Attribution of the 2015 drought in Marathwada, India from a multivariate perspective

Abstract

The agricultural region of Marathwada in India incurred significant loss of crops and lives during the summer of 2015, which was characterised by persistent hot and dry conditions. We use observations and large ensembles of regional climate model simulations to understand and attribute this joint occurrence of deficient rainfall and high temperature in a novel, multivariate framework. Highly unlikely in a world without anthropogenic climate change (1-in-256 year), the event is found to be frequent (1-in-38 year) in the actual world. Thus, the risk of this event is found to be atleast quintupled due to anthropogenic factors, implying that the 2015 event is more likely due to anthropogenic climate change. Interestingly, the 2015 dry event is not unprecedented (∼1-in-15 year) based on observed records and for either of the model scenarios, suggesting that risk assessments based on rainfall alone may not be enough to reconcile the observed impacts. Further, such compound drought events are projected to become even more frequent under future end-of-the-century warming targets of 1.5 °C and 2 °C above pre-industrial levels, with expected doubling and tripling of the probability of the 2015 hot-dry conditions, respectively. Our findings highlight the role of human-induced warming on increased incidences of compound extreme events, thereby warranting adaptation strategies that aim at alleviating associated risks.

Keywords

• Probabilistic event attribution
• Drought
• Climate change
• Multivariate return period
• Copula