Detecting trend and abrupt changes in mean and extreme climate variables and assessing the impacts on agriculture

Climatic variables, temperature and precipitation, in particular, play critical roles in water resources management and agriculture management. To assess the impacts of climate change on human and natural environment, as well as to understand the causes of climate change, it is essential to investigate historic and current climate changes and predict potential future climate changes. This thesis is to develop a multiple change points detection method in time series to identify the change patterns of climate in the Continental US. The method will then be used to detect the changes in linear state changes and changes in variance slope with change time for multiple change points without knowing the number of change points before detection. It is found that abrupt changes occurred more frequently with precipitation than with temperature. Hot spots of identified changes in climate show closely correlated spatial-temporal patterns. The identified spatial-seasonal variation and correlation of changes highlights the uncertainty and information loss of averaging climate variables over years and the assumption of linear trends. As global agricultural area has decreased recently, which may continue in the future, while population is increasing particularly in the developing counties, increased food production per unit area, i.e. yield, is required to meet growing food demand. Climate change is happening globally, with a general warmer trend and spatial difference in precipitation change. Particularly, more frequent extreme weather events (e.g., floods and droughts) have become a serious concern for agriculture, as well as human life and natural environment. Understanding how the change has affected crop yield to date will provide insights on its future possible impacts on food availability. The thesis will assess the impact of change in average monthly climate variables during the growing period on irrigated and rainfed crop yield to analyze whether impacts on yields are different during different months, also whether different impact on two types of yields. In general, we find that the changes in mean climate variables during the past half century benefited less or hurt more the irrigated maize than the rainfed maize in Nebraska. Besides mean change in climate, extreme climate change (e.g., drought and flood) may have changed or is changing. This thesis will focus on drought. Drought is a recurrent extreme climate phenomenon. It can last for weeks, months, even years, and the spatial extent of droughts is usually larger than other natural hazards (e.g., floods and hurricanes), resulting in devastating impacts on agriculture, water resources, environment and human lives. Meteorological drought events vary significantly from one region to another due to different climate characteristics. Additionally, internal variability in the Earth’s climate causes the temporal variation of droughts, which has also been linked to climate change. Understanding the temporal and spatial characteristics of droughts can help in evaluating future drought risk and in choosing appropriate drought mitigation strategies. The thesis will investigate the spatial and temporal patterns of multiple drought characteristics (duration, severity and intensity) under different return periods in India and the Continental U.S (CONUS). In India, the temporal and spatial comparisons based on the univariate return period show different change patterns of duration, severity and peak intensity in different areas. Generally, in the areas which plant wheat more than rice, drought has been alleviated in duration and intensity after 1955; while in the areas which plant more rice than wheat, drought have been aggravated in duration, severity and intensity (except for area 8, a coastal area). In U.S., we find two significant patterns: Pattern I shows persistent droughts in Western & Eastern U.S., and the Great Plains, which experienced large variations in the drought characteristics over long time; Pattern II shows transient droughts in the interior of CONUS, which experienced short-term variations in drought characteristics; trends in these drought characteristics at long and short return periods are different at some locations, showing the different trends of extreme and mild droughts