Date of Award
Master of Science (MS)
Atmospheric and Earth System Science
Walter L. Ellenburg
Thomas L. Sever
Timothy John Mayer
Earth sciences--Remote sensing., Drought forecasting--Pakistan., Drought management--Pakistan.
Evaluation of drought, leveraging multiple variables focused on the onset, persistence, and intensity provides the most complete accounting of all of the different aspects of the hydrological cycle. The overall objective of this study is to evaluate the ability of a monthly Customized Composite Drought Index (CCDI) to provide a localized assessment of agricultural droughts in Pakistan. Agricultural drought is defined as when prolonged dry conditions impact crop growth and sustenance. Pakistan’s economy and communities rely heavily on the agricultural sector, and many areas are at high risk of crop failure due to the dependence on precipitation. The CCDI was developed using 10 geophysical variables which are customized for each district, using a Random Forest approach that selects statistically relevant input variables. A Principal Component Analysis is used to calculate weights of each input variable and combine the data. District level wheat production data, for the years 2005-2017, is used to derive a dynamic approach to determine a single index value for each growing season and to evaluate the performance of the CCDI. In rainfed districts, the CCDI has the ability to identify agricultural droughts with a true negative rate of 0.65 while irrigated districts had a true negative rate of 0.56. The results show the performance of the CCDI is more reliable in rainfed districts, when compared to irrigated districts. The values of the CCDI capture known droughts, as well as neutral years, well throughout most districts. This demonstrates the potential of using the CCDI methodology to enhance preexisting drought monitoring and forecasting systems in the region.
Schwartz, Caily, "Applications of earth observations for the development of a Customized Composite Drought Index in Pakistan" (2021). Theses. 356.