ASSESSING NONSTATIONARITY IN PRECIPITATION AND ITS IMPACTS

Abstract

Nonstationarity in climatic variables is one of the primary responses to climate change that is largely observed as shifts in mean and intensified extremes. With the increasing evidence of nonstationarity in precipitation and its associated devastations around the world in the recent decades, identifying the precipitation changes and assessing their impacts became a vital part of the scientific efforts towards limiting its catastrophic impacts. Over the last half century, researchers have made significant achievements in developing tools and methodologies for detection of changes in historical observations, simulation of future changes in climate, assessing their impact at regional to continental scales. Yet, there remains a wide scope of work for improved understanding and reliable management policies. This thesis contributes to the science of climate change impact assessment by addressing a few critical research gaps in the existing literature. The overall aim of the thesis is to study the changing precipitation patterns, identify potential drivers of observed changes, and develop a robust methodology to model the changes for accurate impact assessment. Specifically, the work is defined under three main objectives – (a) To investigate space-time changes in the frequency of rainfall events in different seasons over India; (b) To investigate prominent natural variabilities and large-scale climate teleconnections of rainfall occurrence during ISM; and (c) To develop a comprehensive framework for assessment of urban flood response to historical and future changes in extreme precipitation. The first objective focuses on an important but previously unexplored characteristic of rainfall, i.e., the frequency of rainfall (or rainfall occurrence) over India. A deficit in rainy days can affect crop production and escalate groundwater abstraction leading to groundwater depletion, thus may threaten food and water security in an agrarian economy like India. Existing literature exhibits a large void of information on rainfall occurrence over vast agriculturally active parts of India. Therefore, space-time patterns are examined in the seasonal count of rainfall events and their count in different rainfall classes for all meteorological seasons of India. The results show a widespread decline in monsoon rainy days over major crop-growing regions including north-central, central, and northeast India. These trends are mainly observed in events with rainfall less than the 75th percentile (light to heavy rainfall events) in all the seasons. The occurrence of extreme events (above 95th and 99th percentile) exhibits contrasting trends over different zones with mainly an increase over central and south-central India and decline over centralnorth India. The results suggest a possible linkage of observed trends in rainfall occurrence with lately observed soil moisture deficit and the decline in groundwater reserves of India. The study findings identify potential avenues for further explorations to address major water-related issues of the country.