RAMON FRANCISCO VERDUGA VERA

Guidelines Homework1, Hydrology,

1PAO2020 Hydrological Statistics I (12pts)

Objective: Practice the methods of hydrological statisctics (normal distribution), using mean annual discharge data
of various rivers in Latin America. Data source: Global Runoff Data Center (GRDC), Koblenz, Germany.

Tasks:

1. Open the file assigned to every student (table in sidweb).Localize the basin on a map (Google Earth etc.) (1pt)

In the file we got this data of the Granobles river location:

We found the river putting the coordinates of latitude and longitude on Google earth and got this:

The river we are working on is in Ecuador, it is a little one so it cannot be saw on Google maps.
RAMON FRANCISCO VERDUGA VERA

2. Select the mean annual data corresponding to at least 10 months of a year. Elaborate a basic statistic
(mean, standard deviation, skew) (1pt), indicate outliers and eliminate them (analysis of outliers). (1pt)

There were some years with not 10 months at least counted so we only take these years shown in the
table. The we made the basic statistic with those.

Year MQ months 1978 5,209 12

(m3/s) 1979 3,240 12
1963 2,805 12 1980 3,806 12
1964 3,269 12 1981 3,793 12
1965 4,865 12 6,224 12
1982
1966 5,501 12 1983 3,970 12
1967 4,560 12 1984 4,796 12
1968 4,243 12 1985 4,243 12
1969 5,579 12 1986 6,225 10
1970 6,465 12 1987 5,039 10
1971 6,535 12 4,918 12
1988
1972 4,083 12 1989 4,955 12
1973 3,572 12 1990 3,762 12
1974 5,377 12 1991 3,657 11
1975 10,108 12 1992 2,434 12
1976 7,592 12 1993 3,174 12
1977 4,465 12 5,140 12
1994
RAMON FRANCISCO VERDUGA VERA

1995 2,319 12 2002 2,776 12

1996 4,178 12 2003 4,418 10
2000 16,316 12 2004 4,293 12
2001 2,590 10 2005 3,722 11

With these data we calculate the basic statistics with these results:

Sample Mean S. Deviation Cs

40 4,9 2,4 3,18109563

But we need to confirm our basic stadistics so it was needed to make an extra analysis for the outliers
values.

To take out the outliers values of our data we use this formula for calculate the high and low limits:

Using the following table:

RAMON FRANCISCO VERDUGA VERA

Having the mean, the standard deviation and the Kn we calculated our limits and these were our results:

Sample Mean S. Deviation K32 H. Outlier L. Outlier Cs

40 4,9 2,4 2,682 11,3 -1,5 3,18109563

Now we can take out the outliers values

MQ Comply 3,806 YES 5,039 YES

2,319 YES 3,970 YES 5,140 YES
2,434 YES 4,083 YES 5,209 YES
2,590 YES 4,178 YES 5,377 YES
2,776 YES 4,243 YES 5,501 YES
2,805 YES 4,243 YES 5,579 YES
3,174 YES 4,293 YES 6,224 YES
3,240 YES 4,418 YES 6,225 YES
3,269 YES 4,465 YES 6,465 YES
3,572 YES 4,560 YES 6,535 YES
3,657 YES 4,796 YES 7,592 YES
3,722 YES 4,865 YES 10,108 YES
3,762 YES 4,918 YES 16,316 NO
3,793 YES 4,955 YES

Now we have only a sample of 39 and our Kn will change.

RAMON FRANCISCO VERDUGA VERA

Now our data will be these:

Sample Mean S. Deviation Max Min Range Amplitude Cs Cv

39 4,6 1,5 10,1 2,3 7,8 1,1 1,38733632 0,33243456

3. Draw (dibuja) a histogram of the data. Apply a rule (regla) that defines the number of bins
(categorías). (1pts)

To draw the histogram of the data we have to use the Sturge Rule for determine the number of bins:

k=1+3.3log39
k=7
Bins will be 7
After it we made a table of frequency with the respective bins and this is what we got:

Absolute Relative Cum Abs Cum Rel

Bins Freq. Freq Freq Freq
[2,319-3,432) 8 0,20512821 8 0,20512821
[3,432-4,544) 14 0,35897436 22 0,56410256
[4,544-5,657) 11 0,28205128 33 0,84615385
[5,567-6,770) 4 0,1025641 37 0,94871795
[6,770-7,883) 1 0,02564103 38 0,97435897
[7,883-8,995) 0 0 38 0,97435897
[8,995-10,108) 1 0,02564103 39 1
TOTAL 39 1

16
14
12
10
8
6 Histogram
4 Frequency Polygon
2
0
RAMON FRANCISCO VERDUGA VERA

4. Probability analysis: assign teoretical probabilities, compute the theoretical (Gauss) and real
standard normal variables, and cumulative real probabilities. Draw (dibujar) the probabilities and
discharges in a probabilistic paper of the Gauss distribution (similar to Fig.3 in the Spanish book 2b1).
In the description of the axis (eje)of probability indicate exactly which probability is there
(probability to equal or exceed the value? to not reach (no alcanzar) the value?(3pt)

We use for the Hazen Teoretical Cumulative Probability the following formula:

With n=39

We use for Teoretical Discharges Values the following formula:

𝑥 = 𝑥̅ + 𝑧𝑆𝑥

And for Real Z Values this:

So we got this table:

Q (m3/s) Probab. Hazen z teor Q teor z real Probability. Real

2,32 0,01 -2,23 1,25 -1,52 0,06
2,43 0,04 -1,77 1,95 -1,44 0,07
2,59 0,06 -1,52 2,32 -1,34 0,09
2,78 0,09 -1,34 2,59 -1,22 0,11
2,81 0,12 -1,20 2,80 -1,20 0,12
3,17 0,14 -1,08 2,99 -0,95 0,17
3,24 0,17 -0,97 3,15 -0,91 0,18
3,27 0,19 -0,87 3,30 -0,89 0,19
3,57 0,22 -0,78 3,43 -0,69 0,25
3,66 0,24 -0,69 3,56 -0,63 0,26
3,72 0,27 -0,62 3,68 -0,59 0,28
3,76 0,29 -0,54 3,79 -0,56 0,29
3,79 0,32 -0,47 3,90 -0,54 0,30
3,81 0,35 -0,40 4,01 -0,53 0,30
3,97 0,37 -0,33 4,11 -0,42 0,34
RAMON FRANCISCO VERDUGA VERA

4,08 0,40 -0,26 4,21 -0,34 0,37

4,18 0,42 -0,19 4,31 -0,28 0,39
4,24 0,45 -0,13 4,41 -0,24 0,41
4,24 0,47 -0,06 4,50 -0,24 0,41
4,29 0,50 0,00 4,60 -0,20 0,42
4,42 0,53 0,06 4,70 -0,12 0,45
4,47 0,55 0,13 4,79 -0,09 0,46
4,56 0,58 0,19 4,89 -0,03 0,49
4,80 0,60 0,26 4,99 0,13 0,55
4,87 0,63 0,33 5,09 0,18 0,57
4,92 0,65 0,40 5,19 0,21 0,58
4,96 0,68 0,47 5,30 0,24 0,59
5,04 0,71 0,54 5,41 0,29 0,62
5,14 0,73 0,62 5,52 0,36 0,64
5,21 0,76 0,69 5,64 0,41 0,66
5,38 0,78 0,78 5,77 0,52 0,70
5,50 0,81 0,87 5,90 0,60 0,73
5,58 0,83 0,97 6,05 0,65 0,74
6,22 0,86 1,08 6,21 1,08 0,86
6,23 0,88 1,20 6,40 1,08 0,86
6,47 0,91 1,34 6,61 1,24 0,89
6,54 0,94 1,52 6,88 1,29 0,90
7,59 0,96 1,77 7,25 1,99 0,98
10,11 0,99 2,23 7,95 3,67 1,00

"Straight Line of Henry"

Probability Discharge
0,5 4,6
0,01 1,1
0,99 8,1

5. Perform a test of goodness of fit, using the Kolmogorov Smirnov or Chi square method, for the
significance levels of 0.05 and 0.2.Detailed tables of the critical values of the tests are in the internet.
(2pts). Comment on the results. Do you accept or reject the Zero hypothesis and why? (1pt)Does the Gauss
distribution represent well your river at both significance levels? Look at the histogram, at the skew of the
distribution and the result of the test . Would you recommend other distribution?(2pts).

Using Kolmogorov Smirnov we can got this data:

RAMON FRANCISCO VERDUGA VERA

Q (m3/s) Probab. Teor. Hazen Probab. Real Kolmogorov Smirnov Test

2,32 0,01 0,06 0,0514
2,43 0,04 0,07 0,0359
2,59 0,06 0,09 0,0260
2,78 0,09 0,11 0,0222
2,81 0,12 0,12 0,0003
3,17 0,14 0,17 0,0299
3,24 0,17 0,18 0,0156
3,27 0,19 0,19 0,0049
3,57 0,22 0,25 0,0286
3,66 0,24 0,26 0,0212
3,72 0,27 0,28 0,0099
3,76 0,29 0,29 0,0067
3,79 0,32 0,30 0,0252
3,81 0,35 0,30 0,0479
3,97 0,37 0,34 0,0346
4,08 0,40 0,37 0,0323
4,18 0,42 0,39 0,0338
4,24 0,45 0,41 0,0428
4,24 0,47 0,41 0,0684
4,29 0,50 0,42 0,0811
4,42 0,53 0,45 0,0739
4,47 0,55 0,46 0,0871
4,56 0,58 0,49 0,0876
4,80 0,60 0,55 0,0506
4,87 0,63 0,57 0,0581
4,92 0,65 0,58 0,0699
4,96 0,68 0,59 0,0859
5,04 0,71 0,62 0,0900
5,14 0,73 0,64 0,0902
5,21 0,76 0,66 0,0988
5,38 0,78 0,70 0,0843
5,50 0,81 0,73 0,0817
5,58 0,83 0,74 0,0903
6,22 0,86 0,86 0,0015
6,23 0,88 0,86 0,0239
6,47 0,91 0,89 0,0171
6,54 0,94 0,90 0,0344
7,59 0,96 0,98 0,0154
10,11 0,99 1,00 0,0127