Analysis of Present and Future Changes in Extreme Rainfall Events Linked to Food Security in the Igad Region of Eastern Africa

Abstract

The observed and projected changes in extremes of daily rainfall have an impact on rain-fed agriculture, crop productivity, food security and other socioeconomic sectors in the IGAD region. Understanding the patterns of changes in extreme rainfall events is important for sustainable food security. This study investigates the patterns of observed and projected changes in extreme rainfall events linked to food security in the IGAD region. The extreme rainfall indicators examined include Rainfall Onset Dates (RODs), Rainfall Cessation Dates (RCDs), Length of Rainy Season (LRS), while extreme rainfall events include wet/dry days, wet/dry spells, and the Standardized Precipitation Index (SPI) as a drought or flood indicator. The study also examines the observed and projected implications of these extreme events on present and future food insecurity hotspots. Data from rain gauges, Satellite Rainfall Estimates (SRE), gridded temperature, Coupled Model Intercomparison Project (CMIP6) historical simulations, and projections are analyzed using various statistical approaches (changes in means, 24 continuous, categorical, and volumetric statistical metrics, scatter plots, Cumulative Distribution Function (CDF), and colored code portrait). Additionally, extreme rainfall criteria and thresholds were used for RODs, RCDs, LRS, wet/dry days, and wet/dry spells. Future changes in extremes and vulnerability to food insecurity are appraised for near and far future time frames under different socioeconomic scenarios. The future changes in extremes and vulnerability to food insecurity are appraised for the near future (2021–2050) and the far future (2071–2100) under three different shared socioeconomic pathways (SSP): SSP1-2.6, SSP2-4.5, and SSP5-8.5 scenarios. The changes are assessed on an annual basis, as well as during the March–May (MAM), June–September (JJAS), and October–December (OND) seasons. The quality check, outliers check, and homogeneity test on 105 stations revealed significant outliers, irregularities, and inhomogeneity in the time series (inconsistent time series). Seven different interpolation approaches were used to create 15 spatially interpolated grid datasets for the region, with deKadal (10-day) and monthly intervals. Among these 15 datasets, the ones generated using Modified Shepard interpolation were found to be the most effective for rain gauge spatial interpolation and are suitable for climate studies in the IGAD region. The validation of nine Satellite Rainfall Estimates (SREs) indicated that CMORPH RT, CHIRPS v2.0, CPC-RFEv2, and GPM-IMERG were the top four best-performing SREs. CMORPH RT performed the best among the SREs, followed by CHIRPS v2.0. The quality of CHIRPS v2.0 decreased from 2016 to the present due to a reduction in the number of stations used for merging. The rainy season in central and northern Sudan, Eritrea, Djibouti, Somalia, southeastern Ethiopia, and eastern and northeastern parts of Kenya has been characterized by a late onset, early cessation, short length, prolonged dry days, and dry spells. In Kenya, the early cessation of rainfall in over 30 counties has resulted in a shortened rainy season by 10-20 days during the MAM season. The patterns of wet days and dry spells show a decrease in the number of wet days and prolonged dry spells in the 1980s, followed by an exceptional increase in wet days in the subsequent decades (2011-2020) during the MAM, JJAS, and OND seasons. The probability of exceeding 7, 14, 21, and 28 days (1, 2, 3, 4 spells) consecutive wet days in northeastern Kenya, Somalia, southeastern Ethiopia, Eritrea, and Djibouti was less than 5% and with an 80% probability of exceeding over Uganda, South Sudan, and central and western Ethiopia. In addition, floods in 1997, 2018, 2019, and 2020 (droughts in 1983, 1984, 1985, and 2021) were triggered by early/late onset and cessation, an increase/decrease in wet/dry days, wet/dry spells. The findings of current food security trends, hunger, and food insecurity indicators, as well as the frequency and duration of food insecurity hotspots at national and sub-national levels, reveal that political instability, civil wars, and governance issues are the primary non-climatic factors causing food insecurity in Sudan, South Sudan, and Somalia. Crop failures due to semi-arid and semi-desert vi climates are potential drivers of food insecurity across northern, eastern, and southern Kenya and southeastern Ethiopia. By 2021, all countries in the IGAD region have not utilized opportunities and available resources to produce enough food locally for the rural population (72.3%) and urban population (27.7%). The results of yields and quantities of cereals being produced in Sudan, Ethiopia, Uganda, Kenya, and South Sudan, compared to available arable land, agricultural land, and permanent crops, meadows, and pasturelands, show that the countries have not utilized the available resources to their full capacity to produce cereals and increase agriculturally based Gross Domestic Product (GDP). The unbalanced policies of commodities being produced, imports, and exports at all levels in the region are another driver of high food prices, hunger, and food insecurity. Although there has been good rainfall in Greater Upper Nile (Unity, Upper Nile, and Jonglei states) in South Sudan (2011–2019), South Kordofan, and Blue Nile in Sudan (2012–2019), the frequency and duration of food insecurity hotspots have been very high due to violence and political instability. The validation of 23 CMIP6 GCMs shows that most models over Arid and Semi-Arid Lands (ASALs) in Kenya, Somalia, Ethiopia, and Sudan scored poorly, with high bias and overestimated rainfall during JJAS compared to MAM and OND. A high percentage of bias exceeding 80% was found over ASALs, with a correlation coefficient ranging between 0.6 and 0.7 across Ethiopia's highlands, and a lowest Root Mean Squared Error (RMSE) of 5–40 over most of the region. Out of the 23 models, the best 10 performing models over the IGAD region are INM-CM5-0, HadGEM3-GC31-MM, CMCC-CM2-HR4, IPSL-CM6A-LR, KACE-1-0-G, EC-Earth3, NorESM2-MM, GFDL-ESM4, TaiESM1, and KIOST-ESM. The projected extreme patterns using EnsMean of the best 10 performing models show a decrease of 10–20% in the number of wet days during the MAM season across Sudan, South Sudan, and central and northern Ethiopia, and an increase of 30–50% in JJAS across central and northern Sudan. However, during the OND season, increases are projected over Uganda, Ethiopia, and Kenya under SSP1-2.6, SSP2-4.5, and SSP5-8.5 scenarios. Arid and semi-arid land (ASALs) in the northern parts of Sudan, Eritrea, and Djibouti, southeastern Ethiopia, most parts of Somalia, and northern, eastern, and northeastern Kenya are projected to experience high levels of exposure to extreme events, Food Insecurity Hotspots (FIH) frequency and duration under SSP1-2.6 and SSP5-8.8 scenarios. Most parts of Uganda, South Sudan, the highlands of western Ethiopia, and western Kenya are projected to have very low exposure, vulnerability to food insecurity, and low FIH frequency and duration under SSP1-2.6 and SSP5-8.5 scenarios during MAM and JJAS for the near (2021–2050) and far future (2071–2100). These findings emphasize the importance of selecting the best climate models for mapping present and future vulnerability to extreme rainfall events and food insecurity hotspots. Additionally, the findings contribute to the advancement of scientific knowledge in rain-fed agricultural planning, food security decision-making, and the development of adaptation and mitigation strategies in the region. The study recommends good governance and policies to better utilize available natural resources such as arable and fertile agricultural land, freshwater resources, and the projected increase in rainfall for sustainable food security in the IGAD region.