31-Oct-19

Kwame Nkrumah University of

Science & Technology, Kumasi, Ghana

GED 475: PRINCIPLES OF HYDROGEOLOGY

(UNIT 4: GROUNDWATER FLOW)

Ing. Dr. E. K. Appiah-Adjei

Department of Geological Engineering
Faculty of Civil and Geo-Engineering
College of Engineering
KNUST, Kumasi, Ghana

92

4. GROUNDWATER FLOW
 The flow of groundwater takes place in porous media
- the pores through which flow takes place are generally very small
(except areas of large openings); hence the flow is mostly laminar
 Flow of groundwater cannot be measured directly, however,
groundwater head can be used to identify groundwater conditions
 Groundwater head (or hydraulic head) is the height to which water
would rise in a piezometer at any location in an aquifer
 The hydraulic head is given by: h = z + P/ρg

93

93

1
 31-Oct-19

4. Groundwater Flow –cont’d

 In its natural state, groundwater is always moving and flows from a
point of higher groundwater head to a lower head

 The factors of importance in groundwater flow include:

1. The boundary conditions
2. The density and viscosity of the water
– density of freshwater varies very slightly with temperature and
its effect may be ignored in groundwater flow
– viscosity is a measure of the shear strength of a liquid; the
lower the viscosity value, the more mobile the liquid
3. The media through which the water moves (i.e. its permeability)
94

94

4. Groundwater Flow –cont’d

 The fundamental law for groundwater flow is Darcy’s law (1856)
– it states that flow rate through a porous media is proportional to
the head loss and inversely proportional to the length of flow path

– for a particular aquifer cross sectional area A (i.e. area at right

angles to flow), the discharge Q is given as:

𝑄 = −𝐾𝐴 ⇒𝑣= = −𝐾 ⇒ 𝑣 = 𝐾𝑖 ……..(a)

where v is called the specific velocity or Darcy velocity

– Eq. (a) follows that rate of groundwater movement is governed by

hydraulic conductivity of the aquifer media and hydraulic gradient

– the Darcy velocity is not the true velocity of fluid flow through the
medium since the cross-sectional area of the porous medium
where water actually moves is smaller than that of the aquifer
95

95

2
 31-Oct-19

4. Groundwater Flow –cont’d

– hence, the average real or pore velocity representing the rate at
which groundwater actually moves is: vp = Q/(ne.A) = v/ne

 Force potential, f – is the driving impetus behind groundwater flow

- given by the product of hydraulic head and acceleration of gravity
- hence, i can be written as df/dl, l being distance along a flow line,
and f the potential head

 The analysis and solution of problems relating to groundwater

movement and well hydraulics began after the Darcy's law

 The basic simplifying assumptions used in expressing groundwater

flow in a mathematical form are that:
– the material is homogeneous and isotropic,
– there is no capillary zone, and
– there is a steady state of flow (i.e. conditions do not change with time)
96

96

4.1 Groundwater flow direction (Three Point Problem)

 With only three groundwater elevations known from wells,
estimates of local groundwater contours and flow directions can be
determined

- first, water level contour map is constructed from static water

levels of measured from the field

- flow lines sketched perpendicular to the contours show direction

of groundwater flow

- contour maps of groundwater levels together with the flow lines

are useful in locating new wells

- areas on groundwater contour maps with wide contour spacing

possess higher K than those with narrow spacing; hence the former
provide a better prospect for production wells than the latter
97

97

3
 31-Oct-19

4.1 Three Point Problem – cont’d

 To determine the flow direction in three wells of known elevations:

1. Join the 3 wells with straight lines and identify the well with
intermediate water level

2. Along the straight line of wells with maximum and minimum water
levels, identify the point of the same water level as in step 1

3. Draw a straight line between the intermediate well and the point in
step 2

4. Draw a line perpendicular to the equipotential line passing through

the well with minimum water level.

5. The direction of the line indicates groundwater flow direction

whilst the slope of the perpendicular line is the hydraulic gradient
98

98

99

99

4
 31-Oct-19

100

100

4.2 Steady Flow in Confined Aquifer

 The energy required to move water through the pores continually
uses up the available head, and so the potential head declines

 From Darcy’s law: and if

 [1] and [2] are fundamental 1-D equations for steady flow in
confined aquifers 101

101

5
 31-Oct-19

4.2 Steady Flow in Confined Aquifer – cont’d

 When suitable boundary conditions are introduced, many problems
in this case can be solved using the equations [1] and [2]
- For example if there are two observation wells in a confined
aquifer where the hydraulic head can be measured, then the
quantity of flow per unit width (q) is:

q   Kb
dh h  h 
, Integrating  q  Kb 1 2 ..................[a]
dx L
where b is aquifer thickness; K is hydraulic conductivity; L is distance
between two wells; h1 and h2 are heads in the well

- Also, at any point (or distance) x between h1 and h2 (i.e., 0 and L),
the head (h) is:

q
h  h1  x.......... ...[ b ]
Kb
102

102

• Sample Question 9
A confined aquifer has a thickness of 33 km and 7 km wide. Two
observation wells are located 1.2 km apart in the direction of flow.
The head in well 1 is 97.5 m and well 2 is 89.0 m.
a) What is the total daily flow through the aquifer? (Hydraulic
conductivity of the aquifer is 1.2 m/day)
b) What is the elevation of the potentiometric surface at a point
located 0.3 km from well 1 and 0.9 km from well 2?
c) Estimate the time taken for a water molecule to move from one
well to the other assuming effective porosity 0.3.

• Try Question
If the distance and observed potentiometric surface between two
adjacent wells are 1000 m and 3 m, respectively. Find an estimate
of the time it takes for a molecule of water to move from one well
to the other. Assume steady 1-directional flow in a homogenous
silty sand confined aquifer with hydraulic conductivity of 3.5 m/day
and effective porosity of 0.35.
103

103

6
 31-Oct-19

4.3 Steady Flow in Unconfined Aquifer

 In unconfined aquifers, the water table is the upper boundary of the aquifer and
the gradient of water table is not constant; hence flow is complicated

 Flow in such aquifers is solved using Dupuit’s (1863) assumptions:

i. Hydraulic gradient is equal to slope of the water table
ii. For small water table gradients, streamlines are horizontal and
equipotential lines are vertical (i.e. df/dx= dh/dx)
dh
- Hence from Darcy’s law: q   Kh .......... .......... .......... .......... ...[ 3]
dx
- Differentiating [3] with respect to x gives:

- By continuity rule: ; then

- [3] and [4] are fundamental differential equations for solving problems in
unconfined aquifers except where Dupuit’s assumptions are not valid
104

104

4.3 Steady Flow in Unconfined Aquifer

 Integrating [3] with the boundary conditions above (at x = 0, h = h1

and at x = L, h = h2):
h2
L
 h 2  h 12 
0
q dx   K 
h1
hdh  qL   K  2
 2


1  h 2  h 22 
 q  K  1  .......... .......... .......... ......[ c]
2  L 
where q = the flow per unit width; K = hydraulic conductivity; h1 = head at the origin (0 from L); h2 = head at L; L = flow length

 The head (h) can be solved for any point x from h1 (i.e., substitute L
with x and h with h2)
105

105

7
 31-Oct-19

• Sample Question 10
A stratum of clean sand and gravel between two channels has K of
0.1 m/s and is supplied with water from a ditch of 6.5 m deep that
penetrates to the bottom of the stratum. If the water surface in the
second channel is 4 m above the bottom of the stratum and its
distance to the ditch is 150 m (which is also the thickness of the
stratum), estimate the unit flow rate into the gallery.
b) What is the elevation of the water table at a point located 100 m
from the ditch?
• Try Question
An unconfined aquifer in a stratum of clean sand and gravel has a
hydraulic conductivity 0.01 cm/s. From two observation wells 200 m
apart, the observed water table elevations are 11 and 7 m measured
from the bottom of the stratum.
a) If the width of the aquifer is 1km, determine the total flow
through the aquifer
b) If another observation well is situated 150 m from the first well,
estimate the water level in the well and the time taken for water
molecule to reach this well (assume effective porosity is 0.27).
106

106

4.4 Recharge to Unconfined Aquifer

 Consider an aquifer being recharged by rainfall on the ground
surface at a net infiltration rate of w
Here: dq = w . dx

net infiltration w

q
q + dq
dx

 If

the; hence,

107

107

8
 31-Oct-19

4.4 Recharge to Unconfined Aquifer – cont’d

• For an unconfined aquifer being recharged at a rate of w, the flow is

determined by integrating equation [5]:
d 2 (h 2 ) 2w wx 2
i . e ., 2
  • Integrating  h 2   c1 x  c 2
dx K K
• Substituting the boundary conditions h = h1 at x=0 and h=h2 at x=L,
the head (h) at any point x from the origin:

(h12  h22 ) x w
 h  h12   ( L  x) x ................[6]
L K
where w is recharge rate and all symbols have their usual meanings
108

108

4.4 Recharge to Unconfined Aquifer – cont’d

• If there is no infiltration (or recharge), w = 0 and then [6] becomes:
( h12  h22 ) x
h h12  .......... .......... .......... .....[ 7 ]
L

• The discharge per unit width at any section x distance from the
origin is given by: K ( h12  h22 ) L
qx   w(  x )......... .......... .......[ 8 ]
2L 2

- If the water table is subject to infiltration, then there may be a

water divide with crest in water table. At water divide, qx = 0 and x =
distance (d) from origin to water divide; substituting these in [8]:
L K  h12  h22 
d   .......... .......... .......... [9]
2 w  2 L 

- If the distance from origin to water divide is known, then the

elevation of water table at the divide, hmax is:
( h12  h22 ) d w
hmax  h12   ( L  d ) d .......... .......... ......[ 10 ]
L K 109

109

9
 31-Oct-19

Sample Question 11
An unconfined aquifer of clean sand and gravel is located between
two fully penetrating rivers and has K = 0.01 cm/s. The aquifer is
subject to uniform recharge of 1.6 m/yr. The water surface
elevations in A and B are 8.5 and 10 m, respectively, above the
bottom, and the distance between A and B is 460 m. Estimate (a)
the maximum elevation of water table and location of
groundwater divide, (b) the travel times from groundwater divide
to the rivers if effective porosity of aquifer is 0.35, and (c) the daily
discharge for 1 km width from the aquifer into both rivers.

Try Question
Rainfall of 2.50m per annum falls on a strip of land 1km wide lying
between two parallel canals, one of which (canal A) is 3m higher
than the other (canal B). The infiltration rate is 80% of the rainfall
and there is no runoff. The aquifer that contains the canals is 10m
deep below the level of canal B and both canals fully penetrate it.
It is underlain by a horizontal impermeable stratum. Compute the
discharge per m length into both canals, assuming their
boundaries are vertical and the aquifer coefficient of permeability
is 10m/day.

110

110

4.5 Steady Radial Flow to a Well in Confined Aquifer

 When a well is pumped, water is removed from the aquifer
surrounding the well, and the water table or piezometric surface is
lowered
 The drawdown at any given point is the distance by which the water
level is lowered and the cone of depression shows (in 3-D) the
variation of drawdown with distance from the well
 The outer limit of the cone of depression (i.e., zero drawdown)
defines the area of influence of the pumping well

111

111

10
 31-Oct-19

4.5 Steady Radial Flow to a Well in Confined Aquifer

 Assuming that the aquifer is homogenous and isotropic, the well

fully penetrate the aquifer, and Dupuit’s assumptions apply:
- then using Darcy’s law, the relation between the well pumping rate
and the drawdown at any distance r from the well is given as:

(i.e., neglecting the –ve sign) 112

112

4.5 Steady Radial Flow to a Well – cont’d

 Integrating eqn. [11] for the boundary conditions at the well, h = hw
at r = rw , and the limit of cone of depression h = ho at r = ro:

- Eqn. [12] enables the T or K of a confined aquifer to be determined

from drawdowns in two observation wells at different distances
from a well pumped at constant rate:

- In practice, the drawdown(s) rather the head (h) is measured in

pumping test; hence eqn. [13] is modified (Thiem, 1906) to be:

113

113

11
 31-Oct-19

4.6 Steady Radial Flow to a Well in Unconfined Aquifer

 In an unconfined aquifer, if Dupuit’s assumptions hold and the well
is pumped at constant rate until equilibrium is reached, then the
relation between discharge and drawdown at distance (r) is:

 Integrating eqn. [15] for the boundary conditions h = h1 at r = r1 ,

and h = h2 at r = r2, which corresponds to two observation wells at
radial distance from the pumping well and re-arranging gives:

 Transmissivity is then approximated by:

 Drawdown is what is measured in pumping test, hence replacing the

head (h) with drawdown (s) in eqn. [16] gives:

114

114

Sample Question 12
A well fully penetrates a 25 m thick confined aquifer. After a long period of
pumping at a constant rate of 0.05 m3/s, the drawdowns at distances of 50
and 150 m from the well were observed to be 3 and 1.2 m, respectively.
Determine the hydraulic conductivity and transmissivity of the aquifer.
[Ans: K = 16.8 m/day; T = 420 m2/day]

Try Questions
1. A well of diameter 1m penetrate vertically through a confined aquifer,
which is 30 m thick. When the well is pumped at 113 m3/hr, the drawdown
in a well 15 m away is 1.8 m; in another well 50 m away, it is 0.5 m.
Compute (a) the transmissivity of the aquifer, (b) the approximate head and
drawdown in the pumped well for steady-state conditions, and (c)the radius
of influence of the pumping well. Take the initial piezometric level as 40 m
above datum. [Ans: T= 400 m2/day; hw = 34.5 m; sw = 5.5 m; ro = 79 m]

2. A well penetrates an unconfined aquifer with a water level of 25 m prior

to pumping. After a long period of pumping at constant rate of 0.05 m3/s,
the drawdowns at distances of 50 and 150 m from the well were observed
to be 3 and 1.2 m, respectively. Compute the hydraulic conductivity of the
aquifer and the radius of influence of the pumping well.
[Ans: K = 18.3 m/day; ro = 327 m] 115

115

12
 31-Oct-19

4.7 Non-Equilibrium Pumping Equation

 When a well penetrating an extensive confined aquifer is pumped at
constant rate:
i) the influence of the discharge extends outward with time
ii) the rate of decline of head times the storage coefficient summed over
the area of influence equals the discharge
iii) since the water must come from a reduction of storage within the
aquifer, the head will continue to decline as long as the aquifer is
effectively infinite; therefore unsteady or transient flow exists
iv) the rate of head decline, however, decreases continuously as the area
of influence expands
 The applicable differential equation for such flows (i.e., unsteady) is:

 C. V. Theis, by using the boundary conditions h=ho for t=0 and h→ho as
r→∞ for t≥0 , solved eqn. [18] to be:

116

116

 The infinite series in eqn. [20] is called well function W(u), hence:

 Eqn. [21] is known as the Theis or non-equilibrium equation

 Using eqn. [21], the time-drawdown data from an observation well

can be used to determine both the T and S of an aquifer; the data is
analyzed graphically with the aid of a type curve

• In the absence of an observation well, the time-drawdown data

from the pumping well can be used to determine only T

117

117

13
 31-Oct-19

4.8 Theis Recovery Method

 When the pump is stopped at the end of pumping test, the water
levels in both pumping and observation wells begin to rise; this is
known as recovery of groundwater levels

 Drawdowns below the original static water levels (i.e., before

pumping) measured during the recovery period is known as residual
drawdown

 Theis showed that the residual drawdown (s’) is given as:

where t is time when pump was started and t’ is the recovery time

 For large t’ and small r, the well function can be approximated by

the first two terms of the infinite series in eqn. [20] so that eqn. (a)
becomes:

118

118

4.8 Theis Recovery Method –cont’d

 It implies that a plot of the residual drawdown (s’) against the log of
t/t’ forms a straight line with a slope equal to 2.303Q/4πT

- Therefore, the change in residual drawdown per log cycle of t/t’ is:

- Hence, the T of an aquifer can be determined from measurements

of residual drawdown

- Eqn. (c) provide an independent check on the T value obtained

from pumping test data

- In the absence of an observation well, the recovery method of

estimating T is more preferred than all the other methods because
the recovery data is of better quality since there are no errors due
to well losses and fluctuations arising from pumping are absent
119

119

14
 31-Oct-19

Example (Theis Recovery Method)

120

120

4.9 Cooper and Jacob Method

 Cooper and Jacob observed that for small values of r and large
values of t, u is small; therefore the higher powers of the infinite
series become negligible

 Thus, the Theis solution (i.e., eqn.[20].) can be modified as:

- Rewriting and changing to decimal logarithms reduce eqn. [22] to:

- Plot of drawdown against log of time forms a straight line, which if

projected to zero drawdown (i.e., s=0) at t=to implies eqn.[23] is:

121

121

15
 31-Oct-19

- From such a graph, times t1 and t2 will have drawdowns s1 and s2

respectively. If t2>t1 and s2> s1, then from eqn. [23]:

- If t2/t1=10 implies that log(t2/t1)=1, which means the drawdown change

over one log cycle on the graph is given by:

- Using eqns. [24] and [26], the T and S of an aquifer can be determined
from time-drawdown data from an observed well

- In the absence of an observation well, data from the pumping well is

used to determine only T; this method is often used in Sub-Sahara Africa 122

122

Example (Use of Cooper – Jacob Method)

123

123

16