Regional scale analysis of trends in rainfall using nonparametric methods and wavelet transforms over a semi‐arid region in India

Abstract

Space‐time variability of rainfall at local scale is affected by several regional factors such as aerosol concentration, greenhouse gases, land cover changes, etc., along with large scale atmospheric circulations. Predictive ability of regional circulation models can be significantly improved and efficient management of water resources can be assured by identifying dominant variables controlling spatiotemporal variations in rainfall among aforementioned factors. The present study aims to investigate dominant climate system(s) controlling trends in rainfall over Chhattisgarh state (a semi‐arid region) in India over the period of 115 years (1901–2015). Discrete wavelet transform in conjunction with Mann–Kendall test is applied to the rainfall data series at different time scales (monthly, seasonal, annual, pre‐monsoon, monsoon, post‐monsoon and winter) in order to identify the long term trends and dominant periodic components influencing the trend. In the results, negative trends are found to exist in all rainfall time series at majority of districts (except for annual and monsoon rainfall at Bijapur and Sukma district). In addition, analysis of trend in actual evapotranspiration and soil moisture in the region does not exhibits the effect of anthropogenic variables (such as land cover change, irrigation projects, etc.) on the rainfall as significant negative trend are also observed in soil moisture for majority of districts. Overall, 2‐year and 4‐year periodic components have been detected to be dominating the trends in most of the rainfall time series (annual, monsoon, post‐monsoon and winter). On comparing the identified dominating components with the existing climate systems (Atlantic multidecadal oscillations, Indian ocean dipole, Madden‐Julian oscillations, Inter‐Tropical Convergence Zone, etc.), El Niño‐Southern Oscillations has been recognized as predominant climate circulation influencing the rainfall trends over the study region. The study outcomes are expected to improve the regional precipitation forecasts and should be useful in various hydro‐meteorological analyses and decision making at regional scale.