Compiling non-parametric tests along with CA-ANN model for precipitation trends and variability analysis: A case study of Eastern India

Abstract

In the present research, mean monthly precipitation data have been used to analyze the existing trends of the precipitation for the periods of 1981–2019 using non-parametric methods as the Mann-Kendall test and Sen's Slope (β). The results show a positive increasing trend in annual precipitation in the south whereas the central part of the study area recorded insignificant declining tends of precipitation over the last century. The maximum decrease in annual precipitation is detected in the Raiganj station (−0.286 ​mm/year). Alipore station recorded the maximum increase (1.354 ​mm/year) and the maximum decrease of monsoonal precipitation recorded in Malda station (−1.155 ​mm/year). The average coefficient value of all stations of winter is 72.83% whereas seasonally the coefficient of variation is low as 19.21% in monsoon time. The high coefficient value in winter (72.83%), pre-monsoon (39.70%), and post-monsoon (34.68%) depicting the weather severity in the study area. For forecasting potential rainfall patterns CA-ANN method was utilized and the high kappa value of the forecasted maps illustrates the authenticity of the findings. The annual rainfall will notice a little of 0.26% of rises in rainfall quantity from 2023 to 2027 and the post-monsoon rainfall will experience a major decline of 8.53% by the year 2023. For the policymakers, this study significantly enhances the climatic knowledge regarding the study region and that can help in managing resources and overall development of this area.

Keywords

• Precipitation
• Non-parametric test
• Trend analysis
• Future prediction
• West Bengal