Spatiotemporal Variation in Precipitation Extremes in South Ethiopia Based on PERSIANN-CDR and Gauge Data and Their Impact on Regional Sorghum and Teff Yields

Abstract

Extreme precipitation events affect crop production and threaten food security, particularly in regions that depend on rainfed agriculture. Recurrent droughts and flash floods in Ethiopia have severely affected the farmers' ability to grow crops to meet their dietary needs, necessitating food assistance during climate shocks. The present study aimed to assess the effects of climatic extremes on crop yields in South Ethiopia. Particularly, the objectives were to assess spatiotemporal variation in extreme precipitation indices in South Ethiopia and analyse the relation between the indices and sorghum and teff yields. Long-term (1990–2021) data from 18 gauge stations and gridded PERSIANN-CDR data were used to compute extreme precipitation indices, viz. Standardised Precipitation Index (SPI), Standardised Precipitation-Evapotranspiration Index (SPEI), annual precipitation (PRCPTOT), highest amount of rainfall over one and five days (Rx1d and Rx5d), cumulative dry days (CDD) and cumulative wet days (CWD). For SPI and SPEI, three seasons relevant for sorghum and teff production were considered: Belg season (MAM), JASO and JJASON. The relationship between precipitation indices and crop yields (2003–2021) was assessed through correlation and relative importance analysis. Drought events were identified during 1992, 1999, 2000, 2009, 2011–2012, and 2015 based on MAM; and 1990, 1993, 2002–2004, 2009–2010, and 2015–2016 based on JASO and JJASON months. The number of drought years matched for majority of the cases between PERSIANN-CDR and gauge data. The PERSIANN-CDR, however, underestimated the magnitude of rainfall (mean PRCTOT 1099 vs. 1200 and Rx1d 34 mm vs. 55 mm) compared to the gauge data. Among the precipitation indices, sorghum yield was influenced by SPI-MAM, CDD, Rx5d, and Rx1d; whereas teff yield was influenced by SPI-MAM and SPI-JASO. Crop yields had a negative correlation with SPI-MAM and a positive correlation with CDD, which could indicate a detrimental effect of wet or flooding conditions on crop yields, or the result of farmers' adaptation of re-planting short-season sorghum variety after a dry Belg season. The study recommends further investigation into the relation of Belg season rainfall with crop yields, preferably with crop yield data at a finer spatial resolution representing different agroclimatic zones. The role of adaptation strategies such as flood-based farming and late planting of short-season sorghum varieties may be further explored. The study provides new insights enhancing the understanding of the complex relationship between precipitation extremes and crop yields, which may help prioritise areas for further monitoring and targeted studies for climate-smart agriculture strategies.