the atmoshere directly facing the sun, ~, depends on the sun - earth dis-
tance R and thus varies significantly throughout the year. The solar con-
stant, S, is then defined as So when the earth is at mean distance from the
sun; that is, when Ra = 149.6 X 108 km. The average value of the solar con-
stant determined from balloon, rocket and satellite measurements over the
period 1976 - 1981 was 1367 Wm- 2 with an uncertainty of about ±0.3 per cent
(Frohlich and Brusa, 1982 and World Climate Programme, 1984). Since the· solar
irradiance ~ is proportional to the inverse square of R and the day-to-day
variation of R is known very accurately, the daily value of So can easily be
computed. The following formula gives an empirical determination of So accu-
rate to 0.01 per cent. (Spencer, 1971),

2:2
CHAPTER 3 RADIATION QUANTITIES IN METEOROLOGY

R~
S = So . R2 = so' (1.000110 + 0.034221·cos 8 0
( 3.1 )
+ 0.001280 • sin 80 + 0.000719 • cos 28 0 + 0.000077 . sin 28 0 )

where Ra is the mean sun - earth distance and 8 0 is defined in terms of the
Julian day of the year - that is, of day d counted from 1 to 365 or 366,
starting on January 1st. Thus,

2nd
= 365
(3.2)

S presently ranges from So +3.3 per cent on about January 3rd (perihelion) to
So -3.3 per cent on about July 5th (aphelion).

For most purposes the value of S can be regarded as constant with time,
although satellite measurements have revealed that it varies at the level of
a few tenths of a per cent on time scales of days to weeks (see Fig. 3.3). It
may be that there are similar variations over the period of a solar cycle,
but these have not yet been verified.

0.15
MEAN
0.05
1367.7
-0.05 %
Wm- 2
0.15
,--"-~--'--.L..-..L-"""""""'---'---'---!.-~-'--L..-...J-....L."...,J,~~~~~--,L-J---'--.L..-..L--'----'~....I.....&-O.
2 5
100 200 300 100 200 300 100
1980 DAY 1981 DAY '1982 DAY
Figure 3.3 Solar "Constant" as measured with ACRIM on the Solar Maxim urn
Sa telli te from 1980 to 1983 (Willson, 1984).

3.3 SOLAR POSITION AND RELATIVE AIR HASS

A common problem in radiation meteorology is the calculation of the ze-

nith and azimuth angles of the sun (8 and t/J respectively) at a given time of
day, time of year and locality. The basic equations for calculation of 8 (or
the elevation angle r) and t/J in terms of latitude ~, declina.tion lJ and hour
angle t h :

sin(n/2-8) = sinr = sinlJ'sin~ + coslJ'cos~'coSth (3.3)

sint/J =- coslJ'sinth/cosr (3.4)

Latitude and declination are conventionally positive north of the equa-

tor and negative to the south of it. Declination varies with season. To an
accuracy of 0.05°, delta is obtained for any given day of the year from the
em pirical rela tion (Spencer, 1971):

lJ = 0.006918 - 0.399912 cos 80 + 0.010257 sin 8 0

- 0.006758 cos 28 0 + 0.000907 sin 28 0 , (3.5)

where again 8 0 is in terms of Julian day according to Eq. 3.2. The hour
angle (0 to 2n:) is counted from midday of local apparent or true solar time

23
CHAPTER 3 RADIATION QUANTITIES IN METEOROLOGY

LAT and can be obtained from the clock time of a given place from the rela-
tion:

LAT = clock time + longitude correction +Eq (3.6)

Here clock time is the time referred to any of the standard meridians
about the world; the longitude correction is 4 minutes for every degree of
the given place east (positive) or west (negative) of the standard meridian;
and Eq is the equation of time for which an empirical relation is as follows
(Spencer, 1971):

Eq = 0.000075 + 0.001868 cos 80 - 0.032077 sin 8 0

- 0.014616 cos 28 0 - 0.040849 sin 28 0 (3.7)

For most scattering and absorption processes in the atmosphere which

concern the direct solar beam it is necessary to know the total mass or
'optical path' of atmosphere (or of relevant substance in the atmosphere)
which the beam traverses on its· way down to the level in question - usually
to the ground. Apart from certain minor corrections which might be required
in highly specialized applications, the relation between the vertical optical
path m and the actual slant optical path ms of the solar beam is a
straightforward function only of zenith angle - namely:

ms/m = sec 8 = mR (3.8)

This quantity mR is known as the relative air mass, and ranges from 1.0
for an overhead sun, through 2.0 at a zenith angle of 60·, to infinity with
the sun on the horizon. This simple formula is valid for a plane parallel
atmosphere. Taking into account a spherical earth atmosphere with refraction,
mR never really reaches infinity, but is limited to about 36.5 as given by
Bemporad's formula (see, for instance, List, 1968). If more accurate values
are needed, e.g. for sunphotometry at high air masses, the formula of Kasten
(1966) may be used:

1
(3.9)
sin-8-+-0~15(8+3~885)~~-

with 8 in degrees. Most measurements of optically active constituents in the

atmosphere are in the form of vertical profiles, and it is therefore easy to
correct to the slant profile by multiplication by the relative air mass of
the time.

304 SOLAR INPUT AND SOLAR SPECTRUM

The solar input at the top of the atmosphere is entirely direct radia-
tion so that the instantaneous input per unit area at the top of the atmo-
s p her e is si mply So' cos 8, wher e 8 is t he I 0 c a Iso I a r zen i t h an g I e a t a
particular time. Calculation of the total daily input involves integration of
So . cos 8 over a 24-hour period. Calculations of this sort for all locations
and all times of year lead to the contour plots of daily input as a function
of latifude and time given in Fig. 3.4. It is notable that the maximum daily
total input occurs at the poles in midsummer. This is because there. is 24
hours of daylight per day at the poles during that time. Note also that, at
present, the solar irradiance during the summer is slightly higher at polar
latitudes in the Southern Hemisphere than in the Northern Hemisphere. This is
because the Southern-Hemisphere summer solstice currently occurs at a time
close to perihelion.

24
CHAPTER 3 RADIATION QUANTITIES IN METEOROLOGY

Daily-Total Input Into The Top Of The Atmosphere

N.P

MJ m- 2

EQ.

J F A M J J A s o
Figure 3.4 Contours of daily total input of solar radiation on a horizontal
surface at the top of the atmosphere as a function of latitude and time
of year.

1I'l...- ...,

-
N

E
C
....... 1I'l
(lE ...:
.......
::
-

3BB 4BB 5BB 6BB 7BB 8BB SBB IBBB

Wavelength (nm)
Figure 3.5 Extra-terrestrial solar spectral irradiance

25
CHAPTER 3 RADIATION QUANTITIES IN METEOROLOGY

The spectrum of solar radiation at the top of the atmosphere is given in

Fig. 3.5. Numerical values can be found in Appendix A.1, where also the refe-
rences for the compilation are listed.

3.5 SOLAR RADIATION IN CLEAR SKIES

3.5.1 Direct Beam, Total Irradiance

There are three major mechanisms which modify solar radiation on its way
down through the non-cloudy atmosphere.

Ozone exists at all atmospheric levels, but has its maximum concentra-
tion at about 20 - 25 km. It absorbs solar ultraviolet and visible radiation.
The absorption is concentrated in three wavelength bands, the Hartley and
Huggins bands in the ultra-violet (UV) up to about 0.36 tim, and the Chappuis
band from about 0.42 tim to 0.74 Jim in the visible region of the spectrum.
Within each band the absorption is a strong function of wavelength. The UV
absorption by ozone ensures that no radiation of wavelength shorter than
0.29 tim reaches the ground. The absorptance a of each band (the fraction
absorbed of the total spectrum of solar radiation impinging on the top of the
atmosphere) can be expressed to fair accuracy as an empirical function of the
ozone path, x, (in cm STP) traversed by the solar beam (Lacis and Hanson,
1974). Thus

a 0>
YIS (3.10)

and

1082x 0.0658x
a UY
0>
-- ---------------- + -------------
(1 + 1 38.. 6 x ) 0.8Il5
(3.11)
1 + (103.6xP

Since the ultra-violet and visible bands are spectrally separate, the
total absorptance is simply the sum of the absorptance in both regions. Fur-
ther, it is a fairly good approximation to assume that most of the ozone
absorption takes place in the high atmosphere - that is before scattering
processes have modified the direct nature of the solar irradiance.

Water vapour also absorbs solar radiation, but at wavelengths in the

visible and near infrared which are different from those of ozone absorption.
This separation is convenient for computations, since the individual absor-
ptances of total solar radiation by water vapour and ozone are again addi-
tive. A formula for water-vapour absorptance in terms of water-path y (in cm
of precipitable water) has been derived by Lacis and Hanson (1974):

aWY = 2.9y/[(1 + 141.5y)0.635 + 5.925y] (3.12)

There is a slight complication here, if high accuracy is required, in

that the absorption by water vapour is both pressure- and temperature-sensi-
tive. Various techniques have been proposed· to take this sensitivity into
account by modification of y to an "effective" path, but for normal work such
refinement is probably not required.

The molecules of the atmosphere scatter ("Rayleigh" scatter) radiation

out of the direct beam. Again, the process is wavelength-sensitive. It de-
creases in proportion to the 4th power of the wavelength so that it is effec-
tively zero above about 0.8 tim. Thus, on the one hand, the scatter has no
great effect on water-vapour absorption (which is thereby conveniently con-
fined to direct beam abs0rption). On the other hand, since Rayleigh scatter

26
CHAPTER 3 RADIATION QUANTITIES IN METEOROLOGY

is the main source of diffuse radiation in clear skies, diffuse radiation is

confined primarily to the UV and the visible parts of the spectrum. Again,
when referred to the total spectrum of the impinging direct beam, the scat-
tering coefficient of the total atmosphere is, to a good approximation, a
function only of zenith angle or air mass, and an empirical relation is as
follows (after Robinson, 1966):

OR = O. 01 + O. 0911 . mR - O. 0082 . mR 2 + O. 00032 . mR 3 (3.13)

Thus a first-order expression for the ground-level value of the direct-

beam-flux density (effectively as a function of the solar zenith angle and
the observed vertical paths of ozone and water vapour) ca n be writ ten in
terms of the various relative extinction coefficients above. That is

S = So (1 - aw v - ao Z - OR) . (3.14)

There are, however, a number of minor constituents of the atmosphere

which absorb or scatter the direct solar beam. The most significant are
aerosols, and certainly, on occasion, the extinction of the direct bea m by
dust particles can be of the same order as that by Rayleigh scatter. The
difficulty is that the character of aerosols is highly variable (as are its
absorption and scattering properties) and in any practical situation can
rarely be defined with enough accuracy. (See also the discussion in Section
2.4.2). Thus, as far as extinction of the total-spectrum direct beam is
concerned, any method of incorporating aerosol extinction will have an inac- ~
curacy at least of the order of the errors otherwise incorporated in
Eq. 3.13. However, there were attempts to use the total irradiance to dia-
gnose the atmospheric aerosol content, in other words, to determine the atmo-
spheric turbidity. Linke and Boda (922) were the first to define a turbidity
factor, which was later termed 'Linke's turbidity factor T'. It is defined
as the ratio between the total optical depth of the atmosphere, ~T, along a
path and the optical depth due to molecular scattering only along the same
path, ~R, averaged over the entire solar spectrum (0.3 - 3 gm).

(3.15)

If So is the extraterrestrial solar irradiance, T can be determined from

ground-based measurements of the direct radiation S by

S
-
So
<3.16)

where m = air mass. T is not only determined by the turbidity due to

aerosols, but is also influenced by absorption by ozone and water vapour.
Although some corrections for the latter have been proposed, T is still a
very rough measure of atmosphere turbidity.

3.5.2 Direct Beam, Spectral Irradiance

It is more or less straightforward to calculate the direct-beam inten-

sity when the atmospheric composition is known. The inverse problem, however,
is much more difficult to solve, and from the above it seems extremely diffi-
cult to use a measure of total irradiance for diagnostic purposes. For this
reason measurements in more or less narrow spectral bounds have to be used.
As a guide to the spectral changes introduced by ozone, water vapour and
Rayleigh extinction one can refer to Fig. 3.6.

27
CHAPTER 3 RADIATION QUANTITIES IN METEOROLOGY

Lt')
,
C\I

Extraterrestrial
,..
I
-""'
C\I
,

E
c
N
I

E Lt')
,
;: ~ Elevation

H2 O
(1) 0,
(J ~
CO 2
C
<0
.- - 03
"'0
<0 - O2
-
"-
"-
Lt')

0
,

o,
o 300 500 1000 2500
A (nm)

Figure 3.6 Theoretically-computed spectral distribution of the solar ra-

diation reaching the ground.

If aerosols are added to the atmosphere, the scattered and reflected

part· will increase. From Fig. 3.6 it is clear that for a narrow spectral
band the separation of the different effects may be possible, and the aerosol
optical depth can be used as a straightforward definition of turbidity, For a
reasonably narrow spectral band the spectral distribution of the solar irra-
diance is

-mR [~R 0.) + ~"O.) + ~6 0.) ]

SO.) = So 0.) . e (3.17)

where A is the wavelength, and ~R,"'6(A) the optical depth for Rayleigh
scat tering (R), aerosol extinction (A) and gaseous absorption (G) respective-
ly. The Rayleigh optical depth can be calculated for the total atmosphere
from Eq. 2.11 and can be expressed by the following approximation (Frohlich
and Shaw, 1980 and Young, 1980).

-(3.916 + 0.074'1. + 0.050/1.)

~R (A) = (p/Po) '0.00865'1. (3.18)

with p being the barometric pressure at the observing station, Po =

1013.25 mb and A the wavelength of the radiation in ,um. If the wavelength is
chosen where no gaseous absorption occurs, ~,,(A) can be inferred from S(A).

Wider spectral bands can also be used if the wavelength dependence of

B" is assumed to be known. This was the ba~is for Angstrom's and SchUepp's
turbidity coefficients, (Angstrom, 1927 and SchUepp, 1949), At the time
these methods were develDped only broad-band filter measurements were feasi-

28
CHAPTER 3 RADIATION QUANTITIES IN METEOROLOGY

ble (for the technique of these measurements see 4.2.4.1). Angstrom assumed
the following behaviour of ~A:

~A (A) = /3 • A-1.3 , ( 3 . 19)

which means that /3, the Angstrom coefficient, is the aerosol optical depth at
a wavelength of 1 /-lm and the extinction follows a power law as shown in
Fig. 2.8 in Chapter 2. SchUepp replaced the exponent 1.3 by parameter a and
changed from natural to decadic logarithmS 5. Furthermore, his coefficient is
referenced to the optical depth at 500 nm:

<3.20)

SchUepp's B is only valid for wavelengths below about 850 nm, whereas
Angstrom's coefficient has no wavelength limitations. From Eq. 3.18 and 3.19
it is easy to calculate /3 from B and a and vice versa:

/3 = B·(ln10-0.51.3) = 0.935-B
B = /3/ (l n 10·2-«) = 0.434 -/3 - 2« (3.21)

3. 5 • 3 Solar Aureole

The scatter by dust particles in the atmosphere tends to be "forward

peaked" so that much of the radiation scattered by dust out of the true
direct beam is concentrated at angles close to the sun. This raises a slight
practical problem concerning the definition of the true direct solar beam as
measured by instruments whose aperture is much greater than that required to
view only the disc of the sun. On the other hand, since this circum-solar
radiation or "aureole" is primarily due to dust, the phenomenon potentially
provides a very sensitive measure of atmospheric dust loading.

As an order of magnitude guide for an atmosphere of normal turbidity

(say an Angstrom /3 of 0.05) each degree-wide annulus away from the solar disc
"emits" a total of scattered radiation of the order of 0.1 per cent of the
radiation in the true direct beam. That percentage is very roughly a linear
function of the dust loading as measured by /3. (e.g. see Box and Deepak,
1981).

3.5.4 Diffuse and Global Radiation

Even in dust-free clear skies the diffuse solar radiation Rayleigh-

scattered from the direct beam is not isotropic (see Fig. 3.7) However, since
in clear skies the flux density of downward diffuse radiation D is much less
than that of the direct beam, accuracy limitations normally do not justify
any more complicated assumption than isotropy_ The standard expression for
the flux density of global radiation, Ell, on a "horizontal surface in terms of
the direct and diffuse-flux densities is then

Ell = S cos 8 + Ed (3.22)

The downward diffuse flux density for a dust-free atmosphere over a non-
reflecting ground can be written as:

Ed = (S. -S) cos 8 - So cos PR (3.23)

where PR is given by an empirical relation for the albedo due to Rayleigh

scatter as a function of the zenith angle.

PR = 0.28/(1+6.43 cos 8) (3.24)

29
CHAPTER 3 RADIATION QUANTITIES IN METEOROLOGY

~~~~o~... ~
~ ~

Figure 3.7 Distribution of the sky radiance at 368 nm over the hemisphere
from measurements in Switzerland at an elevation of 400 m a.s.l. norma-
lized to the zenith radiance (Heimo, 1985).

In the real situation, how'ever, one must also take into account that the
direct-beam radiation, which is initially reflected by the ground, is, in
turn, scattered back to the ground by Rayleigh scatter so as to augment the
diffuse flux. Fig. 3.8 demonstrates that the actual total flux reflected up
from or down from such a multiple reflection system is (in either case) the
sum of an infinite series. If Et and Et are the normalized flux densities
upwelling and downwelling from the layer, extension of Fig. 3.8 shows that

Et/E D t = P + (1-p)2 P6 {1 + P6·P + (P6 p)2 + ...... } (3.25)

Et/E D t = (1-p) {1 + P6 P + (P6 p)2 + ....... }
These series can be expressed analytically as:

Et (1-p)2 P6 Et (1 - p)
= P + --------- and = --------- (3.26)
E co (1 - PP6 ) E co (1 - PP6 )

p (1-p)2p (1-p)2p 2p
o o

( 1-p)

·Figure 3.8 Illustration of a typical "double reflection" situation, where

the initial downward radiation is reflected by a layer of albedo P and
also by a surface below the layer of albedo P6.

30
CHAPTER 3 RADIATION QUANTITIES IN METEOROLOGY

In many practical applications where the accuracy of the calculation of

Eg is set (say) by the limi ta tions of the cloud observation, the associated
calculations involving the global flux in clear skies need not be highly
accurate. In such cases it may be that empirical relations giving S, Ed (and
Eg ) as a function of zenith angle only (or equivalent solar elevation angle r
in degrees) are adequate (Paltridge and Platt, 1976). Thus one can write:

S = 1000 [1 exp (-0.06 r)] (3.27)

Ed = 96 [1 exp (-0.05 r)] (3.28)

10 + 1410 . sin r - 310 Jsin r (3.29)

These relations yield the flux densities in Wm- 2 • They are derived from
data of a typical mid-latitude station with average turbidity coefficient f3
equal to about 0.04, and could be in error for an instantaneous individual
value by as much as 20 per cent. The spectral distribution for clear sky
conditions are shown in Fig. 3.9.

Cl)
(J
C
.-ca
"......
2.5
ca

--...ca 2.0

...
(J 15
Cl)
Q.
un
en 10

Cl)

...ca
>

-a::
(1)
500 600 700 om

Wavelength A

Figure 3.9 Spectra of sun and clear sky (E g {-) and of northern sky (E d )
(i.e. diffuse) radiation (each normalized to the irradiance at 550 nm).

3.5.5 Polarization in the Real Atmosphere

The degree of polarization in the sun's vertical at wavelengths of 0.32
and 0.70 pm as measured in a very clear atmosphere from an altitude of 3400 m
on the island of Hawaii is shown in Fig. 3.10. The variation differs from
that for a Rayleigh atmosphere only in detail (c.£. 2.4.3). While the maximum
degree of polarization is lower than that for an ideal model, it is still
about 90· from the sun. As with the curves of Fig. 2.10 in Chapter 2, the
neutral points merge or disappear at the longer wavelength, but are very
distinct at 0.32 pm. Surface reflection and aerosol effects must be inherent
in these data, but they do not appear to be pronounced.

A quite different polarization behavior is shown by the data of

Fig. 3.11 for a wavelength of 0.3.2 pm at eight different sun elevations in
the case of moderate to heavy air pollution in Los Angeles. The buildup of
aerosol loading from moderate values in the early morning to heavy pollution
by noon resulted in a strong decrease of the polarization maximum and a
migration of the Babinet and Brewster points toward the sun with a virtual
disappearance of the negative branch of the polarization curves. Since these

31
CHAPTER 3 RADIATION QUANTITIES IN METEOROLOGY

features are more pronounced here than in cases of less extreme aerosol con-
centrations, they are indicative of the effects produced by aerosols in gene-
ral and demonstrate the diagnostic potential of polarization measurements.
8o,..-------'------------,
70 !-
...-...•
•
~ :
•• ••
•• o.70~m
~ 60-
z
o
•
•
• .............•
....
....
'
i=
< 50 ~ •
!:::! •
a:
<
..J 40~
•
• .. ..
~
&I-
o 30 .... :
'.'.
~
w
w •
a:
<:)
w
•••
•
o
10 - '.•'.
•, 9
..
,.. ..
••
.'
••

.~ of .'
ol--------'---.,---~~.-~*---t
BA/"" ""'BR

I I I I I I I. I I
-10 L-aJ..0--6.LO--4.LO--2..J.O--I.0---:-L20:-----:J40'::-""'""'=6'::'0--:S"=0.....
ZENITH ANGLE le)

Figure 3.10 Degree of polarization as a function of angle in the sun's ver-

tical as measured at two different wavelengths from an altitude of
3400 m on the island of Hawaii (Coulson, Private Communication).

60..--""'T"'--.,.....----,..-----..---r---,.----,..---"'-,---""T"""--t

50

;z40
o
.-
«
~O
0:
«
...I
o
Q.20
I&.
o
~IO
a:
(!)
w
o 0

-10---=8~0----=6'-::-0----J4..."..P-........."2L,,O--0~--:2~O~-4,....,O:----,6:"-:O~-8~0~.....
ZENITH ANGLE (.)

Figure 3.11 Degree of polarization in the sun's vertical at 0.320 J.Lm in a

case of moderate to heavy air pollution. The angles indic.ated are the
'. solar elevation (from Coulson, 1971).

32
CHAPTER 3 RADIATION QUANTITIES IN METEOROLOGY

J.6THERHAL RADIATION IN CLEAR SKIES

A 'black body' for thermal infrared radiation is one which is a iOO-per-

cent absorber or emitter at all wavelengths in the interval 4 tim - 100 ,urn.
The radiative laws pertaining to such bodies are well ]{nown. In particular
in the present context, their total infrared emission E is given by the
Stefan - Boltzmann relation

(3.30)

where a is the Stefan - Boltzmann constant (5.67 . iO- 8 Wm- 2 K-4)· and T is
the .absolute temperature (K). The actual distribution of this radiation as a
function of wavelength is also a function ·0£ temperature (see Fig. 3.12), If
a body absorbs only a fraction f; of thermal radiation (and f; being the same
at all wavelengths), it is said to be a grey body of emittance f;, and for
such cases E = f;aT 4 •

The earth's surface is generally assumed to be a black body and there-

fore to emit an upward flux density of IR radiation appropriate to the sur-
face temperature according to Eq. 3.29. Nevertheless there are many occasions
when this is not true. The emittance of different substances is discussed in
Section 3.8.4. Desert surfaces may have a value of f; as low as 0.7, and even
water has an emittance closer to 0.95 than to 1.0. The atmosphere also emits
infrared radiation, but is certainly not a black body. Water vapour and car-
bon dioxide are the prime infrared emitting gases of the atmosphere, and this
emission is in specific wavelength bands. There are parts of the spectrum
(e.g. the so-called atmospheric window from 8 to 14 tim) in which there is

FREQUENCY (Hz)
15 1
,-.. J 10 , 10'\ 4 10 ' • 4
, ,(--;--"rj;'1Irr-I'I-1--.--;"--",,;,.",,-,,'-:;-,--.-,---:r--""T-'" " .-,(--"
E 10·
::L
, '0
,
6000b..L
~OOo':::'J
-----
I 1 ,

I! I
~ II
4000
~
(/) V/ ,/
,/ -~ooo
• 10
N
I

E '0• / I
/
V
/
,
20~'- ~ ~
--..... , ._-_.- i, l!
~
•
1/ /
1
I1
/
11
~~~ K- ~ ~~ -, ...
I
UJ
U '0
JI / / 1/ I
!/
/'
IOO<L::".
800~ ---
,

- ~
~
Z
<C • / / I
1
1// ' 60~
~
~"~
...... " ~
-
Cl
<C
Cl:
I()

10
I , 1/
fll
I
1
I
~--
I / ,
400

27
~ ~~
- ~ ............. + I

I, ~" " '::-..:: ~~

1 /

2~---: -......::.
I
I /
..J
<C .0
I 1 1/ I 1/
~ ,,~~ '.
"
0:
I- ,I IO- ,
I
I 1/ 1 / 1
/ / ,
"" ' '" , ,
'
U ,1
w-
0.1()
,I 1 I
I
I I 1/ I
/ I
I

"'"
C/)
, I I
I
1/
/
I

I
1
""" "-
<
.6
{
I
I
I
I

/ I
I
I "
...J 16"
01
o. 0.4 0.1 0.' 1.0 6 • 10 '0
WAVELENGTH (lJm)
I" , , I " I I I I I I! I I I

~4
•• ~J
•
WAVENUMBER (cm-')

Figure 3.12 Spectral radiance of a black body, at the temperature in K shown

on each curve; beyond 100tim the curves continue to be linear (plotted on
logari th mic scales). The diagonal line, intersecting the curves at their
maxima, shows Wien's displacement law. (Subdivisions of the ordinate
scale are at 2 and 5).

33
CHAPTER 3 RADIATION QUANTITIES IN METEOROLOGY.

very little emission of thermal infrared by the gases of the atmosphere and,
by Kirchhoff's law, in which there is little absorption. In the case of clear
skies more of the up welling IR flux from the ground in this spectral region
can be radiated directly out to space.

To a first approximation, the downcoming IR flux density from the clear

atmosphere Eolr is of the order of 0.7 of the upwelling black-body radiation
from the surface. In other words the net loss of IR energy fro·m the surface
under clear skies is about 0.3 of its black-body radiation, and most of this
loss is via the atmospheric window. Under atmospheres containing little
water vapour (e.g. at a very high altitude or in very dry desert regions) the
net loss would be larger.

Slightly more refined empirical formulae have been derived which to

reasonable accuracy give Eolr as a function of surface temperature and humidi-
ty, or even of surface temperature alone. This latter assumption has been
justified by anaJrgument to the effect that the emission bands of water
vapour in the atmosphere aJre close to being "saturated", so that changes in
watel.~ vapour concentration do not make much difference to the downcoming
flux. One such formula is:

Eolr· = 53 .. 1 . 10- 1 3 • T6 (3.31)

w.here Eolr is in Wm- 2 and, T in K fSwinbank, 196·3).

Accurate numerical computations of the downward flux density at the

surface are possible, given information on the detailed vertical profiles of
t.emperatureand. wateJr vapour in the atmosphere.

3,.7 CLOUDY SKIES

3.7.1 Effect on Solar Radiation

On the one hand, clouds scatter much solar radiation directly back to
space and thereby reduce the total solar input to the ground. There can be
rare and very specific exceptions to this - as, for instance, when two adja-
cent clouds funnel solar radiation downwards. In terms of reasonable ave-
rages, however, it seems that the 55 per cent cloud cover of the world re-
duces the solar energy absorbed by the planet by about 20 per cent. On the
other hand, the scattering of the direct solar beam by clouds is normally so
great as to ensure that for cloudy skies the total global input of radiation
to the ground (E.. ) is entirely made up of diffuse radiation Ed •

Detailed calculations show that the albedo of clouds depends on the

zenith angle () of the sun, on the distribution of liquid water content within
the cloud, and on the actual size distribution of the drops or ice particles.
The significance of size distribution is still somewhat obscure, and it has
been suggested that, for most· practical applications concerned with horizon-
tally uniform clouds, the cloud albedo can be expressed as a function only of
the zenith angle and the total liquid-water (or ice) path in the vertical
column of the cloud. Fig. 3.13 gives one such parameterization as was derived
for water clouds (i.e. spherical particles), but appearing to be not too far
in error for cirrus ice-clouds either.

Even this simplification is not sufficient for most practical applica-

tions, since the liquid-water path will not be known. The state-of-the-art at
this time is perhaps to use a table of albedos corresponding to a given cloud
type (e.g. Table 3.1) and ignore such factors as 'the variation of albedo with
zenith angle. It should be noted, however, that the visual cloud albedo va-
ries significantly not only with cloud type, but vlithin each cloud type. The
values given in Table 3..1 therefore represent average values for each of the

34
CHAPTER 3 RADIATION QUANTITIES IN METEOROLOGY

major cloud types. The mathematics of the calculation of the albedo of "bro-
ken" clouds is enormously difficult and only now being attempted. No simple
and practically useful rules have yet been devised.
. 10000 r----,----y-~--r----,r----o,...,._____,..-T'"""1
Cloud
N
I
E
Ci1
1000

L-
W
13
3

20 40 60 80
Solar Zenith Angle (degrees)

Figure 3.13 . Theoretical (and approximate) values of cloud albedo as a func-

tion of the" liquid water in the vertical column of the cloud and of
solar zenith angle. The values are for the cloud by' itself - Le. based
on the assumption of zero reflection by atmosphere and ground below the
cloud. (Paltridge, Private Communication)

The change to completely diffuse radiation is an obvious consequence of

cloud cover. Generally, the spectrum of the downward diffuse radiation in
cloudy skies is not greatly different in shape from that of the total global
radiation in clear skies. It is, however, vastly different from the equiva-
lent· clear-sky diffuse-flux spectrum. (For clear-:sky- conditions see
Fig. 3.9). It may be, therefore, that the cloud-induced change in the spec-
trum of the diffuse radiation is relevant in certain applications concerned
with the. penetration of photosynthetic radiation in plant canopies.

Table 3.1 The average short-wave albedo p and long-wave ~mittance £ as-
sumed to be associated with the major cloud types. These ·values show
relatively good agreement between observations and th.eoretical models.
They are appropriate for average sun height· and cloud thickness.

Cloud Type p £

Cirrus 20 0.3 - 1.0

Altostratus 50 0.9 - 1.0
Cumulus 70 1.0 .

stratus 60 1.0

Cumulonimbus 70 1.0

Nimbostratus 70 1.0

35
CHAPTER 3 RADIATION QUANTITIES IN METEOROLOGY

Clouds also absorb solar radiation, since liquid water and ice have ab-
sorption bands in the near infrared. However, these bands to a large extent
overlap those of water vapour, so the change in absorption by the whole
atmosphere is not very great. The main effect is to concentrate the absorp-
tion in the upper levels of the cloud (primarily because scatter in the
cloud increases the effective optical paths of water vapour).

3.7.2 Effect on Thermal Radiation

The simplest concept of the effect of clouds on the downward IR flux

density is of "filling in·' of the atmospheric window. Most clouds radiate as
black-bodies at their ~urfacetemperature~ and. the downward black-body ra-
diation (or that part' of it from 8-14 tim) from the cloud base reaches the
ground. Thus one can" imagine that if the cloud base is very close to the
ground (and its temperature is very close to ground temperature) the net flux
density at the surface is close to zero. For a high cloud, since its radia-
tive temperature is very low, the net flux density at the surface will be
closer to that of a clear sky. .

This very simple concept can then be expressed in the follow.ing formula
for downcoming IR flux density under cloudy skies:

(3.32)

where Et e s is the downcoming flux density in clear skies (computed say by

Eq. 3.28 or its equivalent), Te is the estimated (or measured) cloud-base
temperature, and ge is the estimated cloud emittance. Estimating ge is
reasonably straightforward, since most clouds are optically thick in the IR
and hence have an emittance close to 1.0 (see Table 3.1).

As far as cloudy skies are concerned, the empirical relationship given·

in 3.3.1 for the computation of Et at the ground is reasonably adequate for
most practical purposes. More detailed methods, however, are generally used
for specific meteorological problems (e.g. weather and climate predictabili-
ty). .

3.8 SURFACE PROPERTIES

3.8..1 Albedo

The short-wave reflectance or albedo of any given surface is normally a

function of both wavelength and zenith angle, and these functions are in turn
strongly dependent on surface character. See, for instance, the examples
shown in Figs. 3.14 for the solar elevation and 3.15 for the wavelength de-
pendence. Most available measurements of albedo are of the reflectance to the
total solar spectrum - and it is often the case that the quoted values do not
specify the solar zenith angle at which they were obtained. Indeed, on many
occasions it is even not obvious whether the measurements were taken under
clear or cloudy skies, and, therefore, whether the albedo is relevant to
direct or diffuse solar radiation.

The main difficulty is that, except over the ocean and over certain
highly u'niform land surfaces, such as extensive snow cover on flat country,
the variation of surface character leads to highly variable albedo. Even if
specific spot measurements are a:vailable, it is rare that these measurements
will be generally useful for any purpose other than that for which they were
taken.

To a certain extent this is true even for the ocean. Oceanic surface
albedo depends heavily on wind speed among other things, and is presumably

36
CHAPTER 3 RADIATION QUANTITIES IN METEOROLOGY

also a function of the extent of organic and inorganic 'contaminants' of the

sea.
\'0
09
0'8
0'7 -------r-- Snow (fresh)

00'6 - White sand

-0
~ 0·5 +-- Range for
snow with age
c:r0-4
O' 3 Desert sand
02 Dry grassland-semi desert
O' I Eucalypt forest
L--L-L--l-=L:::c::t:::r=C~Fre sneI (s m 00 t h sea)
'0 30 60 90
Solar elevation (deg,)

Figure 3.14 A schematic example of surface albedo as function of solar ele-

vation.
90

80 .-.-.-.-.- .........
--- --
/'-

,,'-.
".

/'
~ --..... '.
"
70

60 "-

p%
50

40
-·-1
--2
---3
l'"
.. -.-
'""".

-"-4

:: ~ ------~-
--5
30 ----6
---7
.-----
;: / ~~r-:" - - - --
··········8 J 'I
---9
20 - - - --10

10

.300 400 500 600 700 800

A(nm)
Figure 3.15 Spectral reflectance of different natural underlying surfaces at
various solar elevations r. 1 = snow with ice crust. r = 38°; 2 = large-
.g r a i n e d wet s now, r = 3 r; 3 = wate r sur f ace 0 f a 1a k e , r = 56 0; 4 =
soil after thawed snow, r = 24°30'; 5 = silage corn, r = 54°; 6 = tall
green corn, r = 56°, 7 = yellow corn, r = 46°; 8 = Sudan grass, r = 52°;
9 = chernozem, r = 40°; 10 = stubble of cereals, r = 35°. (Paltridge and
Platt, 1976).

37
CHAPTER 3 RADIATION QUANTITIES IN METEOROLOGY

Given the albedo p(e) of a surface as a function of zenith angle 8, it

is possible to calculate a fairly, accurate albedo PD to the downcoming dif-
fuse radiation which exists (say) under completely cloudy skies. That is

11.12
PD - 2 J p(8) cos8 sin 8 d8 . (3.33)
o

The spectral dependence of the reflectance of different surfaces is

highly variable as, shown in Fig. 3.15. In general, snowsur£aces have peak
reflectances in the visible, but vegetated surface reflectances increase
sharply toward the near infrared. '

3.8.2 Typical Albedo Values

An extensive series of measurements by j{rinov (1947) established the

general reflection characteristics of natural surfaces, particularly those of
geologic formations and certain types of plants. Later measurements have been
more detailed in some respects, but that pioneering work is still a valuable
source for this type of information.

3.8.2.1, Snow an9. Ice

Most of the available information on the albedo of snow is the result of

many investigations in the Antarctic. An annual average of 84 per cent for
the Antarctic plateau was obtained by Kuhn et a1. (1977), and this appears to
be typical for other locations as welL Snow albedo' varies strongly, how-
ever, with conditions of illumination and age of snow. Data (which have been
smoothed somewhat) from various authors on the change of snow albedo with sun
elevation under different amou'nts of cloudiness 'are shown in Pig. 3.16.
100
0 __

,,
,,
--- --- --~oULTRAVIOLET
,
,, (FITCH)
,,
........ ',,/N<0.7} CAR ROLL AND FITCH"
......... ,N::::O.7
..... .X
90 ~ ",
\. .......... "
. \. .... ........ '"
\.. ..... ,
o \.. ANNUA .........." ...
o
UI
" " , AVERAGE ,..-~,-=.-:.-:.-:.-=.
ID
...J
et
" ..... ..... .....
80 ............
.......
.......
.......
.....
......
............ KORFF ET AL

10 20 30 40
SUN ELEVATION (0)
Figure 3.16 The albedo of snow as a function of sun elevation. The curve for
Kuhn et a1. (1977) represents' the average of four Antarctic stations,
that of Korff et a1. (1974) is for a snow field in the Rocky Mountains,
and the data of Carroll and Pitch (1981) and of Fitch (1976) are for the
South Pole.

38
CHAPTER 3 RADIATION QUANTITIES IN METEOROLOGY

Although the absolute values of snow albedo show some variability, the
change with sun elevation is quite similar at the different locations. From
an extensive ,series of measurements at the South Pole, Carroll and Fitch
(1981) developed the dependencies of albedo on cloudiness N and sun elevation
r given by:
peN < 0.70) = 0.98 - 0.007~j (3.34)
for cloud cover N < 0.70, and

peN > 0.70) = 0.93 - 0.0050r (3.35)

for near ly overcast conditions (see also Fig. 3.16).

Contrary to the situation for most other types of surfaces, the albedo
of snow is higher in the ultraviolet than at visible and near-infrared wave-
lengths. The data of Fitch (976) in Fig. 3.16 show a difference of 10 to 11
per cent between the albedo for the region 0.30 - 0.385 tLm and that for total
solar radiation. Spectral measurements of Kuhn and Siogas (978), summarized
in Table 3.2, show that surfaces composed of frozen water in some form have a
general decrease of albedo from visible to the near infrared. One exception
to this trend is a secondary maximum at about 1.4 tLm, which the authors
ascribe to inclusions of bubbles and snow inclusions in sea ice and to den-
dritic forms of crystals in fresh snow.

Table 3.2 Spectral albedo of surfaces composed of frozen water in some

form, as summarized from Kuhn and Siogas (1978), The data are for
freshly fallen snow, sea ice containing air bubbles and snow crystals,
and ice on a fresh-water lake which was free of such inclusions.

Albedo (%)

Wavelengths (tLm) Fresh snow Sea ice Fresh-water ice

0.4 88. 47 22
0.5 88 47 19
0.6 88 40 17
0.7 87 30 10
0.8 87 20 6
0.9 86 16 5
1,0 77 11 4
1,1 74 11 5
1,2 61 9 4
1,3 55 8 5
1,4 66 15 -
1.5 25 4 -
1, 55 12 4 -

3.8.2.2 Vegetated Surfaces

The albedo of vegetated surfaces is extremely variable, depending on type

of vegetation, stage of growth, altitude of the sun, and other variables.
Typical ranges of values, as obtained by many different authors, are shown in
Table 3.3.

Of perhaps more interest than total albedo is the spectral distribution

of the reflectance of green vegetation. Although there are large differences
in this respect among different types of plants, and even within a given type
at different stages of growth, chlorophyl and other plant pigments reflect

39
CHAPTER 3 RADIATION QUANTITIES IN METEOROLOGY

ultraviolet, blue, and red wavelengths only weakly, making for very lO~o1 re-
flectance in these regions. These features are shown in the corresponding
reflectance spectrum given in Fig. 3.15. A slight increased reflectance in
the green, combined with the spectral distribution of sunlight and the spec-
tral sensitivity of the eye, produce the green colour of most growing plants.

Table 3.3 Typical albedo values of different types of green vegetation.

Type Albedo (% ) Remarks

Coniferous forest (no snow) 3-10

Coniferous forest (with snow) 10-25
Decidious forest (no snow) 15-20
Short grass turf, lush 14-33
Wheat and rye 14-27 Sharp drop on ripening
Alfalfa (lucerne) 23-27 Lush, ground entirely covered
Broad-leaf, vegetables 15-27 Highest for horizontal leaves

The near-infrared wavelengths are absorbed very little by most plant pig-
ments, a fact which permits this type of radiation to enter the main cellular
structure of plant tissues, where large differences of refractive index bet-
ween cells and cell walls are effective in producing a high reflectance at
infrared wavelengths. High values persist throughout the near infrared ex-
cept in the main absorption bands of liquid water. These bands are' strong
enough to cause a significant reduction in total reflectance of the plants,
as well as within the ba nds them selves. In addition to those shown in
Fig. 3.15, mean values of the reflectance of various agricultural crops at
visible and near-infrared wavelengths, as measured by Kondratyev et al.
(1964), are given in T'able 3.4.

Table 3.4 Reflectance of various types of agricultural crops at selected

wavelengths (Kondratyev et, al. , 1%4).
.
, '
Wavelerigth (pm)

Plant type .55 .:60 .65 .67 .70 .71 .75 .80 .85 .90
, ;

Lucerne (June). lupine 7.4 5.8 4.8 4.6 10.8 20.0 40 43 45 45

Lucerne (June). lupine : 7.4 5.1 4.4 4.3 10.9 19.6 ,40 40 43 42
;Sudan grass 6.0 4.8 3.5 3.4 5.8 11.2 43 51 52 50
Sudan grass , 6.0 4.7 3.4 3.2 5.8 10.7 ,43 51 51 ' 50
Lucerne (July) 5.5 4.9 4.8 4.7 8.5 13.5 27 32 34 34
Clover and maize for
silage, beet 3.9 5.0 4.5 4.8 8.6 12.7 '25 30 30 32
Sunflower and high maize 4.7 7.3 7.6, 7.7 7.2 10.4 19 26 28 28

The reflectance of agricultural crops during the growing-season gene-

rally decreases with increasing sun elevation, although the reverse appears
to apply at sun elevations of less than about 10·. For instance, the mea-
surements of Coulson and Reynolds (1971) show a relatively sharp increase of
reflectance as the sun rises to tOo. after which a gradual decrease pe'rsists
to at least 70· sun elevation. The magnitude of the change varies with
wavelength, being as much as 22 to 27 per cent at ;\=0.795 J1m for crops, such
as alfalfa, sugar beets, and bluegrass. Much smaller changes of only 5 to 7
per cent occur in green and red wavelengths, and still smaller in the blue
and ultraviolet, bu,t of course these changes are superimposed on basically
low values at the shorter wavelengths.

40
CHAPTER 3 RADIATION QUANTITIES IN METEOROLOGY

3.8.2.3 Mineral Surfaces

The albeqo of soils, sands, rocks, and other mineral-type surfaces has a
wide variation from as low as 2 to 3 per cent for wet black soil to as much
as 70 to 80 per cent for gypsum or quartz sand, certain clays, limestone, and
other light-coloured materials. More usual values for bare soils in farming
areas are 15 to 30 per cent when dry and 8 to 15 per cent when wet. Deserts
in general are higly reflective, as shown by the measurements of Otterman and
Fr a se r (1 976) l i s t e din Tab I e 3 . 5.

Table 3.5 Albedo of various deserts as deduced from satellite measurements by

Otterman and Fraser (1976).

Area Albedo Area Albedo

Sinai, Fall 0.39 Thar Desert 0.38

Sinai, Winter 0.41 West Coast of Africa 0.44
Sahel, overgrazed 0.41 Afghanistan 0.34
Sinai, near sea 0.38 Northern Mexico 0.34

The spectral distribution of the reflectance for most mineral-type sur-

faces shows a gradual increase of reflectance with increasing wavelength from
the ultraviolet through the visible and into the near infrared. Broad ab-
sorption bands due to water in many rocks and minerals cause decreases of
reflectance at about 0.85, 1.0, 1.4, and 1.9 ,um. The most important soil
property in determining its reflectance, in addition to mineral composition,
is particle size, the reflectance increasing strongly with decreasing size of
the particles.

Table 3.6 Spectral reflectance (per cent) of sands and soils (Data from
. Condit, 1970).

Wavelength (#m)

Type state 0.31 0.40 0.50 0.60 0.70 0.80 0.90 1.00

Sand of quartz and dry 12 15 20 32 38 41 42 44

carbonate (Florida) Yet 6 £, 11 21 25 28 29 30

Gypsum sand (White dry 25 43 60 70 72 74 75 71

Sands Nat. Monument) Yet 19 31 45 57 59 60 60 52

Clay Missouri dry 10 20 30 37 42 47 51 56

Yet 8 17 24 30 34 39 42 45

Chernozem soil dry 3 4 6 10 13 17 20 25

(Nebraska) Yet 2 2 3 4 £, 9 11 14

Chernozem soil dry 4 6 . 10 14 19 24 30 35

(Oklahoma) Yet 2 2 3 5 8 10 13 16

Pedocal soil dry 10 23 32 39 41 42 42 42

(Nebraska) Yet 5. 17 2,6 33 37 38 38 38

Pedialier silt dry 9 11 20 30 36 41 43 46

(Arkansas) Yet 3 5 9 13 17. 20 21 24

Pedal£er soil dry 8 10 20 42 51 SS SS 57

(Missouri ) Yet 3 4 8 22 30 31 32 36

Quartz sand dry 5 17 26 40 47 50 51 54

(Oregon) Yet 10 6 10 20 24 27 28 28

41
CHAP.TER 3 RADIATION QUANTITIES IN METEOROLOGY

Very high reflectances are exhibited -by the small particles of many
clays and by the naturally light-coloured sands of quartz and gypsum, whereas
the dark-coloured soils prevalent in the midwestern part of the United
States, southern parts of the Soviet Union, and other rich agricultural areas
generally show low reflectances. Typical spectral reflectance values for
some of the different types of soil in the region 0.31 to 1.0 J.lm as measured
by Condi t (1970), are listed in Table 3.6.

3.8.2.4 Sea Surface

The albedo of a perfectly smooth water surface can be computed from

Fresnel's laws for known index of refraction and distribution of incident
radiation. Surprisingly, even for the simple case of completely diffuse ra-
diation and a still surface, computations by various authors yield albedo
values varying from 5.6 to 17.0 per cent.

Typical values of sea-surface albedo as measured by various authors are

shown in Table 3.7. It is clear from these data that the theoretical expecta-
tions of a strong dependence of albedo on sun elevation, most pronounced at
low sun elevations, are borne out by the measurements. None, however,. has
approximated the complete reflection that must occur at grazing incidence,
since at very low sun elevations most of the incident radiation is composed
of diffuse skylight and diffuse radiation yields a low albedo.

The evidence on the effect of wind speed on sea-surface albedo is con-

flicting. Whereas Burt (1954) finds increased roughness of the surface in-
creases the albedo at high sun elevations and decreases it at low sun eleva-
tions, Neiburger (1948) found no correlation with wind speed and Payne (1972)
found no noticeable effect in wind speeds of up to 30 knots.

Table 3.7 Sea-surface albedo from data of various authors.

Sun elevation Albed.o Author Remarks

35.6° , 7.9 Willis (1971 ) Clear sky

30.2' 14.3 " "

high 3.0 " "

low 45.0 Payne (1972 ) " "
- 6.1 Overcast sky

- 10.5 Neiburger (1948) " "

- 5.1 Middleton (1952) " ..

- 6.6 Burt (1954 ) Diffuse, smooth surface

- 5.5-5.8 " , rough "
47" 4 Ocean, near shore
20° 14 " ..
"
List (1951)
5.5° 46 " " ..
Variable 3-10 Inland waters

42
CHAPTER 3 RADIATION QUANTITIES IN METEOROLOGY

3.8.3 Polarization by Surface Reflection

The polarization of light reflected from natural surfaces has been stu-
died much less extensively than has that scattered by the atmosphere, and the
theory of reflection from rough surfaces is not as well developed as is the
theory of scattering. Polarization data are available for a few surfaces,
however, from measurements taken in laboratory conditions with artificial
light. The degree of polarization varies with type of surface, angle of inci-
dence, angle of reflection, azimuth angle from the position of the source,
and wavelength of radiation. In addition, the degree of polarization strong-
ly depends on the particle characteristics and roughness of the surface, and
a given surface polarizes reflected light much more strongly when it is wet
than when dry. A highly reflecting surface generally polarizes light less
strongly than does dark-coloured material.

Figure 3.17 Distribution of the degree of polarization over the hemisphere

of light reflected from desert sand (dashed isoplots) and black loam
(solid isoplots) at 492 nm and an angle of incidence of 53° (after
Coulson, 1966).

Polarization values as measured by Coulson (1966) and the distribution

of the degree of polarization throughout the reflection hemisphere is shown
in Fig. 3.17 for surfaces of desert sand and black loam soil at a wavelength
of 0.492 j.lm and an angle of incidence of 53°.

3.8.4 Emittance

The infrared emittance of any surface is also variable with surface

character, though less obviously than for albedo. Furthermore, it is also
variable with wavelength. The variation with wavelength is usually not of
great concern since the emittance measurements are often obtained for the
window region (8-14j.lm), ·where any variation of emittance is most significant
to the total IR-cooling of a surface. (Nevertheless, i f a quotedemittance is
to be used, it should be checked as to the conditions under which the mea-
surements were taken).

Since emittance is normally determined by combining measurements of

hemisphere reflectance with Kirchhoff's law, most of the values available are
for reflectance p. The emittance (; is easily determined by the relation
(; = 1- p. Some typical reflectances of natural materials are given below.

3.8.4.1 Mineral Surfaces

The reflectance of mineral surfaces is strongly wavelength-dependent in
the infrared, the dependance arising from absorption bands of water, hydrated
water, reststrahlen features, and various pigments in the minerals. Varying

43
CHAPTER 3 RADIATION QUANTITIES IN METEOROLOGY

particle size may decrease spectral contrast and even introduce some absorp-
tion features of its own. The infrared reflectance of some typical minerals,
as determined by Hovis (1966), is shown in Fig. 3.18. The reflectance of
Oregon beach sand is particularly variable, but there is better regularity in
the features for the purer minerals. For instance, the curves for serpentine
and basalt show low· and relatively constant values at A < 8 jlm, whereas
those for other minerals have considerable variations in this region.

The one feature prevalent for all samples is the reststrahlen enhance-
ment in the 9-11-jlm region, and most samples have a similar reststrahlen
feature at about 20 jlm. There is evidence that the reststrahlen peak shifts
toward shorter wavelengths with increasing acidity of the mineral.
r.
/ i
I
,, - - SERPENTINE
I
60 I ,
I - - - BASALT
- - - DUNITE
\,
I _ ••- BEACH SAND
'\ (, ..- ...., GRANITE
I , ,1\ ,
,"\I
. ;. i I \ f: I '
<-!V\ \} \
I \

'\' ,
j : H\ ~
,,
.1\ \ r, •

i
I

I \.
{\I \ IV:1
~i I{\~i(',
\i I i \ , :'
\.; i ,/f'\ I I i i\
• •/ ., I I "
'.~: ~
\::
- "~
I 11

Figure 3.18 Infrared minerals.

3.8.4.2 Vegetated Surfaces

Very few data on the reflectance of plants in the wavelength interval of

thermal infrared are available, although it is known that the high values
which occur in the 0.8 to 2.5 jlm are not repea ted a t longer wa velengths.
Representative values of reflectance of several common species, as measured
by Ga tes and Tan traporn (1952) a t selected wa veleng th s, are given in
Table 3.8.

Table 3.8 Reflectance of several species of plants at selected infrared wave-

lengths, as measured by Gates and Tantraporn (1952).
Wavelength (~JlI)

Conllllon name Botanical nalle 3.0 5.0 7.5 10.0 15.0

Acer saccharinum 3.0 4.0 4.5 5.5 5.0

Maple 2.2
Oak Quercus robur 2.0 2.0 2.0 2.0
Ulmus americanus 2.2 2.4 2.4 2.4 2.5
Elm 7.5 7.0
Cottonwood Populus deltoides 5.0 6.0 7.0
Aspen Populus tremuloides 0.2 0.2 1,4 2.3 3.5
Salix babylonica 4.0 4.7 5.3 7.0 6.0
Willow 10.0
Lemon Citrus limonica 6.0 8.2 14.0 17.0
Musa paradisiaca 2.5 2.5 4.2 5.5 9.0
tlanana 5.5
Rubber tree Ficus elastica 4.0 5.0 5.0 6.0
Grass Sestuca elatior 1.0 1,0 1,5 2.0 2.0
Water lily Nymphae 1 :0 3.0 4.0 7.0 8.0
Sumac Rhus glabra' 1,0 1,5 1,2 2.0 ! 2.2
Prickly pear cactus Opuntia 0.5 1,0 1,4 1,4 1.4

44
CHAPTER 3 RADIATION QUANTITIES IN METEOROLOGY

REFERENCES

Angstrom, A. (1929) On the Atmospheric Transmission of Sun Radiation on Dust

in Air. . Geoqr. Annual. 2: 156.

Box, M.A. and A. Deepak (1981) An Approximation to Multiple Scattering in the

Earth's At mosphere: AIm ucantar Radiance Form ula tion. J. Atmospheric Sci.
38: 1037.

Burt, W. V. (1954) Albedo of wind-roughened water. J. Het. 11: 283.

Carroll, J.J. and B.W. Fitch (1981) Effects of solar elevation and cloudiness
on snow albedo at the South Pole. J. Geophys. Res. 86: 5271.

Condit, H.R. (1970) The spectral reflectance of American soils. Photoqram-

metric £nqineerinq 36: 9. Falls Church, Virginia.

Coulson, K.L. (1966) Effects of reflection properties of natural surfaces in

aerial reconnaissance. App.l. opt. 5: 905.

Coulson, K.L. (1971) On the Solar Radiation Field in a Polluted Atmosphere.

J. fllh'int. Spectrose. Rad.iat. Transfer. 2: 739. Pergamon Press.

Coulson, K.L. (1974) The polarization of light in the environment, p. 444, in

Gehrels, T. (Ed.) Planets Stars and Nebulae Studied with Photopolari-
metry, University of Arizona Press, Tucson, Arizona.

Coulson, K.L. and D.W. Reynolds (1971) The spectral reflectance of natural
surfaces. J. Appl. Het. 10: 1285.

Fitch, B.W. (1976) Albedo of the snow-covered Antarctic Plateau, M.S. Thesis,
Dept. of Land, Air, and Water Resources, Univ. of California, Davis, 59
pages.

Frohlich, C. and R.W. Brusa (1982) Solar Radiation and its Variations in
Time. Solar Physics 74: 209.

Frohlich, C. and G.E. Shaw (1980) New Determination of Rayleigh scattering in

the terrestrial atmosphere. Appl. opt. 19: 1773.

Gates, D.M. and W. Tantraporn (1952) The reflectivity of deciduous trees and
herbaceous plants in the infrared to 25 microns. Sc.ience 115: 613.

Heimo, A. (1985) Diagnosis of the atmosphere through optical measurements

performed with a mobile station. Diss. ETH, in press.

Hovis, W.A. and W.R. Gallahan (1966) Infrared reflectance spectra of igneous
rocks, tuffs, and red sandstone from 0.5 to 22.um. J. opt. Soc. Am. 56:
639.

Kasten, F. (1966) A new table and approximation formula for the relative air
mass. Archiv. Het. Geoph, Biokl. 206: B14.

Kondratyev, K.Ya., Z.F. Mironova and A.N. otto (1964) Spectral albedo of
natural surfaces, PAG£OPH 59: 207.

Korff, H.C., J.J. Gailiun and T.H. Vonder Haar (1974) Radiation Measurements
over a Snow Field at an Elevated Site, Pap.221, Dept. of Atm. ScL Colo.
State Univ., Fort Collins, (NTIS N74-3187813GD.

Krinov, E.L. (1947) Spectral reflectance properties of natural formations,

Laboratory of Aeromethods, Academy of Sciences, Moscow, USSR (Transla:-

45
CHAPTER 3 RADIATION QUANTITIES IN METEOROLOGY

tion to English, G. Belkav, Tech. Trans. TT-439, Nat. Res. Council of

Canada, Ottawa, 1953).

Kuhn, M., L.S. Ku ndla and L. A. Stroschei n (1977) The radia tion budget at
Plateau Station, Antarctica, 1966-1967, Antarctic Research Series, Am.
Ceop.!Jys. t/nJon 25: 41.

Kuhn, M. and L. Siogas (1978) Spectroscopic studies at McMurdo, South Pole

and Siple stations during the austral summer 1977-78. AntarctJc Journal
13: 178.

Lacis, A.A. and J.E. Hanson (1974) A parameterization for the absorption of
solar radiation in the earth's atmosphere. J. Atmos. Sc.l~ 31: 118.

LinkeF. and K. Boda (1922) Vorschlage zur Berechnung des TrUbungsgrades der
Atmosphare aus den Messungen der Intensitat der Sonnenstrahlung. Heteor.
ZeJtschr. 39: 161.

Li st, R.J. (1968) Smi thsonia n Met'eorological Tables, Smi thsonia n Institution,
Washington, D.C., 6th Revised Edition, p. 442.

Middleton, W.E.K. (1952) Note on the reflection of diffuse daylight from

still water. fluart. J. Royal Het. Soc. 78: 627.

National Academy of Sciences (1975) Understanding climatic change - A program

for action, Washington D.C.

Neiburger, M. (1948) The reflection of diffuse radiation by the sea surface,

Trans. Am. Geophys. Union 29: 647.

Otterman, J. and R.S. Fraser (1976) Earth-Atmosphere system and surface re-
flectivities in arid regions from LANDSAT MSS data. Remote Sens. of Env.
5: 247.

Paltridge, G.W. Private Communication.

Paltridge, G.W. and C.M.R. Platt (1976) Radiative Processes in Meteorology

and Climatology, Elsevier, Amsterdam, p. 318.

Payne, R.E. (1972) Albedo of the sea surface. J. A tm. ScL 29: 959.

Robinson, N. (1966) Solar Radiation. Elsevier, Amsterdam, p. 347.

Rozenberg, G.V. (1969) Light scattering in the earth's atmosphere. SovJet

Physics t/spe,khJ 3: 346.

Sekera, L. (1957) Polarization of skylight. EncyclopedJa of PhysJcs 48: 288,

Springer, New York.

SchUepp, W. (1949) Die Bestimmung der Komponenten der atmospharischen TrUbung

aus Aktinometermessungen. Arch. Het., &~eophys. und Bioklima tol, Serie B
1: 257.

Spencer, J.W. ,(1971) Fourier series reprasentation of the position of the

sut:.!. Search 2: 172.
Swinbank, W.C. (1963) Long-wave radiation from clear skies. fluart. J., Royal
Het. Soc. 89: 339.

Willis, J. (1971) Some high values of the albedo of the sea. J. Appl. Het.
10: 1296.

Willson, R.C. (1984) Measurements of Solar Total Irradiance and its Variabi-
lity. Space Sci. Rev. 38: 203.

46
CHAPTER 3 RADIATION QUANTITIES IN METEOROLOGY

Young, A.T. (1980) On the Rayleigh-Scattering Optical Depth of the Atmo-

sphere. .J. App.l. Het. 20: 328.

47
 4. RADIATION MEASUREMENTS

4.1 RADIATION DETECTORS

Radiation detectors transform radiation into energy of other type to

which the detector output device responds, such as electrical, thermal, che-
mical, optic-acoustic, pyroelectric, etc. Thermal detectors are used for non-
selective measurements of radiant energy over a wide spectral range. Most
pyrheliometers, actinometers, pyranometers and pyrradiometers operate on
these principles. Photoelectric devices are mainly used in instruments for
spectral measurements (sunphotometers, lux meters, special pyranometers,
etc. ).

The reference for radiation measurements is normally the black-body

radiation at the temperature of melting gold. The intensity of solar ra-
diation, however, is so much higher, that such a calibration technique for
solar detectors is not possible and the detector itself has to be the re-
ference. Hence, solar radiometry has to be referenced to absolute detectors
and absolute radiators cannot be used for this purpose. This is the reason
why radiometry for meteorology and standard metrology had to go different
ways and absolute detectors are of paramount importance for meteorological
radiation instruments.

4.1.1 Thermal detectors

Thermal detectors work on the principle of converting radiant energy to

heat. Among the variety of techniques for measuring thermal energy the ther-
mometric heat-flow method gives the highest detector sensitivity. The tempe-
rature difference along a thermal resistor is directly proportional to the
amount of heat released in its sensing element. This dependence is described
by the equation of detector-heat balance. Thermocouples and resistance ther-
mometers are widely used at present, because electrical techniques are cur-
rently the most suited for temperature measurements. An important factor
common to all thermal detectors is their ability to absorb radiation. Thus,
the properties of absorbing coatings define their basic operational features,
primarily the width of the operating spectral interval A2 -Ai' Furthermore,
the absorption coefficient of absolute detectors should be as close as pos-
sible to one and the thermal resistance should be sufficiently low to ensure
a small temperature difference and therefore small systematic errors of abso-
lute measurements due to infrared losses of the surface of the detector. The
black coating of detectors operating in the laboratory are platinum, gold, or
bismuth black, which are deposited by evaporation techniques. For detectors
used in the field varnish-and-paint coatings, such as Parson's optical lac-
quer, enamel AK-243, 3M-paint, varnish-soot compositions, are used, since
these are more resistant to atmospheric effects.

For direct solar spectrum measurements, the integrated absorption coef-

ficient, a, is estimated from spectral values of the reflection PI or absorp-
tion coefficient ai, which are measured at a specific wavelength AI:

a = =---- (4•1)

where n is the number of spectral intervals, SI is the solar radiation in the

spectral interval AAI.

49
CHAPTER 4 RAD I AT 1ON MEASUREMENTS

Some results of measurements of the spectral reflection coefficients of

present actinometric coatings (Kmito et aI., 1976; Kmito, 1979) are shown in
Fig. 4.1. It can be seen from these data that the optical properties of
coatings depend substantially on their thickness: the selectivity for suffi-
ciently thick coatings (density >' 5 mg/cm 2 , at approximately 20 Ilm thickness,
e.g. curves 2 and 5) is considerably less pronounced than that for thinner
ones (e.g; curves 1 and 4). The type of coating and application technique (at
s u ff i ci e n t t h i c k ne ss) h a v e li t t le a Hec ton its pro per tie s . The in t e g r a 1
absorption coefficient for Parson's optical lacquer with 5.6 mg/cm 2 (see
curve 5 in Fig. 4.1.) gives a = 0.981 and for enamel AK - 243 with 4.8 mg/cm 2
(see curve 2 in Fig 4.1.) gives a =: 0.978, which are reasonably high values
for operational radiation detectors. Moreover, the substantial difference in
the solar spectrum between extraterrestrial and aurface values does not aub-
stantially affect the value of a: for Parson's black, for example, it
increases only by about 0.1 per cent. The value of a for thin coatings,
however, is more sensitive to solar spectrum changes because of higher
spectral selectivity. As an example, a thin AK-243 coating 0.4 mg/cm 2 )
varies from 0.971 to 0.974. Such a sy.stematicuncertainty in a limits the
accuracy of present pyrheliometer's and other absolute instruments with flat
detectors. Cavities are used as absorbing elements in modern accurate thermal
detectors to reduce this component of systematic errors. As a result, the
value of aK and its accuracy increase considerably. According to Willson
(1980) a cone with specular black coating can achieve an a = 0.99988
± 0.00002 with an absorption coefficient of the coating of about 0.95.

40
1/ I

35
I

30

-
-
e1< 25
Q)
o
c
ca 20
I1 r r h ./
IV
i

~
o
If\ 1/
-
Q)

Q)
15

~
CC

10
I h .-. 1/

\~
'-' ! \ 11
I

I ,\ 1'-' _J p\\ / ~~
1

5
~/ V -,
11 '1/
/--
L\ff '-.-. - ~

_. -'-'
_2

HV ~
, I' --~
\ t--
"
"-
-....... ~
.b. ::.
/
\1
\
.- ...... r;:'.c':=-.
\
~~
---
I
t._-=- " ::::::.- --~":. -~~:::
::.~....
::'::.::: --- ~
l!
't-, t'- -- ',: ~. f:,-.: ..... .-._. -~
..--'-

o
OA 1.0 3.0 4,0 6.0 8.0 10.0 12.0 14.0 16.0
Wavelength (IJm)
Figure 4.1 Optical properties of black coatings. For identification of cur-
ves see text.

The thermal resistances of black coatings, R, ha've not been studied

thoroughly so far. According to the data (Blevin and Brown, 1966) for Par-
son's black c.oating of' about 10 Ilm thickness the resistance is 2.5 10- 4

50
CHAPTER 4 RADIATION MEASUREMENTS

Km 2 W- 1 and 7.10- 4 Km 2 W- 1 for gold black of thickness of about 5 ,urn. Such

values lead to systematic errors of thermal detectors up to about 0.3 to 0.4
per cent (Kmito, 1979). This effect is also considerably reduced in cavities
(approximately by an order of magnitude). Thus, application of ca vi ty thermal
detectors ensures considerably higher operational stability of the sensitivi-
ty, and substantially higher accuracies of absolute radiation measurements.

4.1.2 Quantum Detectors

Matter reacts to the photoelectric effect of radiant energy in many

different ways. Thus there is a great variety of modern photoelectric detec-
tors. They are divided into two large groups according to the type of photo-
effect: those with external and those with internal photoeffects: (that is
photoemission and internal production of electron-hole pairs). Detectors
based on the internal photoeffects are mainly used for current application in
meteorology. In these detectors the radiation quanta excite free charges,
which can move inside the photosensitive layer and result in a change of its
conductivity .. The photocells acting on the basis of this principle are called
photoresistors. The time constant of such detectors is determined by the life
time of carriers C"r !:::: 10- 3 sec), and the spectral sensitivity is determined
by the transparency and quantum efficiency of the photosensitive layer. Pho-
toconductivity appears if the radiation energy, hv, is sufficient for an
electron to pass the forbidden gap AW. With AW for typical materials between
0.2 and 2eV the long-wave cutoff for photoresistors, given by Ac = c h/AW,
lies between 0.5,u and 5,um.

For example, selenium, Se, (AW = 1.55eV, Ac = 0.8,um); lead sulphide,

P b S, ( AW = 0.4 e V, Ac = 3 .1,u m); in diu m ant i mon i de, I n As, ( AW = 0.3 e V,
Ac = 4.1,um); gold-doped germanium (AW = 0.15eV, Ac = 8.2,um) etc. are widely
used in the labora tory. Seleniu m ha s mai nly been used in meteorological
measurements because of the spectral sensitivity peaks at about the same
range as the spectral peak of solar radiation.

In the internal photoeffect in a system consisting of a junction between

different types of conductors (metal-semiconductor, p-n semiconductor) an emf
is established across the barrier, if it is illuminated or a current can flow
through an external circuit. The barrier-layer photocells operating on the
basis of p-n junctions are referred to as photodiodes. The basic substances
used for manufacturing short-wave photodiodes, are silicon (Ac !:::: 1.0,um), ga-
lium arsenide(Ac !:::: 0.8,um), and germanium (Amax. !::::1.8,um). The time constant
of photodiodes of about 10- 5 to 10- 7 seconds, depends strongly on the thick-
ness of the n-p junction and the doping profile. For short-wave detection
silicon diodes are most widely used and, due to the advanced technology in
producing such cells, they have proven to be very stable over long periods of
time. Another advantage of the photodiodes is that they are inherently linear
over a large dynamic range (several orders of magnitude). Moreover, it has
been shown that silicon diodes can be used as absolute quantum detectors,
because their internal quantum efficiency is very close to one and is a known
function of wavelength (Zalewski and Geist, 1980). Thus, if the reflectance
on the surface of the diode is known, the radiation of a monochromatic beam
can be measured very accurately.

4.1.3 Radiation Scales

Radiation measurements in meteorology have to be very homogenous in time

and space in order to yield useful information about the subtle differences
in radiation climates. From the beginning of the establishment of radiation
networks some concern existed on how this homogeneity may be achieved. The
national metrology institutes were not in a position to help, as their tech-
niques for the establishment of radiation standards were limited to low in-
tensity levels and, therefore, not adapted to the measurement of solar radia-
tion. This was the main reason for different and independent development of

51
CHAPTER 4 RADIATION MEASUREMENTS

radiation standards in meteorology by use of a pyrheliometer, an instrument

used to measure direct solar radiation.

In order to maintain a stable basis for the worldwide radiation mea-

surements some standard instruments were used to define a reference, which
was called a "scale". This concept, however, is misleading because an abso-
lute scale in Wm- 2 or in cal cm- 2 S-1 was already defined on the basis of a
standardized physical system.

Before 1956 two "scales"· were used as references for meteorological ra-
diation measurements: the original Angstrom scale and the Smithsonian
scale 1913 revised. The Angstrom scale is based on the calculated perfor-
mance of the Angstrom compensation pyrheliometer A 70. The reference of the
Smithsonian scale is the Waterflow instrument with a self-calibrating cavity-
type. receiver and the Silverdisc instruments as transfer· standards.
The 1956 International Radiation Conference at Davos established the Interna-
tional Pyrheliometric Scale OPS 1956) by correcting the Angstrom and the
Smithsonian scales as follows:

IPS 1956 = AS 1905 + 1.5%

(4•2)
IPS 1956 = SS 1913 2.0%.

These corrections were based on the assumption that the difference between
the Angstrom scale and the Smithsonian scale was 3.5 per cent as determined
by Angstrom (919), in spite of the fact that Kimball (1910), found this
correction to be 5.4 per cent, 5.3 per cent and 5.1 per cent respectively,
(Morikofer, 1937 and Hinzpeter, 1956).

The IPS 1956 was defined and used during the First International Pyrhe-
liometric Comparisons OPC 0, organized by the World Meteorological Organi-
zation at Davos 1959. The reference instrument was A 158, the Stockholm
standard, (traceable to A 70) corrected to IPS 1956 by changing its original
calibration factor by 1.5 per cent. According to the results of the compari-
sons, the calibration factors of all participating instruments were adjusted
in order to give the same readings as the reference. At IPC II in 1964 the
same procedure was applied. During the Regional Comparison 1969 of the Re-
gional Association VI at Carpentras, a substantial change in the readings of
the Stockholm instrument was noticed. This change was confirmed during
IPC I I I in 1970. A 158 measured 1.2 per cent higher than during IPC I and
IPC 11. Later, some of the discrepancies were explained by an irregularity
in the auxiliary equipment used during IPC I and 11. Therefore, the
"IPS 1956" used after 1959 did not represent the scale as defined by the
Radia tion Com mission in 1956.

In order to clarify this situation the results of many comparisons made

at different places, including the measurements with modern absolute radiome-
ters, were critically analysed by Frohlich (1973a). The results are shown in
Fig. 4.2 and Table 4.1.

Table 4.1 Differences between different scales in use for meteorological ra-
diation measurements

Smithsonian Absolute: 4.6% ± 0.4%

Angstrom Absolute: 2.4% ± 0.7%
IPS 1956 Absolute: 2.0% ± 0.4%
Smithsonian Angstrom 5.0% ± 0.7%
IPS 1956 Angstrom: 0.4% ± 0.4%
Smithsonian IPS 1956: 4.6% ± 0.4%

These results demonstrate that the assumed differences for: the defini-'
tion of the IPS were wrong, and that· it is difficult to establish such a
reference with the traditional pyrheliometers. Moreover, it suggests that

52
CHAPTER 4 RADIATION MEASUREMENTS

absolute radiometers should be used to obtain a reliable reference for meteo-

rological radiation measurements. This was done through the introduction of
the World Radiometric Reference (WRR).

5 _~...- SO SI 13 (Oavos 1973)

Smithsoni an _________ SO SI 5 (IPC I 1959)
Scale 1913 revised-------H SO SI 50 (Oavos 1973)
======~SD SI 50 (Oavos 1934/46)
4 SO SI _13 (Oavos 1934046)
SO T7 (Oavos 1973)

World Radiometric ___~--WILLSON (Table Mountain 1968/72)

Reference "'~~WATERFLOW (Table Mountain 1932/52)
2 PHO (Oavos 1973)
Absolute Scale ------H:l
PACRAO (IPC III 1970,Oavos 1972/73)
NBS (IPC III 1970)
TINGWALOT (Oavos 1930)

International A 210 (IPC III 1970)

Pyrheliometri~cale 210 (Oavos 1934)
0 .....- - - " ~ 158 (IPC III 1970-1.5%)
~ngstrom Sc~le------ff A 212 (IPC III 1970)
~_::::-A 158·(IPC III 1970-1.5%)
A 149 (Toronto 1955/59)
-1

Figure 4.2 Sum mary of the relations between different scales in use for
meteorological radiation measurement s (Frohlich 1973a).

Since the first participation of absolute radiometers at IPC III (1970)

many comparisons between such instruments and pyrheliometers representing
"IPS 1956" were performed at the World Radiation Centre Davos. The results
of all the intercomparisons are summarized in Fig. 4.3. PACRAD Ill, an in-
strument developed by Kendall and Berdahl (970), is used as reference, main-
ly for historical reasons. The results of the comparisons show close agree-
ment among the 10 different types. All lie within a range of ±0.8 per cent,
and half of them lie in the range of ±0.15 per cent, centered around a value
which is about 0.2 per cent higher than PACRAD. The concentration of the
individual results within a narrow range is an indication that the most pro-
bable value in SI (Systeme Internationale) units will also fall within this
range, and that a weighted mean of these results can be used to fix the WRR.
The following ratio has been determined to enable conversion from the
"IPS 1956" to the World Radiometric Reference.

WR R
= 1. 022. (4.3)
"IPS 1956"

With this value and those given in Table 4.1, results of previous stu-
dies can now easily and clearly be reduced to the WRR.

Mandatory use of the WRR has been put into the regulations of the World
Meteorological Organization in 1981. Its absolute uncertainty is estimated
to be ±0.3 per cent, which means that the true value lies within this range.

53
CHAPTER 4 RADIATION MEASUREMENTS

0.995 1.005 RA TIO TO PACRAD HI

ACR 310
ACR 311
ACR 701
CROM
ECR 10
ECPR NS21-11A
~~-ljEPAC 11402
EPAC 12843
EPAC 13617
PMO 2
ifii'.imR*t--_..-------. PMO 3
PMO 5
PVS-5
TMI 67502
PACRAD III

Figure 4.3 Summary of the results of comparisons of 15 absolute radiometers

of different design and origin and the definition of the World Radio-
metric Reference (WRR). The shaded area represents a ±O.2 per cent range
about the WRR (Frohlich, 1975). .

4.2 TECHNIQUES FOR THE MEASUREMENTS OF TOTAL RADIATION

4.2.1 Pyrheliometers

A pyrheliometer is an instrument for measuring the direct component of

solar radiation at normal incidence. This instrument classification is di-
vided here into two categories: self-calibrating and common instruments. The
latter category refers to those instruments which obtain their calibration
constants by comparison with instruments of the self-calibrating type. Fur-
thermore, a distinction is made between first- and second-class reference-
standard pyrheliometers according to their specified accuracy (see also CIMO
Guide, 1983). The reference standard instruments are the basic standards for
calibration purposes. The field of view of pyrheliometers should be small
enough to minimize the registration of circumsolar radiation, but still be
consistent with sufficient tracking tolerance. Experience has shown that a
field of view of a 2,5°-half angle circular cone and a slope angle of 1° with
appropriate collimating diaphragms is a satisfactory compromise (see
Fig. 4.4). This geometry is also recommended by WMO fora circular geo.metry
(CIMO Guide, 1983),

For instruments with square apertures, such as the Angstrom pyrheliome-

ter, the influence of circumsolar radiation is more difficult to assess and
the above criteria not easy to apply (see e,g. Angstrom and Rodhe, 1966).

54
CHAPTER 4 RADIATION M~ASUREMENTS

Front Aperture

Full viewAngle
Slope Angle

Receiver

Figure 4.4 Definition of the view-limiting geometry.

4.2.1.1 Historical Reference Pyrheliometers

The most important types in the class "Reference Standard Pyrheliometer"

are the Angstrom Compensation Pyrheliometer and the Silver-Disk Pyrhelio-
meter.

The Angstrom pyrheliometer is probably the most widely used instrument

at the reference standard level. The instrument was developed by Knut
Angstrom (1899) and it has undergone a number of modifications over sub-
sequent years. Although originally intended as an absolute instrument, the
newer models have generally been calibrated against earlier instruments, for
which the calibration had been calculated from physical principles of the
instrument design. The principle of operation of the instrument is electri-
cal compensation. The receivers are strips of manganin which are coated with
a black absorber. Some recent models are coated with Parson's black, while
the earlier versions were coated with camphor soot over platinum black.
While one receiver is exposed to sunlight, the other (which is shuttered) is
heated electrically, until a null reading is achieved for the signal from the
differential thermocouple attached to both strips. The accurate determi-
nation of the current gives the calculated radiation according to the for-
mula:

S = kP, (4.4)

where S is the solar irradiance, i the measured current and k is the in-
strument constant. In principle k can be calculated according to

k = cr/ab, (4.5)

where c is a conversion factor, r is the resistance per unit length of the

strip, a is the absorptance and b is the width of the strip. In this case
the instrument can be called an absolute radiometer. However, deviations
from the expressed ideal behaviour above may be substantial and thus the
uncertainty quite large. Hence its use as an absolute instrument is no longer
recommended.

55
CHAPTER 4· RADIATION MEASUREMENTS

The geometry of the Angstr6m detector requires that it has a rectangular

field of view in contrast to the circular field of most other instruments.
Present commercial units have a 4.5°. by 1.5· field. A considerable number of
studies have been performed to assess the influence of the circumsolar radia-
tion on the. measurement of the dire'ct beam _PeCaUl;le of the differences in the
view-limiting geometry' :6f the, different models of Angstr6m Pyrheliometers and
instruments with' Circular fields (e.g. Frdhlich, 1973b). In clear-day condi-
tions at mountain stations this effect is' at most O. 1 per cent and thus
within the limits of the day-to-day varial;lility.

The Silver-Disk Pyrheliom et er was developed by Abbot (1922) at the

Smithsonian Institution. It is calibrated ag.ainst a primary standard instru-
ment, such as the water-flow pyrheliometet,' which will be described below,
and was not intended' 'to be used in an :abSblute mode. This instrument has
generally been the device by which the .Smithsonian scale was transferred to
working instruments. The origin.aJ inst:ruments had opening angles of 10·,
while the later units have a fieid of view of 5°. The principle' of operation
is the measurement of the temperature rise and fall of the receiver, a sil-
ver-disk, which is coated with -lamp-black during irradiation and shading,
using a. mercury thermometer. The thermometer is mounted radially in a hole
in the disk, with mercury used as a heat transfer agent between the disk and
thermometer. The instrument is operated with a two-minute interval of solar
exposure followed by a shuttered periud'of two to four minutes. The tempera-
ture readings are corrected for ambl'eht air as well as corrected for bulb
and stem temperatures. The constant supplied with the instrument is then
applied to obtain the radiation value. Detailed studies have shown that there
is a significant difference in the results, depanding on the timing perfor-
mance of the shutter. These differences were to some extent responsible for
the confusion which arose with use of the revised Smithsonian scale 1913
(Fr6hlich, 1973a).

The basic standard instruments of the Smithsonian Institution were a set

of calorimetric pyrheliometers, the Water-Flow Pyrheliometers (Abbot and Al-
drich, 1932; Aldrich and Abbot, 1948) developed over the first half of this
century. The first instrument was a single cavity unit, while the later
improved instrument had dual cavities. The basic principle is to measure the
temperature change of the water which flows through the walls of the irra-
diated conical radiation receivers. These cavities were also fitted with
electrical heaters which were energized to calibrate the instrument when no
sunlight was entering the sensor. In the single tube instrument solar
measurement and electrical calibra tion were per for med sequentially, while
during operation of the double-tuba instrument one sensor was exposed to
solar radiation, while the other, the shuttered one, was electrically heated.
In this case the output of a differential thermophile, measuring the tempera-
ture difference between the outlets of the two cavities, was monitored to
achieve a null reading by adjusting the electrical input to the shuttered
cavity. In many respects this device is similar in design to the modern
cavity pyrheliometers discussed below and in operation to the Angstr6m pyr-
heliometer. The dual-cavity instrument was claimed to have an accuracy of
±0.2 per cent and constituted the reference for the absolute scale of the
Smithsonian Institution. It is very close to the current reference, as shown
in Fig. 4.2.

4.2.1.2 Self-Calibrating Pyrheliometers

Many new self-calibrat'ing instruments have been developed since the

early 1960's. They are similar in that most employ a cavity receiver for the
absorption of incident solar radiation and a heat-flux transducer to monitor
the power. Also, the measurement of the electrical power in a heater during
the radiation measurement constitutes the method· of traceability ·to SI-units.
Most of these devices have. been adapted· for use at atmospheric pressure,
although many were developed for use in space experiments under vacuum condi-
tions. A few are available commercially. A number of correction terms asso-

56
CHAPTER 4 RADIATION MEASUREMENTS

ciated with each type of instrument, due to deviations from ideal behaviour,
must be applied in order to convert the observed electrical power to a true
irradiance quantity. The most important items cover the accurate measurement
of the area of the precision aperture, errors in the electrical power mea~
surement, the emission from the cavity through the entrance aperture, heat
losses through the power or sensor leads, and a number of items which are
generally grouped under the term "non-equivalence" and which relate to the
differences in the temperature distribution and, thus, the thermal paths of
the radiatively and electrically produced heat in the cavity. The absorp-
tance of the cavity receiver over the spectral range of the incident solar
radiation must also be accurately assessed.

The estimated accuracy of all these factors determines the uncertainty

of the radiation measurements with a given instrument. It is, therefore,
very important to design the radiometer so that it minimizes the individual
errors and to devise accurate and reliable methods to evaluate their corre-
sponding uncertainties. An example of such a procedure, called the charac-
terization of a radiometer, by dedicated and independent experiments is de-
scribed in detail by Brusa (1983). The present state-of-the-art of such cha-
racterization methods may be defined by an achieved uncertainty of the radia-
tion measurement to within ±O.2 per cent. The different instrument types re-
cently developed have all slightly different designs so that the characteri-
zation in each case has to be different in detail; Prom the above it becomes
clear that only fully characterized instruments can be used as "true" abso-
lute instruments. If only typical values of the correction factors are used,
the uncertainty will increase to the extent that the factors may vary from
one instrument to the other.

Table 4.2 Summary of self-calibrating cavity pyrheliometers

Identification Sensor and mode Cavi ty type Cav i ty-heat fl Oil Aperature Character i zat ionl )

.. eter aITanqelll.ent'· area Cc.')

ACR: resi stance thenaometer classical cone Nodi- E and H

Active Cavity Radiome- sensor operated in £ ied to remove menis-
developed by R.e. Willson. active mode cus dt bottoll'l point.
Jet Propulsion Labor:atory, specular paint.
Pasadena. eaU f.. USA

CROH: flux Itteter ~ith inte- cylinderleal in each

developed by O. Crolllllelynck. qrated therJllopile ope- of two cavities \11th
Institute Royale Heteoroloqique ro.ted 1'n active. lQode flat battolll. and dif-
de Belqique, Bruxel1es. fuse paint.

H-f: circular plated and inverted cone within 0.5 E and T

developed by .J. Hickey and "irevound therllopi le a cylinder, specular
R. Frieden. operdted in passive pa.inL
The Eppley Laboratory. I ne .. Mode.
Newport. R.1.. USA

PACRAD: discrete junction Classical cone with LD H

Prhlary Absolute Cavity therMopi le ,operated truncdted inverted
Rad<iollleter. in passive Mode cone above, diffuse
developed by J.ff. Kendall. pa.int.
Jet Propulsory laboratory,
Pasadena, Col if., USA

PHD: .ark 2 has d discrete inverted cone uithin 0.2 E

developed by R.W.8rusa and therlQopi le, mark (, a a cylinder, specular
C.Frohlich. resistance thermometer
Phys i ka lisch-Keteorol 09 isches both are operated in
Observatori Ul'll. Dovos. Swi tzer-Iand active mode

rHI: discrete junction ther- i overied cone in cy- - I. 0 E and T

deve loped by J. H. Kenda 11 , Jlopile. usually operated 1 i nder, semi -specula
Technical Keasurements Inc. i npass i ve Mode paint.

The drawings ill ustrate the cavi ty-heat f 10" JIleter arrangements:
- cavity. =·heat-flowaeter. '7777777 heat sink

E chorocter i tlttion by experillental J1\ethods

H: characterization band on _odel calculations
T: characterization -band on typical values-

As examples of such radiometers the following types will be described in

s om e de t ail: P ACRAD (K end a 11 and Ber d a h I, 1970) ; ACR (W ill so n , 1 973); P M0
(Brusa and Prohlich, 1972); CROM (Crommelynck, 1973); H-P (Karoli et aI.,
1983) and TMI (Reid et aI., 1978). The general design features are summarized
in Table 4.2.

Each individual instrument has its own set of correction factors just as
the older pyrheliometers had their own calibration constants. The absolute

57
CHAPTER 4 RADIATION MEASUREMENTS

accuracy for all described instruments is between ±O.3 and ±O.2 per cent,
depending on the realiabilityof the characterization. Some instruments, such
as the PACRAD and ACR, have been built by different manufacturers. The TMI,
H-F, and PMO radiometers are also in commercial production. The ACR, how-
ever, is no longer commercially available (19·84).

The terms. "active" and "passive" have been employed with regard to the
operation of absolute radiometers. Generally speaking, "passive" devices are
operated in a sequential manner by alternately exposing the receiver to the
sun and then electrically calibrating the heat-flow transducer with no elec-
trical power being applied during t.he sunlit phase. "Active" operation gene-
rally refers to the maintenance of a constant temperature difference along
the heat.-flow resistor during both exposure and reference phases of a se-
quence controlled by a servo or other control system which automaticalJy
adjust.s t.he power to the heater. None of the inst.ruments, except CROM, is
truly bifurcated as was the case with the Angstrcim and the double-tube water--
flow instruments. But Toast are employed with a compensating receiver to mini-
mize the effects of changes in ambient conditions on the instrument perfor--
mance. A number of automatic cont.rol mechanisms and computer interfaces have
been developed for operation of these devices.

4.2.1.3 Opera tion a 1 Pyrheliomet.ers

Contrary to the i nstrum ents described under 4.2.1.1 and 4.2.1.2, which
are mainly used as reference standards for calibrating the instruments de-'
scribed below, the inst.ruments discussed here are used operationally. The
basic sensing element in these instruments is a thermopile used as a thermal
flux meter mount.ed on a heat sink and painted black on the surface exposed to
the incident radiation. They have to be calibrated by comparison with a
reference standard pyrheliometer.

Figure 4.5 Eppley Normal Incidence Pyrheliomet.er mounted on an equatorial

tracker.

58
CHAPTER 4 RADIATION MEASUREMENTS

The Eppley Normal-Incidence Pyrheliometer (NIP) is such an instrument.

The early devices, many of which are still in use, had a copper/constantan
eight-junction ther mopile. The more recent versions have been fitted with
wirewound thermopiles of copper-plated const.ant.an wire. The first. unit.s of
t.his t.ype were designated by the let.t.er "D" in t.he serial number. The most
recent. unit.s are designated wit.h "E" in t.he serial number. All of t.he more
recent. unit.s and some of t.he older ones are t.emperature-compensat.ed. The new
units have a double-walled construct.ion t.o dampen t.he effects of external
wind and temperat.ure changes. Each unit is fitted with a crystal-quart.z
\vindow (unless otherwise requested by the user) which is sealed int.o t.he
front. apert ure. The blackening of the original units was la mp-black, the. t
of the later units is Parson's black. The new types have 3M, 101 NEXTEL. black
velvet paint. The instrument can also be fitted with a filter wheel for the
use of the classical broad-band filt.ers. Fig. 4.5 shows a NIP mount.ed on an
equatorial tracker and the _performance specifications of the NIP are given in
Table 4.3.

The Kipp & Zonen Pyrheliomet.er corn monly referred t.o as t.he Linke - Feus-
sner Act.inomet.er ("Panzer--Akt.inomet.er"), is shown in Fig. 4.6. It. has a
Moll-t.ype t.hermopile element. wit.h fort.y junct.ions of manganin-const.ant.an ar-
ranged in two groups of twent.y each ina. circle. This arrangement. offers
compensat.ion for inst.rument.-t.emperat.ure fluct.uat.ions. The unit. is fitt.ed with
a Eil t.er wheel for t.he selection of either no iil ter, a quartz glass, or one
of t.he classical Schott filters. There is no t.emperat.ure compensat.ion, but
the calibration constant can be corrected by the factor 1 +a·(t.-20), with t
being t.he instrument temperature and a = O.002K-1. A thermometer is mounted
on the instrument body for establishing the temperatur~at the time of mea-
surement. This pyrheliomet.er is usually supplied on a mount allowing an ad-
just.ment for solar elevation and azimut.h to be made by hand .• The performance
specifications are summarized in Table 4.3.

Fiqure 4.6 Kipp [" Zonen Pyrheliometer '"

59
CHAPT'ER 4 R AD I AT ION MEASURE ME'NTS

The pyrheliometer used in the· Soviet Union and E'astern European coun-
tril;!s is the actinometer A80 developed at the Main Geophysical Observatory in
Leningrad CYane.shevky, 1957), The receiver is a thermopile, soldered in the
form of a st.ar, of manganine and constantan bands CFig.4.7a:J. The cold
junctions of the thermopile are fixed on a, copper ring with a thin layer of
electric insulator inbetween. This ensures g.ood thermal contact with the
a·ctinometer body. A thin silver-foil disk of 11 mm in diameter is glued to
the top of the hot junctions, which is painted with a mat black lacquer.

aJ
Figure 4.7 a) Construction of the copper-m.anganin thermopile of the A80
actinometer. 1: hot junctions, 2: cold junctions, 3: copper ring, 4:
silver foil disc. b) A80 actinometer. 1: actinometer, 2: coarse a.nd 3:
fine adjustment of hour angle 4: support, 5: cover, 6: base plate.

The general view of the actinometer is given in Fig. 4.7b and the per-
formance specifications are summarized in Table 4.3. The instrument is nor-
mally fixed on a clock-driven equatorial mount.

Tahle 4.3 Performance specifications of three typical normal incidence pyr-

heliometers.

Eppley NIP 'Kipp So. Zonen Actinometflr A80 units

Sensivity 8 16 10 /lV/Wm-'

Response time 1.0 2.7 4.0 sec

1/e

View limiting
geometry:
full view angle 5.7' 9.6 10 . deg
slope- angle 0.8·' 2.4 2;3 deg

Tempe.ralure. temperature compensated

coefficient to better than i1:\: -O •.2%/K +O.08%/K
in the range -20' f+40'C

60
CHAPTER 4 RADIATION MEASUREMENTS

4.2.1.4 Mounting of Pyrheliometers

Since pYl;'heliometers need view only a small field centered on the solar
disk, the mounting site requirements are less restrictive than for most types
of radiation instruments. The tracker carrying the instrument should be
mounted on a stable platform from which a view of the sun is not restricted
by buildings, towers, cables, or other obstruction$ throughout the sunrise -
sunset period. In the choice of a mounting site, the change of - solar azimuth
with season must, of course, be taken into account. But obstructions of the
horizon have no other significance for measurement of the direct solar beam.

The mount itself is normally of the clock-driven equatorial type, al-

though computer-controlled mounts or solar-sensing devices, both of which may
have two degrees of freedom, are sometimes employed. The proper orientation
of the main axis of an equatorial mount can be a frustrating process, unless
accomplished in a systematic .stepwise manner, and, even then, a certain
amount of trial and error may be necessary. The steps in orienting the mount
are outlined below, the objective always being to get the axis of the mount
precisely parallel to the rotational axis of the earth. A tolera nee of
± 0.25° in angular positioning is permissible for an instrumental slope angle
of 1°. In the following, the axis of rotation of the mount will be termed
'the hour-angle axis', orientation around the vertical is 'azimuth', and the
declination axis is 'the axis of the mount' which is normal to the hour-angle
axis.

Step 1: Set the mount on a stable and approximately level surface with the
hour-angle axis in approximately the north-south plane. Level the mount
by means of a spirit level.

step 2: Set the inclination of the hour-angle axis as precisely as possible

to the latitude angle of the site, the slope being down to the south in
the northern hemisphere and the reverse in the southern hemisphere. A
convenient method of setting the angle is with a leveling protractor,
preferably one with a vernier scale. This setting should b~ accurate to
at least 0.25°.

Step 30: . Orient the pyrheliometer to view the zenith by adjust'ing the decli-
nation and the hour angle.

step 4: Precisely at solar noon, rotate the instrument around the decli-
nation axis to approximately view the sun. By small adjustments of the
instrument around the declination axis and the mount in azimuth but not
around the hour axis, orient the instrument precisely to the solar posi-
tion. The mount should now be in the proper orient a tion, but it is well
to check the incl~nation angle of the hour-angle axis again to be sure
it has not been disturbed.

Errors in either the inclination of the hour-angle axis or the azimuthal

setting will result in tracking-errors. The best way to detect an azimuthal
error is to compare the tracking-errors near sunrise and sunset. If the
error is in one direction at sunrise and the opposite direction at sunset,
the azimuthal setting needs adjust ment in the obvious direction. If, how-
ever, the sunrise- and sunset-errors are in the same direction and of the
same magnitude, and the orientation appears correct at noon, then the incli-
nation of the hour-angle axis needs adjustment. Obviously, a misalignment of
both axes will cause a combination of these symptoms. It should be remem-
bered that, except at the solstices, the declination of -the sun itself is
slowly changing, but the change over one day is too small to be of much
significance in the orientation process. Once the mount has the required
orientation, it is advisable to carefully mark the outline of the base on the
platform and securely fasten the base to the platform so that its orientation
cannot be inadvertently disturbed.

61
GHAPTER 4 RADIATION MEASUREMENTS

Signal cables from the instrument to the data-acquisition system should

be of twin-conductor stranded and shielded wire of No. 16 gauge or larger.
All outside connections should be weatherproofed, and in thunderstorm-prone
areas the cables should be put underground, if the instrument is located a
considerable distance from the data-acquisition system. Cable lengt.hs of up
to 200 meters may be used if the impedance of the recording system is suit-
able, but long lines are subject to stray noise pickup a.nd may promote light-
ning damage to the equipment. For the usual case, in which the instrument
makes one rotation per day, twisting of the signal cable is a problem. One
method of avoiding such twisting is the use of slip rings for transfer of the
signal from sensor to cable. Slip rings, however, tend to introduce noise in
the signal, and are usually avoided for that reason. Probably the best solu-
tion is to simply untwist the cable every few· days by either disconnecting it
momentarily or counter-rotating the instrument a few times by slipping the'
clutch. Care in this aspect of the operation is required, as good data has
frequently been lost by the signal cable hanging up or breaking because of
excessive twisting..

4.2.1.5 Maintenance of Pyrheliometers

Because of the weatherproof design of both instruments and mounts, pyr-

heliometers have only minimal maintenance requirements. Besides the checking
of orientation and i.mtwisting of cables mentioned above, the principal task
is a daily inspection to keep the entrance aperture and window, if fitte.d,
clean and free from dew or frost. rn most cases, the optical surfaces are
readily cleaned with a camel's~hair brush, a soft lint-free cloth or a lens
cleaner, but in polluted conditions the use of a solvent, e.g. alcohol or
acetone, may be required. Under no circumstances should a solution which
leaves a deposit on the surface, or, which might cause a deterioration of the
seals, be employed. Frost, snow, or dew on the surfaces interferes with the
measurements, so steps should be taken to alleviate the problem where it is
significant. Probably the most satisfactory solution is to install an air
blower with a discharge across the aperture, but it is necessary to assure
that the temperature of the air stream is only slightly higher than that of
the ambient air. If this. is not the case, thermal gradients may introduce
errors in the measurements.

4.2.1.6 Calibration of Fyrheliomet.ers

Operational pyrheliometers are calibrate.d by direct comparison with an

absolute radiometer or a secondary standard pyrheliometer. For this method,
the operational and standard instruments are installed side by side and ope-
rated simultaneously for periods of two to three hours near noon for each of
a few days. A relatively large number (at least 100) of pairs of simul-
taneous values, Ai and A2 respectively, should be obtained from the records,
and the average R of the individual ratios A2 /Ai computed. If C1 and C2 are
the calibration factors for the operational and standard instruments, we have
C1 = C2 • R. The method can also be used for checking for possible changes
of instrument-response characteristics with sun elevation by appropriate
choice of times for performing such comparisons. The fields of view of the
two instruments should not be significantly different, otherwise the effects
of different amounts of circumsolar radiation reaching the sensors would be
important.

4.2.2 Pyranometers

Instruments in this category are employed for measurement of the total

radiation falling on a surface from the entire hemisphere above the measure-
ment plane. Generally, these instruments are used with the radiation recei-
ver placed horizontally so that the incident radiation is that coming from
the sun and sky over t.he full elevation range from horizon to horizon and

62
CHAPTER 4 RADIATION MEASUREMENTS

over an azimuth range of 360 degrees about the zenith vector. Pyranometers
are usually used for routine measurements and left permanently outside. For
this reason, they must be able to withstand the effects of all types of
meteorological condit.ions. Most. inst.rument.s are of t.he thermoelectric typE'
with the detector protected by glass hemispheres, which also define the spec-
tral range of the radiat.ion being measured. In the case of a black receiver
two domes -are required in order to shield the infrared radiation from the
outer dome, the temperat.ure of which may vary quite rapidly with changing
wind speed and air temperature. Pyranomet.ers have recently been developed
which employ sensors of various t.ypes of silicon diodes (see 4.2.2.8).

4.2.2.1 Most Widely Used Pyranomet.ers

A short description of the represent a ti ves of the mo~.t. widely used in-
strument t.ypes: Eppley PSP, Kipp & Zonen eM5 and CM10, the Schenk stern
Pyranometer and the Russian instrument M-BO will be given below; This list is
by no means 'exhaustive, but it is representative. Furthermore, these intru-
ments have been tested unde~ a great variety of conditfons and were compared
many times with each other and with other types during int.ernational trials.
Therefore, a great deal of useful information is available to illustrate the
behaviour of these instrument types. This has helped considerably in their
selection for field use. The general specifications of these instruments are
sum marized in Table 4.4.

Figure 4.B Precision Spectral Pyranometer (PSP) of Eppley.

Table 4.4 Specification of various Pyranometers

Type Sensitivity Temperature Time constant

flV IWm- 2 coefficient lie (se c . )
.~

Eppley PSP 9 compensated 1

Kipp & Zonen CM5 11 -0.1%IK 6

Kipp & Zonen CM10 4.5 compensated 5

Schenk 15 -0.1%IK 6

M - 80M 11 -0.15%IK 8
--------_._-
63
CHAPTER 4 HAD IATI ON MEASUREMENTS

The Eppley PSP (Precision Spectral Pyranometer, Fig. 4.8) has a circular
multijunction t.hermopile of t.he t.ype wound wit.h copper-plated const.ant.an
wire. The receiver is coat.ed with Parson's optical black lacquer. It has
t.wo precision-ground and polished hemispheres of Schott glass WG295 (short-
wave cutoff, 50 percent transmission at 295 ,um),which may be replaced bya
coloured glass dome, hence the name "precision spectral" (see 4.2.4.1).

Figure 4.9 CM-10 Pyranometer of Kipp & Zocnen.

The Kipp & Zonen CM5 comprises a Moll-type thermopile of very simple and
rugged construction. The domes are made fromSchott glass K5" which has a
short-wave cutoff at about. 310 nm. The thermopile is coated with lamp-black.
The recentlydevelopedCM10 has a thermopile of hundred thermocouples printed
on a ceramic substrate, using Cl thick-film technique. The blacj{ coating is
the 3M velvet black. The material of the domes is the same as that for the
CM5. A CM-iD is shown in Fig. 4.9.

The SchenJ{ 'stern' Pyranometer has six thermopiles arrang.ed radially,

like a star (stern), with the hot junctions painted black and the cold junc-
tions painted white (Fig. 4.10). The glass dome is made from a glass with a
cutoff at about 330 nm.Only one dome is needed, since both the black and the
white surfaces react the same way to the infrared radiation of the dome.

Figure 4.10 Pyranometer black and white of Schenk.

64
CHAPTEf~ 4 RADIATION MEASUREMENTS

Figure 4.11 View of the pyranometer M-80M .. 1; level, 2; pyranometer, 3:

pivot with the screen, 4: fixing screws, 5; cover, 6; hinges.

The Russian instrument is also of the black-and-white type and was deve-
loped at the Main Geophysical Observatory in Leningrad (Fig. 4.11). The re-
ceiver is a flat square thermopile built up of pieces of manganine and con-
stantan strip soldered together. The black and the white paints have the same
factor 0.£ absorption in the long-wave spectral region, which makes the in-·
strument sensitive .only to short-wave radiation. The thermopile is protected
by a thin hemispherical glass dome.

Comparison of Calibration Constants (mV/kWm- 2 )

Calibration Constants Manuf Cahbrallofl

Pyr anO'Tl(>ler Instrurnent °/0 • Devlahon = - 1
Type (mV/kWm') {mV/kWm-', WAC Calibr~lion
Manuf;lclurer Number
(Manufaclurer) (WAC Davos e1) -3 -2 -1 0. 0 + 1 +2 + 3 1 4-.4 +5
EKO STAR 81901 8.24 8.12 ~
EKO STAR 81903 7.85 7.88
EKO STAR 81906 689 7.09
EKO' STAR 81907 7.25 7.40
EKO STAR 81908 9.61 9.62
EKO STAR 81909 7.42 7.45

EPPLEY PSP 14806F3 9.81 9.78

EPPLEY PSP 17750F3 9.15 9.27
EPPLEY PSP 18135F3 8.78 8.92
EPPLEY PSP 20523F3 9.95 9.90
EPPLEY PSP 20524F3 10.10 10.01
EPPLEY PSP 20655F3 10.28 10.24

KIPP & ZONEN CM5 773656 11.94 11.72

KIPP & ZONEN CM5 773992 12.62 12.16
KIPP & ZONEN CM5 774120 . 13.41 12.80
KIPP & ZONEN CM5 785017 10.59 10.35
KIPP & ZONEN CM5 785047 12.23 11.87

KIPP &
KIPP &
KIPP &
KJPP&
KIPP &
ZONEN
ZONEN
ZONEN
ZONEN
ZONEN
CM10
CM10
CM10
CM10
CM10
790059
810119
810120
·810121
810122
5.58
4.58
4.54
4.66
4.24
5.65
4.59
4.52
4.62
4.22 ..
'\.
SCHENK STAR 1626 14.26 14.49
SCHENK STAR 2186 14.94 15.15
SCHENK STAR 2209 15.36 15.29
SCHENK STAR 2217 14.16 14.17
SCHENK STAR 2221 15.24 14.97

Figure 4.12 Comparison of calibration factors of different types of pyrano-

meters.

65
 Outdoor and indoor tests have been performed in order to provide infor-
mation of the performance of pyranometers. A very exhaustive investigation
has been completed recently (TEA Task III (984)), during which the perfor-
mance of more than 30 instruments of eight different types was studied. The
outdoor tests covered a wide range of conditions: winter, summer, clear
skies, cloudy skies, and overcast skies. The temperature dependence, lineari-
ty, influence of tilt and directional response (cosine 'law) were meas'ured in
the laboratory for all instruments. The ultimate goal of the trials was to
determine hot\! -a.ccurate or how variablepyranometers ,can be under different
conditions and i f the accuracy can be substantially improved by introducing
corrections based on the laboratory results.

A first, very interesting result ispre-sent.ed in Fig. 4.12, in which a

comparison between the manufacturer's and the "real" calibration constants
is presented. This information is very important as systematic differences·of
up to several per cent were found. The mean value of the calibration con-
stant isonea.spect, but the variation within one day and from day to day may
give some insight into the physical reasons for, the variability. The results
of the calibrations for a wide range of temperature, intensity and incid;ent
angles are shown in Fig. 4.13 for 2 instruments of different types: Eppley
pSP (20523), Kipp &Zonen CM5 (113656), ,

't. " ',JuI2,9 6 .Jul 30 Jon·13 Jan 14 '" Jan ~8; ",Ju129 6JuI 30 + Jan13 x Jan 14 " Jan18

. .:.'::,:'"=~:'I
.,0 + x

-10'"

15.00
6 q-C%"
' - .~.'..
0- . ~

1.2~OO
• ..

V·,
( ",,~
~.

.14.00
'" ".

118.00
-~ ' & ~' 'ii

18;00
']==,:"
-6:00
'£1:"":=4<:'"
10iOO
.~= 1'8.00

l'ime(hours) Time -(hours)

+10,'" Of- - - - - - - - - - - - - - - -

c
j .. ":.~. 15 20 25 30 '0 25
6

30
10 15 .20
.2
'1Q
:; Instrument Temperature {O'C) Instrument Temperature (OC)
III
o

"°:l 6
1!l

x~""~l ~--,----'
","10'1.1--1~-~x- _ _- - _ _-"-~_~
o 200 400 600 ::-800 1200

2
:Intensity {Wm c ) --hilensity(Wm '2)

'20 -40 aD 80

'Solar .Angle: (0) 'Solar Angle (0)

Figure 4.13 Comparison of the performance of an Eppley PSP (lef.t) and a Kipp
& Zonen eMS (right) as a function of time of day, temperature, intensity
and' incident angle.

From these results it follows that the Eppley PSP and Kipp & Zonen CMiO
perform quite well and give measurements within 2-3 per cent of each other
over a very large range of conditions. On the other hand, the GMS and the
Schenk stern show systematic differences of lIP to 6 to 9 per cent. This
deviation is partly due to the temperature compensation in the former two in-
struments, hut for the latter instruments the linearity and directional re-

66
 CHAPTER 4 RADIATION MEASUREMENTS

sponse characteristics may also be important. This is illustrated in

n Fig. 4.14, in which the results of the laboratory evaluation of the linearity
and temperature dependence for a number of these instruments is shown
(Frohlich, 1984). The linearity is influenced by three terms and is related
to the temperature behaviour of the sensor as follows:

d(ln;\.) d(lnV) _d_(_l_n_aJ) J

VDut = K·Q·[l - Q.(------ - ------ - (4.6)
dT dT d(llT)

with VDut : thermopile output

Q: radiative input power
A: thermal conductivity of the heat-flow meter
a: convective heat-transfer coefficieht
K: conversion factor

The classical, non-compensated instruments, such as CM5, Swissteco and

PMOD-6703, show a behaviour mainly determined by the combined temperature
coefficient of the thermopile materials and is strongly influenced at high
intensities by convective losses. The compensated instruments show a com-
pletely different behaviour. Their sensitivity increases with increasing in-
tensity. This increase is influenced by how well the temperature compensa-
tion is taken into account. Thus the non-linearity of the sensivity may
change from one individual instrument to another. On the other hand, the non- ,
linearity of the classical instruments is very similar for all individual
instruments. With respect to the black-and-white sensors the situation is
quite different and more complex.

Such investigation contributes to a better physical understanding of the

instrument and thus helps the manufacturer improve the pyranometer and the
user choose an adequate instrument. The normal case, however, is that the
user has to cope with existing instruments -and may have only the choice of
the type of pyranometer used in his network. For the latter, it is important

----- ----
PSP

~o+---=="'~~:---===-------

w.
a.
o-l
Ul
<l -1

-2

-4

-J +---Ir----+--+---f---!----+--+--~I- -s-J-t-----i--+---f---!---+--r---l-
100 200 JOO ~OO ~OO 600 700 BOO -JO -20 -10 0 " ·-10 20 30 40

INTENSITY [Wm' 2 ] TEMPERATURE (OC) •

Figure _4.14 Linearity and temp~rature dependence of different types of pyra-

nometers.

to know what variability is to be expected. To assess the variability daily

sums can be compared as shown in Table 4.5 for the four types of instru-
ments. It is obvious that the variability is reduced when compared with

67
CHAPTER 4 RADIATION MEASUREMENTS

instantaneous values and it is surprlslng that the CMS variability is less

than what one would expect from its performance specifications. This seems
to be the result of a subtle compensation of different effects, such as the
combination of low temperature and low sun.

Table 4.5. Comparison of variability of daily sums for different conditions.

The per cent deviation is with respect to the m.ean of a group of PSP and
CMiO.

Maximum deviations (%)

Condition Eppley PSP K&Z CH10 K&Z CMS Schenk

Clear, summer -0.5 +0.3 -0.2 +0.7 ~2.2 -0.3 +0.4 + 2.2
Cl ear, wint,er -2.4 +0.8 -0:.7 +1.8 -5.7 -3.5 +7.7 +10.4
C.loudy, summer -0.5 +0.1 +0.2 +0.5 -1.7 +0.5 +0.8 + 1. 9
Cloudy, winter -3.0 -1.2 +1.1 +2.0 -1. 8 +0.7 +4.4 + 8.2
Overcast, winter -1.3 +0.6 +1.2 +2 •. 0 +0.6 +3.9 +3.0 + 3.,8

Up to now npthing has been said about long-t.erm stability. and about how
often the different types of instruments need to be calibrated. As shot.ln in
Fig. 4.15, the long-term variability is not very important compared with the
short-term variations due to changing conditions during. the measurements. As
to the calibration stra.tegy it is importantt.hat they are always performed
under similar conditions, otherwise the results will be inconsistent. This
holds true mainly for the less accurate types. Ag.ain, it is ast.onishing how
well the CM-5 performs at the level of daily sums.

1981 [982 1983

: : : (7'";>==::=====:::=--==~;;::-'- "'-~<l::----=. _-=====---1

..:....-If-'--:::;;7'_-="",EP::=PL='E:::Y=::P:::,::SP:::::..::::::.....

::]~._=_~_=_~_~__: : : = = :=-:~:.-q. . .: :.~ CM~6~':-= - -: : :f=>:~" " " ': ~.~ ~'~:·~:=I
: : : .: .. .:.. . ::K:I:P:P:": .•

Figure 4,.15 Long-term behaviour af the calibratiOll constant of three types

of pyranometers (Flowers, 1984).

68
CHAPTER 4 RADIATION MEASUREMENTS

4.2.2.2 Installation of Pyranometers

Since global radiation is made up of radiation from the entire hemi-

sphere of the sky, the site for installation of the pyranometer should be
free from any 'significant obstructions extending above the horizon. From a
practical standpoint it is not always possible to obtain a perfectly free
horizon, in which case emphasis should be on avoiding obstructions from east-
northeast through south to west-northwest in the northern hemisphere, and the
reverse of these directions in the southern ·hemisphere. No shadows of the
direct sun should fall on the instrument at any time of the day. Nearby
light-coloured walls extending above the level of the sensor' should be
avoided, as should artificial-light sources of significant radiant emission.
Finally, the site should be easily accessible to personnel maintaining the
instrument, and, for minimizing disruption of the data by maintenance per-
sonnel, the sensor should be at a height of about 2 m above the walkway.

For proper record-keeping and possible future: reference, the charac-

teristics of the site should be carefully documented. The documentation
should include not only the latitude, longitude, and altitude of the site,
but all obstructions of the horizon should be listed and characterized as to
type of material, angular extent in altitude and azimuth, directional orien-
tation of plane surfaces, and if possible even an estimate of the reflection
values of the surfaces'. Such information can be used for the estimation of
errors in the measurements ·due to the existence of' sueh obstructio.ns. In
addition, any changes of the site which occur, such as the construction of
buildings nearby or the growth of trees in the vicinity, should be noted in
the record.

In case a significant portion of the sky is obstructed, a correction to

the global radiation value given by the pyranometer may be made. The magni-
tude of this correction can be determined from the mean irradiance Ed of
diffuse radiation at the site and a map ·of the outline of the obstruction on
an azimuth (tjJ) versus elevation (8) diagram.

The change of' irradiance produced on the sensor by an·eleme·ntaL area of

the obstructions ~s given by the relation

(4.7)

and the total irradiance change by

N
AE' = ~ l1E n , (4.8)
n=1

The geometry of the problem is such that obstructions of only 5 to 7"

above the horizon have little influence on the vertical component of the
irradiance. In such cases, the irradiance correction is negligible.

The pyranometer should be mounted on a platform of sturdy construction

and free from vibrations. The instrument should be fastened securely, but
not so tightly as ·to inhibit i'-ts being levelled by the levelling screws
provided. Most pyranometers are delicate instruments, so mechanical shocks
should be avoided. As to the electrical wiring see 4.2.1.4 ..

4.2.2.3 Maintenance of Pyranometers

Pyranometers in continuous operation should be inspected at least daily.
Dust or moisture should be removed from the cover by a 'softbrush or soft,
li nt-free cloth, care being exercised to a void scra tchi n g of t he surf ace.
Dew, frost, or other material on the optical surfaces invalidates the mea-

69
CHAPTER 4 RADIATION MEASUREHENTS

surements, so every effort should be expended to avoid such interference.

Perhaps the most satisfactory method of eliminating dew or frost from the
instrument is by a blower which discharges a current of air over the dome.
Such a device is readily constructed from a small fan and flexible tubing,
feeding air jets which discharge over the dome. Such a device was described
by Peterson et a1. (1973) and similar ones are used in the US, Canadian and
some European networks. Care must be exercised, of course, to keep reflecting
surfaces below t.he plane of the sensor. The strength of the air £Iow required
depends on environmental conditions, and a certain amount of trial and error
at any particular location may be necessary. The temperature of the airstream
should be only a little higher than that of the ambient air to prevent con-
densation of dew or frost on the dome, but prevent measurement errors due to
artifical heating.

One problem frequently encountered in pyranometer operation is the. con-

densation of moisture inside the instrument itself. The' normal manifestation
of this trouble is an obvious moisture deposit on the inside of the optical
dome, with a consequent decrease of signal from the instrument. Either one or
two cures m{ty be effective, depending on the seriousness o.f the problem.
Most pyranometers are fitted with a small chamber filled with a desiccant,
and if the desiccant loses its effectiveness it may build, up moisture. If a
desiccant is provided, it is normany visible from the outsid.e of the instru-
ment, and' when fresh it is a deep blue in colour. A change of colour to very
pale blue or pink means that the material is no longer effective, and it
should be replaced. Bxcept in unusual circumstances a renewal of the desic-
cant should cure the condensation problem.' In severe cases, however, it m.ay
be necessary to remove the cover, dry out the various par·ts, renew the de.sic-
cant, and reassemble the instrument ina.s dry conditions as available. In
case of deteriora.tion of the seals, sensor, or other components, the in-
strument should be returned to the factory for refurbishing and recalibra-
tion.

4.2.2.4 Calibration of Pyranometers

Pyranometers should be calibrated at lea.st once per year, and more fre-
quently if there are reasons to sus.pect unusual changes of instrument perfor-
mance. The most precise calibration, and the one g.enerally recommended, is
provided by anyone of several national or regional radiation centres which
have special equipment and expertise in radiometer calibrations. This way of
calibration, of course, requires transport of the instrument to and from the
centre, and its use in the field is interrupted for a period of time. A
second, more appropriate method is to transport a secondary standard pyra-
nometer to the field for a direct comparison with the operational instrument.
In this method the secondary standard is installed alongside the operational
instrument and a statistical sample of their records is obtained under a
variety of radiation conditions, including clear and cloudy skies, over a
period of about two weeks. As for the case of calibration of pyrheliometers"
a relatively large number (at least 100) of simultaneous samples of thesig-
nals Ai and A2 aTe- taken, and the average R of the individual ratios A2 /Ai
computed. If Ci and, C2 are the calibration factors of the operational and
standard pyranometers, respectively, then we determine the calibration factor
for the operational instrument as Ct = C2 • R. This should be checked
against previous calibration factors for the instrument, and any discrepan-
cies not,ed. In case there has been a significant change since the previous
calibration, the process should be repeated within a. feww'eeks to detect any
progressive deterioration of the instrument. If such deterioration conti-
nues, the instrument should be replaced and a new calibration performed.

A third method of checking pyranometer calibrations, and the one re-

corn mended for establishing the calibration of secOndary standards, is by the
use of a pyrheliometer and an additional. pyranometer shaded from the direct
s'olar beam with a disc. Simultaneous measurements are made with the shaded
pyranometer, the pyrheliometer and the pyranometer under test. From the read-

70
CHAPTER 4 RADIATION MEASUREMENTS

ings of the shaded pyranometer Vs and the pyrheliometer Vp and their calibra-
tion factors Cs and Cp the global radiation incident on a horizontal surface
can be calculated as a function of ro the solar elevation by:

(4.9)

From this value and the reading Vc of the pyranometer under test its calibra-
tion factor can be calculated:

Cc = E';'/V c . (4.10)

Details of this procedure are described in WMO Guide to Meteorological In-

struments and Observing Methods (1977).

4.2.2.5 Measurement of Diffuse Radiation

In order to measure the radiation incident at the surface due to the

sca t tering of sunlight by the at mosphere (the diffuse component of global
radiation) with a pyranometer, it is necessary to shade the sensor from the
direct solar beam. This may be done by either of two different methods. The
more traditional method has been to utilize a shadow ring positioned over the
pyranometer and oriented parallel to the ecliptic so that the sensor is kept
in shadow throughout the day. A general discussion of shadow rings is given
by Coulson (1975). The shadow ring is relatively inexpensive and simple to
install, but its application suffers from the serious disadvantage of requir-
ing a correction to the data due to the portion of the diffuse sky light
itself intercepted by the ring. Furthermore, the magnitude of the correction
varies strongly with several quantities, such as the geometrical dimensions
of the ring, the sun elevation, atmospheric turbidity and cloudiness, albedo
of the surrounding surfaces, and reflectivity of the inner surface of the
ring.

Several authors have proposed formulae for operational corrections to

the sky radiation taking into account the part not measured due to the shade
ring _(Drum mond, 1956; Dehne, 1980; Le Baron, 1980). For a ring with b/r<0.2
the radiation Dv lost during a day can be expressed as:

Dv ~ - cosJ~·f ED~(t)'sinr(t)dt (4.11)

where ~ :declination of the sun

t :hour angle of the sun, trlu, tut the hour angle at
sunrise and sunset respectively
ED Ht) :sky radiance during the day
r :solar elevation
b/r :ratio' of the width b to the radius r of the shadow ring,

With this expression and some assumptions on the sky radiance, a correction
factor f can be determined by:

f = (1 - Dv I ED ~ ) - t (4.12)

D being the unobscured-sky radiation. In Fig. 4.16 an example of this

correction factor is given for a clear and an overcast sky, compared with the
corresponding empirical curves. It is evident that deviations from the theo-
retical curves depend on climatological factors of the station and should be
determined experimentally by comparing the instrument having a shading ring
with an instrument shaded by a continuously tracked disk. If no experimental
data are available for the station, data computed for the overcast case with
the corresponding b/r should be used. Thus

(Dv lED .j.)Dvmast = (b/r) 'cos J ~(t541t-trISQ)sin~sin~+

cos~cos~' (si ntut-si nt rls.,) , (4.13)
where ~ is the geographic latitude.

71
CHAPTER 4 RADIATION MEASUREMENTS

1.15 ; - - " ' - - - - - L - - -_ _---IL- ----l._----l..--ri-

...0 1.10
u
-IV

-......
0
u
11 1.05
0
()

1. 00 +--r-------r----~---__,...----_.__--+
-23.5°.20°
Declination

Figure 4.16 Comparison of calculated and empirically determined correction

factors for a shading ring with b/r=0.169. f and F indicate calculated
and empirical curves, respectively CDehne, 1980).

A much more satisfactory shading device is a small opaque disk mounted

on a slender carrier arm and made to follow the apparent movement of the sun
by a clock-driven mechanism of some kind. The disk should have a blackened
surface and be of such size as just to shade the glass domes of the pyrano-
meter, with a few millimetres tolerance. It should be positioned on the car-
rier arm at the correct distance from the sensor to subtend the same solid
angle as that viewed by the calibration pyrheliometer. If a is the half-angle
of a circular cone subtended at the sensor by a circular disk of radius r,
the required distance d between sensor and disk is given by d = r ctg a. For
instance, for a 2.5 half-angle cone, a com mon field of view for pyrheliome-
0

ters, a disk of 3.5-cm radius would have to be approximately 80 cm from the

sensor.
The positioning of a shading disk on a clock-driven mount for shading a
pyranometer is a problem in geometry which can be frustrating unless it is
approached properly. As with the pyrheliometer orientation discussed above,
the rotating axis of the clock-driven mount is termed the hour-angle axis.
The principal requirements are to get the hour-angle axis parallel to the
rotational axis of the earth and a line extending along the hour-angle axis
to pass through the centre of the sensor. In some commercially available
units, this latter requirement is incorporated into the design of the mount,
a feature which simplifies the alignment problem. The alignment of the in-
stallation on site is analogous to the one described for the pyrheliometer
tracker (4.2.1.5).

For construction of the pyranometer mount, the following has to be kept

in mind: The pyranometer has to be mounted on a separate structure which is
firmly secured to that of the tracker. This structure should be located in
the same north-south plane as the hour angle axis and have an arm extending
away fro'm the pedestal toward the tracker far enough to permit the disk-
carrier arm to make a complete circuit around the pyranometer. For most
convenient positioning of the pyranometer, the height of the extended arm
should be adjustable.
The horizontal and vertical positions of the pyranometer have to be
adjusted in such a way that aline extended along the hour-angle axis of the
tracker passes through the centre of the sensor and the sensor is at the
centre of curvature of tpe disk-carrier arm. This assures that the distance

72
CHAPTER 4 RADIATION MEASUREMENTS

between disk and sensor will be constant, thereby making the solid angle
subtended by the disk independent of solar declination.

4.2.2.6 Measurement of Solar Radiation on Inclined Surfaces

The increase of activity in utilizing the sun· as an alternative energy
source has resulted in the need to measure solar radiation incident on sur-
faces of different orientations. Although such a measurement is conceptually
very simple, there are two potential problems in using pyranometers mounted
at angles other than horizontal. First, the cosine response errors of the
sensor may be more pronounced for the tilted position than the normal orien-
tation, although this factor will obviously depend on tilt angle and lati-
tude. The second problem, that of apparent changes of sensitivity with angle
of tilt, is more critical for most installations than is the cosine response.
Tests performed by Flowers (1978) and by lEA Task III (1985) on a group of
pyranometers from different manufacturers showed a great variability of the
tilt effect, both among instruments from different manufacturers and dif-
ferent instruments from the same manufacturer. The tilt effect also depended
on the season prevailing during the test. Generally, it was found that the
errors produced by instrument tilt were not more than about 2 per cent for
most instruments, and the error could be in either the positive or negative
direction. In extreme cases, however, the error amounted to as much as 15 per
cent. For instruments showing such extremes, the error was always in the
direction of decreased sensitivity, and it reached its maximum value at tilt
angles of 60 to 70· from the horizontal.

For practical reason, the best approach when choosing a pyranometer for
use in a tilted position would seem to consider the problem and calibrate the
instrument in the tilted position with a shading technique.

An alternative method of acquiring data for insolation on a tilted sur-

face is to utilize measurements for a horizontal surface together with theo-
retical or empirical models to compute that for the tilted surface. A number
of such models, with varying degrees of sophistication, have been developed
(e.g.. Liu and Jordan, 1962; Klein, 1977; Temps and Coulson, 1977; Dav~ 1978;
Beckman et al., 1977). The majority of such models give approximations to
average energy fluxes over an hour, day, or month. They are very useful for
many engineering design studies of solar-energy systems, but it is clear that
for precise information actual measurements are needed.

4.2.2.7 Special Pyranometers

In recent years &everal attempts have been made at Eppley (Karoli et
al., 1983), TMI (Reid et al., 1978) and PMOD to design self-calibrating cavi-
ty pyranometers. The detectors are very similar to the ones used in absolute
radiometers described in 4.2.1.2. The main difference is the shape of the
cavity and the position of the heater which has to be optimated for the hemi-
spheric acceptance of a pyranometer. Thus, the non-equivalence of such an
instrument becomes more important than for a pyrheliometer and most likely
limits the absolute accuracy to about ±0.5 per cent. However, if a good co-
sine response and overall system linearity can be achieved, such instruments
will substantially simplify the calibration of operational pyranometers and
improve the reliability of the international network.

Instruments with silicon detectors have been used in order to provide

simple, rugged, and inexpensive pyranometers. The main difference between
such an instrument and a "real" pyranometer is its spectral response. If the
detector is unfiltered, the peak response is at about 900 nm, decreasing to
50 per cent of that value at 400 nm and 1000 nm, respectively. It is obvious
that this spectral weighting introduces systematic differences, arising from
the state of the atmosphere; the solar elevation, etc. It may be that daily
sums are quite representative, if they are normalized using a pyranometer
73
CHAPTER 4 RADIATION MEASUREMENTS

calibrated und&r average climatic conditions at th~ station where the instru-
ment will be used. But this is very difficult to quantify and, therefore,
these instruments are not· recom mended for accurate climatological measure-
ments.

4.2.3 PyrradiolReters and Pyrgeometers

Both types of instruments are for the measurement of long-wave, Le.

terrestrial or atmospheric radiation. A pyrradiometeris sensitive to radia-
tion of wavelengths from 0.3 to 60 ·J./m, whereas a pyrgeometer filters the
solar radiation and is only sensitive to terrestrial radiation ("'2 - 60 J./m).
As in case of the pyranometer the angular acceptance is hemispheric.

The problems with the measurement of long-wave radiation' are somewhat

more involved than those encountered with solar radiation. Although the
measurement principle is also based on a heat-flow transducer, e.g. a thermo-
pile, the protection of the sensor from the environment is more difficult,
because it has to be transparent for wavelengths of about 2 - 60 J./m. Fur-
thermore, the absorptance of the sensor itself has to be flat over a very
wide wavelength range of about a factor of 30 (2~60J./m), whereas the solar
radiation only covers a single decade range <O.3-3.0J./m). Thus the selection
of the mater.ials to be used for long-wave radiation measurements is much more
difficult.

When interest is mainly in the net radiation, Le. the difference be-
tween the incoming and the outgoing radiation fluxes, rather than in the two
quantiti&s themselves~ one heat-flow transducer can b& used by exposing both
sides, the "cold" and the "hot", to the radiation. In this case the instru-
ment is called a net pyrradiometer.

4.2.3.1 Most Widely Used Pyrradiometers

The instruments in use are of two different types: the unshielded and
the shielded types. It is obvious that the un shielded radiometer has an
inherently better optical performance, but as the sensors are bas&d on ther-
mal heat-flow meters, varying losses to the surrounding air introduces er-
rors. Courvoisier (1950) and others (e.g. Geir and Dunkle, 1951) tried to
overcome this problem by blowing an "air curtain" of constant speed over the
sensors. Although the influence of wind is drastically reduced, it is not
eliminated. The construction of such devices is rather involved and they are
normally not used for operational measurements, but as a reference for cali-
bration. The Yanishevsky net pyrradiometer (Fig. 4.17a), which is also un-
shielded, but not artificially ventilated, uses individually determined cor-
rection factors for the influence of wind. These typically range from about
1.03 at lms- i to 1.22 at 10ms- i • The sensing element is a thermopile of
silver and constantan which consists of ten sections connected in series. One
section consists of a rectangular copper bar (Fig. 4.17b) around which a thin
ribbon of constantan is wound over a thin insulating layer. One half of each
loop is electroplated with silver. Thus the junctions between the silvered
and the non-silvered parts, the hot and cold junctions, are situated on oppo-
si te sides of the bar.

Ten bars - side by side - form the complete thermopile. A copper foil is
glued onto each of the two surfaces with an insulating spacer and painted mat
black. The sensivity is about 12 J./V/Wm- 2 and the temperature difference
between the hot and cold junctions at 1 kW/m- 2 is only about lK. Under' such
conditions the infrared radiation of both sides of the receiver can be consi-
dered the same. It is obvious that this instrument can only be used occasio-
nally and not continuously. Th& wavelength dependence of the absorptance of
the black coating is one of the sources of errors of the unshielded instru-
ments. In case of most coatings, the absorptance decreases with increasing
wavelength and may eveD have some line features at wavelengths where the

74
CHAPTER 4 RADIATION MEASUREMENTS

thickness of the coating is comparable to multiples of a quarter of the

wavelength. These effects are very difficult to correct by calibration or in
the evaluation procedure and may give rise to systematic errors of up to
several per cent because of the complicated structure of the atmospheric
infrared spectrum.

a) b)
n

Figure 4.17 a) External view of the balance meter M-l0M. 1: receiving sur-
face, 2: frame, 3: knob, 4: case. b) One section of the thermopile of
t.he balance meter M-l0M. 1: constantan strip, 2: copper bar, 3: electric
insulator, 4: silvered sections

From the above it is obvious that shielding of the instrument is needed,

if it must be exposed continuously. Schulze (1953) replaced the glass domes
of pyranometers by polyet.hylene domes and thus expanded the spectral range of
the instrument to about 60 ,urn. This material has a good transparency over
the whole spectrum apart from some narrow absorption features due to it.s

Figure 4.18 Photograph of the Swissteco net pyranometer Type S-1.

75
CHJ'lPTER 4 RADIATION MEASUREMENTS

organic structure. As in the case· of"" the black coating, this introduces
further uncertainties which are difficult .to quantify. In order to minimize
this effect very thin polye.thylene sheets have to be used ~h:i.ch are no longer
self-supporting. Therefore the domes have to be ~ept in shape by pressurizing
the interior. The sa me technique is used for the net pyrradiometer, developed
by Funk (959) and manufactured by Middleton, Australia, and Swissteco,
Switzerland. 'An exanrPle of such an instrument is shown in Fig. 4.18 and ,the
specifications of the-'two types are summarized in Table 4.5. ,.

Table 4.5 Specification of Funk type pyrradiometers and the Eppley pyrgeo-
'-meter'

Middleton Swissteco Eppley

,

SensH i vi ty 40JlV IWm- 2 " SJlV/Wm- 2 6J1V/Wm- 2

Time constant 15 s 6 s 2 S
(1 le)

Temp. coefficient
of sensitivit.y )+0.012%/K +1% for -20 to 40· C

Linearit.y <1% <1% <1%

-"
Wavelength range 0.3 to 60 J.im 0.3 t.o 60 Jlm 4 to 50 J.im
-

The main disadvantage of the un shielded and polyethylene-shielded in-

struments is that. their sensivity is not' limited to the infrared, but in-
cludes the whole shortwave region. Thus, the pyrradiometer can be used at
night only'when no shortwave radiation is present or it can be used together
with a pyranometer to determine the short wave part during the daytime. In
t.hat case pyranomet.ri8 values must be subt.racted from the reading of the
pyrradiometer. This lat.ter method may int.roduce significant. errors. due t.o
uncertainty of the calibrat.ion of both type of instrument.s and due to pos-
sible wavelength dependence of t.he response in t.he infrared, "especially, if
t.he calibration of each instrument is performed in the presence of short-wave
radiation (e.g. the sun).

Fiqure -4. t 9' ·Pyrgeometer. with Silicon Dome" from Eppley.

This problem can be overcome by using "domes made of material which is

opaque in·, the short.-wave region. Silicon is such a mat.erial and. combined
,with a vacuum-deposited int.erference filter, such domes can be manufactured

76
CHAPTER 4 RADIATION MEASUREMENTS

to have a nearly grey transmission from 4 to 50 ~m. These domes are used in
the Eppley pyrgeometer shown in Fig. 4.19 The sensor and design of the in-
strument is the same as for the Precision-Spectral Pyranometer and the gene-
ral specifications of the pyrgeometer are listed in Table 4.5. In contrast to
the net pyrradiometers a single instrument needs to take its own infrared
emission into account. The corresponding signal can be calculated from the
instrument's temperature according to the Stefan - Boltzmann law or is, as in
the Eppley pyrgeometer, automatically eliminated from the signal by applying
a battery voltage which is precisely controlled by a thermostat sensing the
temperature bf the instrument.

4.2.3.2 Maintenance and Calibration of Pyrradiometers

For operational purposes, none of the existing pyrradiometers or net

pyrradiometers is without problems, but most will give reliable data, if
properly installed and maintained. The most important operational tasks with
instruments with unshielded sensors are to keep the sensing surfaces clean
and dry, to renew the surface coatings whenever degradation is apparent, and
to calibrate them semiannually. For instruments with shielded sensors, the
domes should be kept clean, dry, and distended when necessary, and at least
annual calibrations should be performed.

The simplest method of checking the calibration of a pyrradiometer which

responds to solar radiation is by the sun-and-shade method already described
for the pyranometer (see 4.2.2.5).

A better method of calibrating a pyrradiometer is by comparing its re-

sponse with that ofa secondary standard instrument. This method also is the
equivalent of that for pyranometer calibration described in 4.2.2.5. The
most accurate method, which has the added advantage of being an absolute
calibration,_ is to use a special calibration chamber. Such a chamber was
described by MacDowell (1955), and modified forms of it were developed by
Fritschen (1963). The theory of such calibrations was given by Coulson
(1975).

A final and very si mpIe met hod, s u g ges ted by Id so (1971), utilizes a
large temperature-controlled flat black plate. The sensing element is posi-
tioned directly over or under the plate. If surrounding laboratory condi-
tions are constant, then the net irradiance at the sensor is E =Ee - ~€aT~.
Here Ec is the irradiance (considered constant) from the surroundings, € is
the plate emittance, a is the Stefan-Boltzmann constant, T is plate tempera-
ture, and ~ is the angle factor given by Idso (1971). The output signal V is
V = C~ - K'€a~, where C and K are the calibration factors for the radia-
tion from the room and from the plate, respectively. In performing the cali-
bration, values of V ~re obtained as a function of T. The value of K is
given by the slope of the curve of V plotted versus the quantity ~€aT~. Idso
estimated the accuracy obtainable by the method to be ±5 to ±3 per cent.

4.3 TECHNIQUES FOR THE MEASUREMENTS OF SPECTRAL RAD'IATION

Operational spectral measurements in meteorology started with coloured

glass filters in front of thermopile instruments. The main reason for using
this type of fH ter was the need for a high throughput, that is, a wide
wavelength band, due to the limited sensitivity of the detectors. These
filter.s were mainly used in conjunction with pyrheliometers and as
pyranometers domes.

-The first quantum detectors in use were the cadmium cells for the mea-
surement of ultraviolet radiation. As the operation was very cumbersome, they

77
CHAPTER 4 RADIATION MEASURRMENTS

were never used operationally. Selenium cells were later introduced for de-
termination of the spectral extinction of the direct solar beam. The modern
sunphotometers were developed from these later instruments.

4',3.1 Spectral Keasurements with Pyrheliometers and Pyranometers

The traditional method of making spectral measurements of direct

solar radiation with a pyrheliometer is by means of coloured glass or
broad-band absorption filters. The filters normally used are cut-off
filters manufactured by Schott. The transmission characteristics of the
most common ones are given in Table 4.6, also included are the wave'-
lengths of the short wave cut-off. One problem with these filters is their
calibration. It does not depend so much on the physics involved, but rather
on the answer sought from the measurements. An absolutely calibrated ra-
diometer gives irradiance in SI-units (e.g. Wm- 2 ). If the transmittance of
the fi1ter.over the whole bandpass is one, the result of the measurement of
the ra.diometerwith the filter in front is jus~t the incoming radiation in the

Table 4.6 Transmittance values at various wavelengths for Schott glas's fil-'
ters (2 mm thickness, +25'C) as given by the Eppley Laboratory, Inc.
(CC: 50 per cent point of cutoff; FF: filter factor; wavelength in #m)

WG29'5 ecc: .281; FF: 1:.09), Original designation WG7

.27 .28 .29 .30 .32 .34 .36-.44 A6-1,5 2.0' 2.5
.22 ,41 .60 .74 .83 .90 .91 .92 .91 ,85

G'G395 (CC: .385; FF: 1. n ) Original designation GG22

X ,36 .3'8 .40 .42 .44' .46 .48 .50 .52-1.5 2.0 2.5
T • 0'1 • 31 • 68 .82 • 87 • 88 . 88 • 89 • 90 • 89 .86

GG40'lJ (CC: 397; FF. 1.10)

.38 .40 .42 .44 .46 .48 ·50-.64 .66-1.5 2.0 2,5
.01 .52 .78 .83 .87 .89 .90 .91 .90 .85

GG495 (CC: 485; FF: 1. HI) Original designation GG1:4

X .46 .48 .50 .52 .5,4 .56 '.58-1.5 2.0 2.5
T •. 01 .27 • Ba .88 .90 .90 .91 .9'0 .85

OG53'Q (CC: .529; F'F: 1.0,9) Original designation OGl

X .52 .54 .5-6 .58 .. 60-.74 .76,-1.5 2.Q 2.5
r. • 0'5 • 76 • 88 • 90 • 91 . lJ2 • 90 .87

OGS7'fr (CC: .562; FF 1.09) Original designation OGl:

X .56 .58 .60 .62-1.5 2.0 2.5
"[ .. 38 .a3 . 89 .,91 .90 .86

RG;61U (CC: 603; FF: 1.10) Original designation RGl

X .60 .62 .6,4 .66 .Gc8-2.0 2.5
T "J3 .83 ,89 .90 .91 ' .87

RG6JO (CC: .623; FF: 1.0'9) Original designation RG2

A .62 .64 .66-.72. .74-1,52.0 2.5
"[ . 26 • 85 .,9'1 • 9'2 • 91 • 87

RG.695 (GC: .686; FF: 1.09) Original designation RG8

k .,6,8 .,70' .72 .74-1.5 2.0' 2.,5
"[ .27 .84 .91! .92 .9'1.87

RG715 (CG: .707; FF: 1.09) Original designation RG10

A .68 .70 .72 .74 36 .78-1.5 2.0 2.5,
T .EH .2.5 .74 .89 .91 .92 .91 .87
CHAPTER 4 RADIATION MEASUREMENTS

wavelength band covered by the filter in Wm- 2 • If the cut-offs are less
steep and the transmittance is lower than one, or even varies in the
transmitting band, the interpretation of the reading is more difficult. Be-
cause the result will depend on the spectral distribution of the incoming
radiation, the so-called filter factor, FF, has been introduced. This compen-
sates for the attenuation of radiation in the transmission band. The FF of
a specific filter can be calculated by convoluting the spectral
transmission of the filter with the spectral irradiance of the incoming
solar radiation and comparing the result with that found with an ideal fil-
ter. Such a FF, however, will obviously depend on the state of the atmo-
sphere as does the spectral distribution. But if a typical condition is cho-
sen for the calculation, the FF may be representative for some range of
conditions as long as the variation is small enough for the purpose of
the measurements. The first FF was calculated as described above, by using
measured filter transmissions and a typical solar spectrum as mea-
sured, for instance, by the Smithsonian Institution from mountain tops in
their programme to determine the solar constant. This is a very cumbersome
way to calibrate individual filters. The FP's were therefore normally deter-
mined by comparison with standard filters, which had their FF's derived
from calculations. This method was applied at the Davos Observatory for
many filters as a service. One severe drawback to this method is that
the intrinsic transmittance curve of the filters to be calibrated may be
different from those of the standard. This may happen if the filters are from
different batches, e.g. the standard filters, manufactured before
World War Il, have a substantially different transmittance towards the in-
frared cut-off than those delivered later. But they have still been used for
calibrations. Thus most of the more recent FP's should be applied with some
caution and a calculated redetermination should be made from the individua-I
transmittance curve and a representative- terrestrial solar spectrum. Another
problem is the shift of the short-wave cut-off with the temperature
of the filter. The FF changes by 0.02 %/K for OG530, by 0.03 for RG630 and
by 0.04 for .RG695 (e.g. Angstrom and Drummond, 1961). As it is very diffi-
cuI t to measure the temperature of a filter accurately, corrections are
difficult to apply.

The situation is even more complicated if the cut-off glass filters

are used a.s domes to filter the input to the pyranometer. First, the spec-
tral distribution of the global radiation from the direct sun and the sky
is, in- the short-wavelength range of the cut-off, even more variable than
the direct sun alone. Second, the transmission of a spherical dome onto
an extended detector depends on the direction of the incoming radiation~ as
the path length through the dome changes. Therefore the· determination of-an
equivalent FF -is even more difficult than for the filter used ~ith a
pyrheliometer and the evaluation and interpTetation of such measurements is
somewha t questionable.

4.3.2 Sunphotometers

A sunphotometer is a simple instrument for the measurement of solar

spectral - irradiance mainly used to determine the spectral optical depth of
the atmosphere •. It consists of a photo-electric detector - formerly of sele-
nium, but presently of silicon and a filter.of more or less narrow-band, cur-
rently of the interference type. The first sunphotometerswere designed . and
used. by Volz (1959). The World Meteorological Organization has- issued recom-
mendations for the w_avelengths to be used and the performance to be met for
different kind of uses. The recom mended wavelengths are chosen according to
the f ollowi ng criteria:

- . no interference of gaseous absorption should be present, for the

determination of the aerosol optical depth,

- the place in ,the· solar spectrum should be in a local minimum or

maximum in order to minimize the effect of changes of the wavelength due

79
CHAPTER 4' RADIATION MBASUREMEN'TS

to· temperature effects,

- due to these rather tight specifications the centre wavelength of the

filters should be accurate to within 2 nm.

The band-width of the filters is specified to be 5 nm for high-

accuracy measurements and not more than 10 nm for less demanding applica-
tions. An important requirement is the out-of-band rejection, that is the
ratio of the signal in the band to the signal leaking through at wavelengths
where the transmission should be zero. For a 368-nm channel with a
peak transmission oE 40 per cent and an out-oE-band transmittance as low
as 10- 4 the out-of-band signal is still about 0.1 Wm- 2 compared to the
real signal of about 1.5 Wm- 2. Thus the signal-to-noise ratio is only
15:1. The out-oE-band transmittance for reliable results m;ust be at least one
order of magnitude lower. The situation gets even . ~orse for large air
masses as the peak of the solar spectrum is shifted towards the red and the
signal in the blue becomes very weak. Due to this very demanding specifi-
cation it is difficult to get good Eilters for the blue, and below about 350
nm it is nearly impossible even. with the most advanced techniques to
manufacture interference filters which achieve a reasonably high bandpass
transmission and still a sufficient out-of-hand rejection. As it is easier
to block the short~wave side of the filter than the long-wave side, the
blue filters have potentially . more problems than the red ones and should be
tested more carefully.

WMO recommends a mlnlmum of three wavelength intervals centered at 368;

SUO and 778 nm, in the following referred to as channels. These channels
are used for the determination of the aerosol optical depth, although
the SOO-nm channel is slightly influenced by ozone and NOi may have an
effect on the 368-nm channel. The optical depth of ozone (0.3 cm STP) is
0.011 for the 500-nm band and its real concentration can be deduced from
climatological values with sufficient accuracy. A N0 2 concentration of 10- 3
cm STP, found only in highly polluted areas, yields an optical depth of
0.03 in the 368 nm band. The 778-nm channel is between the oxygen and water-
vapour bands and is not influenced by gaseous absorption. Other infrared
channels barely influenced by water-vapour absorption are at 862 and 1024 nm.
The 1024-nm channel is not recommended with silicon cells as detectors be-
cause of the strong temperature coefficient of these detectors at this wave-
length.

Measurements at 610 and 675 nm can be used together with the 500-nm
channel for the in situ determination of ozone from absorption in the Ghapuis
hand, To improve the information in ozone-free spectral regions, channels at
412 nm and 450 nm may be added. These channels may be used as reference for
N02 determinations and together with the 778-nm channel as reference for
ozone determination. Channels at 719, 817 and 946 nm have been used to
determine the columnar content of water vapour by comparing their optical
depth with those at the 675-, 778-, 862- and 1024-nm channels (e.g. Koepke
and Quenzel, 1978).

A fully equiped set-up would therefore have twelve channels, but if only
aerosol optical depth is sought, the three mandatory WMO-channels are
sufficient. Table 4.7 lists all the wavelengths which can be used in sunpho~
tometry available with choices for different applications and possible
combinations for three-, six- and nine-channel instruments. Fig. 4.20 shows
the extraterrestrial spectrum together with the distribution of the channels
and a curve indicating the sig.nal received from a typical silicon cell.

One of the most important performance specifications for sunphotometers

is the demand for high temporal and operational stability. Silicon detec-
tors operated in current mode have a positive t~mperature dependence of up
to 0.3%/K at 1030 nm, an almost zero dependence at the central visible
wavelengths and a small negative dependence of about O.l%/K in the ultravio-
let. It is therefore aqvisable to either measure the temperature of the

Sll
CHAPTER 4 RADIATION MEASUREMENTS

detector and apply a correction or thermostatically control the detector, or

better, the detector-filter combination. The first method is used most in
simple, easy-to-use hand-held instruments, the second in instruments for
high precision work and calibrations. An example of an instrument based on
the first method, developed by d'Almeida et al. (1983), is shown in
Fig. 4.21 and demonstrates the simplicity of the instrumerit, although it is
a twelve-channel version. Fig. 4.22 displays the more sophisticated ver-
sion of a temperature-stabilized triple-sunphotometerdeveloped by PMOD/WRC.
Instruments with state-of-the-art high-quality filters, detectors and elec-
tronic components have proven to be stable within about 1 per cent at 368 nm
and within 0.2 per cent at 778 nm over the period of several years. This is
a substantial improvement over the instruments available about ten years
ago.

111
N
r--------------------------,~

rl
E
c
IS)

N
,..,
a:
"<
(\j

E
--
111
:J
L..J

111

"
~
L..J
IS)
+'
C
ID
L
ID L
tJ IS)
:J
C tJ
I'd 0
+'
"U 0
Itl ..!:
L
L
a...
H
ISl
ISl
III -
ISl
ID
U1
J'.
ID - -
lTl
J'.
aI
J'.
J'.
J'.
aI
N
ID
aI
ID
'Of"
lTl

300 500 700 900 1100

Wavelength [nm]

Figure 4.20 Extra terrestrial spectrum, the corresponding silicon-detector

signal and the wavelength for the twelve sunphotometer channels.

Table 4.7 Wavelengths used for sunphotometry.

Channel wavelengths (nm)

Application 368 412 450 500 610 675 719 778 817 862 946 1024

Aerosols
minimum
extended * * *
Ozone * * * * * *
Water vapour
* * * * *
weak * * *
* *
strong
Nitrous Oxide
* * *
* * *
3 Channels
6 Channels * * *
9 Channels
* * * * * *
12 Channels
* * * * * * * * *
* * * * * * * * * * * *
81
CHAPTER 4 RkDIATIOW MEASUREMENTS

Figure 4.21 Hand-held Sunphotometer developed by d'Almeida et a1. (1983).

A further· specification required is that involving the view-limiting

geometry, which may influence the measurements. A big acceptance angle is
liable to yield high signals during conditions of high turbidity and hence
high aureole intensities. Thus" the calculated optical depth may be underesti-
mated substa-ntially (see e;g. Frohlich and Quenzel. -1973). A good compro-
mise is a full-view angle of 2.5 and a slope of 1 (for definition of these
0 0

angles see Fig. A.4Y.

Sunphotometers were first manufactured by Volz in h~s basement. A se-

cond source for the instrument was a set of some fifty instruments made by
Eppley Laboratory for a us turbidity n~twork operated by the Environmental
Protection Agency. These intruments, which used Selenium cells and filters
built up from a combinatiotl of coloured glass and Wratten filters with a
band-width ranging from 10 to - 50 nin, are no longer available. As the WMO
need for sunphotometers to be used at the Background Airpollution Monito-
ring (BAPMon) stations of the United Nations Environmental Programme (UNEP)
network increased continuously, other manufacturers started to build sun.

Shutter Interference Filter

(1 of 3) (1 of 3) Si-Diode (1 of 3)

/ /29405 /
/ //
~ Electronics

III

,''''-
Shutter Motor Hea~Wln~
(1 of 3) Temperature Sensor

Figure" 4.22 Schematic drawing of the PMOD Sunphotometer-detector.

82
CHAPTER 4 RADIATION MEASUREMENTS

photometers and WMO issued building-instructions for interested Institutes

(WMO, 1979). These new sunphotometers are based on silicon diodes as de-
tectors and thin-film dielectric interference filters for selection of the
wavelengths. Today, EKO Trading, Tokio, Japan and NolI GmbH, Undenheim Fede-
ral Republic of Germany, manufacture multichannel sunphotometers. Plans also
exist for manufacturing high-precision sunphotometers of the type developed
at the PMOD/WRC, Davos, Switzerland.

Although calibration procedures are described below, some remarks are

appropriate at this point. Spectral sensors can be calibrated in terms of
absolute units using a standard source, such as a black-body radiator of
known temperature or a calibrated lamp. Or they can be calibrated by
using a variable monochromatic source from a laser, the power of which
is measured with a reference detector such as an absolute radiometer or an
absolute silicon detector. The calibration result is irradiance per unit
output, e.g. Wm- 2 IV for the whole transmission band for the case of a
lamp or at many wavelengths in the band for the calibration by the
laser-radiometer method. In the case of calibration with a lamp, the indivi-
dual form of the transmission curve can no longer be taken into account and
the value is actually valid only for the type of spectral distribution of the
source used in the calibration. This is not the case for calibration with a
variable monochromatic source. In that case the sensitivity is deter-
mined as a function of wavelength and can later be convoluted with any
spectrum. However, due to the relatively small band-widths used in sun-
photometry, the difference between a measurement with a lamp and the sun may
be small enough to allow the use of lamps to test the long-term stability
of these instruments.

4.3.3 Calibration of Sunphotometers

Sunphotometers may be calibrated either directly by the so-called

Langley method or by comparison with a standard instrument. The latter
method is straightforward as long as the centre wavelength of the filter and
its band width are comparable. For sunphotometers complying with WMO stand-
ards this is the case and a few comparisons at different air masses will
suffice to determine the V. value from the V. of the standard instrument and
the mean of the ratios of the readings of the two instruments. If the centre
wavelength of the two instruments differs significantly, corrections have to
be applied based on knowledge of the wavelength dependence of the atmospheric
transmisson. As a first approximation Rayleigh extinction may be assumed
(Eq. 3.12).

The Langley method is based on the fact that the logarithm of the irra-
diance is a linear function of the air mass (see also Eq. 3.15). Thus V. can
be obtained by extrapolating the values at different air masses to air mass
zero. This method, although very simple in principle, relies on the stabili-
ty of the atmosphere during at least half the day, which is very difficult to
check. Thus, this method can only be applied from high-altitude stations and
under conditions with very clear sky. Moreover, a great deal of experience is
needed to judge the quality of the results obtained even under more or less
ideal condition s.

In recent years much progress has been made in sunphotometry and its
calibration from high-altitude balloons and rockets (Brusa et aI., 1983) so
that reliable reference instruments which are directly calibrated with extra-
terrestrial sunlight may be available in the near future. In any case, it is
recom mended that comparisons be made with a stable reference instrument to
standardize field instruments so that the V. can be traced back, whenever a
reference of higher accuracy becomes available.

83
CHAPTER 4 RADIATION HEASUREMENTS

4.3.4 Measurement of Ultraviolet Radiation

The mea.surement of ultraviolet (UV)radiation at the ground is ,go-

verned by the steep decrease of the radiation below 300 nm,due to the
absorption by ozone. The UV is divided into regions of different biol.o-
gical importance: UV-A, 320 - 400 urn; UV-B, 280 - 320 nm; and UV-C, below
280 nm. The UV-B is the most important to be monitored, as it is very sen-
sitive to changes in the ozone content. Twodifferettt types of instrument
exist for thi.spurpose and these can he classified by their spectral
response.as ~f.ollows:

- the respon:se is in accordance with a speciftc action curve, such as,

forexa'm.ple, t.heerythemal resp.onse of the human skin (sunhurning:ef-
fect,seealso Fig.3.2 in Chapter 3), or

- the re.!:!])O'nse divides the UV"'Bregioninto several more or less narrow

band.s,andme.asures the actual .energy in each band.

It is difficult to obtain a relatively good re.sponse curve, since the uncer-

tainty of the weighted-spectrum instrufltents tsa function of the .spectral
distribution, which m.ain1ydepenrls on the solar elevation .and the ozone con-
.tent. In the 'second case the accuracy of the results greatly depends on the
precision of spectral selection and, considering the steep slope of the
terrestrial solar spectrum in this spectral rang,e, on the shape and the band-
width of the filter functions or on the resolution 'of themonochromator,
respectively. At 300nm the spectral width of the filters or thespectrome-
tershoul'd notexce:ed 1 nm.

Many instruments belonging to the first group were proposed .and

some were built in the past. They are :b:a'sed on chemical effects, fil-
ter-detector combinations or 'on fluorescent coatings. ChemicaJ systems do
not allow continous monitoring of variationsa·ndare thus not very convenient
for meteorological applica tions.Currently the most widely used instrument
is that ;designedby :Rober'tson (1972) and further developed by Berger
(1976). It is based on the use of a fluorescent phosphor,whichchan-
ges the DV radiation below about 320nminto visible light of 400-600 nmand
acts a s a diffuse recei vingsurface. The visible light is measured with a
photom ul tiplier. 'The combina tionoE the spectral response of thep.hos-
phor and the pre- andpo:stfiltersdetermines the over-all spectral re-
sponse, which is adjus.tedto the erythemal response of the human skin.
Choice of the type of action curve adopted is.ba.sed on concern about skin-
cancer and relatcedeffects of UV on the hum.an skin. The syst.em has proven to
be reasonably sta'ble and has been used in a world-·wide netw.orksince
the mid seventies. :·Asimilarinstrument has been developed by 'Dehne
(1975) and uses a specially designed interference filter with a quartz
diffuser in front of it and aphotomulttplier as the detector .

Thefl:ccurac-y of these in·struments 'should be consid.ered.in.two

different categories. On one hand tbeinstrum:ent'spectral response basan
inherent .characteristic different from that of the idealized action
curve. 'This can be referred to as a deviation. Ontheoth:er handtheinstru-
ment chara:cteristics might vary with temperature, angle of .incidenceand
aging. Thes'e effects represent measurement errors. The deviation is inherent
to the instrument, the errors will depend on the operating conditions.
The calibration of these instruments i,s normally performed with
lamps, assuming that the relatives,pectralsensitivity is knownfro·m other
measurements. From the experience of the networks presently· in .operation for
several years, it .is known that the instruments yield good results for inoni-
toringpurposes,Hthefield instruments are c.arefully ma.intainedandcali-
brat·edabout om;:e a year.

Instruments of the second group have mainly been used in research~type

investigations (e.g. Bener, 1972; Goldbergan.d Klein, 1974) and w.ill not
be described here. The problems are similar to thoBe encountered in the
CHAPTER 4 RADIATION MEASUREMENTS

"action curve" instruments and in sunphotometers.

4.4 TECHNIQUES FOR RESEARCH-ORIENTED MEASUREMENTS

4.4.1 Measurements of Surface Albedo

The albedo of a surface is taken in the present context as the ratio of

the flux density Et reflected from a surface to the density E.Jr incident on
the surface, hemispheric integration being assumed in both cases. Thus p =
Et/E.Jr.

Two different techniques have been developed for measuring the albedo of
natural surfaces. The prevailing method involves the use of an inverted
pyranometer, whereas the use of an integrating sphere has been mainly for
special studies. In the following only the pyranometer method. is described.

Since the albedo is the ratio of upward- to downward-flux den si ties of

solar radiation, it is readily determined by two pyranometers. One pyra-
nometer is mounted in the normal upright position for measuring the downward
irradiance through a horizontal surface, and the second pyranometer is di-
rectly inverted for measuring the corresponding upward irradiance.

Although the measurement concept is simple, the measurements are subject

to the same errors as are pyranometers in general (see 5.1.4.) plus some
additional errors due to the combination of instruments. Since two instru-
ments are involved, their errors are additive in the albedo computation.
Thus the two must be well calibrated, preferably with respect to the same
calibration source, and many unforeseen problems Can be avoided by using the
same model devices from one manufacturer. The response characteristics must
be the same for the two orientations, a criterion which is not necessarily
fulfilled by the available instruments. The reason for the observed depen-
dence of response on instrument orientation is not well-understood, but it
appears to be caused by air currents inside the sensor chamber. In any case,
this factor should receive careful consideration in selection of the instru-
ments. The sensor surfaces of the two must be closely parallel, a criterion
which is satisfied by firm attachment of the two to the same rigid mount. A
most important requirement is for accurate levelling of the sensor surfaces.
This is most critical for the upward-facing instrument, since the largest
component of the downward irradiance has a monodirectional character and the
irradiance on the sensor surface is proportional to the cosine of the angle
incidence.

The quasi-diffuse character of the reflected radiation makes levelling

of the inverted pyranometer less critical, but gross misalignment may permit
the sensor to be irradiated by the direct solar beam near sunrise or sunset.
The same difference in directional characteristics of the two fields makes it
particularly important that the upward-facing pyranometer have a good cosine
response and lack nf azimuth sensitivity. Otherwise, the albedo measurements
will be subject to an apparent but unreal dependence on position of the sun.

Another error in albedo measurements with tt-70 pyranometers may be caused

by the two devices being at different temperatures, but the high-quality
temperature compensation available in modern instruments minimizes this ef-
fect. A factor of much greater concern has nothing to do with the instru-
ments themselves, but arises from the site and mounting arrangement. It is
important that the surface for which the albedo is being measured be as level
as possible. The directional nature of the solar beam makes a surface which
slopes toward the sun receive more radiation than a level surface would, and
the consequent increase of reflected irradiance would make the albedo appear
too high. The converse \<10uld be true for a surface sloping away from the

85
CHAPTER. 4 RAD IA T I ON MEASUREMENTS

sun. Theoretically, this effect could be . counteracted by·mounting the pyra-

nometers out of level and parallel to the surface, but this artifact may
introduce other problems. Finding a level site for the measurements is much
to be perferred. In some surfaces, such as the snow surface of the Antarc-
tic, the roughness elements may be of such size that there are several slopes
in the region directly below the pyranometers, making the surface illumina-
tion highly non-uniform and time-dependent. In such a case the instrument
must be high enough above the surface to effectively integrate over many
roughness elements, or the data may be analyzed for approximating and re-
moving the effects· of the surface features.

Artifical shadowing of the surface by instruments and mount is an un-

avoidable source of error, although under normal circumstances such errors
are of small magnitude and may be minimized by adequate height of the instru-
ments and construction of the mount. The one instance where shadowing from
the instrument mount has proved to be very troublesome is in high-latitude
sum mers,during which time the sun remains above the horizon throughout the
twenty-four hours and the shadow of the. mount· passes directly beneath the
pyranometers.· For such an installation, slender uprights in the mounts are
to be used when possible. Shadowing from the instruments themselves is un-
avoidable in low latitudes where the sun may be directly overhead, but cor-
rections for such conditions are readily applied.

404.2 Measurements of Surface· Emittance

While most of the foregoing discussion has been confined to wavelengths

of most importance in the solar radiation regime, there is considerable inte-
rest in the terrestrtal radiation regime as well. Obviously, tche ehergy ba-
lance of surfaces at normal terrestrial temperatures depends on both the
absorption of solar radiation and the emission and absorption of terrestrial
radiation. Thus considerable effort in the past has been devoted to deter-
minations of the emittance of natural and engineering-type surfaces at wave-
lengths at which significant terrestrial radiation is emitted (Le., at
S<X<4°tLm).

The emittance e. of surfaces is most readily determined by measuring the

hemisphere reflectance p of the surface and computing the emittance from
Kirchhoff's law, written as e = 1 - p. Spectral discrimination in measure~
ments of p yields similar spectral values of e, whereas total emittance may
be determined by either spectral or total measurements of p.

A number of different techniques have been devised for measurements of

reflectance at infrared wavelengths. Some of these are based on the Goblentz
hemisphere (Coblentz, 1913) in combination with an infrared spectrometer.
The optical principle for theCoblentz hemisphere rests on the fact that
there are conjuga.te focal points on either side of the centre of curvature of
a hemisphere.. If the interior of the hemisphere is specularly reflecting and
light throug,h a hole in the hemisphere faUs on a sample surface at the first
conjugate point, a defocused image of the illuminated area falls on the se-
cond conjugate point. Absolute measurements of reflectance of a sample are
made by placing a detector at the first conjugate point to measure the inci-
dent flux, and then moving the detector to the second conjugate point while
placing the sample surface at the first conjugate point. Then the reflectance
is computed as the ratio of the two detector signals.

Another type of reflectometer for use in the infrared is the 'heated

cavity' concept of Geir et al. (1954). In this device the sample is placed in
a black-body cavity, which assures diffuse illumination of the sample. The
sample and Cl spot on the wall of the cavity are viewed alternately through a
peephole in the cavity wall. Thus, measures of the reflected and incident
flux densities are obtained, from which the absolute reflectance of the
sample may be computed. In principle, the 'heated cavity' method has the po-
tential for very accura,te reflectance measurements, but in practice t.he

86
CHAPTER 4 RADIATION MEASUREMENTS

inability to distinguish between energy which is reflected from the sample

and that which is emitted by the sample tends to limit the accuracy sub-
stancially.

Blevin and Geist (974) have developed a very simple reflectometer with
a large-area pyroelectric detector made of Polyvenylfluoride and mounted
inside a cone. For the measurements, a thin cone is placed on the surface,
the reflectance of which has to be determined. Chopped infrared radiation
from a spectrometer illuminates the sample through a small hole on the top of
the cone. The reflected radiation is then collected by the large-area pyro-
electric detector. Thus a direct determination of the reflectance is al-
lowed. As pyroelectric detectors are only sensitive to chopped radiation,
the thermal radiation from the sample will not be measured.

4.4.3 Measurements of the Polarization of Sunlight

The techniques of measuring polarization have been discussed in sorne

detail by Coulson (1975). Briefly, they may be divided into visual and elec-
tronic methods. The first of these has a long and distinguished history, but
the natural limitations of the eye severely limit the wavelength range and
speed with which such measurements can be made. Thus visual techniques have
been supplanted mostly by electro-optical methods which are much more versa-
tile and less subject to the former limitations. Of course, these modern
devices have their own problems, but they are of a less fundamental nature
than those associated with the human eye.

The two most frequently employed devices for environmental measurements

are polarimeters with a rotating analyzer and those with fixed analyzer and
rotating retardation plate. The rotating analyzer type is the simplest, and
the minimization of polarization effects in modern photomultiplier tubes and
other detectors makes it very satisfactory for measurements of linear polari-
zation. For this case, the radiance Ld ('f) of light passing through the ana-
lyzer, oriented with respect to the plane of transmission at angle '1', and
falling on the detecteor is, except for instrumental constants, given by the
relat;.ion .

Ld ('I') = Ltcos 2 '1' + Lr sin 2 '1' + (U/2) cos 2'1'. (4.14)

Here, La and Lr are, respectively, the radiance components of the incident

light along the orthogonal D and r axes (D for parallel and r for perpendicu-
lar to the plane of reference), and U is a parameter defined by:

u = (La - Lr ) tan 2X , (4.15)

where X is the angle of the plane of polarization of the incident light. The
degree of linear polarization P is given by the relation:

(4.16)

and the radiance of the incident light is:

L = La + Lr • (4.17)

Theoretically, the complete set of L, P, and X of the incident light

could be determined by three determinations of Ld ('f) taken at three angles 'I'
of the analyzer. In practice, however, the ease and speed with which Ld ('f)
can be determined with an electronic system make it feasible to increase the
number of measurements for increased precision of the data.

The simple rotating analyzer system outlined above is applicable for

measurements of only linear polarization, but if elliptic or circular pola-
rization is to be measured, a different system is required. Fortunately, the

87
CHAPTER 4 RADIATION MEASUREMENTS

main interest for remote-sensing purposes is attached to the linear polariza-

t.i,on field, the elliptic or circular polarization in the natural environment
being so small as to be at the limit of detectability. However, the system
of fixed analyzer and rotating retardation plate has the capability of mea-
suring all polarization parameters.

The use of such a system is outside the scope of the present discussion.
One such system has been described byCoulson (1975), and an extensive set of
measurements using the system was made by Sekera et al. (1955).

Abbot, C.G. (1922) The silver disk pyrheliometer. Sm.ith. H.isc. -Coll. 56: 19.

Abbot, C.G. and L.B. Aldrich (1932) An imp.roved water~flowpyr'heliometer and

the standard scale of solar. radiation. Smith. Hisc. ColJ. 87: 15.

Aldrich, L.B. and C.G. Abbot (1948) Slllithsonian pyrheliometry and the
standard scale of solar radiation. Smit1J.H.isc. Coll. 110~ 5.

d'Almeida, G.A., R. Ja.enicke, P. R'oggendorf and D. Richter (1983) New sun-

photometers' for network operation. Appl.Op.t. 22: 3796.

Angstrom, A. (1899) Ueber, absolute Bestimmungen der Warmestrahlung mit dem

elektrischenCompensationspyrheliometer, nebst ei nigen Beispielen der
Anwendung dieses Instrumentes. Ann. Phys. ClJem. 67: 633.

Angstrom, A. \1919) Note on comparisons between pyrheliometers and on the

difference between the Angstrom standard and the Smithsonian standard.
Honth. {leather Rev. 47: 798.

Angstrom, A. and A.J. Drummond (1961) Bas.ic Concepts Concerning Cutoff Glass
Filters used in Radiation Measurements. J. Het. 18: 360.

Angstrom, A. and B. Rodhe (1966) Pyrheliometric Measurements with special

Regard to the Circum solar Sky Radiation. Tellus 18: 25.

Beckman, W.A., S.A. Klein and J.A. Duffie (1977) Solar Heating Design by the
f-Chart Method.Wiley and Sons, New York.

Bener, P. (1972) Approximate Values of Intensity of Natural Ultraviolet Ra-

diation for Different Amounts of Atmospheric Ozone. Final Techn. Report,
European Res. Office, H.S. Army, London, Contract N·o. DAJA37-68-C-1017,
June 1972.

Berger, D.S. (1976) The Sunburning Ultraviolet Meter: Design and Performance.
Photochem.istryand Photo.b.iology 24: 587.

Blevin, W.R. and W.!. Brown (1966) Black coatings for absolute radiometers.
Hetrologia 2: 140.

Blevin, W.R. and J. Geist (1974) Infrared Reflectometry with a Cavity-shaped

Pyr-oelectric Detector. Appl. opt. 13: 2212.

Brusa, R.W. (1983) Solar Radiometry. Dissertation ETH No. 7181.

Brusa, R.W. and C. Fr.ohlich (1972) Entwicklung eines neuen Absolutradiome-

ters. Technical Note 1, World Radiation CentElr Davos.

Brusa, R.W., C. Frohlich and Ch. Wehrli (1983) Solar Radiometry from High-
Altitude Balloons. 6th ESA Symposium on European rocket & balloon pro-

88
CHAPTER 4 RADIATION MEASUREMENTS

gram mes (ESA SP-183), p. 429.

CIMO Guide (1983) Chapter 9: Solar Radiation. In: Meteorological Instruments

and Methods of Observation, 5th edition. WMO, Geneva.

Coblentz, W.W. (1913) Instruments and methods used in radiometry. If. Bul.
Ha t. Bur. std. 9: 7.

Coulson, K.L. (1975) Solar and Terrestrial Radiation: Methods and Measure-
ments, Academic Press, New York.

Courvoisier, P. (1950) Ueber einen neuen Strahlungsbilanzmesser.

Sehwe.iz. Haturforseh. Cesellseh. 130: 152. .

Crommelynck, D. (1973) Th80rie instrumentale en radiometrie absolue. Inst.

Royal Metrorologique de Belgique, Bruxelles, Pub!. Sera A, No. 81.

Dave, J.V. (1978) Performance of a tilted solar cell under various atmosphe-
ric conditions. Solar £nerqy 21: 263.

Dehne, K. (1978) Design and Performance of a New Instrument for Measuring

UV-B Global: Radiation. Proc. WMO Technical Conference (TECIMO) Hamburg,
August 1977, WMO, Geneva.

Dehne, K. (1980) Vorschlag zur standardisierten Reduktion der Daten verschie-

dener nationaler. Himmelstrahlungs-Messwerte. Ann. d. Meteor. (neue Fol-
ge), 16: p. 57, Offenbach a m Main.

Drummond, A.J. (1956) On the Measurement of Sky Radiation. Arch. Het. Ceo-
phys. B.io.kl. Ser.ie B, 7: 413.

Flowers, E.C. (1978) Solar Radiation Facility - Report for 1978, National
Oceanic and Atmospheric Administration, Boulder, Colorado (unpublished).

Flowers, E.C. (1984) Longterm behaviours of· pyranometers. Proc. lEA Meeting,
Norrkopping, Sweden, January 1984.

Fritschen, L.J. (19.63) Construction and elevation of a miniature net radiome-

ter. J. Appl. Het. 2: 165.

Frohlich, C.· (1973a) The relation between the IPS now in use and ·Smithsonian
Scale 1913, Angstrom Scale and Absolute Scale. Proc. Sym. Solar Radia-
tion, Smithsonia Institution, Rockville MD, Nov. 13-15.

Frohlich, C. (1973b) Absolute radiometry and the internationalpyrheliometric

scale 1956. CIMO IV, Scientific Discussions, WMO Helsinki.

Frohlich, C. (1984). The need for the characterization of pyranometers. Proc.

lEA Meeting, Norrkopping, Sweden, January 1984.

Fro h li ch, C• and H. Que n z e 1 (1974) In fl u e n c e 0 f the Sun's· Au re 0 1 e 0 nth e

Determination of Turbidity. fIIHO Pu.blJcC1t.ion 368: 538.

Funk, J.P. (1959) Improved polyethylene shielded net radiometer. J. Sei.

Inst. 36: 267.

Geir, J.T. and R.V. Dunkle (1951) Total hemispherical radiometer.s. Trlins.
Am. Inst. £lee. £nq. 70: 339.

Geir, J.T., R.V. Dunkle and J.T. Bevans (1954) Measurement ,of absolute spec-
tral reflectivity from 1.0 to 15 microns. J. apt. Soc. Am. 44: 558.

Goldberg, B. and IV.H. Klein (974) Radiometer to Monitor Low Levels of Ultra-
violet Irradiance. Appl. opt. 13: 493.

89
CHNP'fER 4 RADIAT ION MEASUREMENTS

Hinzpeter, H. 0956) Ueber die Repr~sentanten der Anstrdm-Skala und der

Smithonian-Skala des Observator-iums Potsdam. Z. J/eterol. 10: 257.

Idso, S.B. (1971) A simple technique for the calibration of long-wave radia-
tion probes. Aq.r..ic~ . Heteor. 8: 235.

lEA Task HI, (1985) Results of an Outdoor and Indoor Pyranometer Comparison.
In press.

Karoli, A.R., J.R. Hickey and R.. G. Frieden (983). Self calibrating Cavity
Radiometers at. the Eppley Laboratory: Capabilities and Applications.
Froc. SPIE, East Arkington, April 19'83.

KendaU, J.M., Sr.' and C.M. Berdahl (1970)· Two black body radiome.ters of high
accuracy. Appl. opt. 9: 1082.

Kimbal1, H.H. (1910) Solar radiation, atmospheric absorption, and sky polari-
zation at Washington, D.C. B.uIl. /fount W'eatl1er Ohs. 3: 69.

Klein, S.A. (1977) Calibration of monthly average insolation on tilted SUF-

faces. Solar Energy. 19: 32S.

Kmito, A.A. (979) Estimation· of the present accuracy of pyrheliometric

measurements,in "Use of Data on Solar Radiation in National Economy",
Gidrometeoizdat, p.p. 128.

Kmito, A.A., V.A. Parfinsky, M.M. Seredenko and V.A. Klevantsova (1976) Spec-
tral coefficient of blackened surface reflection. "HGO Proceed.inqs" 370:
39.

Koepke, P. and H. Quenzel (1978) Water vapor: . spectral transmission a.t wave-
lengfits between 0.7 J.lm and 1 J.lm. Appl. opt. 17: 2114.

Le Baron, B.A., W.A. Feterson and 1. Dirmhirn (1980) Corrections for diffuse
irradiance measured with shadowbands. Solar Enerqy 25: 1..

Uu, B.Y.H. and R.C. Jordan (1963) Daily insolation on surfaces tilted toward
the equator. ASJlRAE Journal 3: 53.

McDowell, J. (955) Total radiation fluxmeter. Heteor. Haq. 84: 65.

Mhrikofer, W. (1937) Ueber die zwische·n Kompensations- und SHverdisk-Pyrhe-

liometern beobachteten Differenzen. Intern. Strahlungskomm. Oxford 193&,
p. 49.

FeterS'on, J .T., E.C. Flowers and J.H. Rudisill (973) Dew and fros·t deposi-
tion on pyranometers.. J. App.l. Heteor. 12: 1231.

Reid, M.S., C.M. Berdahl and J.M. Kendall (1978) Calibration standards and
field instruments for precision measurements of insolation. Solar li'nerqy'
20: 357.

Robertson, D.F. (1972) Solar Ultra-violet Radiation in Relation to Human

Sunburn and Skin Cancer, Ph.D. Thesis, University of Queensland.

Schulze, R. (1953) Ueber ein Stt'ahlungsmessgerat mit ultrarotdurchlassiger

Windschutzhaube am Meteorologischen Observatorium Hamburg. Geof.is. Para.
AppJ; 24: 107.

Sekera, Z., K.L. Coulson, D. Deirm.endjian, R.S. Eraser and C. Seaman (1955)
Investigation of polarization of skylight. Final Rep., Contr. AF19(122)-
239, Univ. o·f Calif., Los Angeles, California.
CHAPTER 4 RADIATION MEASUREMENTS

Temps, R.C. and K.L. Coulson (1977) Solar radiation incident upon slopes of
different orienta tions. ~(.701ar E'nerg-y 19: 179.

Volz, F. (1959) Photometer mit Selen-Photoelement zur spektralen Messung der

Sonnenstrahlung und zur Bestimmung der Wellenlangenabhangigkeit der
Duns t trubung. Arch. meteor. G'eophys. Bio.klim. B10: 100.

Willson, R.C. (1973) Active cavity radiometer. AppJ. opt. 12: 810.

Willson, R.e. (1980) Active cavity radiometer type V. AppJ. opt. 19: 3256.

World Meteorological Organization (1979) Building instructions for WMO/PMOD

Sunphotometers. Geneva, Nov. 1, 1979.

Zalewski, E.F. and J. Geist (1980) Silicon Photodiode Absolute Spectral Re-
sponse Self-calibrating. Appl. opt. 19: 1214.

91
 5. - DATA PROCESSING, VALIDATION, AND QUALITY-CONTROL

5.1 INTRODUCTION

A radiation measurement program me must include prOVISIon both for the

validation of the observed data to ensure that they are of the highest pos-
sible quality and for the processing of the data into a form which is appro-
priate to the intended uses. Validation or quality-control attempts to ensure
that each value achieves a certain level of reliability and the data are as
complete as possible, while data processing will typically involve a reduc-
tion in the volume of data through the generation of summaries and use-
specific statistics. These statistics are frequently needed to develop par-
ticular algorithms used to generate such information as solar heat loads,
evaporation rates, etc.

An overview as well as specific detail information on the topics of

radiation-data validation and processing can be found in Wendler and Eaton
(1980), World Meteorological Orga niza tion (1981), Sch mid (1976), Foulk es
(1980) , Wolf f et a 1. (1980), La t i mer and Won (1981) and the In t er n a t ion a 1
Energy Agency (1980).

5.2 QUALITY-CONTROL AND VALIDATION

There is a variety of approaches which may be used to verify the quality

of measured data. These techniques vary in their degree of rigour, the com-
plexity of the operation and in regard of the amount of data required when
implementing them. In the approach of choosing the most appropriate method,
therefore, one has to consider the degree of accuracy required, the capabili-
ties of the data processing system and the availability of supporting data.

5.2.1 Checks Using Independent Measuring Systems

The optimum approach is to use entirely independent sensing and re-

cording systems, to perform the intercomparisons at frequent (at least
monthly) intervals and to collect data under conditions which approach a
steady state (e.g. cloualess skies, small solar zenith angle and low atmos-
pheric turbidity). Discrepancies between the two systems should be no larger
than typical measurement errors, which are usually less than 4 per cent for
sola r (H a y and War dIe, 1982; La t i mer, 1972) and 1e s s t h a n 5 per c e n t for
long-wa ve and net-radiation measurements (La timer, 1972).

5.2.2 Consistency Checks Using Observed Horizontal-Surface Irradiances

A relatively convenient method for verifying the quality of a given

measurement is to check for consistency between a number of radiation vari-
ables recorded over the same period of time. As a result of the comparisons
aberrant values can frequently be isolated.

5.2.2.1 Global Solar Radiation (E g {»

Global solar-radiation values measured under cloudless skies should fall

wit hi n rea sonably well-defined li mit s. These lim its ma y be deter mi ned by

93
 CHAPTER 5 DATA PROCESSING, VALIDATION, AND QUALITY-CONTROL

reference to published values for the specific location and time or by refe-
rence to a long series of observed data, in which near-maximum values are
usually connected with clear skies (see for example Fritz. 1949). Kondratyev
(1969) has tabulated representative clear-sky values of global solar radia-
tion, expressing these as a function of latitude. Sample values are included
here as Table 5.1.

Table 5.1 Possible Global Solar Radiation in MJ m- 2 day-1 referred to WRR

(modified from Kondra tyev, 1969J.

LatH ude. Jan. Feb. Marc. Apr. May June Suly Au-g. Sep I. Qcl. Nov. Decc.
de-s,·

; Nor Ih
90 0.0 0.0 0.2 14.0 30.7 36.6 33.3 18. 1 3.3 0.0 0'.0 0.0
85 0.0 0.0 1.0 14'.3 30.6 38·. 1 32.9 18.4 4.3 0.0 0.0 0.0
80 0.0 0.0 2.9 15. 1 30.1 35.4 32.2 18.7 6.0 0.. 6 0.0 O. O·
75 0 .. 0 O. 8 5.6 16.4 29.5 34.4· 31.0 19.4 8.2 1.9 0.0 0.0
70 0.0 2.2 8.5 18.4 28.8 33.0 29.9 20.5 10.6 3.8 0.7 0.0
65 1.0 3.9 11.3 20.4 28.7 32. 1 29.5 28.2 13.3 6.1 1 .9 0.3
80 2.5 6. 1 13.9 22.5 29.2 32.2 30.0 23.5 15.8 8.5 3.8 1.6
55 4.4 8.7 18.4 24.3 30.2 32.8 30.8 25.2 18.1 11.0 5.7 3.0
50 6.8 11 .5 18.7 26.0 31 . 1 33.3 31.7 26.8 20.2 13.8 8.1 5.8
45 9.4 14.5 16.6 27.4 31.9 33.8 32.1 . 28.3 22.2 14.4 10.9 8'.2
4·0 12.4 n:.2 23.0 28.5 32.4 33.7 33.0 29.0 23.9 18.5 13.6 11.1
35 15.0 19.6 24.8 29.4 32.6 33.8 33.1 30.1 25.4 20.6 16.0 13.7
30 17.5 21.7 28.2 30.0 32.8 33.3 32.9 30.8 26.8 22.6 18.4 16. 1
25 19.8 23.8 27.3 30.3 32.2 32.8 32.5 30.7 27.9 24.4 20.6 18.4
20 21.8 25.2 28.3 30.3 31.8 32.0 31.7 30.6 28.7 26.0 22.6 20.7
15 23-.7 26.6 29. 1 30. 1 30.8 30.9 30.8 30.3 29.4 27.2 24.4 22.6
10 2.5.4 27.8 29.7 29.8 29.7 29.5 29.6 29.8 29.8 28.2 26.0 24.6
5 27.1 28.7 30.1 29.4 28.5 28.0 28.3 29.0 29.9 29.1 27. 5 26.4
0 28.4 29.4 30.2 28.7 27.1 26.4 26.8 28.2 29.8 29.7 28.7 28.0
5 29.7 30.1 30.1 27.9 25.7 24.8 25.2 27. 1 29.3 30.1 29.8 29.6
10 30.8 30.5 29.6 26.9 2.4.2 22.8 23.5 25.7 28.6 30.1 30.6 31.0
15 31.8 30.8 29.0 25.7 22.5 21.0 21 .6 24.3 27.7 29.9 31.3 32.2
20 32.5 31.0 28.2 24.2 36.1 18.9 1-9.8 22.7 26.6 29.5 31.9 33.0
25 33.1 31.0 27.3 22.5 18.7 16.9 17.8 21.0 22.2 29.0 32.2 33.9
30 33.6 30.7 26.1 20.9 16.7 14.9 15.6' 19. 1 23.8 28.3 32.5 34.6
35 33.8 29.9 24.7 19. 1 14.6 12.7 13.5 17.0 22.3 27 .4 32.5 35.1
4'0 33.8 29.0 23. 1 17.1 12.3 10.3 11.3 14.9 20.6 26.3 32.1 35.4
45 33.7 27.9 21.3 15. 1 9-.9 7.7 8.8- 12.7 18,.7 24.9 31.5 35.5
50 33.3 26.6 19.4 12.9 7.6 5.3 6.4 10.3 16.8 23.4 30.7 35.2
55 32.5 25.0 17.3 10.5 5.3 3.2 4.2 7.7 14.5 21.6 30.0 34.6
60 3'1.7 23.4 15. 1 7.9 3.4 1 .4 2.2 5.3 12.0 19.8 29.5 34.3
65 31 .4 21.6 '12.7 5.4 1.5 0.1 0.6 3.2 9.4 17.8 29.2 34.5
70 31.7 20.0 10.2 3.2 0.3 0.0 0.0 1.4 7.0 16.0 29.4 35.0
75 32.6 18.9 8.0 1.3 0.0 0.0 0.0 0.2 4.9 14.6 29.9 35.8
80 33.8 17.9 6.0 0.0 0.0 0.0 0.0 0.0 2.9 13.6 30.8 36.6
85 34.6 17.4 4. 1 0.0 0.0 0.0 0.0 0.0 1.3 12.9 31.4 37. 3
"0. 35.0 17.3 2.4 0.0 0.0 0.0 O. o· 0.0 0.1 12.8 31.7 31.8
Sou-Ih

5.2.2.2 D-iffuse Solar Radiation (Ed·H

Under conditions with clear sky the ratio (Ed..ji I Eg..ji) will generally be
a smooth function of solar zenith angle allowing unusual measurements to be
readily recognized. However, if direct (S) and global (Eg..ji) data are avail-
able, a preferable approach is to verify the consistency of the data, using
the rela'tionship:

(5.1')

Differences should be less than 5 per cent after the appropriate corrections
have been ma.de to the diffuse reading (see 4.2.2.6). Schmid (1976) demon-
s:trates the effectiveness of this approach (Fig. 5.1).
CHAPTER 5 DATA PROCESSING, VALIDATION, AND QUALITY-CONTROL

- N
I

E
200

-
3:
UJ
0
lOO

I" 0
UJ

-100
0 5 10 15 20 25 ao as .0 u 10 15 10 15

Solar Elevation (0)

Figure 5.1 Difference between global and diffuse solar radiation for a
horizontal surface when there is no bright sunshine. Data are plotted
against the solar elevation. (Schmid, 1976).

5.2.2.3 Direct Solar Radia tion (S)

Under cloudless skies, the value of the direct radiation should, as

with global and diffug.e radiation, stay within well-defined limits, which
might be established by reference· to ·published values or examination of a
long series of observed data. A preferable approach is to ensure that the
difference between. Sand E9 {- - (E d {- / cos 8) is less than 5 per cent. For
example, Wendler and Eaton (1980) illustrate the capability of this tech-
nique to isolate erroneous values (Fig. 5.2).

'" r=-+--!-+-~f-----"'----t---'--'--t-----j!
I I
150 I i "
!
- 125 ~_+__--_-+I----+------+-----+--I
~ tower .ffect
\.A

-"
3=
W
100 1==-=JJrf=====:::t====:t====+::::::=::=fc=t=""'1
~
75l--4--+__---_l_---'----+---____1----t___;
tower .ffect
50 ~-H-+__---_l_---_+---____1----t___;

25 ~..JI-~---_+_---_+---____1~.---t___;

10 11 12 13 HI

Local Standard Time

Figure 5.2 Temporal variation over one day of the ratio of -global radiation
determined by combining of Sand Ed to the global radiation Eq deter-
mined by the pyra nometer. (Wendler and Ea ton, 1980).

5.2.2.4 Reflected Solar Radiation (Er t)

Given that measured values of E9 {- would normally also be available, the

most appropriate consistency check is to ensure that values of the ratio
Er t / E9 {- (i.e. the albedo) do not without reason depart subs ta ntially from
published values for the surface in question (see, for example, 3.8.2). This
approach to detect the anomalous behaviour of one of the sensors is illu-
strated in Fig. 5.3.

95
CHAPTER 5 ' DATA PROCESSING, VALIDATION, AND QUALITY-CONTROL

15. Oet 1978

- cP.'9 0 ! - - - - - - 1 l - - - - - + - - - - - - - - - - - t
_
0, 80 I'-~-~_I+_----+------.--­
"C:
CD· 7 0 1 ' - - - - - 1 + - - - - + - - - - - - - - - - ;
.a
_. 80
~. . =='N~====:::::::::::;;:::::::::::::;====i
t=:===j1~fi.=
9h

~200 I ' - - - - - - - - - - t - - - - - = " , . . - = ' t f - - - ;

.

w: 50' 1'-'- - - - - - - = - - - - - - t , - - - - - - - - -

al); H,

Ffgure 5.3 Two hours of alhedo (upper curve), global (m.iddle curve) and
reflected (lower curve) radIation values. (Wendler, 1981.)...

5.2.2.5 Downward Atmospheric Radiation (E,·H

Again, under conditions of cloudless sky it is poss.ible to compare mea-

sured values with either published data or values abstracted from an observa-
tional record covering a long period of time. Another consistency check is
to ensure that differences between the observed irradiance E, -I- and either
(5.2)

are less than about 10 per cent (Dav.ies et al., 1970). E*' is the net and Et
the downward solar Eg and terrestrial El irradiance.

5.2.2.6 Upward Terrestrial Radia.tion (E, +)

A gTOSS check on the consistency of a measurement may be made through a.
comparison with aT. ~, where a is the Stefan - Boltzmann constant and T. is
air temperature (K). A mare accurate approach would be to use a realistic
value for the emittance in eaT. ~ wherever possible.. A cons.istency check is
achieved by comparison of. the measured irradiance Et t with either
(5.3)

and e'nsuring that differences are less than about 10 per cent.

5·.2'.2'.7 Net Radiation (E*)

Althoug.h it is unlikely that all the components of the· radiation balance

are separately measured, the only possible consistency check using obs'erved
values is to determ·ine whether the difference between measured E* and

E-I-· - E1t' + E, -I- - E" l' (5.,4)

IS less than about 15 per cent (Davies et al.,. 1970). Wendler and Eaton
(1980) also utilized this me·thod.
CHAPTER 5 DATA PROCESSING, VALIDATION, AND QUALITY-CONTROL

5.2.3 Consistency Checks Usinq Calculated Horizontal-Surface Irradiances

If other measured irradiances are unavailable, it is necessary to resort

to the use of calculated irradiances. However, in many cases the calculations
are lengthy, the data requirements exacting and the errors in the calculated
values so large that their usefulness for verification purposes is highly
questionable. Thus these techniques must be used with caution.

5.2.3.1 Global Solar Radiation (E g {o)

An attractive approach involves the estimation of the global radiation,

using locally available data together with tables and graphs. Machta (1978,
1980) and World Meteorological Organization (1981) detail such procedures and
Machta notes that they provide a calibration accuracy of within 5 to 10 per
cent for the prescribed conditions. Greater accuracy is precluded owing to
the difficulty in specifying turbidity and precipitable-water correction fac-
tors.

Alternatively, it is possible to use empirical relationships of the

types originally suggested by Angstrom (924) and Kimball (928):

(5.5)

where E. is the extraterrestrial global radiation, nlN the relative sunshine

duration, c the cloud amount and ai, a2, b 1 , and b 2 empirical coefficients.
However, the error of these relationships () 25 per cent) for individual
hourly and daily values makes them unpractical for verification purposes.
Even the use of more elaborate, theoretically-based models does not lead to
sufficient reduction in the errors at these time scales, primarily due to an
inability to specify adequately the radiative properties of clouds. Thus the
use of the more elaborate models such as those of Machta 0978, 1980), Davies
et a1. (1975), Atwater and Ball (978) and Hoyt (1978) is best restricted to
cloudless-sky conditions, when fewer input quantities are required and their
specification is more accurately defined (see Davies and Hay (1980) and
Davies (981) for a comprehensive review of these and other solar radiation
models).

5.2.3.2 Diffuse Solar Radiation (E d t)

Numerous studies (e.g. Liu and Jordan (1960), Collares-Pereira and Rabl
(1979), Erbs et a1. (1982)) have demonstrated a relationship between
Ed t I Eg {o and Eg {O, which has potential for the verification of diffuse radia-
tion data. However, on an individual hourly and even daily basis the scatter
in the relationship is' large and consequently this represents little more
than a crude verification method.

5.2.3.3 Direct Solar Radia tion (S)

Calculation of the direct radiation over short periods with cloudless

skies makes it possible to use adapted versions of the models based on the
principles of radiative transfer, since the need to specify cloud charac-
teristics can be avoided (e.g. Lowtran). This approach, however, cannot be
used when cloud obstructs the sun for any part of the observation interval.

5.2.3.4 Reflected Solar Radiation (Er 1-)

Tabulated values of albedo and the knowledge of the global solar radia-
tion would normally make it possible to verify the measured reflected solar
radiation. On the other hand, both the considerable variability in the albe-
do for a given cover type and the zenith angle dependence (see 3.8.2) suggest

97
CHAPTER 5 DATA PROCESSING, VALIDATION, AND QUALITY-CONTROL

that verification involving the specification of albedo is subject to sub-

stantial error. Arnfield (1975) has suggested a functional relationship
which accounts for the zenith-angle dependence, and any verification using
this approach implies an approximate a priori knowledge of the albedo of the
surface.

5.2.3.5 Downward Atmospheric Radiation (E e t)

Sections 3.6 and 3.7.2 describe equations which may be used to calculate
this irradiance. However, the accuracy of these equations is frequently not
sufficient to estimate Ee t.

5.2.3.6 Upward Terrestrial Radiation (E e 1')

Since this irradiance is a function of the temperature and emittance of

the radiating surface (see 3.8.4), a knowledge of these terms and their in-
corporation in the expression E aT4 would generally result in an estimate
which is within about 10 per cent of the measured value. Nevertheless, cau-
tion is required. Since there is spatial variability of E and there is a need
to express the surface temperature as a function of the air temperature (with
the possibility of a correction due to the lapse rate of temperature), this
approach may produce less useful results than might firs·t appear.

5.2.3.7 Net Radiation (E*)

Empirical relationships of the forms:

(5.6)

have been described in the literature (e.g. Davies (1967), Viswanadham and
Ramanadham (1967)). Potentially these represent a viable verification tool
and they can be used to a·dvantage for time periods of a day or more. At
shorter time intervals the phase shift between the short- and long-wave quan-
ti ties could lead to invalid conclusions about the quality of the observed
data.

5.2.4 Consistency Checks Using Measured Inclined-Surface Irradiances

A typical configuration in slope radiation measurements is to have a

number of sensors positioned at different slopes and azimuth angles. With
this arrangement there should be a relatively smooth variation, as the irra-
diance responds to changes in such parameters as the angle of incidence of
the direct radiation and the view factors of the sky and the underlying
surface. An example is provided in Fig. 5.4, where the changes with slope
and azimuth are consistent for all seven sensors even though the conditions
were partly cloudy. Aberrant values would be fairly easily isolated.

5.2.5 Consistency Checks Using Calculated Inclined-Surface Irradiances

If the global and other component solar irradiances are known for a
horizontal surface, it is possible to perform some relatively rigorous veri-
fications of the slope radiation data. The individual components are con-
sidered in the following sections, except for the reflected solar and upward-
terrestrial irradiances, which may be treated in the same way as in the
earlier horizontal-surface cases.

98
CHAPTER 5 DATA PROCESSING, VALIDATION, AND QUALITY-CONTROL

~
:)
1.4
0
..c:
co
'E
1.2 • -E
..., 1.0
~
z 0.8 ·s
0
i= 0.6

N::~
ci
0
ci
a:
a:
0.4
0.2
t
ci
..J
0 0.0
en 0 30 60 90

SLOPE ANGLE

Figure 5.4 Measured global solar radiation for the hour ending 0800 Local
Apparent Time, June 11, 1980 at Vancouver Canada, Data for given slope
angles and orienta tions (E=East, W=West, S=South, N=North) under partly
cloudy skies. (Hay, pri va te com munication),

5.2.5.1 Direct Solar Radiation on an Inclined Surface (Si)

The comparison should be between the observed irradiance and a value

calculated, using the expressions for the projected solar radiation or an in-
clined surface

S cos i / cos 8 = Si (5.7)

where

cos i = cos s cos 8 + sin s sin 8 cos (a-b) (5.8)

and s is the slope angle, 8 is the solar zenith angle and a and bare
the solar and slope azimuths, respectively (see Fig 5,5). It is desirable to
use values valid for an hourly interval or less due to non-linearities in the
above expressions. Agreement between calculated and observed values should
be within 2 and 5 per cent, with the higher accuracy being associated with
the use of a pyrheliometer to measure the direct radiation.

NORMAL TO THE
INCLINED PLANE

WEST_.
HORIZONTAL
SURFACE _ _ /

PROJECTION ON A HORIZONTAL
PLANE OF NORMAL TO THE
INCLINED SURFACE

Figure 5.5 Definition of the irradiance geometry of an orbitrarly inclined

surface,

99
CHAPTER 5' DATA PROCESSING, VALIDATION, AND QUALITY-CONTROL

5.2.5.2 Diffuse Solar Radiation on an Inclined Surface (Bd,1~)

and agreement between calculated and obse.rved irradiance should be within 10

per cent even for hourly values.

5.2.5.3 Global Solar Radiation on an Inclined Surface (Eg,1~)

Fig. 5.6 p!:'ovides an example of quality-control procedures applied to

g.lobal solar radiation on an inclined surface. The irradiance is calculated
making use of the relationships de.scribed in 5.2.5.1 and 5.2.5.2. This pro-
cedure has proven very effective not only in isolating. dispara.te slope-:-radia-
tion mesurements but also for verifying the quality of the direct, diffuse
(and hence global) solar~radiation data, which are the input values for the
calculation. Errors in these latter variables will typically be amplified
when the slope irradiances are caTcula ted (Hay, 1979a).

5.0;r---..,......---r------.---,----,r---~,

3:0:
• .'

• ..

0.0.
0.0 l'.0 2:0 3,0 4.0 5~0

MEASURED TOTALSHORTWAVE RADIATION

Figure 5.6 Res.ults of quality-control of observed hourly global solar ra-

diation on a 60-degree, south-facing slope at Vancouver, Canada. Analy-
sis revealed four anomalous observations associated with malfunction of
equipment.

5.2.5.4 Downward Atmospheric Radiation on an Inclined Surface (Et,1~)

On the basis of the assumption of isotropy in the downward and upward

and reflected long-wave irradianc'e, i.t is possible to provide a tractable
mathema,tical expression for this quantity:

E~,1~' 91, 0.5 Et~ (i-cos s) + O.S 0-[;) Ett (l+cos s) + 0.5 Et,t(i-cos. s) C5.10)

100
CHAPTER 5 DATA PROCESSING, VALIDATION, AND QUALITY-CONTROL

Departures from isotropy will limit the usefulness of this expression in

isolating unusual observations, though experience gained with a lengthy data
set should improve it as a quality-control. Unsworth and Monteith (1975)
have developed a model which incorporates the anisotropy of the long-wave
radiation over the sky hemisphere.

5.2.5.5 Net Radiation on an Inclined Surface (E*I)

The observed net radiation should agree with the sum of the observed or
calculated component fluxes, namely:

(5.11)

However, the accumulated effects of errors in each of these terms limit the
applicability of this relationship.

5.2.6 Quality-Control of Broad-Band Spectral Data

The most common and effective method for verifying the accuracy of spec-
tr~l measurements is based on the relatively conservative nature of the ratio
of the radiant intensities between two given spectral intervals. Wendler and
Eaton (1980) lean on this fact to assess the quality of data generated by
pyranometers equipped with OG1, RG2 and RG8 filters (see Fig. 5.7). Stewart
(1981) has shown that over short time-intervals numerically modelled values
may be used both for verifying the quality of data and replacing erroneous or
filling in missing data associated with instrument malfunction (Fig. 5.8).
Less success will be achieved with quality-control of the ultraviolet-radia-
tion data by the use of the ratio of UV to global solar radiation. The
wavelength dependence of the at mospheric at tenua tion processes results in
large variations for this ratio, precluding the application of strict quali-
ty-control limits (Wendler and Eaton, 1980).

15.0ct 1978

300 --~---- - ~ - ~ -

250

-C'l
I
E
200

-==
ISO

~ 100
W
50

01 IIH 09 10 11 17 13 11 16

Time Ch)

Figure 5.7 Daily variation of ultraviolet radiation and radiation measured

by pyranometers equipped with WG-7, OG-1, RG-2 and RG-8 filters. (Wend-
ler a nd Ea ton, 1980).

101
CHAPTER 5 DATA PROCESSING, VALIDATION,. AND QUALITY-CONTROL

£Total 'insolaUon

..0
..
.... -2

.......
·111

...
.

Time 1(ll1'h-12h) .

Fig.ure 5.·S Go'mp.ar~;son of total solar radiation,RG-2 filtered 's,olar ra:dia-

tionand computer -m,odelled RG-2 filtered :solar radiation(x),~{st.ewart.,
198B.

'5.3 lfI'SSI1I:G DA.TA

Form,any analyse's (e.g. the calcula;tion and comparison of ,average' va-

lues) and applications, such as the simulation 0.£ solar heating systems, i t
is desirable, if not essential, to hav'e a complete set of data. However, ·even
the best maintained operating system will have breaks in the data, even if
the observations are only interrupted because of the need to verify thecali-
bration factors of the monitoring equipment. There is a variety of methods
which will ensure the completeness ofa data set and each has its limitations
a.nd advantages..

5.3.1 Interpolation

This procedure involves the use of data observed both before and after
the break in the record to estimate the missing data values. Many methods can
be used, the simplest being to use linearly weighted averages of the two
observations adjacent to the data gap. More sophisticated techniques make
use of advanced curve-fitting procedures (e.g. cubic splines) and a longer
data record on either side of the dis.continuity. All interpolation proce-
10~ ....

:5 06\
:.;i .-\
.~ D.6,.1-;-\-------------'1
·0 '
'u .0" I--'\-,-------------'1
.. '~02; ;.-1\__
(
.'....."
<: Flrst-orderregresslon
..< . . '
;
;
o .~- •

'0·2·~_""""'--2~-~.3--.•~-""'~.....J

;D1s.placement. days
Figure. 5. 9 Averag:e autocorrelation for U1'lited Kingdom global solar-radia-
tiondata. (BriIlkw<?rth, 1977).
CHAPTER 5 DATA PROCESSING, VALIDATION, AND QUALITY-CONTROL

dures invoke assumptions about the temporal changes in the variable under
consideration (e.g. that it varies linearly with time). However, only under
clear-sky or heavily overcast conditions does radiation have simple temporal
variations. Thus interpolation schemes must be chosen with reference to the
temporal characteristics of the radiation prevailing at the time of the break
in observations. The temporal variability (primarily the auto-correlation)
of the data series and the method used to approximate the time series (e.g.
autoregressive-moving average, maximum entropy, maximum likely hood models)
will also strongly influence the length of the data gap, which may be filled
by interpolation. Fig. 5.9 and the results of Armstrong (1980) and Valko
(1980) indicate that it is of very little value to use interpolation tech-
niques for time periods as long as a day. In an analysis of data from a
single site in continental North America (st. Paul, Minnesota) Skaggs et a1.
(1982) found that the characteristic time between uncorrelated observations
ranged between slightly more than one day to just over 3 days.

960

~)
..6f"1"'bb01.tOfd
~-AirpOfl

Only 1000 m contour Ihown.

o I Km 10
l' , , !: : '

• BMihot

Figure 5.10 Interstation correlations (x 1000) of hourly solar radiation at

Vancouver, Canada for station pairings showing maximum correlation va-
lues. Data from June 1978 to May 1980.

5.3.2 Calculations Using Variables Coincidently Observed vith the Data Break

The use of a linear regression relationship to provide estimated values

for data that would be otherwise missing introduces a heterogeneity into the
dispersion of the data series, because such an interpolation procedure fails
to take into accou1}t the random component associated with the regression
(World Meteorological Organization, 1960). A recommended alternative ap-
proach is to establish a joint frequency distribution and use, for each of
the missing values, the value in the incomplete series which has the same
probability as the corresponding value in the associated data set. This
approach still relies on a high correlation between the two variables in the

103
CHAFTER 5 DATA PROCESSING, VALIDATION, AND QUALITY-CONTROL

joint distribution, these being different variables observed at the same site
or identical variables observed at adjacent sites. Fig. 5.10 demonstrates
the high interstation correlations for solar radiation that exist in Vancou-
ver, Canada, and which have made it possible to estimate missing data for a.ny
of the station pairings (Hay, 1984).

Numerical models provide an alternative method for filling in data gaps.

Chapter 3 and Section 5.2 have described the range of models available, and
the success with their use in this context depends on a variety of factors
including required accuracy, availability of input data, computational re-
sources and the length and number of data gaps. A procedure which will
usuany improve the quality of the calculated data is to "tune" the model to
the atmospheric conditions prevailing during the period of the missing data.
This "tuning" is most readily achieved by calculating values for perio.ds
adjacent to the data gap and determining a correction factor' which win, on
an average, minimize the difference between the ohserved and calculated ra-
diation data. This same correction ("tuning") factor can then be applied to
the values calculated for the duration of the data gap.' The amount of obser-
vational data used to "tune" the. model is again difficult to define. Except
in extremely persistent wea.ther conditions it is unlikely that there will be
any benefit if the model is tuned for periods longer than one day (McLaren et
al., 1980).

5.4 DATA ARCHIVE

Since archive.d data are the basis for all subsequent studies, including
those of both the originator and others who may may have access to the ar-
chive, it is necessary that the archive be accurate, accessible. and unambi-
g.UOliS. The quality-control procedures outlined in 5.2 will do much to ensure

F •••• 151 VAtUEl"FFI".,

LLL..L..LL...L....l~L..L.L.L.L..L...LJ.:.~~~~' U I I I I LJ
repe.ated 24. t.i mes

DaHy Record· of Hourly Values (HLY·) length·,- ),85

IS I VALUE If.; I
U I I11 LJ
repeated J!' times.

Honthly Record of DaHy Values (On): length. -'·2)2,

~. VALUE ~.
. I I I I .•
repeated: 12' t; mes

Abbrev i at,i",,'

STN ID Stat'ion Ident'i,ficat,ion 7 aJphanumer:i'c

YR Year., le'.., 9,8t =. 198,1 3: nume,riG.
HO' Month. ie." 01: = January 2' numerJc
nume.r-i-c
DY, Day, 2
Hement Numbe r 3 nume,r'tc
ELEH
ne9~1- iv•.
le... I-
S· 5.;9'1' of Parameter I'
'0.:' posit h,.,
Dan Via"'" 5 nurner.,ic.,
VALUE,.
F n:all' l' al'pharilUne-rric.

Fig'ure 5.11 Format of archived hourly, daily and monthly d.ata at the Cana-
dian At mosp:heric' Environm,ent. Service.

104
CHAPTER 5 DATA PROCESSING, VALIDATION, AND QUALITY-CONTROL

that the first objective is met. The other desirable attributes require that
the archive be organized in a systematic manner. An example of a convenient
format is shown in Fig. 5.11. The flag fields may be used to designate such
information as specific sensor and recording systems, techniques applied for
adjusting or interpolating data or to indicate the quality of the data in
case the quality has deteriorated through such factors, as, for instance,
moisture on the sensor. The archived programme of the World Radiation Data
Centre at the Main Geophysical Observatory, Leningrad, USSR (Fig. 5.12) is
associated with considerable expertise, Additional information on archiving
is contained in World Meteorological Organization <1960, 1981), In 1985 the
archiving at the WRDC will be fully computerized. Thus the data sent to the
WRDC may be delivered on standard magnetic tapes and the data sets can be
requested on tape.
REGION VII PEn10H VI EUROPE! EBPonA
CO\..OOn, iUlD
~D TlJRJ<n
CT!'A.... iIIl&.llt'I T1P'Jitq

nATIC:>HS
CTAHUM!<
U=ne
Kxpyxa
1AI.lal
J1,.190
Stockbola
CToJtronM "no"
.woe
1.000=0
JOIltaIlfO
8aaou.n
CawcfJI
btanbul
~T8»6y.l
4nkarli
AHRap8.
hair
H3l<Zp

Detle,,o,a T '
I ,80 '80 9? ''''2 . 872 920 Wl 868 -
2 "3 Z09 536 953 812 9}3 895
..... 1162 -
3 ,84 131 3"- 849 '90 8}1
,., 850 -
•5 158 16 68 553 244 111 190 -
54 25 86 soe 792 '/13 }64 805 -
6 ,08 '44 252 805 81. 573 ~oo 2.} -
7 50 '7 240 888 2'1/' 75 206 2}3 -
8 112 50 }1} 900 783 72 150 3'5 -
9 90 160 '4C 87' 759 139 1e:6 '64 -
10 ,01 ,6<; 2}<o
.,. 879 77' 21' ,"" 1t}1 -
11 90 '56 810 78' 609 }08 792 -
12
13

"15
94
6'
}2
50
.,
'69

""
&,
263
382
m
2,,",
806
875
263
8.18
630
7&
'86
56'
159
&50
001
858
'·9
507
696
760
525
10!-5
'079
849
-
-
-
-
16 &, '2. 22~ 795 '95 1b; ,80 ;05 -
17 32 79 '32 828 309 32} Ho 09} -
18 50 97 ''/1 185 &39 '08 ,7] 32' -
19 12 83 80Z 61~ ,.} .2} -
""25 Wl
-
20
21 }&
22
58
2"
205 ,.,
105 570
&10
20}

...'"
&<2
&&5
.,8

""" -
2'2
23
25
22
'01
&5
29Z
281
151
63}
584
501 44
-.0
699
158
699
--
2'
25
25 50 '5' - 544 521 113 800 -
-
26
22
22
}6
zz
58
03
259
15'
}58
Z9Z
196
190
8}<o
no .,.
9'lO
-
27 22 90 '22 80 68 151 8}5 651 -
28 22 '01 '48 '71 589 184 1161 590
,., -
29 2.2 25 ,"" 261 6'6 '52 '168 -
JO
31
22 22 2,2 208 621 190 }O} '59 -
IT
-
..
'969 2681 6829 20109 '58" 122&0 15119 '7}52
T 66 228 612 -
..
90 528 '09 50" 518
III
SS
}O
'.0 , .5
15·
2.5
1"-2-·
0.1
,.0
0.1
5'
,.1
>.
5''.8
82
2.1
115
}

Figure 5.12 Sample table for November 1983 from the published USSR radiation
data archive (USSR State Committee). Global radiation is given in
10- 2 MJm- 2 , and LT: 'monthly total, T: monthly meq.n, Lss: monthly total of
sunshine duration in hours, SS:, monthly mean.

5.5 SUMMARY STATISTICS AND GRAPHS

For many practical purposes, and as a result of the large amounts of

data that may be generated (see Section 5.6), it is necessary to summarize
the original data set through the generation of statistical and graphical
composites. This section draws attention to some of the assumptions and limi-
tations of these techniques and provides examples of appropriate summaries of
the original data.

5.5.1 Distributional Characteristics

Many of the statistics used to summarize the characteristics of radia-

tion data assume that the data are normally distributed. This is frequently

105
CHAPTER 5 DATA PROCESSING, VALIDATION, AND QUALITY-CONTROL

not the case and Fig. 5.13 gives one example which indicates the departure
from normality. Therefore, the distributional characteristics are best de-
scribed using non-parametric statistics including the median, quartiles and
percentiles. Differences between the two measures of central tendency (the
mean and median) indicate the skewness shown in Fig. 5.15 which is a typical,
feature of the radiation data.

b) 05
a)
OA ~---.-----r----y----' OA Vancouver, January. 1977
Vancouver, U. B. C., K160 degrees, /
0.3 0.3
South facing KI

0.2 January 1977 0.2

0.1 0.1

0.0 0.0

>-
u
~
::J
ow
"C i f~r;u: ."~ I
0.3
0.2

0.1
~_

. " ]
. "/
: 0.0 _.
0: 0.0
u.

::C=.-I ~ ~~lt:::
::.L.'.· ~~L:"
0.0

0.1. z . - .
O.Ci.obeo

. 0.1
/
,'. .

I
1
0.0 '. . . '.
0.0 1.0 2.0 3.0 , 4.0 0.00.0 1.0 20 3:0 4,0

HOURLY SOLAR RADIATION CMJ m· 1 hour·) HOURLY SOLAR RADIATION (MJ ni ., hOUf 1 1

Figure 5.13 Frequency dist'ributions 'of observed hourly values of global so-
lar radiation at Vancouver, Canada for a) the horizontal and three
south-facing surfaces in January 1977 and b) the 60-degree south-facing
surface in January, April, July and October, 1977. (Hay, 1979a.).

100

~ 80
>-
u
z
w
0 60
w
a:
u.
w
> 40
i=
<{
..J
:;)
~ 20
:;)
u

2 3 4 6

GLOBAL SOLAR RADIATION (MJ m'2 hour'l)

Figure 5.14 Cumulative frequency distribution of observed hourly global so-

. lar radiation for a horizontal surface in January at Vancouver, Canada
for the period 195~-1975.

106
CHAPTER 5 DATA PROCESSING, VALIDATION; AND QUALITY-CONTROL

5.5.2 Frequency Distributions

Frequency distributions of the type illustrated in Figs. 5.13 and 5.14

are very useful in summarizing the distributional characteristics of radia-
tion. The benefit of the approach represented in Fig. 5.14 is that the ef-
fects of persistency in the radiation may be seen by an examination of the
change in the frequency distribution as the averaging period increases. The
data presented here demonstrate the substantial reduction in extreme values
that occurs as a result of temporal averaging.

VANCOUVER January

Averaging Period {day&)

Figure 5.15 Percentiles and quartiles of measured global solar radiation for
a horizontal surface at Vancouver in January for the period 1959-1975
and for averaging times of one to ten days. (Hay, 1979a).

5.5.3 Joint Frequency Distributions

Radiation conditions are seldom considered alone, since other variables,

such as temperature and wind, have an important impact on the use that can be
made of the radiant energy. The combining of the frequency distributions for
two or more quantities can lead to a more beneficial use of the radiation
data. This should be considered at the time of formulating the data collec-
tion as well as the archiving programme. Examples of joint frequency distri-
butions are presented -Ftg. 5.16. In the ~xample from Table 5.4 of an average
'fo~ the~ month of J'aD,uary ther~ were 6.2 hours during whicp.tJ!~ global solar
irradi~hce w~s between 300 W m- 2 and 400 W m- 2 , and the ambient temperature
being during the same h:ours between -15 and -10 degrees Celsius.

5.5.4 Threshold Analyses

Before some uses of the solar energy (e.g. solar heating) are possible,
the radiant intensity must be above a minimum level. In many cases the mini-
mum threshold is also a function of temperature. For this reason tabulations
which consider radiative-energy inputs above specified thresholds and for
specified temperatures can be of immense use. Table 5.2 is one such example
and provides the mean monthly value of global solar radiation, which is
greater than the energy-threshold level given at the top of the column and
which was received while the ambient temperature was within the temperature
range noted at the left of the table. To illustrate this, the average values
in January were 27.37 MJ m- 2 of global solar radiation above a threshold
intensity of 300 W m- 2 that occurred at the ambient temperature between -15
and -10 degrees Celsius.

107
CHAPTER 5 DATA PROCESSING~ VALIDATION, AND QUALITY-tONTR6L

-50 V-'-"T
. 1.
10
'/'
1"'\1=l~UIY"
!50'/' 1,10'10 . '

0,0 I, -11-~-4----1
-, I ,
X / Stintis I ]

5,0 J, ,~L:'/_+
C-~ I
/ +
I
-+-;I
--1
','I'

I,
o

.. oJ?I.
U
10

"
~

~ 0
.o
CL
I
r:
. -+---=-/-+--.I-
E
t-
15,0

25.0 -t- -----:'+-I--'---'-+-----;~__l'

3.9,0";-,..L-l-l-"'-!-'---J'--"Y'If---.'L.'L... 1-'

o 8 12 16
•• +' • ~

Duration of lun,hin. (h)

Figure' 5.16 combined frequency distribution of daily mean temperature and

'daily 'total hours of sunshine, 'for Z,Uric.h, (lower) and Santis (upper).
(Valko, 1980).

Table 5.2 An example of the analysis of radiation data for selected thresh-
olds of global solar radiation intensity and air-temperature classes.
(McLaren et aL. 1979).
SUDBURY (OMT> (60HUO)

"~.n RDnthly lIediet ion abov. Thr••ho"ld Le"•••

(HJ I Squftr. "_t'r.)

Month: .Jan-

.lln tOO 200 JOO .00 SOD ... 711--' . aD wo ,UM

-3S
30.-: 3S
2S-: 30
20-: 2S
IS-: 2D-
to-: IS
S.: 10
0.02
0.-: ~ 4.43 4.43 1.24 0.36 0.13 0.05
-5+:
-10+: -S
0 n."
38.19
22.6'-
38.19
10.82
25.08
5.94
17,28
3.47
11.93
2.03
7.94
1.13
4.94
0.54
2.65
0.21
1.14
0.07
0.25
0.01

51. 57 38.13 27.37 \9.03 12.48 7.15 3.37 1.08 0.14 0.01
-15+ : -10 68.89 68.89
82.51 64.13 49.49 36.82 26.35 17.59 10.11 4.75 1.37 0.15 0.00
-20+ : -IS 82.51
57.82 47.24 37.69 28.75 20.93 14.35 8.74 4.45 1.39 0.07
-25+ : -20 57.82
37.28 37.28 30.79 24.81 \9;08 13.90 9.81 6.16 3.40 1.28 0.11
-25-

Honthl y Dat. lor A11- r••per_lure.

266.00 200.30 175.30 148.20 119.40 101.1 D 79.50 51. 20 33.00 6.30 1.00
Total 90 PCTL 744 .00 0.20
744.00 265.60 181.80 141.00 11 6.20 90.60 77.10 61. 30 41. 20 23.80 4.30
Hour. Hean
744.00 264.90 154.70 111.00 85.80 69.10 55.40 41.00 26.90 15.90 I. 90
10 PCTL
-13.75 -15.16 -15.36 -15.60 -15.62 -16.07 -16.11 -15.94 -13.91
"ean 90 pcrt -10.40 -9.48 -12.02
-14.33 -15.72 -16.71 -17.02 -17.23 -17.57 -17.19 -18.58 -20.13 -16.35
Air Hean -12.80 -12.14 -18.79
-15.80 -16.86 -17.60 -17.89 -18.60 -18.13 -\9.91 -ZO.13 -21.15 -24.50
Te.p 10 Pcrt -16.01
(D 8 9. eel.lus)

219. SO 162.32 114.99 76.39 45.92 22.77 7.85 0.93 0.05

Total 90 pcrt 310.66 310.66 286.52 0.01
230.86 127.55 90.23 60.31 35.35 17.33 5.44 0047
Global Kean 311.81 311.81 173."
170.49 126.52 91.57 64.84 42.81 25.00 12.28 3.09 0.06
Rad 10 PCTL 246.26 246.26
(H,lISq "etre)

228.40 228.40 167 • .,. 1\7. SI 74.93 47.88 25.10 9.64 I. 85 0.02
TolaJ 90 PCTL \,16 0.01
Hean 192.93 191.93 137.62 94.93 61.42 36.87 18.3£ 6.88
Direct
10 PCTL 135.57 135.51 91.61 61.47 39.09 22.81 11. 09 4.37 0.71
Rad
(HJ/Sq "etre)

293.88 221. 67 17'.24 153." 120.40 '1.91 50. " 10.n 1.34
Tot_J 90 pcn IOU.'" 377.82 2S•• " , .80 0.2'
336.86 248.23 200.68 170.17 134 .04 114.13 '2.32 £2.30 36.84
De9ree Hean 964.20 2.53
889.74 306.83 204.23 152.24 127.93 103.29 82.3S 60.79 39." 23.50
Day. 10 PCTL

108
CHAPTER 5 DATA PROCESSING, VALIDATION, AND QUALITY-CONTROL

5.6 TEM POR AL REP RESENTA TIVENESS

In many applications single values are used to represent long-term ave-

rage conditions. It is necessary to study the variability of the data in
order to specify the length of the record required to produce a sample mean
that will be within certain limits of the population (Le. true) mean. Black
(961) and Budyko and Drozdov (966) suggested a· technique which yields a
preferred duration for the length of the record (N) to be used to generate a
mean. But their approach ignores the serial correlation which is a typical
feature of the radiation regime (see 5.3.1). On the basis of their suggestion
N may be determined from:

N = e
(5.12)

where z' is the critical value, a is the standard deviation and e is

the re qui r ed a c cur a c y . B1 a c k (1 961 ) and Hay (1 979 b) h a ve use d E q • 5 .12 to
determine the length of required record for a variety of locations including
the southern United states (Fig. 5.17). Skaggs et a1. (1982) assessed the
impact of serial correlation and the non-Gaussian nature of the frequency
distribution of solar radiation data upon the ability to estimate a reliable
mean. Their results (Table 5.3), based on 17 years of data from a single
station (st. Paul, Minnesota,· U.S.A'), indicate that these factors substan-
tially increase the required length of record.

Figure 5.17 Map of the southern United states showing the number of years of
observation required for an estimate within 5 per cent of the true mean
95 per cent of the time. (Based on data for July). (Black, 1961).

Table 5.3 Sample size needed to estimate the true monthly mean solar radia-
tion within 1 MJm- 2 day-t at a confidence level of 95 per cent. "Random"
means the persistence has been removed from the data, "Real" means the
real data has been used.(after Skaggset a1.1982)

Total nUllber of Total nUllber of

Day. needed Year. needed
Month Random Real Random Real
Jan 25 38 0.8 1.2
Feb 52 57 1.9 2.0
Mar 135 301 4.4 9.7
Apr" 221 307 7.4 10.2
Hay 223 455 7.2 14.7
Jun 184 418 6.1 13.9
Jul 130 277 4.2 8.9
Aug 160 446 5.2 14.4
Sap 137 441 4.6 14.7
Oct 77 164 2.5 5.3
Hoy 35 93 1.2 3.1
Dec 19 47 0.6 1.5

109
CHAPTER 5 DATA PROCESSING; ·VALIDATION, AND QlIALITY-CON'fROL

5.7 SPATIAL REPRESENT·ATIVENESS

In many cases the radiation data are needed for a site other than the
measurement location. This requires a knowledge of the spatial coherency of
the radiation field. Variations in this coherency will have an influence on
the distance for which data may be extrapolated or interpolated, and hence on
the density requirements for an observational network. The spatial coherency
is a function of the integration period of observation (e.g. hour, day,
month) as well as of locational and seasonal factors. These effects have
been demonstrated in a number of papers including those of the World Meteoro-
logical Organization, 1969 and 1981; Gandin, 1970; Alaka, 1970; Pivovarova,
1978; Wilson and Petzold, 1976; Hay and Suckling, 1979; Latimer and Won,
1981; Pivovarova and Stadnik, 1969 and Hay, 1984. The results presented in
Figs. 5.18a, band 5.19 are from meso- and macroscale networks and show that
an increase in observational integration period increases the spatial cohe-
rency and hence decreases interpolation and extrapolation errors. Fig. 5.20
illustrates the change in extrapolation error, as distance increases, while
Figs. 5.21 and 5.22 provide an indication of the considerable impact that
topography has on the coherency of the solar radiation field. Distance-
dependent correlation variations such as those discussed above, are an essen
I- I-
zQ Zc::J
W •
W •
&.;.-...t ..... a) .-. .... I f , __ b)
U
.-. . I f 111
U
.-. .,lJlf"II . . **
IfIlfs
IJ...
IJ...
.~.
•
5
III • .. IJ...
IJ... -+ t-
5 , "If
11 i'J'-ltw
W
0
..1II;ifl lf ,
11
W
0 -+
t-
T • If
't'+
III It
-+ 11 If
~
t.-~ u~ T+
u~ + t- Will 111 11
z::Q III~ •
zc::J t-
o
.-. TT t- t- + + -+
0
.-.
T
I- + I-
a: a:
.-J .-J
WlQ WlQ
0:: • 0::- .•
0::0
o 0.0 20.0 40,0 60,0 80.0 ~';Jo.o 20.0 40.0 60.0 80.0
u STRT ION SEPRRRT10N LK/1) Ll STRTlON SEP8RRTION (KM)
Figure 5.18 a) Correlation 'of hourly global solar-radiation data for indi-
vidual station pairs versus the distance between that station pair. Data
involving Grouse Mountain (see Fig. 5.8) shown by (+). All other station
pairs shown by (*). Data from June 1979 to May 1980. b) Same as a) but
for daily data.

,,
,
- ' .... ....
..........
........
-"---3

0.6

0.4

0.2

o 400 1200 1600

Figure 5.19 Distance (::orrelation function for (1)" monthly and for (2) daily
global solar radiation for Jllne-Julyand for (3) samples for the entire
year. From World Meteorological Orga niza tion, 1981.

110
CHAPTER 5 DATA PROCESSING, VALIDATION, AND QUALITY-CONTROL

tial component in the determination of optimum network densities. An appro-

priate methodology for application of such information has been suggested by
Alaka (1970) and Gandin (1970). Table 5.4, which is -based on the 'optimum
interpolation' concept proposed by Gandin and the data presented in
Fig. 5.18a, describes both absolute and relative interpolation errors as a
function of station separation over mesoscale distances.

-.'" 8
...
•
'\lI
~

.,E 5
/'

-
::E
Z
0
4
.
,,
I
I
I "

j: I

~
3 I
t
> /
w I
C
2
c ,
a:
« 1
,
I

C t
Z
«
I- 0
f/l
0 200 400 800 800 1000 1200
DISTANCE BETWEEN STATIONS (km)

Figure 5.20 The standard deviation of the daily differences in' global solar
radiation forgiven station pairs in British Columbia and Alberta, Cana-
da, as a function of the distance between the station pair. Broken line
for data from Southern Canada by Wilson and Petzold, 1972. (Suckling and
Hay, 1976).

Strait
of
Georgia

Ol'l1y tDOOm contour Ihown.

•: ' , •
, ';'
It"' 10

Pit"

Figure 5.21 Correlations (xl000) of hourly values of measured global solar

radiation with values observed at Vancouver, UBC. Data for June 1979 to
May 1980 (see, also Fig. 5.10),

111
CH~PTER 5 DATA PROCESSING, VALIDATION, AND QUAL.ITY-CONTROL

Monthly. totol I March. April. Hay}

'" 1.0 ~ • 1 _. T .---.----_.,.----.--- ~~---_.
~~. a •• "'ha.. ,--
f I • .,~ • •'" '" .:a.. a· ••• ••. ;- G CANAE:iA:
~ O.!lr .~:"•• ' • • •" -~-.----- .,--
j 0.9 f .~ ~11Jt~ .' • --------

; r
ij
U 0.7
CQHIS.WGadt
'_n 1
Wo,,~""
1 ---==1: - - IL-
• Central Swit.",land
--II_. -J
o 50 100 ISO DJ 2SO 300
Stollon •• porollan di.lonce 0 (km!

Figure 5.22 Correlation coefficients for monthly sunshine totals for given
station pairs in Switzerland. For some' categories only the regression
curves are drawn. Also shown (dashed' .line) is the regression curve for
Southern -Canadian global solar -radiation "data (see Fig. 5.21). (Valko,
1980). .

Table 5.4 Errors (E) associated with the spatial interpolation of hourly and
daily values of global solar" radiation at Vancouver, Canada. Based on
the technique of Gandin (1970) and Alaka (1970).. Interpolation error (E)
in per cent of mean, kJ m- 2 h- 1 (for hourly data) and MJ m- 2 day-1 (for
daily da tal, .

Distance Hourly Daily

(km) E % E %

0 25.8 2.9 0.332 2.7

5 78.8 8.9 0.637 5.7
10 120.6 13.6 0.878 7.2
20 168.7 19.0 1.427 11.7
30 222.4 25.1 1.574 13.0
40 226.3 25.5 1.591 13.1
50 234.8 26.5 1.786 14.7

e--
". -----

a'

Figure 5.23 The relationship between the optimum interpolation error in the
. centre of a square and of a triangle and distance for different values
of the observation error. (Gandin, 1970).

112
CHAPTER 5 DATA PROCESSfNG, ",VALIDATION, AND QUALITY-C'ONTROL

The relative role of station separation, network configuration and mea~

surement error on the 'optimumoi,nterpol-ation' uncertainty has been studied by
Gandin (1970), and is illustrated in Fig. 5.23. Table 5.5 presents the inter-
polation errors for anomalies in daily radiation totals as a function of the
interpolation distance. Such results can be used for recommendations of
distances between stations. The World Meteorological Organization (1969)
suggested a mean spacing of some 500 km, when stations are located at or near
grid points. Contrarily, Pivovarova and Stadnik (1969) suggested a spacing
of 100 km (see Table 5.6) if the interpolated values were to have the same
accuracy as data at an observation point.

Table 5.5 RMS errors of linear interpolation between anomalies of daily ra-
diation totals for given mean distances between stations (Pivovarova,
1978).

Units: Megajoules Per Square Metre Per Day

Component Distance in km

January July

100 200 400 100 200 400

Daily Totals, Individual Values

DiffusB Solar 1.2 1.3 1.9

Global Solar 0.3 0.5 0.7 2.2 2.7 3.1
Net Radiation 0.6 0.8 0;8 1.4 1.8 2.2
Daily Totals Averaged Over a Month

Direct Solar 0.1 0.1 0.2 0.8 0.9 1.0

Diffuse Solar 0.2 0.2 0.2 0.3 0.3 0.5
Global Solar 0.2 0.2 0.2 0.8 0.9 1.0

Table 5.6 Optimal distance between stati·ons of an actinometric network (Pivo-

varova and Stadnik, 1969).

Units: Kilometres

Component Radiation Total

January July Year

Monthly and Yearly Totals

(By Standard Observation)

Direct Solar 50 130 115

Diffuse Solar 90 60 75
Global Solar 100 100 110
Net Radiation - 150 100

Daily Totals (Integrated)

Diffuse Solar 150

Global Solar 50 100
Net Radiation 50 60

113
 ,DA'FA P'ROCE&SDfG, llALIDA'tlON,' ANI) OUALITY-~OlfTROt. '

5.8 COIfCLUSIOltS

Radiation-data quality control requires an intensive effort if desirable

standards are to be achieved. This involves the use of both contemporary
data and numerically calculated values if unusual values are' to be isolated.
The complexity of such a scheme is readily apparent in Fig. 5.24. Considera-
ble attention must be paid to the archival process and statistical analyses
in order to ensure that the data may be fully utilized. Consideration of
spatial (network design) and temporal (observation frequency) sampling stra-
tegies will also aid in the acquisition of quality data.

Figure 5.24 , Flow diagram' .showing' interrelationship of 'thee radiation vari-

ables ~n;a qual:j;:ty c!:H;d;,rol procedure.' (Wendler; and, Eaton','t980). '

, REFERENCES

Alaka, M.A. (1970) TfleoreTlcal and practical considerationS; of network

design. Heteoroloq.ical !folJoqrapll..q .11:' ,;20.

Angstrom, A. (1924) Sol<V:" ~nd terrestrial radiation. {)uart.,J;.Royal Het. Soc.

50: 121.
~./ .. ,
Armstrong, P.R. (1980) How weather persistence affects perform~nce of solar
heating sys:t,eml?' Ut;lpublifshed M.,S,c'.:' The:sis, Department of Mechanical
Engineering, 'Colora,do state" Universiity, Fort Collins, Colorado, p. 57.

Arnfield, A.J. (1975) A note on the diurnal, latitudinal and se<jlsonal varia-
tion of the surface reflection coefficient. .1;, AjJp.l. He{. 14: 1603.

Atwater.' M.A. and J.T. Ball (1978) A nUinerical solar ra:diat'ioni model based
on standard meteorological observations. Sohfr-Bnerqy 21: ,163.

Bahm, R.J. (1978:) Reg,iqnaL 'differences in' solar radiation availability. So-
lar Age, June 1978, p. 34.
Berland, T.G. (1962) Geographical principles of the: solar radiation regime.
Proceedings All-Union Scientific Meteorology Conference, 1961: Vol. IV.

114
CHAPTER 5 DATA PROCESSING, VALIDATION, AND QUALITY-CONTROL

Black, J.N. (1961) Some aspects of the climatology of solar radiation. Paper
E/CONF. 35/S13, U.N. Conference on New Sources of Energy, Rome, p. 311.

Brinkworth, B;J. (1977) Autocorrela tion and stochastic modelling of inso-

la tion sequences. So.lar Enerqy 19: 343.

Budyko, M.I. and O.A. Drozdov (1966) On the use of means in climatological
studies. Het. and G.idr. 10: 3.

Collares-Pereira, M. and A. Rabl (1979) The average distribution of solar

radiation - correlations between diffuse. and hemispherical and between
daily and hourly insolation values. So.lar Enerqy 22: 155.

Da vies, J .A. (1981) Models for estim a ting incoming solar irradiance. Unpub-
lished Manuscript. Canadian Climate Centre Report No. 81-2, Atmospheric
Environment Service, Downsview, Ontario, p. 120.

Davies, J.A. and J.E. Hay (1980) Calculation of solar radiation incident on
a horizontal surface. In Proceedings First Canadian Solar Radiation
Data Workshop, J.E. Hay and T.K. Won eds., Canadian Atmospheric Environ-
ment Service, Downsview, Ontario, p. 32.

Davies, J.A., P.J. Robinson and M. Nunez (1970) Radiation Measurement Over
Lake Ontario and the Determination of Emissivity. First Report, March
1970, Department of Geography, McMaster University, Hamilton, Ontario,
Ca nada, p. SS.

Da vies, J .A., W. Schertzer and M. Nunez (1975) . Estimating global solar ra-
diation. Boundary layer Heteoro.loqy 9: 33.

Erbs, D.G., S.A. Klein and J.A. Duffie (1982) Estimation of the diffuse
radiation fraction for hourly, daily and monthly average global ra-
diation. So.lar £nerqy 28: 293.

Foulkes, C.W. (1980) Controlling the accuracy of solar radiation data from a
-low cost, 30 station network. Proceedings of the Sixth National Passive
Solar Conference, American Section, International Solar Energy Society,
Newark, Delaware, p. 377.

Fritz, S. (1949) Solar radiation on cloudless days. Heat.ing and Vent.i.lat.inq

46: 47.

Gandin, L.S. (1970) The planning of meteorological station networks. World

Meteorological Organization Technical Note No. 111, Geneva.

Hay, J.E. (1976) A rev'ised method for determining the direct and diffuse
components of the total short wave radiation. Atmosphere 14: 278.

Hay, J.E. (1978) Measurement and modelling of shortwave radiation on in-

clined surfaces. Preprint Volume, Third Conference on Atmospheric Ra-
diation, June 1980, Davis, California. American Meteorological Society,
Boston, Mass., p. 150.

Hay, J.E. (1979a) Study of Short wave Radiation on Non-horizontal Surfaces.

Canadian Climate Centre, Report No. 79-12, Atmospheric Environment Ser-
vice, Downsview, Ontario, p. 141.

Hay, J.E. (1979b) An analysis of solar radiation data for selected loca-
tions in Canada. Climatological Studies No. 32, Atmospheric Environment
Service, Downview, Ontario, p. 158.

Hay, J.E., (1984) An assessment of the mesoscale variability of solar ra-

diation at the earth's surface. So.lar Energy 32: 425.

115

~
/ -
CHAPI:ER 5 DATA PROCESSING, VALIDATION, AND QUALIT'l-CONTRftL

Hay, J.B. and J.A. Davies(1980) Calculation of solar radiation incident on-
an inclined surface. In Proceedings of the First Canadian Solar Radia-
tion Data Workshop, J.E. Hay and T.K. Won, eds., Canadian Atmospheric
Environment Service, Downsview, Ontario, p. 59.

Hay, J.E. and P.W. Suckling (1979) An assessment of the networks for
measuring and modelling solar radiation in British _Columbia and adjacent
areas of Western Canada. Canad.ian G'eoqrapker 23: 222.

Hay, J.E. and D.L Wardle (1982) An assessment of the uncertainty in

measurements of solar radiation. So.lar £nerqy 29: 271.

Hoyt, D.V. (1978) A model for the calculation of solar global insolation.
So.lar £nerqy 21: 27.

International Energy Agency (1980) An introduction to Meteorological

Measurements and Data Handling for Solar Energy Applications. United
States Department of Energy, DOElER-0084, Washington, D.C.,p. 133.

Kimball, H.H. (1928) Amount of solar radiation that reaches the surface of
the earth on the land and on the sea, and the methods by which it is
measured. Honth.ly fiea ther Rev. 56: 393.

Kondratyev, K. Ya, (1969) Radiation in the _Atmosphere. Academic Pres~, New

York, p. 471.

Latimer, J.R. (1972) Radiation Measurement. Technical Manual Series No. 2,

International Field Year for the Great Lakes, Information Canada, Otta-
wa, Ontario, p. 53.

Latimer, J.R. and T.K. Won (1981) Recommendations concerning -meteorological

networks for solar energy applications. International Energy Agency,
Task V, Atmospheric Environment Service, Downsview, Ontario, p. 29.

Liu, B.Y.H. and R.C. Jordan (1960) The interrelations ship and characteristic
distribution of direct, diffuse and total solar radiation. So.lar £nerqy
4: 1. -

McLaren, J.F., J.E. Hay and J.A. Davies, (1979) Define, Develope and Estab-
lisha Merged Solar and Meteorological Computer Data Base. Report to the
Atmospheric Environment Service, Downsview, Ontario, p. 250. '~

Machta, L. (1978) Workbook for the Approximate Calibration of Solar Radiation

Sensors. National Oceanic and - Atmospheric Administration,TM-ERL-ARL-70,
Silver Spring, Maryland.

Machta, L. (1980) -An ,approximate method for quality control-of solar radia-
tion in'struments. In an Introduction to Meteorological Measurements and
Data Handling for Solar Energy Applications; International Energy Agen-
cy, Task IV, United States Department of Energy, DOE/ER-0084, Washing-
ton, D.C., A-II-1 to A-II-28. -

Pal tridge, G. W. and C.M .R. PIa t t (1976) Radia ti ve Processes in Meteorology
and Climatology, Elsevier, Amsterdam, p. 318.

Pivovarov..a, Z.I. (1978) Accuracy of solar radiation data and density of net-
work of solar radiation stations. Report to the World Meteorological
Organization, Solar Energy Meeting, Geneva, October 1978, Appendix F
(b), WMO, Geneva, p. 7.

Pi v 0 v a r 0 v a , Z. L and V. V. S t a d n i k (1 969) The ace u,r a c y 0 fob s e r vat ion d a t a

from an actinometric network and optimal distances between stations.
Trudy, Gla vnaia Geofizichesk aia Obsera toriia (Leningrad) No. 249: 3.

116
CHAPTER 5 DATA PROCESSING, VALIDATION, AND QUALITY-CONTROL

Schmid, W. (1976) Aufbereitung und Qualitatskontrolle langjahriger Messun-

terlagen der Globalstrahlung und Himmelstrahlung. Working Report of the
Swiss Meteorological Institute, No. 59, Zurich, P. 47.

Sellers, W.D. (1965) Physical Climatology. University of Chicago Press,

Chicago, Illinois, p. 272.

Skaggs, R.H., D.G. Baker and J.E. Ljungkull (198~) The influence of per si-
stance and variability on the required solar radiation record length.
Solar £nerqy 28: 281.

Suckling, P.W. and J.E. Hay (1976) The spatial variability of daily values of
solar radiation for British Columbia and Alberta, Canada. Climatological
Bulletin, No. 20, McGill University, Montreal, Quebec, p. 1.

Stewart, R. (1981) Region II: Atmospheric Sciences Research Center. In Solar

Energy Meteorological Research Program, First Annual Report, Solar Ener-
gy Research Institute, Golden, Colorado, p. 11.

Unsworth, M.H. and J.L. Monteith, (1975) Long-wave radiation at the

ground. 11. Geometry of interception by slopes, solids and obstructed
planes. Quart. J. Royal Het. Soc. 101: 25.

USSR State Committee for Hydrometeorology and Control of the Natural En-
vironment: Solar Radiation and Radiation Balance Data (The World Net-
work), Leningrad.

Valko, P. (1979) Meteorological aspects of, solar energy applications.

Working Reports of the Swiss Meteorological Institute, No. 91, Zurich,
p. 11.

Valko, P. (1980) Some empirical properties of solar radiation and related

parameters. In an Introduction to Meteorological Measurements and Data
Handling for Solar Energy Applications. International Energy Agency,
Task IV, United states Department of Energy, DOE/ER-0084, Washington,
D.C., 8-1 to 8-46.

Viswanadham, Y. and R. Ramanadham (1967) The dependence of net-radiation on

short-wave radiation at Waltair. Indian J. Het. and Geophys. 18: 527.

Wendler, G. (1981) Region VII: Geophysical Institute. In Solar Energy Me-

teorological Research and Training Site Program First Annual Report,
Solar Energy Research Institute, Golden, Colorado, p. 59.

Wilson, R.G. and G.E. Petzold (1976) A solar radiation network evaluation for
the Churchill Rive'r Basin. Papers in Climatology - The Cam AlIen Memo-
rial Volume, Department of Geography, McMaster University, Hamilton,
Ontario, p. 84.

Wolfe, W.J., G. Clark and J.E. Rudzki (1980) On-site quality control of
insolation measurements. Proceedings of the American Society Mechanical
Engineers, 1980 Winter Annual Meeting, Chicago, Illinois, p. 1.

World Meteorological Organization (1960) Guide to Climatological Practices.

World Meteorological Organization WMO, No. 100. TP.44, Geneva, p. 200.

World Meteorological Organization (1969) Radiation Climatology, Suggestions

for Increasing the Number of Actinometric Stations. WMO Commission for
Climatology, Working Group on Radiation Climatology, CCI - V/DOC 30,
Geneva, p.14.

World Meteorological Organization (1981) Meteorological Aspects of the Uti-

liza tion of Solar Radiation as an Energy Source. World Meteorological
Organization Technical Note No. 172, WMO-No. 557, Geneva, p. 298.

117
 A.1. EXTRATERRESTRIAL SOLAR SPECTRUM

This compilation of the extraterrestrial solar spectrum is based on the

following publications:

200 .. 310 nm: G.Brasseur and P.C.Simon (1981) L~ Ceophys. Res. 86: 7343
310 •. 330 nm: J.C.Arvesen, R.N.Griffin and B.D.Pearson (1969) Appl. opt.
8: 2215
330 .. 869 nm: H. Neckel an D.Labs (1984) Solar PhysJcs 90: 205
870 .. IR nm: E.V.P.Smith and D.M.Gottlieb (1974) Space Sc.i.Rev. 16: 771

At the wavelength limits between the different spectra the values are forced
to be equal. After these adjustments the total spectrum is scaled in order to
yield a solar constant of 1367 Wm- 2 •

A(nm) A(nm)

199.5 0.005 5.498E-3 243.5 0.065 1.707E+0

200.5 0.007 1.253E-2 244.5 0.062 1.769E+0
201.5 0.007 1.971E-2 245.5 0.051 1.820E+O
202.5 0.008 2.758E-2 246.5 0.051 1.872E+0
203.5 0.009 3.629E-2 247.5 0.057 1.929E+0
204.5 0.009 4.575E-2 248.5 0.045 1.974E+0
205.5 0.010 5.531E-2 249.5 0.058 2.033E+0
206.5 0.010 6.549E-2 250.5 0.059 2.092E+0
207.5 0.011 7.696E-2 251. 5 0.047 2.138E+0
208.5 0.015 9.152E-2 252.5 0.044 2.182E+0
209.5 0.024 1.158E-l 253.5 0.055 2.238E+0
210.5 0.028 1.433E-1 254.5 0.061 2.299E+0
211. 5 0.034 1.771E-l 255.5 0.OS9 2.387E+0
212.5 0.030 2.073E-l 256.5 0.107 2.494E+0
213.5 0.032 2.391E-l 257.5 0.129 2.623E+0
214.5 0.041 2.801E-l 258.5 0.134 2.758E+0
i15.5 0.037 3.170E-l 259.5 0.108 2.865E+0
216.5 0.034 3.512E-l 260.5 0.102 2.967E+0
217.5 0.036 3.873E-l 261. 5 0.103 3.071E+0
218.5 0.045 4.322E-l 262.5 0.121 3.191E+0
219.5 0.048 4.799E-l 263.5 0.175 3.367E+0
220.5 0.048 5.280E-1 264.5 0.274 3.641E+0
221. 5 0.039 5.673E-l 265.5 0.280 3.921E+0
222.5 0.051 6.179E-1 266.5 0.260 4.181E+0
223.5 0.066 6.838E-1 267.5 0.270 4.451E+0
224.5 0.058 7.417E-l 268.5 0.260 4.711E+0
225.5 0.054 7.956E-l 269.5 0.252 4.964E+0
226.5 0.041 8.366E-1 270.5 0.293 5.257E+0
227.5 0.041 8.774E-l 271.5 0.232 5.489E+0
228.5 0.054 9.314E-l 272.5 0.215 5.704E+0
229.5 0.048 9.792E-l 273.5 0.204 5.908E+0
230.5 0.056 1.036E+0 274.5 0.137 6.045E+0
231.5 0.050 1.085E+0 275.5 0.200 6.245E+0
232.5 0.055 1.140E+0 276.5 0.258 6.503E+0
233.5 0.046 1.186E+0 277.5 0.240 6.744E+0
234.5 0.039 1.226E+0 278.5 0.166 6.910E+0
235.5 0.057 1.282E+0 279.5 0.089 6.999E+0
236.5 0.049 1.332E+0 280.5 0.112 7.111E+0
237.5 0.053 1.385E+0 281.5 0.231 7.342E+0
238.5 0.042 1.427E+0 282.5 0.307 7.649E+0
239.5 0.046 1.473E+0 283.5 0.330 7.979E+0
240.5 0.043 1.516E+0 284.5 0.244 8.223E+0
241.5 0.052 1.569E+0 285.5 0.141 8.364E+0
242.5 0.072 1.641E+0 286.5 0.320 8.684E+0

119
APPENDIX 1 EXTRATERRESTRIAL SOLAR SPECTRUM

A(nm) SA (Wm- 2 nm- t ) So -' A(Wm- 2 ) A (nm) . SA (Wm- 2 nm- t:) So _A(Wm- 2 )

287.5 0.371 9.055E+0 325.6 0.682 3.142E+l·

288.5 0.307 9.362E+0 326.0 0.852 3.176E+l
289.5 0.456 9.817E+0 326.4 1. 049 3.218E+l
290.5 0.623 1.044E+l 326.8 1.111 3.262£+1
291.5 0.600 1.104E+1 327.2 1.10R 3.307E+l
292.5 0.545 1.159E+l 327.6 1.050 3.349E+l
293.5 0.545 1.213E+l 328.0 0.965 3.387E+l
294.5 0.509 . 1. 264E+l 328.4 0.914 3.424E+l
295.5 0.548 1.319E+l 328.8 0.913 3.460E+l
296.5 0.492 1.368E+l 329.2 0.952 3.498E+l
297.5 0.531 1.421E+l 329.6 1.043 3.540E+1
298.5 0.413 1.462E+l 330.0 1.144 3.586E+l
299.5 0.485 1.511E+1 330.4 1.137 3.597E+l
300.5 0.403 1.551E+1 330.5 1.006 3.698E+l
301.5 0.445 1.596E+l 331.5 0.968 3.795E+l
302.5 0.484 1.644E+1 332.5 0.921 3.887E+l
303.5 0.631 1.707E+l 333.5 0.905 3.977E+l
304.5 0.610 1.768E+l 334.5 0.940 4.071E+l
305.5 0.580 1.826E+1 335·.5 0.982 4.169E+l
306.5 0.575 1.884E+l 336.5 0.765 . 4.246E+l
307.5 0.645 1.948E+l 337.5 0.866 4.332E·f1
308.5 0.613 2.010E+l 338.5 0.916. 4.424E+l
309.5 0.484 2.034E+l 339 .. 5 0.937 4.517-E+l
310.0. 0.495 2.054E+l 340.5 0.992 4.617E+l
310;4 0.507 2.074E+l 341.5 0.936 4.710E+l
310.8 0.588 2.097E+l 342.5 0.995 4.810E+l
311.2 0.707 2.126E+l 343.5 0.985 4.908R+l
311. 6 0.7A7 2.156E+l 344.5 0.7n 4.980E+l
312.0 0.707 2.184E+l 345.5 0.96'7 5.077E+l···
312.4 0.644 2.210E+1 346.5 0.919 5.169E+1·
312.l;l 0.663 2.236.E+l 347.5 0.902 5.259E+l
313.2 0.7l0. 2.265E+l 3"48.5 0.948' 5.354E+l
313.6 0.691 " 2.2928+1 349.5 0.865 5.440E+l
314.0 0.689 2.3201:;+1 350.5 1.119 5.552E+l
314 .4 0.722 2.34.9E+l 351.5 0.993 5.651E+l
314.8 0.673 2.375E+l 352.5' 0.871· 5.738E+l
315.2 0.695 2.403E+l 353.5 1.115 5.850E"+1
315.6 0.765. 2.434E+l 354.5 1.133 5.963E+l
~16.0 0.6·75 2.461E+l 355.5 1.058 6.069E+l
316.4 0.569 2.484E+l 356~5 0.938 6.163E+l
316.8 0.623 2.509E+1 357.5 0.891 6.252E+l
317.2 0.749 2.539E+l 358.5 0.627 6.314E+l
317.6 0.830 2.572E+1 359.5 1.136 6.428E+1
318.0· 0.813 2.604E+l 360.5 0.979 6.526E+1
3ULA. 0.673 2.631E+l 361.5 . 0.894 6.615E+l
3~8.8 0.642 2.657E+l 362.5 1.175 6.733E+l
319.2 0.768 2.688E+l 363.5 0.958 6.829E+l
319.6 0.759 2.718E+l 364.5 1. 015 6.930E+l
320.0 0.712 2.746E+l 365.5 1. 263 7.056E+l
320.4 0.778 2.778E+l 366.5 1. 249 7.181E+l
320.8 0.844 2.811E+l 367.5 1.214 7.303E+1
321. 2 0.847 2.845E+l 368.5 1.088 7.412E+l
321.6 0.736 2.875E+l 369.5 1.331 7.545E+1
322.0 0.695 2.902E+l 370.5 1.075 7.652E+l
322.4 0.773 2.933E+1 371.5 1. 307. 7.783E+1
322.8 0.758 2.964E+l 372.5 1.065 7.889E+l
323.2 0.646 2.989E+l 373.5 0.838 7.973E+l
323.6 0.603 3.014E+l 374.5 0.878 8.061E+l
324.0 0.604 3.038E+l 375.5 1.141 8.175E+l
324.4 0.618 3.063E+l 376.5 1.101 8.285E+l
324.8 0.654 3.089E+l 377.5 1.291 a.414E+l
325.2 0.646 3.115E+l 378.5 1.341 8.549E+l

120
APPENDIX 1 EXTRATERRESTRIAL SOLAR SPECTRUM

A(nm) SA (Wm- 2 nm- 1 ) So _ A (Wm- 2 ) A(nm) SA (Wm- 2 nm- 1 ) So _ A (Wm- 2 )

379.5 1.000 8.649E+l 441.5 1.933 1.790E+2

380.5 1.289 8.777E+l 442.5 1.982 1.810E+2
381.5 1.096 8.887E+l 443.5 1. 911 1.829E+2
382.5 0.733 8.960E+l 444.5 1.975 1.849E+2
383.5 0.684 9.029E+l 445.5 1.823 1.867E+2
384.5 1.027 9.131E+l 446.5 1.893 1.886E+2
385.5 0.954 9.227E+l 447.5 2.079 1.907E+2
386.5 1.071 9.334E+l 448.5 1.975 1.927E+2
387.5 0.966 9.430E+l 449.5 2.029 1.947E+2
388.5 0.912 9.522E+l 450.5 2.146 1.968E+2
389.5 1.227 9.644E+l 451.5 2.111 1;990E+2
390.5 1. 223 9.767E+l 452.5 1.943 2.009E+2
391.5 1.398 9.906E+l 453.5 1.972 2.029E+2
392.5 0.955 1.000E+2 454.5 1.981 2.049E+2
393.5 0.489 1.005E+2 455.5 2.036 2.069E+2
394.5 1.101 1.016E+2 456.5 2.079 2.090E+2
395.5 1.378 1.030E+2 457.5 2.102 2.111E+2
396.5 0.650 1.036E+2 458.5 1.973 2.130E+2
397.5 1. 040 1.047E+2 459.5 2.011 2.151E+2
398.5 1.538 1.062E+2 460.5 2.042 2.171E+2
399.5 1.655 1.079E+2 461.5 2.057 2.192E+2
400.5 1.649 1.095E+2 462.5 2.106 2.213E+2
401. 5 1.796 1.113E+2 463.5 2.042 2.233E+2
402.5 1.803 1.131E+2 464.5 1.978 2.253E+2
403.5 1.658 1.148E+2 465.5 2.044 2.273E+2
404.5 1.602 1.164E+2 466~5 1. 923 2.293E+2
405.5 1. 672 1.180E+2 467.5 2.017 2.313E+2
406.5 1.624 1.197E+2 468.5 1.996 2.333E+2
407.5 1.545 1.212E+2 469.5 1.992 2.353E+2
408.5 1.824 1.230E+2 470.5 1.879 2.371E+2
409.5 1.706 1.247E+2 471.5 2.020 2.392E+2
410.5 1.502 1.263E+2 472.5 2.043 2.412E+2
411. 5 1. 819 1.281E+2 473.5 1. 993 2.432E+2
412.5 1.791 1. 299E+2 474.5 2.053 2.452E+2
413.5 1.758 1.316E+2 475.5 2.018 2.473E+2
414.5 1.739 1.334E+2 476.5 1.958 2.492E+2
415.5 1.736 1.351E+2 477.5 2.077 2.513E+2
416.5 1.844 1.369E+2 478.5 2.011 2.533E+2
417.5 1.667 1.386E+2 479.5 2.078 2.554E+2
418.5 1.686 1.403E+2 480.5 2.037 2.574E+2
419.5 1.703 1.420E+2 481.5 2.092 2.595E+2
420.5 1.760 1.438E+2 482.5 2.025 2.615E+2
421.5 1.799 1.456E+2 483.5 2.021 2.636E+2
422.5 1.584 1.471E+2 484.5 1.971 2.655E+2
423.5 1.713 1.488E+2 485.5 1.832 2.674E+2
424.5 1.770 1.506E+2 486.5 1.627 2.690E+2
425.5 1.697 1.523E+2 487.5 1.832 2.708E+2
426.5 1.700 1.540E+2 488.5 1. 916 2.727E+2
427.5 1.571 1.556E+2 489.5 1.962 2.747E+2
428.5 1. 589 1.572E+2 490.5 2.009 2.767E+2
429.5 1.477 1.587E+2 491.5 1.898 2.786E+2
430.5 1.136 1.598E+2 492.5 1. 898 2.805E+2
431.5 1.688 1.615E+2 493.5 1.890 2.824E+2
432.5 1. 648 1.631E+2 494.5 2.060 2.845E+2
433.5 1.733 1.649E+2 495.5 1.928 2.864E+2
434.5 1.672 1.665E+2 496 .5 2.019 2.884E+2
435.5 1.725 1.683E+2 497.5 2.020 2.904E+2
436.5 1.'931 1.702E+2 498.5 1.868 2.923E+2
437.5 1.808 1.720E+2 499.5 1.972 2.943E+2
438.5 1. 569 1.736E+2 500.5 1.859 2.961E+2
439.5 1.827 1.754E+2 501.5 1. 814 2.979E+2
440.5 1.715 1.771E+2 502.5 1. 896 2.998E+2

121
APPENDIX 1 EXTRATERRESTRIAL SOLAR SPECTRUM

A(nm) A(nm) SA (Wm- 2 nm- 1) So _ A (Wm- 2 )

503.5 1.936 3.018E+2 565.5 1.800 4.176E+2

504.5 1.871 3.036E+2 566.5 1.831 4.194E+2
505.5 1.995 3.056E+2 567.5 1.889 4.213E+2
506.5 1.963 3.076E+2 568.5 1.812 4.231E+2
507.5 1.908 3.095E+2 569.5 1. 862 4.250E+2
508.5 1.921 3.114E+2 570.5 1.772 4.267E+2
509.5 1.918 3.134E+2 571.5 1. 825 4.286E+2
510.5 1.949 3.153E+2 572.5 1.894 4.305£+2
511. 5 1.999 3.173E+2 573.5 1.878 4.323E+2
512.5 1.869 3.192£+2 574.5 1. 869 4.342E+2
513.5 1.863 3.210E+2 575.5 1.832 4.360E+2
514.5 1.876 3.229E+2 576.5 1.848 4.379E+2
515.5 1.902 3.248E+2 577.5 1.859 4.397E+2
516.5 1.671 3.265E+2 578.5 1.786 4.415E+2
517.5 1. 728 3.282E+2 579.5 1.830 4.434E+2
518.5 1.656 3.299E+2 580.5 1.840 4.452E+2
519.5 1.830 3.317E+2 581.5 1.855 4.471E+2
520.5 1.833 3.335£+2 ' 582.5 1.875 4.489E+2
521. 5 1.908 3.354E+2 583.5 1.859 4.508E+2
522.5 1.825 3.373E+2 584.5 1.862 4.526E+2
523.5 1.896 3.392E+2 585.5 1.786 4.544E+2
524.5 1.960 3.411E+2 586.5 1.832 4.563E+2
525.5 1.932 3.431E+2 587.5 1.850 4.581E+2
526.5 1.676 3.447E+2 588.5 1.752 4.599E+2
527.5 1.830 3.466E+2 589.5 1.614 4.615E+2
528.5 1.899 3.485£+2 59'0.5 1.815 4.633E+2
529.5 1. 920 3.504E+2 591.5 1.789 4.651E+2
530.5 1.954 3.523E+2 592.5 1.810 4.669E+2
531.5 1.965 3.543E+2 593.5 1.798 4.687E+2
532.5 1.773 3.561£+2 594.5 1.776 4.705E+2
533.5 1.925 3.580E+2 595.5 1.785 4.723E+2
534.5 1.860 3.599E+2 596.5 1.8D7 4.741E+2
535.5 1.9'92 3.619E+2 597.5 1.783 4.758E+2
536.5 1.873 3.637E+2 598.5 1.760 4.776£+2
537.5 1.884 3:656E+2 599.5 1.777 4.794E+2
538.5 1.906 3.675E+2 600.5 1.748 4,.811E+2
539.5 1.834 3.694E+2 601.5 1.753 4.829£+2
540.5 1.772 3.711E+2 602.5 1.721 4.846£+2
541. 5 1.883 3.730E+2 603.5 1.789 4.864E+2
542.5 1.827 3.748E+2 604.5 1. 779 4.882E+2
543.5 1. 881 3.767E+2 605.5 1.766 4.899E+2
544.5 1. 881 3.786E+2 606.5 1.762 4.917E+2
545.5 1.903 3.805E+2 6'07.5 1.760 4.935E+2
546.5 1.881 3.824£+2 608.5 1.745 4.952E+2
547.5 1.835 3.842E+2 609.5 1. 746 4.969E+2
548.5 1.865 3.861E+2 610.5 1.705 4.987£+2
549.5 1.897 3.880E+2 611.5 1.748 5.004E+2
550.5 1.864 3.898E+2 612.5 1.707 5.021E+2
551.5 1.873 3.917E+2 613.5 1.685 5.038£+2
552.5 1.848 3.936E+2 614.5 1.715 5.055E+2
553.5 1.884 3.954E+2 615.5 1. 715 5.072E+2
554.5 1.900 3.973E+2 61'6.5 1.611 5.088E+2
555.5 1.899 3.992E+2 617.5 1.709 5.105E+2
556.5 1.823 4.011E+2 618.5 1.726 5.123E+2
557.5 1.848 4.029E+2 619.5 1.7'09 5.140E+2
558.5 1.789 4.047E+2 ,62'0.5 1.736 5.157E+2
559.5 1.810 4.065E+2 621.5 1.692 5.174E+2
560.5 1.845 4.084E+2 '622.5 1.715 5.191E+2
561.5 1.826 4. W2E+2 623.5 1.668 5.208E+2
562.5 1.852 4,.120E+2 624.5' 1.658 5.224E+2
563.5 1.863 4.139E+2 625.5 1.634 S.241E+2
564.5 1.856 4.158E+2 626.5 1.699 5.258;E+2

12,2
APPENDIX r EXTRATERRESTRIAL SOLAR SPECTRUM

A(nm) SA (Wm- 2 nm- 1 ) So _ A (Wm- 2 ) A(nm) SA (Wm- 2 nm- 1 ) So _ A (Wm- 2 )

627.5 1.699 5.275E+2 749.0 1.271 7.062E+2

628.5 1. 699 5.292E+2 751.0 1.263 7.088E+2
629.5 1.679 5.317E+2 753.0 1.260 7.113E+2
631.0 1.641 5.350E+2 755.0 1. 256 7.138E+2
633.0 1.653 5.383E+2 757.0 1.249 7.163E+2
635.0 1.658 5.416E+2 759.0 1. 241 7.188E+2
637.0 1.656 5.449E+2 761. 0 1.238 7.213E+2
639.0 1. 653 5.482E+2 763.0 1.242 7.237E+2
641.0 1.616 5.514E+2 765.0 1.222 7.262E+2
643.0 1. 623 5.547E+2 767.0 1.186 7.286E+2
645.0 1.629 5.579E+2 769.0 1.204 7.310E+2
647.0 1.605 5.612E+2 771. 0 1.205 7.334E+2
649.0 1. 560 5.643E+2 773.0 1. 209 7.358E+2
651.0 1.608 5.675E+2 775.0 1.189 7.382E+2
653.0 1.601 5.707E+2 777.0 1.197 7.406E+2
655.0 1. 534 5.738E+2 779.0 1.188 7.430E+2
657.0 1.386 5.765E+2 781.0 1. 188 7.453E+2
659.0 1.551 5.796E+2 783.0 1.177 7.477E+2
661. 0 1. 573 5.828E+2 785.0 1.181 7.500E+2
663.0 1. 557 5.859E+2 787.0 1.178 7.524E+2
665.0 1.562 5.890E+2 789.0 1.175 7.548E+2
667.0 1.537 5.921E+2 791.0 1.159 7.571E+2
669.0 1. 548 5.952E+2 793.0 1.144 7.594E+2
671.0 1. 518 5.982E+2 795.0 1.135 7.616E+2
673.0 1.523 6.013E+2 797.0 1.153 7.639E+2
675.0 1. 512 6.043E+2 799.0 1.136 7.662E+2
677.0 1.510 6.073E+2 801.0 1.143 7.685E+2
679.0 1.500 6.103E+2 803.0 1.130 7.708E+2
681.0 1.494 6.133E+2 805.0 1.116 7.730E+2
683.0 1.481 6.163E+2 807.0 1.121 7.752E+2
685.0 1.457 6.192E+2 809.0 1.096 7.774E+2
687.0 1.469 6.221E+2 811. 0 1.115 7.797E+2
689.0 1.463 6.250E+2 813.0 1.116 7.819E+2
691.0 1.450 6.279E+2 815.0 1.108 7.841E+2
693.0 1.450 6.308E+2 817.0 1.105 7.863E+2
695.0 1.438 6.337E+2 819.0 1.065 7.884E+2
697.0 1.418 6.365E+2 821. 0 1. 081 7.906E+2
699.0 1.427 6.394E+2 823.0 1. 074 7.928E+2
701.0 1.388 6.422E+2 825.0 1.076 7.949E+2
703.0 1.390 6.450E+2 827.0 1.077 7.971E+2
705.0 1.417 6.478E+2 829.0 1. 073 7.992E+2
707.0 1. 402 6.506E+2 831. 0 1.069 8.014E+2
709.0 1.386 6.534E+2 833.0 1.034 8.034E+2
711.0 1. 387 6.561E+2 835.0 1.053 8.055E+2
713.0 1.375 6.589E+2 837.0 1.052 8.076E+2
715.0 1. 368 6.616E+2 839.0 1.042 8.097E+2
717.0 1.355 6.643E+2 841.0 1. 045 8.118E+2
719.0 1.329 6.670E+2 843.0 1. 028 8.139E+2
721.0 1.332 6.697E+2 845.0 1.033 8.159E+2
723.0 1. 349 6.724E+2 847.0 1. 025 8.180E+2
725.0 1.351 6.751E+2 849.0 0.971 8.199E+2
727.0 1.347 6.778E+2 851.0 1. 003 -8.219E+2
729.0 1. 320 6.804E+2 853.0 0.973 8.239E+2
731.0 1. 327 6.830E+2 855.0 0.877 8.256E+2
733.0 1.319 6.857E+2 857.0 1. 011 8.276E+2
735.0 1.310 6.883E+2 859.0 0.997 8.296E+2
737.0 1.308 6.909E+2 861.0 0.997 8.316E+2
739.0 1.279 6.935E+2 863.0 0.999 8.336E+2
741.0 1.259 6.960E+2 865.0 0.970 8.356E+2
743.0 1.287 6.986E+2 867.0 0.880 8.373E+2
745.0 1.280 7.011E+2 869.0 0.967 8.393E+2
747.0 1. 284 7.037E+2 871.0 0.986 8.412E+2

123
APPENDIX 1 EXTRATERRESTRIAL SOLAR SPECTRUM

i\.(nm) S~ (Wm- 2 nm- 1 ) So - ~ (Wm- 2 ) i\. (nm) S~ (Wm- 2 nm- 1 ) So - A (Wm- 2 )

873.0 0.978 8.432E+2 997.0 0.743 9.465£+2

875.0 0.981 8.451£+2 999.0 0.743 9.491E+2
877.0 0.984 8.471E+2 1002.5 0.745 9.528£+2
879.0 0.959 8.490E+2 1007.5 0.737 9.565£+2
881.0 0.960 8.509E+2 1012.5 0.734 9.602E+2
883.0 0.948 8.528£+2 1017.5 0.721 9.638£+2
885.0 0.963 8.548£+2 1022.5 0.704 9.673E+2
887.0 0.947 8.567£+2 1027.5 0.708 9.708E+2
889.0 0.949 8.586£+2 1032.5 0.688 9.743E+2
891.0 0.944 8.605£+2 1037.S 0.692 9.777E+2
893.0 0.934 8.623£+2 1042.5 0.681 9.811£+2
895.0 0.936 8.642£+2 1047.5 0.685 9.846E+2
897.0 0.939 8.661E+2 1052.5 0.661 9.879E+2
899.0 0.912 8.679£+2 1057.5 0.650 9.911E+2
901.0 0.905 8.697£+2 1062.5 0.642 9.943£+2
903.0 0.905 8.715E+2 1067.5 0.643 9.975E+2
905.0 0.893 8.733£+2 1072.5 0.638 1.001£+3
-907.0 0.891 8.751E+2 1077.5 0.630 1.004£+3
909.0 0.861 8.768£+2 1082.5 0.620 1.007E+3
911.0 0.870 8.785£+2 1087.5 0.614 1.010E+3
913.0 0.876 8.803E+2 1092.5 0.612 1.013£+3
915.0 0.866 8.820E+2 1097.5 0.599 1.016£+3
917.0 0.859 8.837E+2 1102.5 0.608 1.019E+3
919.0 '0.858 8.855£+2 11 07.5 0.601 1.02.2E+3
921.0 0.830 8.871£+2 1112.5 0.603 1.025E+3
923.0 0.821 8.888E+2 1117.5 0.589 1.028£+3
925.0 0.825 8.904£+2 1122.5 0.579 1.031£+3
927.0 0.828 8.921£+2 1127.5 0.569 1.034£+3
929.0 0.833 8.937£+2 1132.5 0.566 1.037E+3
931. 0 0.826 8.954E+2 1137.5 0.563 1.039£+3
933.0 0.832 8.971E+2 1142.5 0.557 1.042E+3
935.0 0.818 8.987£+2 1147.5 0.556 1.045£+3
937.0 0.802 9.003£+2 1152.5 0.545 1.048£+3
939.0 0.808 9.019E+2 1157.5 0.554 1.051£+3
941 .0 0.800 9;035£+2 1162.5 0.540 1.053£+3
943.0 0.784 9.051£+2 1167.5 0.530 1.056£+3
945.0 0.799 9.067E+2 1172.5 0.533 1.059£+3
947.0 0 .. 793 9.083£+2 1177.5 0.525 1.061£+3
949.0 0.777 9.098£+2 1182.5 0.514 1.064£+3
951. 0 0.778 9.114£+2 1187.5 0.512 1.066E+3
953.0 0.771 9.129£+2 1192.5 0.511 1.069£+3
955.0 0.760 9.144E+2 1197.5 0.502 1.071E+3
957.0 0.774 9.160£+2 1202.5 0.496 1.074£+3
959.0 0.771 9.175£+2 1207.5 0.494 1.076£+3
961. 0 0.767 9.191£+2 1212.5 0.489 1 .079E+3
963.0 0.767 9.206E+2 1217.5 0.500 1.081E+3
965.0 0.764 9.221£+2 1222.5 0.481 1.084E+3
967.0 0.757 9.236£+2 1227.5 0.481 1.086£+3
969.0 0.776 9.252£+2 1232.5 0.4'84 1.089£+3
971.0 0.763 9.267£+2 1237.5 0.477 1.091£+3
973.0 0.764 9.282£+2 1242.5 0.477 1.093£+3
975.0 0.750 9.297E+2 1247.5 0.466 1.096£+3
977.0 0.768 9.313£+2 1252.5 0.474 1.098£+3
979.0 0.768 9.328£+2 1257.5 0.463 1.100£+3
981.0 0.762 9.343£+2 1262.5 0.444 1.103£+3
983.0 0.766 9.359£+2 1267.5 0.438 1.105£+3
985.0 0.771 9.374E+2 1272.5 0.439 1.107£+3
987.0 0.756 9.389£+2 1277.5 0.453 1.109E+3·
989.0 0~767 9.405E+2 1282.5 0.435 1.11LE+3
991. 0 0.764 9.4201:;+2 1287.5 0.437 1.114£+3
993.0 0.755 9.435E+2 1292.5 0.442 1.116£+3
995.0 0.756 9.450£+2 1297.5 0.438 1. 118E+3

124
APPENDIX 1 EXTRATERRESTRIAL SOLAR SPECTRUM

;\(nm) SA (Wm- 2 nm- 1 ) So _ A (Wm- 2 ) - ;\(nm) SA (Wm- 2 nm- 1 ) So _ A (Wm- 2 )

1302.5 0.438 1.120E+3 1612.5 0.244 1.220E+3

1307.5 - 0.429 1.122E+3 1617.5 0.243 1.222E+3
1312.5 0.419 1.124E+3 1622.5 0.240 1.223E+3
1317.5 0.416 1.126E+3 1627.5 0.244 1.224E+3
1322.5 0.416 1.129E+3 1632.5 0.241 1.225E+3
1327.5 0.411 1.131E+3 1637.5 0.237 1.226E+3
1332.5 0.405 1.133E+3 1642.5 0.234 1.228E+3
1337.5 0.400 1.135E+3 1647.5 0.235 1.229E+3
1342.5 0.398 1.137E+3 1652.5 0.234 1.230E+3
1347.5 0.394 1.139E+3 1657.5 0.234 1.231E+3
1352.5 0.387 1.140E+3 1662.5 0.233 1.232E+3
1357.5 0.382 1.142E+3 1667.5 0.229 1.233E+3
1362.5 0.378 1.144E+3 1672.5 0.228 1.234E+3
1367.5 0.370 1.146E+3 1677.5 0.220 1.236E+3
1372.5 0.369 1.148E+3 1682.5 0.221 1.237E+3
1377.5 0.368 1.150E+3 1687.5 0.219 1.238E+3
1382.5 0.364 1 . .152E+3 1692.5 0.219 1.239E+3
1387.5 0.364 1.153E+3 1697.5 0.214 1.240E+3
1392.5 0.358 1.155E+3 1702.5 0.217 1.241E+3
1397.5 0.357 1.157E+3 1707.5 0.212 1.242E+3
1402.5 0.353 1.159E+3 1712.5 0.203 1.243E+3
1407.5 0.350 1.161E+3 1717.5 0.212 1.244E+3
1412.5 0.346 1.162E+3 1722.5 0.205 1.245E+3
1417.5 0.344 1.164E+3 1727.5 0.196 1.246E+3
1422.5 0.343 1.166E+3 1732.5 0.190 1.24?E+3
1427.5 0.348 1.167E+3 1737.5 0.189 1.248E+3
1432.5 0.337 1.169E+3 1742.5 0.191 1.249E+3
1437.5 0.331 1.171E+3 1747.5 0.185 1.250E+3
1442.5 0.327 1.172E+3 1752.5 0.187 1.251E+3
1447.5 0.318 1.174E+3 1757.5 0.189 1.252E+3
1452.5 0.323 1.176E+3 1762.5 0.184 1.253E+3
1457.5 0.307 1.177E+3 1767.5 0.182 1.254E+3
1462.5 0.317 1.179E+3 1772.5 0.177 1.255E+3
1467.5 0.311 1.180E+3 1777.5 0.173 1.255E+3
1472.5 0.311 1.182E+3 1782.5 0.171 1.256E+3
1477.5 0.307 1.183E+3 1787.5 0.170 1.257E+3
1482.5 0.303 1.185E+3 1792.5 0.169 1.258E+3
1487.5 0.298 1.186E+3 1797.5 0.173 1.259E+3
1492.5 0.303 1.188E+3 1802.5 0.169 1.260E+3
1497.5 0.300 1.189E+3 1807.5 0.168 1.261E+3
1502.5 0.296 1.191E+3 1812.5 0.160 1.261E+3
1507.5 0.295 1.192E+3 1817.5 0.160 1.262E+3
1512.5 0.290 1.194E+3 1822.5 0.159 1.263E+3
1517.5 0.290 1.195E+3 1827.5 0."156 1.264E+3
1522.5 0.286 1.197E+3 1832.5 0.156 1.264E+3
1527.5 0.290 1.198E+3 1837.5 0.150 1.265E+3
1532.5 0.282 1.200E+3 1842.5 0.153 1.266E+3
1537.5 0.274 1.201E+3 1847.5 0.151 1.267E+3
1542.5 0.275 1.202E+3 1852.5 0.148 1.267E+3
1547.5 0.274 1.204E+3 1857.5 0.145 1.268E+3
1552.5 0.273 1.205E+3 1862.5 0.143 1.269E+3
1557.5 0.272 1.206E+3 1867.5 0.143 1.270E+3
1562.5 0.269 1.208E+3 1872.5 0.135 1.270E+3
1567.5 0.263 1.209E+3 1877.5 0.135 1.271E+3
1572.5 0.260 1.210E+3 1882.5 0.140 1 .272E+3
1577.5 0.259 1.212E+3 1887.5 0.138 1.272E+3
1582.5 0.255 1.213E+3 1892.5 0.137 1.273E+3
1587.5 0.252 1.214E+3 1897.5 0.138 1.274E+3
1592.5 0.246 1.215E+3 1902.5 0.133 1.274E+3
1597.5 0.246 1.217E+3 1907.5 0.136 1.275E+3
1602.5 0.247 1.218E+3 1912.5 0.138 1.276E+3
1607.5 0.242 1.219E+3 1917.5 0.136 1.276E+3

125
APPENPIX 1 EXTRATERRESTRIAL SOLAR SPECTRUM

X(nm) S}. (. Wm- 2 nm- 1) So _}. (Wm- 2 ) X(nm)

1922.5 0.134 1.277E+3 2987.5 0.025 1.341E+3

1927.5 0.132 1.278£+3 3025.0 0.024 1.342E+3
1932.5 0,132 1.278E+3 3075.0 0.023 1.343E+3
1937.5 0.131 1.279E+3 3125.0 0.021 1.344E+3
1942.5 0.129 1.280£+3 3175.0 0.020 1.345E+3
1947.5 0.127 1.280E+3 3235.0 0.019 1.346E+3
1952.5 0.126 1.281£+3 3295.0 0.018 1.347E+3
1957.5 0.122 1.282£+3 3355.0' 0.016 L 348E+3
1962.5 0.126 1.282£+3 3425.0 0.015 1.349E+3
1967.5 0.125 1.283£+3 3495.0 0.014 1.350E+3
1972.5 0.125 1.284£+3 3575.0 0.013 1.352E+3
197.7.5 0.129 1.284E+3 3665.0 0.012 1.353E+3
1982.5 0.125 1.285£+3 3755.0 0.011 1.354£+3
1987.5 0.123 1.285E+3 3855.0 0.010 1.355E+3
1992.5 0 .. 121 1.286E+3 3965.0 0.009 1.356E+3
1997.5 0 .. 123 1.287£+3 4085.0 0.008 1.357E+3
2002.5 0.116 1.287E+3 4225.0 0.007 1.358E+3
2012.5 0.114 1.288E+3 4385.0 0.006 1.359E+3
2022.5 0.113 1.289£+3 4575.0 0.005 1.360E+3
2032.5 0.11 0 1.291E+3 4805.0 0.004 1.361E+3
2042.5 0.107 l,292E+3 5085.0 0.003 1.362E+3
2052.5 0.104 1.293E+3 5445.0 0.002 1.363£+3
2062.5 0.10·0 1.294E+3 5925.0 0.002 1.364E+3
2077.5 0.101 1.295E+3 6615.0 0.001 1.365E+3
2092.5 0.098 1.297E+3 7785.0 0.001 1.366£+3
2107.5 0.093 1.298E+3 10075.0 0.000 1.367E+3
2122.5 0.087 1.299E+3
2137.5 0.085 1.301E+3
2152.5 0.081 1.302£+3
2167.5 0.080 1.303E+3
2182.5 0.075 1.304E+3
2197.5 0.073 1.305E+3
2212.5 0.075 1.307E+3
2227.5 0.075 1.308£+3
2247.5 0.072 1.'309E+3
2262.5 0.071 1.310E+3
2282.5 0.069 1.311£+3
2302.5 0.'066 1.313£+3
2322.5 0.053 1.314£+3
2342.5 0.058 1.315E+3
2362.5 0.065 1.316E+3
2382.5 0.055 1.317£+3
2402.5 0.054 1.319E+3
2422.5. 0.057 1.320E+3
2442.5 0.051 1.321E+3
2467.5 0.053 1.322E+3
2492.5 0.054 1.323E+3
2517.5 0.047 1.325£+3
2542.5 0.046 1.326E+3
2567.5 0.044 1.327£+3
2592.5 0.042 1.328E+3
2617.5 0.041 1.329E+3
2642.5 0.039 1.330£+3
2672.5.. 0.038 1.331E+3
2702.5 0.036 1.332E+3
2732.5 0.035 1.333E+3
27G2.5 0.034 1.334E+3
2797.5 0.032 1.33SE+3
2832.5 0.031 1.336E+3
2867.5 0.029 1.338E+3
2907.5 0.028 L 339£+3
2947.5 0.026 1.340£+3

126
 A.2. NOMENCLATURE OF RADIATION QUANTITIES

The terms and definitions of radiative quantities, recommended by the

Radiation Commission (IAMAP) and adopted by IUGG has been discussed in:

E. Raschke (980) Terminology and Units of Radiation Quantities and Measure-

ments, Radiation Commission (IAMAP), Boulder, Colorado, p.17.
CIMO Guide (1983) Chapter 9: Solar Radiation, Meteorological Instruments and
Methods of Observation, 5th ed., WMO, Geneva.

In the following tables a summary of the nomenclature is presented.

Table A.2.1 Definition of derived directional quantities. .Q = (8,4» speci-

fies t he direction of. incidence (i), sca t teri ng (s), reflection (r) or
transmission (t).

5 CAT T E R I N G R EFL E C T ION ~RA N S M I S S ION

2
in a -volume element dV [.m] ] at an area elcmtlut dA [m ) through an area element dA 1m2 }

For directional incoming and outgoing radiation

scattering function reflec tion func tion transmission function

3 . 3
_ d ~s(lli,lls) -i-i
d3~ r ([Ii ,Il ) d ~t ([Ii,ll t )
[sr ~iJ
r [sr-i]
Ys(ll ,Il ) = dV'dll 'dE (Il ) [m sr ]
i s
Yr([li ,Il r ) - 2 'Y t ([Ii ,Il t ) - 2
s i i cos" 'dll 'd x cos "t 'dlll'd x
r r
2
where d x • cos"idEi(lli)dA
scattering indicatrix reflection indicatrix transmission ind1£atrix
(or scattering phase functions p(~s»

4nfl 'Oil TIll

~s(~s) " ~s(lli,lls) "Ts . Ys(lli,ll s ) ~r(lli,rlr) " P(Ilr> 'Yr(lli,ll r ) ~t(rli,llt) " T([Ir> • y t (ll i ,ll )
t

For directional incoming, but multidirectional outgoing radiation

scattered Into a sphere reflect~d ~r transmitted -into a hemisphere

hemispherical reflectance hemispheric"l transmittance

scattering coefficient
for directional incidence for directional lnc.ldence

cr
8
• J
411'-0
Y (fl ,Il ) dO
8 i 8 •
[1I- i ) P(ll i ) - f Yr(lli'rlr) cos"r d!l r T(ll ) -
i f Yt (Ili ,ll t )C08 "tdll t
2TT ·Il 211'rl
0 0 0

- For directional outgoing, but multidirectional incoming radiation

incident from a sphere incident from a hemisphere

scattering source function reflection source function transmission source function

J (H )
s s
• ..!..
0 f Ys(lli,lls)dEr'lli) J (Il )
r- r
. f y/rli,rlr)cos"idF.i(rl i ) Jt(rl ) •
t f y (ll i ,llt)cos"idEi (P. )
t i
s 4T1'rl 21T·1l 2T1 ·Il
0 0 0

[\01 -2
m sr-i) [\01 -2 -h
m sr J [\01 m-2 sr -i)
directional reflectance factor directional transmittance factor
for hemispherical incide~ for hemispherical incidtmc-e- -

P(ll ) •
r
J (ll )'TI'rl
r r
Et
o T(rl )
t
. Jt~llt)'1T'rlo
E
i

where Ei - f
21T'1l
cos "idEi (rl i ) [\01 m-2)

127
APPENDIX 2 NOMENCLATURE OF RADIATION QUANTITIES

Table A.2.2 Radiometric quantities. Symbols in parentheses are proposed for

alternative use.

NAMES SYMBOL UNIT RELATION eIE-no.

radiant
energy Q. (W) J . W 8 ,45-05-130

radiant
flux tIl'. (1') w tI>'. =-
d
dt
Cl
pow-er 45-05-135

radfan,t Radiant flux of any

-2 dtl>.k.
flux CM) , CB) Wm
·dA dA d't origin 45-05~155
density crossing an area element

Radiant flux of any

radiant -2 origin,
M Wm 45-05-170
exitance* emerging from an area
elemen,t

Radiant flux of any

W, -2 dtl> origin
irradiance B m E - dA
- 45-05-160
incident onto an' area
element

-1
2 The radiance is a
-2 d t1>
radiance L Wm, sr L • ""'dO-=-'',---=d;"'A--=-,c-o-s~" . conservative quantity fn
an optical system

radiant -2 May be used for daily

J'm'
exposure sums of global 45-05-165
(per expo-
sure' time) radiation, etc.

rad'iant May be used only

intensity I W sr-1 1 • dtl> for radiation outgoing
dO
fr01ll '''point sources"

T,able A,.2.3 Summary of propeTties of media.

CHARACTERISTICS SYMBOL DEFINITION REMARKS

Me:
'emittance £
£ - M £ - 1 for a black body
£... 1

~ and 4!~, are the

a -&:

absorptance absorbed and incident

radiant flux"
respectively

~.r and
. ~i are the
If!
r
reflectance P • f"""" reflected: and' incident
1 radiant flux,
respectively

<Il :ts the incident flux

i
and llJ is the radiant
tr-ansmi t tance t
flux transmitted through
a.. layer or a surface
APPENDIX 2 NOMENCLATURE OF RADIATION QUANTITIES

Table A.2.4 Summary of meteorological radiation quantities as adopted by

WMO,

No. Symbol Relation Definitions, remarks Units

1,1 ~.j.l ~+=~, ++~. + Downward radiant flux W
0+ 0+=0, ++0. + " radiant energy J (W s)
H+ H+=H, ++H. + " radiant exitance 3 Wm" ,
E+ E+=E, ++E. + " irradiance Wm" 2
L+ L+=L. ++L, + " radiance Wm- , sr- ,
H+ H+=H, ++H. + " radiant exposure Jm- , per
(g=global) for a specif ied time
(k=long-wave) time interval interval
-
1.2 E, + E, +=S'cos8+E d + Hemi spheri cal radiation on Wm-'
horizontal surface
8=solar zeni th-ang le)'
2.1 ~+I ~+=~. ++~. + Upward radiant flux W
0+ 0+=0. ++0. + " radiant energy J (W s)
H+ H+=H. 1'+11. l' " radiant ex i tance 3 Wm" ,
Et Et=E. ++E. l' " irradiance· Wm- ,
L+ L+=L. ++L. t .." radiance Wm" , sr-- •
Ht Ht=H. ++H. + radiant energy Jm- , per
per unit area time
for a specified interval
time interval
2.2 ~d +, ,~d + Subscript d=diffuse As 1
Od 1', Od +
K. 1', Hd +
Ed l' , Ed +
L. 1', L.
H. +, H.
+
+ .
2.3 ~. + ,~. + Subscript R=long-wave As 1 or
0. 1',0. + If only atmospheric radiation As 2
H.+, H. + is considered, the subscript
E. 1', E. + a may be added, e,g. t. , • +
L. +, L. +
H. +, H. +
2.4 ~. + Subscript r=reflected As 2
O. + (the subscripts
H. l' s=specular and d=diffuse
E. + may be used, if a distinction
L. + is to be made between these
H. + two components)
3 41* ~*=~+-~+ The subscripts g or R are to As the
Q* 0*=0+-41+ be added to each of the respec-
H* H*=IH-Ht symbols if only short-wave or tive
E* E*=E+-Et long-wave net radiation symbols
L* L*=L+-L+ quanti ties are considered e.g.in 1
H* H*=H+-Ht
4 S S=S•. r r=atm. transmittance Wm-'
T =e-'/~'" 8=optical depth (vertical)
8=solar zenith-angle
S. - Solar irradiance outside the Wm- ,
atmosphere, normalized to
mean sun-earth distance

I The symbols - or + could be used instead of +, l' (e. g. ~. ;0 ~+) .

, In the case of inclined surfaces 8 is the angle between the normal of
the surface and the direction to the sun,
3
Exitance = r:adiant flux emerging from the unit area; ir:radiance = ra-
diant flux received per unit area. For the flux density in general the
symbol either H or E can be used. Though not specifically recommended,
the sYlllbol F '" 4>/area may alao be introduced.

129
 A.3. CRITERIA FOR THE CHOICE OF INSTRUMENTS

For measuring specific radiation quantities many different types of

instruments are on the marke and a choice among <them can be complicated.

The following general criteria should be considered in making the

choice:

A.J.t Accuracy

The achievement of high accuracy in radiation measurements is an in-

ordinately difficult task, and even moderate accuracy in operational instru-
ments is a goal which is difficult to achieve on a routine basis. Thus one
of the first prerequisites in the choice of radiation instruments is that ~he
demonstrated accuracy (and not just that claimed by the manufacturer) be
commensurate with the measurement requirements.

A.3.2 Sensi ti vity

The sensitivity of the radiation, sensor itself is not as critical as it

used to be, because of the availability of very stable and inexpensive ampli-
fiers. However, other things being equal, an instrument with good sensi ti vi-
ty would take priority over one with lower sensitivity.

A.3.J Reliability and Long Term Stability

Both of these criteria are critical in operational instruments. Very

frequent checks of performance may become a burden in an operational set-
ting, and instrument instabilities unavoidably cause a degradation in data
quality. The best assurance of long term stability is the satisfactory per-
formance of similar devices from a reputable manufacturer.

A.J.4 Simplicity and Convenience

Routine day-after-day operation demands that an instrument be convenient

to use. In addition, the device should be simple in design s,o as to be under-
standable by non-scientist operators.

A.J.5 Time of Response

This factor must be evaluated on an individual I;l.pplication basis. Rou-

tine monitoring at .a fixed location may be more advantageo'usly accomplished
by time averaging with a moderately slowly responding device; The applica-
tion must decide, but, in general, an instrument responding more rapidly than
necessary is better than one which is too sluggish in response.

A.J.6 Temperature Sensitivity

Since operational radiometres are normally exposed to large extremes of

temperature, the temperature sensitivity of such devices is a serious source
of error. Temperature compensation of operational instruments is highly
desirable, if not mandatory, for good quality measurements. Unfortunately,
not all temperature compensation methods are completely successful, in which
case further correction of the data may be indicated. Furthermore, the tem-

131
APPENDIX 3 CRITERIA FOR THE CHOICE OF INSTRUMENTS

perature compensation effectiveness may deteriorate with time. Annual checks

of temperature sensitivity are a small price t.o pay for an assurance that
this potentially large source of error is kept within permissible bounds.

A.3.7Cost
The cost factor tends to be greatly overemphasized in the choice of
instruments. The purchase cost is usually so s,ma.ll relative to the. costoE
long term operations that it pales to insignificance in the overall program.
Only for networks of instruments does the first cost assume a major signifi-
cance, and,even then, economies in the cost of basic sensors are likely to
be mistaken economies. It is normally good judgment to g.et the best quality
instrument available for a particular task, regardless of the first cost of
the device. .

A.3.B Genera1 Remark

In ord'er to avoid dt vergent instrument c'haract·eristics ina given sta-
tion or network, it is a good practice to utilize the same type of instrument
from the same manufacturer whenever possible..
 A.4~ GLOSSARY OF SYMBOLS

B SchUepp turbidity coefficient

B (A,T) Planck black body radiance

CM ,Cs ,Ca single particle cross section for extinction, scattering and
absorbtion respectively

E· irradiance

EH>}) upward (downward) irradiance

El' ,E d , El reflected, diffuse, longwaveradiation

E* net radiation

'FH>}) upward (downward) flux density

J source function

K degrees Kelvin

L radiance

Lb black body radiance

P degree of polarization

Ro mean earth-sun distance

R actual earth-sun distance

S direct solar irradiance

So solar irradiance at·the top of the atmosphere·

T Linke turbidity factor

m complex index of refraction, vertical optical path (air mass)

m,. , ml real part, imaginary part of the index of refraction

ms slant optical path (air mass)

m,. relative optical path (air mass)

p scattering phase function, barometric pressure

Po sea level pressure 1013w25mb

W' absorber amount

g assymetry factor

d Julian day

a absorption coefficient

a. z absorptance for ozone

133
APPENDIX 4 GLOSSARY OF SYMBOLS

a" y absorptance for water vapour

fJ Angstrom turbidi tyc;oe.fficient

r solar elevation angle

3 total optical depth, solar declination

3R Rayleigh optical depth

3p depolarization factor for Rayleigh scattering

8 solar zenith angle, scattering angle

A electromagnetic wavelength

/1 cosine of the solarienithangle (co.s 8)

p reflectance

P6 ground albedo

a standard deviation

extinction, scattering coefficient

Rayleigh scattering coefficient

transmittance

solar hour angle

latitude

solar azimuth -angle

w single scattering albedo

134
 A.5. CONSTANTS AND CONVERSION FACTORS

Constant SI units Nonmetric units

Stefan-Boltzmann co-nstant: a 5.6697 x 10- 8 Wm- 2 K- 4 0.1714 x 10- 8 Btu

ft- 2 h- 1 o R- 4
First radiation constant: C1 3.7419 x 10- 1 6 Wm 2 1.19 x 10 8 Btu Jim 4
ft- 2 h- 1
Second radiation constant: C2 1.4388 x 10 4 Jim K 2.59 x 10 4 Jim OR
Solar constant: So 1367 Wm- 2 433.3 Btu ft- 2 h- 1
4921 kJ m- 2 h- 1 1.96 cal cm- 2 mi n- 1
Astronomical unit: AU 1.496 x 101 1 m 9.3 x 10 7 mile
Radius of earth 6.378 x 10 6 m 3.96 x 10 3 mile
Radius of sun 6.960 x 10 8 m 4.3 x 10 5 mile

Units J m- 2 W h m- 2 cal cm- 2 Btu ft- 2

1 J m- 2 1 2.778 x 10- 4 2.39 x 10- 5 8.81 x 10- 5

1 W h m- 2 3.60 x 10 3 1 0.0860 0.317
1 cal cm- 2 4.187 x 10 4 11. 63 1 3.69
1 Btu ft- 2 1.136 x 10 4 3.155 0.271 1

Units W m- 2 cal cm- 2 min- 1

Btu ft- 2 h- 1

1 Wm- 2 1 1.433 x 10- J 0.317

1 cal cm- 2 mi n- 1 698 1 221.2
1 Btu ft- 2 h- 1 3.155 4.521 x 10- 3 1

135
 INDEX

A
Angstrom
pyrheliometer, 54f
scale, 52

A
Absolute radiometers, 53
active, 58
passive, 58
Absorption, 11, 18
cloud, 36
coefficient, 11, 18
Actinometer, 59
Aerosols
absorption, 8
extinction, 9, 27, 28
scattering, 8
Air mass, 23f
Albedo, 17
cloud, 34f
measurements, 85
Rayleigh scattering, 29
sea ice, 39
sea surface, 42
snow, 38
soils, 41
solar elevation, 37
vegetated surfaces, 39
Atmospheric constituents, 7

B
Black body radiance, 33
Bouguer - Lambert law, 11, 13

C
Carbon dioxide, 7, 33
Cavity receiver, 56
Characterization of a radiometer, 57
Clouds, 10
solar radiation, 34
albedo, 34
thermal radiation, 36
emittance, 35
Criteria for instrument selection, 131

D
Data gaps, 102f
Data statistics, 105f
frequency distributions, 107
threshold analyses, 107
Declination, 23
Detectors
quantum, 51
thermal, 49
Diffuse radiation, 29
measurements, 71

137
 INDEX

E
Emittance, 11, 33, 43
measurements, 86
Equation of time, 24
Equatorial mount. 61
Extinction coefficient, 11, 15

F
Filter factor, 79

G
Gaseous absorption, 28
Global radiation, 29
measurements, 62

I
Index of refraction, 12
International Pyrheliometric Scale, 52

H
Methane-, 8

N
Net radiation, 5
measurements, 74

o
Optical thickness, 12, 28
Oxygen, 8
Ozone, 8, 26
p
Phase function, 13, 15
Planck function, 18
Polarization, 16
aerosols, 18, 32
degree of, 16f, 31
measurements, 87
surface reflection, 43
Pyranometer, 62
calibration, 70
cut-off glass filters, 79
installation, 69
maintenance, 69
performance, 66
self-calibrating, 73
shading disk, 72
shadow ring, 71
silicon detectors, 73
Pyrgeometer, 74, 76
Pyrheliometer, 54
Angstrom, 54f
maintenance, 62
operational, 58
self-calibrating, 56
Silver~Disk, 55f
spectral measurements, 78
Pyrradiometer, 74
calibration, 77
maintenance, 77

138
 INDEX

Q
Quality-control. 93ff
broad-band spectral data, 101
consistency checks based on calculations, 97
consistency checks based on observations, 93
consistency checks for inclined-surface irradiances, 98

R
Radiation, 74
budget, 3, 21
diffuse, 29
global, 29
solar, 8
thermal, 8
scales, 51
transfer, 12
Rayleigh-Scattering, 14, 17, 26, 28
Reflectance, 11, 13
mineral surfaces, 43
natural, 37
soils, 41
surfaces, 37, 44
Representativness of data
spatial, 110f
temporal, 109f

S
Scattering, 13
coefficients, 11
efficiency, 14
indicatrix, 13
Schott glass filters, 78
Silver-Disk pyrheliometer, 55£
Single scattering albedo, 12
Size distribution, 9, 14
cloud droplets, 10
aerosols, 8
Smithsonian scale, 52
Solar constant, 4, 23
Solar radiation, 7
clear skies, 26
clouds, 34
daily total input, 25
inclin&d surfaces, 73
measurement, 54, 73, 77
spectral distribution; 24, 28, 31, 119ff
Source function, 13
stratosphere, 8
Sun - earth distance, 23
Sunphotometer, 77
calibration, 83
recommended wavelengths, 79

T
Terminology,S, 127ff
radiative quantities, 11
radiative transfer, 11
Thermal radiation, 4, 7
clear skies, 33
clouds, 36
measurement, 74
Tracker, 61
Transmittance, 11
Troposphere, 7

139
 INDEX

Turbidity, 27
coefficients, 28

If
Ultraviolet radiation, 26
measurem'ent, 8'4

V
View-limiting geometry, 55

W
Water vapour, 7, 33
absorptance, 2,6
World Radiometric Reference, 53f

140