Drought Projection over Punjab Pakistan, through Copula Approaches

Abstract

The primary objective of this research is to provide a comprehensive drought projection in the Punjab Province of Pakistan. The reconnaissance drought index (RDI) obtains drought duration and severity variables at selected sites with strong correlations. Eleven stations have been selected to measure the prolonged and severe drought periods. Out of the six marginal probability distributions considered, Log-normal and Exponential distributions are identified as the most appropriate for this purpose. Three Copula models, namely Clayton, Gumbel, and Galambos copulas, are utilized to find the joint probabilities of drought variables. The selection of the best copula model for joint distribution is of great importance in effectively capturing drought risk, especially in assessing upper-tail dependency. Therefore, graphical and numerical statistical techniques are employed to identify the optimal Gumbel copula. The assessment of drought hazards involve a comparison of bivariate joint return periods within designated temporal frames. As the return periods increase, droughts' duration and severity also escalate. The contour lines in the severity-duration frequency (SDF) graph are utilized to define the combination of drought duration and severity at the selected return periods. Shorter return periods indicate a higher risk of drought duration and severity reoccurring. This information can be highly instrumental in climate change planning, given the frequent occurrence of droughts and water scarcity in the Punjab province, which is primarily an agricultural region. Consequently, this research can aid in developing irrigation agencies and enhance their planning efforts.