Statistical prediction of the severity of compound dry-hot events based on El Niño-Southern Oscillation

Abstract Compound dry-hot events have attracted much attention due to their amplified impacts on water resources, crop productions, and ecosystems. Thus accurate prediction of such events is of critical importance for issuing early warning to make informed decisions. This study proposed to employ a meta-Gaussian model for the prediction of the severity of compound dry-hot events in Southern Africa, based on the El Nino-Southern Oscillation (ENSO), which has been shown to affect these events in this region. To represent the severity of compound dry-hot events, a Standardized Compound Event Indicator (SCEI) was used to integrate both dry and hot conditions. The SCEI was predicted based on the conditional distribution constructed from the meta-Gaussian model, in which the antecedent SCEI (i.e., the persistence) and ENSO (i.e., the external forcing factor) were used as predictors. The 1-month and 3-month lead prediction of compound dry-hot events in Southern Africa was assessed and results showed good prediction performances. The compound dry-hot event during December 2015 in Southern Africa was used as a case study and the proposed approach performed well in predicting the severity and probability of this event, which was likely due to the strong impact of ENSO in antecedent periods.