versa. On the other hand, trends obtained from MK/mMK are presented in
Figs. 7.3d, 7.5d, 7.5h, 7.6d and 7.6h.
Annual Rainfall
The annual rainfall trends of the 36 meteorological subdivisions are presented in
Table 7.2 and Fig. 7.3. Although the obtained results exhibited few inconsistencies
between the methods, the results of all the methods showed that annual rainfall in
most of the subdivisions (61.11%) is experiencing a decreasing trend (Table 7.2).
About half (27.16%) is statistically significant at 95% confidence interval (Fig. 7.4).
Spatial distribution of the ITA slope shows that the declining trends are circulated
over all India except for some pockets of coastal Karnataka, Arunachal Pradesh,
Gangetic West Bengal, and Konkan & Goa. However, Andaman & Nicobar Island,
Naga Mani Mizo Tripura, and Gangetic West Bengal show an increasing trend in
monotonic fashion during the study period. Also, the subdivisions in northeastern
India (Arunachal Pradesh & Naga, Mani, Mizo, Tripura) show strong slopes (more
than 6 mm/year), whereas the subdivisions in northwestern and southeastern region
show a weak slope (less than 1 mm/year) (Fig. 7.3a). It is also observed from the
results of MK/mMK that more than half (i.e., 52.78%) of the subdivisions had
decreasing trends during the study period; of this, only 25% is statistically significant
at a different significance level. Thus, the few significant decreasing trends ignored
by the MK/mMK test can be identified by ITA, justifying the reliability of this
method to detect unseen trends in the data series. However, the highest increasing
(3.76 mm/year) and decreasing (4.00 mm/year) rates obtained from Sen’s slope are
found in Coastal Karnataka and Naga Mani Mizo Tripura, respectively (Table 7.2).
The percent bias (P BIAS ) showed similar results with ITA with little discrepancy in
spatial distribution, whereas Saurashtra & Kutch show maximum positive
(+11.65%) change and Naga Mani Mizo Tripura show maximum negative
(À13.25%) change in the second half (1959–2015) of the study period. This increase
in extreme rainfall events has increased the potential occurrence of floods in this
region. However, the trends in annual rainfall obtained by using the three methods
are highly correlated; the slope of ITA and Sen’s slope (r ¼ 0.728) is quite highly
correlated compared to the slope of ITA and Z of MK/mMK (r ¼ 0.579) (Fig. 7.4).
Pre-Monsoon Rainfall
The ITA slope, P BIAS , MK/mMK, and Sen’s slope of rainfall in the pre-monsoon
season are denoted in Table 7.3 and Fig. 7.5a–d. The estimated results of all three
methods demonstrate most of the subdivisions had a decreasing tendency in
pre-monsoon rainfall. Results of the ITA show that the majority of the subdivisions
(55%) showed decreasing trends and about 50% of them are statistically significant
at 95% confidence level whereas only 25% registered as significant (α¼0.05) positive trend. The significant positive trend was mainly observed in the subdivisions
166
T. Mandal et al.
Figs. 7.3d, 7.5d, 7.5h, 7.6d and 7.6h.
Annual Rainfall
The annual rainfall trends of the 36 meteorological subdivisions are presented in
Table 7.2 and Fig. 7.3. Although the obtained results exhibited few inconsistencies
between the methods, the results of all the methods showed that annual rainfall in
most of the subdivisions (61.11%) is experiencing a decreasing trend (Table 7.2).
About half (27.16%) is statistically significant at 95% confidence interval (Fig. 7.4).
Spatial distribution of the ITA slope shows that the declining trends are circulated
over all India except for some pockets of coastal Karnataka, Arunachal Pradesh,
Gangetic West Bengal, and Konkan & Goa. However, Andaman & Nicobar Island,
Naga Mani Mizo Tripura, and Gangetic West Bengal show an increasing trend in
monotonic fashion during the study period. Also, the subdivisions in northeastern
India (Arunachal Pradesh & Naga, Mani, Mizo, Tripura) show strong slopes (more
than 6 mm/year), whereas the subdivisions in northwestern and southeastern region
show a weak slope (less than 1 mm/year) (Fig. 7.3a). It is also observed from the
results of MK/mMK that more than half (i.e., 52.78%) of the subdivisions had
decreasing trends during the study period; of this, only 25% is statistically significant
at a different significance level. Thus, the few significant decreasing trends ignored
by the MK/mMK test can be identified by ITA, justifying the reliability of this
method to detect unseen trends in the data series. However, the highest increasing
(3.76 mm/year) and decreasing (4.00 mm/year) rates obtained from Sen’s slope are
found in Coastal Karnataka and Naga Mani Mizo Tripura, respectively (Table 7.2).
The percent bias (P BIAS ) showed similar results with ITA with little discrepancy in
spatial distribution, whereas Saurashtra & Kutch show maximum positive
(+11.65%) change and Naga Mani Mizo Tripura show maximum negative
(À13.25%) change in the second half (1959–2015) of the study period. This increase
in extreme rainfall events has increased the potential occurrence of floods in this
region. However, the trends in annual rainfall obtained by using the three methods
are highly correlated; the slope of ITA and Sen’s slope (r ¼ 0.728) is quite highly
correlated compared to the slope of ITA and Z of MK/mMK (r ¼ 0.579) (Fig. 7.4).
Pre-Monsoon Rainfall
The ITA slope, P BIAS , MK/mMK, and Sen’s slope of rainfall in the pre-monsoon
season are denoted in Table 7.3 and Fig. 7.5a–d. The estimated results of all three
methods demonstrate most of the subdivisions had a decreasing tendency in
pre-monsoon rainfall. Results of the ITA show that the majority of the subdivisions
(55%) showed decreasing trends and about 50% of them are statistically significant
at 95% confidence level whereas only 25% registered as significant (α¼0.05) positive trend. The significant positive trend was mainly observed in the subdivisions
166
T. Mandal et al.