学术文献库

Drought is a costly natural hazard, many aspects of which remain poorly understood. It has many contributory factors, driving its outset, duration, and severity, including land surface, anthropogenic activities, and, most importantly, meteorological anomalies. Prediction plays a crucial role in drought preparedness and risk mitigation. However, this is a challenging task at socio-economically critical lead times (1-2 years), because meteorological anomalies operate at a wide range of temporal and spatial scales. Among them, past studies have shown a correlation between the Sea Surface Temperature (SST) anomaly and the amount of precipitation in various locations in Africa. In its Eastern part, the cooling phase of El Nino-Southern Oscillation (ENSO) and SST anomaly in the Indian ocean are correlated with the lack of rainfall. Given the intrinsic shortcomings of correlation coefficients, we investigate the association among SST modes of variability and the monthly fraction of grid points in Ethiopia, which are in drought conditions in terms of causality. Using the empirical extreme quantiles of precipitation distribution as a proxy for drought, We show that the level of SST second mode of variability in the prior year influences the occurrence of drought in Ethiopia. The causal link between these two variables has a negative coefficient that verifies the conclusion of past studies that rainfall deficiency in the Horn of Africa is associated with ENSO's cooling phase.