13.3 Description of Data
and Methodology
13.3.1 Data Set
The monthly 100 years rainfall series of eight
districts located in BHB covering period of 1901
to 2000 were obtained from the Hydrometeorology Division (Office of the Additional Director General of Meteorology) of Indian
Meteorological Department (IMD), Pune. For the
BHB, mean monthly rainfall values of eight
districts were summed to obtain annual, seasonal
and mean values of rainfall.
13.3.2 Method of Analysis
To study the intra-annual variations (seasonal
analysis) of rainfall over BHB, the year was
divided into four seasons, namely pre-monsoon
(March–May), southwest monsoon (June–October), post-monsoon (November–December) and
winter season (January–February), depending
upon climatic conditions prevailing over the
northeast India, including Gangetic West Bengal
(Rao 1981; Jain et al. 2013). Rainfall characteristics like mean, standard deviation (r), coefficient of variation (CV) and percentage
contribution to annual were computed for annually, monthly and season-wise to find out the
variability during the study period.
The rank-based non-parametric Mann–Kendall (MK) technique (Mann 1945; Kendall 1975)
has been employed in this study to detect the
long-term changing trends of annual, seasonal
and monthly rainfall. The null hypothesis H 0 for
this test is that there is no trend in the series. The
alternative hypothesis is that the trend is either
negative or positive. The MK test is based on the
calculation of Kendall’s tau (s) measure of
association between two samples, which itself
are based on the ranks with the samples. A major
advantage of Mann–Kendall test is that it allows
missing data and can tolerate outliers. Several
researchers from India have used MK test to
identify rainfall trends due to climate change.
The test does not quantify the trend magnitude.
The linear slope (b) using Sen’s slope estimator
was calculated to estimate true slope (magnitude)
of an existing trend (change year
−1 ) (Sen 1968).
Positive values of Z correspond to a positive b,
which indicate an ‘upward trend’, while a negative Z value corresponds to a negative b, and
indicates a ‘downward trend’. To achieve these
analyses the following software were used:
XLSTAT (https://www.xlstat.com/) for homogeneity test, for MK test and Sen’s slope estimator. Further, a linear trend was added as
parametric test to the series for identifying the
trends. Linear trends represented by the slope
(b) of the simple least-square regression line
provided the rate of rise/fall in the variable.
The short-term decadal climate fluctuations
have been studied by applying Cramer’s t test for
the 10-year running means (moving average) to
examine the stability of climate and to establish
climatic variability (WMO 1966; Kale 2012).
The computational procedure is explained by
WMO (1966) and Kripalani et al. (2003). This
method was applied to the monthly rainfall series
to detect decadal changes and to identify multidecade epochs of above- and below-average
rainfall.
13.4 Results
13.4.1 Features of Rainfall
The rainfall characteristics of BHB for the period
of 1901–2000 are presented in Table 13.1.
Considerable variability was observed between
the different seasons with a standard deviation of
288.54 mm, while average annual rainfall was
1417.85 mm. The coefficient of variation
(CV) of mean annual rainfall is 16.12%. Monsoon contributes 83.2% of the annual rainfall in
this basin, whereas rainfall in the winter is least
(2.4%). Looking at the amount of rainfall in
different seasons (Table 13.1), it is evident that
the BHB receives maximum rainfall in southwest
monsoon season. The high amount of rainfall in
the southwest monsoon season in the BHB is due
13 Changing Rainfall Patterns and Their Linkage to Floods …
173
and Methodology
13.3.1 Data Set
The monthly 100 years rainfall series of eight
districts located in BHB covering period of 1901
to 2000 were obtained from the Hydrometeorology Division (Office of the Additional Director General of Meteorology) of Indian
Meteorological Department (IMD), Pune. For the
BHB, mean monthly rainfall values of eight
districts were summed to obtain annual, seasonal
and mean values of rainfall.
13.3.2 Method of Analysis
To study the intra-annual variations (seasonal
analysis) of rainfall over BHB, the year was
divided into four seasons, namely pre-monsoon
(March–May), southwest monsoon (June–October), post-monsoon (November–December) and
winter season (January–February), depending
upon climatic conditions prevailing over the
northeast India, including Gangetic West Bengal
(Rao 1981; Jain et al. 2013). Rainfall characteristics like mean, standard deviation (r), coefficient of variation (CV) and percentage
contribution to annual were computed for annually, monthly and season-wise to find out the
variability during the study period.
The rank-based non-parametric Mann–Kendall (MK) technique (Mann 1945; Kendall 1975)
has been employed in this study to detect the
long-term changing trends of annual, seasonal
and monthly rainfall. The null hypothesis H 0 for
this test is that there is no trend in the series. The
alternative hypothesis is that the trend is either
negative or positive. The MK test is based on the
calculation of Kendall’s tau (s) measure of
association between two samples, which itself
are based on the ranks with the samples. A major
advantage of Mann–Kendall test is that it allows
missing data and can tolerate outliers. Several
researchers from India have used MK test to
identify rainfall trends due to climate change.
The test does not quantify the trend magnitude.
The linear slope (b) using Sen’s slope estimator
was calculated to estimate true slope (magnitude)
of an existing trend (change year
−1 ) (Sen 1968).
Positive values of Z correspond to a positive b,
which indicate an ‘upward trend’, while a negative Z value corresponds to a negative b, and
indicates a ‘downward trend’. To achieve these
analyses the following software were used:
XLSTAT (https://www.xlstat.com/) for homogeneity test, for MK test and Sen’s slope estimator. Further, a linear trend was added as
parametric test to the series for identifying the
trends. Linear trends represented by the slope
(b) of the simple least-square regression line
provided the rate of rise/fall in the variable.
The short-term decadal climate fluctuations
have been studied by applying Cramer’s t test for
the 10-year running means (moving average) to
examine the stability of climate and to establish
climatic variability (WMO 1966; Kale 2012).
The computational procedure is explained by
WMO (1966) and Kripalani et al. (2003). This
method was applied to the monthly rainfall series
to detect decadal changes and to identify multidecade epochs of above- and below-average
rainfall.
13.4 Results
13.4.1 Features of Rainfall
The rainfall characteristics of BHB for the period
of 1901–2000 are presented in Table 13.1.
Considerable variability was observed between
the different seasons with a standard deviation of
288.54 mm, while average annual rainfall was
1417.85 mm. The coefficient of variation
(CV) of mean annual rainfall is 16.12%. Monsoon contributes 83.2% of the annual rainfall in
this basin, whereas rainfall in the winter is least
(2.4%). Looking at the amount of rainfall in
different seasons (Table 13.1), it is evident that
the BHB receives maximum rainfall in southwest
monsoon season. The high amount of rainfall in
the southwest monsoon season in the BHB is due
13 Changing Rainfall Patterns and Their Linkage to Floods …
173