You are on page 1of 10

1) The annual precipitation at station ‘A’ and the average precipitation at 15

surrounding stations are given in the Table 3.19 given below:


a) Determine the consistency of the record at station ‘A’.
b) In which year, there is a regime change indicated?

Year Annual Precipitation Year Annual Precipitation


15 Station Average Station A 15 Station Average Station A
(cm) (cm) (cm) (cm)
1971 71.5 50.5 1981 27.5 36
1972 57 90 1982 60.5 42
1973 27.5 16 1983 55 18
1974 25 21.5 1984 38.5 30
1975 60 50.5 1985 38.5 54
1976 22 62.5 1986 47.5 48
1977 55 69.5 1987 49.5 12
1978 57 36 1988 24 36
1979 36.5 42 1989 44 42
1980 19.5 42 1990 60.5 36

SOLUTION:

Annual Precipitation Annual Precipitation


15 Station
Year Cumulative average rainfall Station Cumulative rainfall value
Average
value of 15 Stations (cm) A (cm) of Station A (cm)
(cm)
1990 60.5 60.5 36 36
1989 44 104.5 42 78
1988 24 128.5 36 114
1987 49.5 178 12 126
1986 47.5 225.5 48 174
1985 38.5 264 54 228
1984 38.5 302.5 30 258
1983 55 357.5 18 276
1982 60.5 418 42 318
1981 27.5 445.5 36 354
1980 19.5 465 42 396
1979 36.5 501.5 42 438
1978 57 558.5 36 474
1977 55 613.5 69.5 543.5
1976 22 635.5 62.5 606
1975 60 695.5 50.5 656.5
1974 25 720.5 21.5 678
1973 27.5 748 16 694
1972 57 805 90 784
1971 71.5 876.5 50.5 834.5

DOUBLE MASS CURVE


900

800 f(x) = 0.99 x − 46.41


Cumulative rainfall value of Station A (cm)

700

600

500

400

300

200

100

0
0 100 200 300 400 500 600 700 800 900 1000
Cumulative average rainfall value of 15 Stations (cm)

Basing on the double mass curve, the points forms a trend line with the equation
y=0.988x-46.41. Yet, it isn’t sufficient to assess the consistency of the curve. For further
proof, we use another method in assessing homogeneity of data by Pettitt Test.

WHAT IS PETTITT TEST?

The Pettitt test is a non-parametric test that has been used in a number of
hydroclimatological studies to detect abrupt changes in the mean of the distribution of
the variable of interest.

PERFORMING PETTITT TEST (via XLSTAT software add in in Microsoft Excel


2016)

For Station A

XLSTAT 2021.2.1.1123 - Homogeneity tests - Start time: 06/05/2021 at 17:51:28 / End


time: 06/05/2021 at 17:51:29 / Microsoft Excel 16.013929
Time series: Workbook = ACT 1 HOMOGENEITY TEST.xlsx / Sheet = Sheet1 / Range
= Sheet1!$C$2:$C$22 / 20 rows and 1 column
Date data: Workbook = ACT 1 HOMOGENEITY TEST.xlsx / Sheet = Sheet1 / Range =
Sheet1!$A$2:$A$22 / 20 rows and 1 column
Significance level (%): 5
Maximum time (s): 180
Number of simulations: 10000
Seed (random numbers): 1920095562

Summary statistics:

Variable Observations Obs. w/ Obs. Min. Max. Mean Std.


missin w/out dev.
g data missin
g data
Station A 20 0 20 12.00 90.00 41.72 18.728
Annual 5
Precipitatio
n (cm)

Pettitt's test (Cumulative rainfall value of Station A (cm)):

K 41.000
t 1977
p-value (Two- 0.757
tailed)
alpha 0.05

The p-value has been computed using 10000 Monte Carlo simulations.
Time elapsed: 0s.
99% confidence interval on the p-value:
] 0.746, 0.768 [

Test interpretation:
H0: Data are homogeneous
Ha: There is a date at which there is a change in the data

As the computed p-value is greater than the significance level alpha=0.05, one cannot
reject the null hypothesis H0.

Year Station A Annual


Precipitation
(cm)
197 50.500
1
197 90.000
2
197 16.000
3
197 21.500
4
197 50.500
5
197 62.500
6
197 69.500
7
197 36.000
8
197 42.000
9
198 42.000
0
198 36.000
1
198 42.000
2
198 18.000
3
198 30.000
4
198 54.000
5
198 48.000
6
198 12.000
7
198 36.000
8
198 42.000
9
199 36.000
0
Station A Annual Precipitation (cm)
100

90

80
Station A Annual Precipitation (cm)

70

60

50

40

30

20

10

0
1971 1976 1981 1986

Year

Station A Annual Precipitation (cm)


mu = 41.725

Summary:
Pettitt
Station A Annual
0.757
Precipitation (cm)
p-values(Pettitt)
12

10

4
p-values

Station A Annual Precipitation (cm)

Variable

For 15 Surrounding Stations

XLSTAT 2021.2.1.1123 - Homogeneity tests - Start time: 06/05/2021 at 17:50:53 / End


time: 06/05/2021 at 17:50:54 / Microsoft Excel 16.013929

Time series: Workbook = ACT 1 HOMOGENEITY TEST.xlsx / Sheet = Sheet1 / Range


= Sheet1!$B$2:$B$22 / 20 rows and 1 column

Date data: Workbook = ACT 1 HOMOGENEITY TEST.xlsx / Sheet = Sheet1 / Range =


Sheet1!$A$2:$A$22 / 20 rows and 1 column

Significance level (%): 5

Maximum time (s): 180

Number of simulations: 10000

Seed (random numbers): 316654567


Summary statistics:

Variable Observations Obs. w/ Obs. Min. Max. Mean Std.


missin w/out dev.
g data missin
g data
15 Station 20 0 20 19.5 71.50 43.82 15.703
Average 0 5
Annual
Precipitatio
n (cm)

Pettitt's test (15 Station Average Annual Precipitation (cm)):

K 29.000
t 1972
p-value (Two-tailed) 0.462
alpha 0.05

The p-value has been computed using 10000 Monte Carlo simulations.
Time elapsed: 0s.
99% confidence interval on the p-value:
] 0.449, 0.475 [

Test interpretation:
H0: Data are homogeneous
Ha: There is a date at which there is a change in the data

As the computed p-value is greater than the significance level alpha=0.05, one cannot
reject the null hypothesis H0.

Year 15 Station Average Annual


Precipitation (cm)
197 71.500
1
197 57.000
2
197 27.500
3
197 25.000
4
197 60.000
5
197 22.000
6
197 55.000
7
197 57.000
8
197 36.500
9
198 19.500
0
198 27.500
1
198 60.500
2
198 55.000
3
198 38.500
4
198 38.500
5
198 47.500
6
198 49.500
7
198 24.000
8
198 44.000
9
199 60.500
0
15 Station Average Annual Precipitation (cm)
80

70
15 Station Average Annual Precipitation (cm)

60

50

40

30

20

10

0
1971 1976 1981 1986

Year

15 Station Average Annual Precipitation (cm)


mu = 43.825

Summary:
Pettitt
15 Station Average Annual
0.462
Precipitation (cm)
p-values(Pettitt)
12

10

4
p-values

15 Station Average Annual Precipitation (cm)

Variable

ANSWERS:
a.) Basing on the double mass curve analysis, supported by the results of the Pettitt
Tests shown above, the given data are consistent and need not to be corrected. Note:
The Pettitt test is a non-parametric test that has been used in a number of
hydroclimatological studies to detect abrupt changes in the mean of the distribution of
the variable of interest.

b.) If we carefully look at the double mass curve, the points we’re not fully aligned,
yet by linear regression, these points have relationship with each other having the
equation y=-0.988x – 40 and a slope of 0.988.
Again, supported with the statistics presented by the Pettitt Tests, we can concur
that there is no regime change that happened. With two sets of tests conducted, for the
Station A and for the 15 surrounding stations, both resulted in accepting the null
hypothesis stating that the data are homogeneous. Meaning there is no date at which
there is a change in the data.

You might also like