Complex Network Theoretic Analysis of Droughts in Future Climate Scenarios Across India.

Abstract:

Droughts are stochastic processes which can propagate spatio-temporally as well as through the hydrological cycle. This can lead to synchronized drought events across different regions which can be captured by analyzing the correlations between time series describing the extreme event. India is particularly prone to frequent drought. There has been significant increase in drought severity and frequency across different Indian regions, with the situation expected to worsen in future climatic scenarios. Hence, it is important to understand drought properties in future climatic projections, especially the drought hotspots and spatio-temporal propagation of drought events from one region to another. The construction of complex networks from spatio-temporal datasets can provide useful information about underlying extreme events through measures like centrality, degree, modularity, propagation length etc. For analyzing the spatio-temporal dynamics of droughts in India under climate change, we construct Event synchronization based complex network using Standardized Precipitation Index for different climate scenarios (RCP4.5 and RCP8.5). Complex networks also aids in providing information about the relative importance of certain regions which promotes drought propagation, thus alluding to their varying drought resilience to mild, moderate or extreme conditions. We find that, in future some regions are more prone to propagate mild drought events to other regions, as compared to past, when the drought events tended to be localized. These regions may act as critical link in the earlier reported migration of drought events from the western to eastern parts of India. We perform the analysis for RCP4.5 and 8.5 scenarios and calculate different network measures to understand the topological changes in the drought network in future. We also observe that extreme drought networks in RCP8.5 scenario have high out-closeness centrality in western India and high in-closeness centrality in Central and Eastern India. This study can thus help to identify hotspots and sources/sinks of drought events along with spatial length scales of propagation in future climate scenarios.