Impacts of climate change and variability on drought characteristics and challenges on sorghum productivity in Babile District, Eastern Ethiopia

Drought is a complex phenomenon, and it originates from a deficiency of rainfall relative to long-term average that extends over a given space and time scale for a specific period, leading to soil water shortage (Mishra & Singh 2010; Van Loon 2015), while others scholars defined drought based on the period of abnormal arid conditions (Eslamian et al. 2017). Based on different scholars, the most common drought would be classified into four types including meteorological (precipitation), agricultural (soil moisture), hydrological, and socioeconomic droughts (Mishra & Singh 2010). Since the 1990s, the frequency and intensity of drought events have been increasing globally under the combined effects of changing climate, environmental conditions, and human activities (Omer et al. 2020). However, sorghum is drought tolerant compared to other cereal crops and can cope with different types of stresses (Ejeta & Knoll 2007). The occurrence of post-flowering drought often causes premature plant death, lodging, reduction in seed size and ultimately grain yield losses, its production is significantly influenced by climate change and variability (Srivastava et al. 2010).

Assessment of the impacts of climate change on drought characteristics at the local level and their future projections by using global circulation models (GCMs) is important for adaption planning (Endris et al. 2013; Ahmed et al. 2018). Thus, several studies have used GCMs for the projection of future climate change impact evaluation under different climate scenarios (Zhou et al. 2016). The impact of climate change on drought characteristics reported based on five GCMs in the Coupled Model Inter-comparison Project (CMIP5) showed a decline in rainfall amount in East Africa in the 21st century that led to an increase in drought intensity and frequency over most of East Africa like Ethiopia (Haile et al. 2020). Even though drought is a complex phenomenon, many drought indices such as Standard Precipitation Evapotranspiration Index (SPEI) and Standard Precipitation Index (SPI) have been used for quantifying and monitoring drought characterization (Vicente-Serrano et al. 2010; Wang et al. 2019). Drought characteristics by SPEI that included the potential evapotranspiration (PET) method are preferable indices suited for studies of the effect of global warming on drought severity and is more preferable to monitor and evaluate water balance conditions and also performed well in measuring droughts characteristics (Vicente-Serrano et al. 2010; Wang et al. 2019).

Our present study evaluates the state of frequency and intensity of drought indices by using SPI and SPEI methods in semi-arid parts of East Hararghe Zone Babile district, where drought frequently affects crop production. Based on the rainfall anomaly index, the severe droughts that occurred in the Babile district during crop production season (JJAS) were 1987, 1991, 2000, 2002, 2005, 2010, 2011, and 2015 among which 2002 and 2010 were the severe drought intensity in recent history (Tolosa et al. 2023). Here, there is a limited study of future climate change impact on drought character interactive impact on the crop productivity in the study area. However, few studies have been conducted in Ethiopia under current and future climate change scenarios over high spatial dimension (Teshome & Zhang 2019). The present study therefore aimed to analyze the impacts of climate change in drought characteristics at local level and its challenges on sorghum crop productivity in the Babile District semi-arid region of East Hararghe Zone, Ethiopia.

In this study, daily rainfall, and maximum and minimum air temperature data recorded at Babile meteorological station from 1980 to 2009 were used as a baseline for future projection. Continuous and completed daily incident solar radiation were obtained from the NASA website (http://power.larc.nasa.gov/). Global climate model (GCMs) data of CMIP5 was downloaded from the Agricultural Model inter-comparison and Improvement Project (AgMIP) website (https://github.com/agmip/Climate-ScenariosGenerator/tree/master/data/).

Historical and future climate simulations of the 20th and 21st centuries produced by the International Protocol of Climate Change (IPCC) Assessment Report 5 (AR5) participating models and analyzed through the Climate Model Inter-comparison Project Phase 5 (CMIP5) initiative were used in the present study as documented by Taylor et al. (2012). The analysis of climate projections is based on moderate Representative Concentration Pathways (RCP 4.5) and high forcing (RCP 8.5) emission scenarios. To make projections for climate change scenarios at the station level, downscaling was performed with historical daily weather data at Babile station of the study area using a simple delta approach technique employed by three global climate models (GCMs) namely; Canadian Earth System Model second generation (CanESM2), Hadley Centre Global Model Earth System (HadGEM2-ES) and Geophysical Fluid Dynamics Laboratory Earth System Model version 2M (GGFDL-ESM2M) under RCP4.5 and RCP8.5 climate change scenarios. The three GCMs Climate Models were selected based on their better performance in resolution and simulation of both historical and long-term climate normal (Chylek et al. 2011; Martin & Bellouin 2011; Dunne et al. 2013).

Scientific descriptions of the GCMs were well documented in various literatures; HadGEM2 (from UK Met Office Hadley Centre has a spatial resolution of 1.2° × 1.8° (Martin & Bellouin 2011); CanESM2 from Canadian Centre for Climate Modeling spatial resolution of 2.8° × 2.8° (Chylek et al. 2011); GFDL-ESM2M from National Oceanic and Atmospheric Administration, USA with spatial resolution of 2.02° × 2.50° (Dunne et al. 2013). Because of having lower spatial resolution of the GCMs result output, GCMs data cannot be used directly for local climate change impact assessment (Sunyer et al. 2012). Thus, GCMs at large spatial scales are translated to scales more representative of local environmental conditions using downscaled techniques based on Ekstrom et al. (2015). Here, future climate projections were created based on baseline climate data from 1980 to 2009 by utilizing the Agricultural Model Inter-comparison and Improvement Project (AgMIP) method of CMIP5 data using a simple ‘delta’ statistical downscaled approach at the station-level using R software version 4.2 for the period of 2010–2039 (near), 2040–2069 (mid) and 2070–2100 end of the 21st century under RCP4.5 and RCP 8.5 climate scenarios. Besides, the ensemble means across GCM of the future downscaled climate data output were used based on Wallach et al. (2016) for future drought characterization.

The monthly average rainfall and temperature in the baseline (1980–2009), near (2010–2039), mid (2040–2069) and end 2070–2100 of the 21st century were computed under RCP4.5 and RCP8.5 climate scenarios. Monthly trends of statistical tests of rainfall and temperature were computed by R software using the Trend package based on Pohlert (2016).

The SPEI with 3-, 4- and 6-month time scales were calculated by fitting a logistic distribution that has been found to fit the monthly water balance data based on Vicente-Serrano et al. (2010), using the R software version 4.2 in the SPEI package tool. The SPEI uses ‘climatic water balance’, the difference between precipitation and reference evapotranspiration (P – ET₀). The use of a particular equation for the estimation of ET₀ is not central to the calculation of the SPEI. Thus, most of the studies have used the Penman–Monteith (PM) equation for the calculation of ET₀ (Allen et al. 1998). However, the PM equation requires extensive data (solar radiation, temperature, wind speed, sunshine hour and relative humidity). However, the meteorological station we utilized in this study does not routinely record all of these variables. To solve such a problem, Droogers & Allen (2002) estimate ET₀ by using the Hargreaves (Hg) method. This method was tested in the east-southern area and results showed satisfactory agreement with ET₀–PM estimates (Droogers & Allen 2002).

In ensemble means across the GCMs model output shows rainfall during October to December (OND) and July are projected to increase, while rainfall in September shows a decreasing trend under the RCP4.5 scenario (Table 1). Similarly, the rainfall amount in OND, MAM, and annual total rainfall as well as monthly rainfall amount in April, July and November showed a statistically significant increasing trend (P < 0.05) under the RCP8.5 climate scenario (Table 2). However, the rainfall trend is not significant during the sorghum growing period from June to September, its amount shows a mixed signal of decreasing and increasing under RCP 4.5 and RCP8.5 climate scenarios, respectively (Tables 1 and 2). The annual maximum temperature showed an increase of 0.31 °C every 10 years and the annual minimum temperature increased by 0.22 °C every 10 years under the RCP4.5 scenario (Table 1), whereas the annual maximum and minimum temperature increases by 0.48 and 0.46 °C every 10 years under the RCP8.5 scenario (Table 2).

The rising temperatures also increase soil evaporation (Abtew & Melesse 2013), and diminish soil moisture availability (Grillakis 2019; Ogunrinde et al. 2022). Thus, the decline in rainfall and an incline in both maximum and minimum temperatures lead mainly to higher moisture stress and reduced crop yield (Singh et al. 2017; Gong et al. 2021). However, sorghum is the most drought resistant crop (Ejeta & Knoll 2007). Further support in the selection of moisture-stressed resistance sorghum varieties, more soil moisture conservation practices and crop management would be promoted by developing climate resilient varieties as suggested by FAO (2010) in the mid and end of the 21st century mainly under the RCP4.5 climate scenario.

The results further revealed drought characteristics such as duration, severity and intensity in SPEI 3-, SPEI 4-, and SPEI 6-month time scales under RCP4.5 and RCP8.5 climate scenarios. According to this pattern of drought indices from 1980 to 2100, Babile station experienced an increasing trend in the frequency and severity of droughts mainly under RCP4.5 climate forcing. The drought events in SPEI-3-month scale were 15, 16, 20 and 22 during observed, near, mid and end of the 21st century under RCP4.5 scenario (Table 3). Similarly, drought events in SPEI-4-month scale result were 13, 13, 14 and 18 during observed, near, mid and end of 21st century, respectively under RCP4.5 scenario (Table 4). Besides, in the SPEI-6-month also show 11, 11, 10 and 14 drought events (Table 5) under RCP4.5 scenario at the Babile District. Similarly, drought events in the SPEI-3-month scale were 14, 25, and 22 during near, mid and end of the 21st century (Table 3) and drought events in SPEI-4-month scale result were 12, 18 and 18 during near, mid and end of 21st century (Table 4), while drought events in the SPEI-6-month scale were 10, 12 and 14 drought events (Table 5) under RCP8.5 scenario at the Babile District.

The average drought durations (D) were 2.8, 3.5, 3.6, and 3.7 months in SPEI 3-month (Table 3), 3.85, 4.54, 4.71, and 4.33 months in SPEI 4-month (Table 4) and 4.64, 5.64, 6.2 and 5 months in SPEI 6-month under RCP4.5 during observed, near, mid and end of 21st century, respectively (Table 5). Similarly, the average drought durations (D) were 3.21, 3.2, and 3.37 months in SPEI 3-month (Table 3), 4.17, 4, and 4.1 months in SPEI 4-month (Table 4) and 5, 6.7, and 4.8 months in SPEI 6-month (Table 5) under RCP8.5 during near, mid, and end of 21st century. The average drought severity (S) was 1.1, 1.53, 1.55, and 1.8 in SPEI 3-month (Table 1) and 1.51, 2.1, 2.38, and 2.29 in SPEI 4-month (Table 4) and 2.15, 2.77, 3.44, and 2.91 in SPEI 6-month under the RCP4.5 scenario during observed, near, mid and end of 21st century (Table 5), whereas, the drought severity (S) was 1.33, 1.37, and 1.79 in SPEI 3-month (Table 3), 1.79, 2.05 and 2.19 in SPEI 4-month (Table 4) and 2.47, 3.19, and 2.69 in SPEI 6-month (Table 5) in near, mid and end of the 21st century under the RCP8.5 scenario.

The overall intensities of SPEI 3-month drought intensity (I) were 0.25, 0.31, 0.36 and 0.36 during the observed, near, mid and end of 21st century (Table 3) under the RCP4.5 scenario. The overall intensities of SPEI 4-month were 0.25, 0.31, 0.36, and 0.38 (Table 4) and SPEI 6-month were 0.29, 0.3, 0.4, and 0.39 (Table 5) during observed, near, mid and end of 21st century under the RCP4.5 scenario. The SPEI 3-month drought intensity (I) was 0.28, 0.249 and 0.401 during the near, mid and end of the 21st century (Table 3) under the RCP8.5 scenario. SPEI 4-month drought intensity was 0.262, 0.343, and 0.374 (Table 4) and SPEI 6-month were 0.308, 0.35, and 0.332 (Table 5) during the near, mid and end of the 21st century under the RCP8.5 scenario. Here, the drought duration, severity and intensity would be increased toward the 21st century due to climate change, which more likely affects sorghum production mainly under RCP4.5 climate scenario at the study area. Similarly, drought frequency, duration and intensity are affected by climate change in East and West Africa (Polong et al. 2019; Ogunrinde et al. 2022). These prolonged dry spells, drought, rainfall variability and temperature increases are the most predominant stresses and have a significant influence on crop yield reduction (Srivastava et al. 2010).

Evaluating drought characteristics in recent and future periods is vital for feasible adaptation strategies. The future rainfall and temperature trends were employed for the input for future drought characteristics analysis. The SPI and SPEI drought indices were used to assess the future impacts of climate change on drought duration severity and intensity under RCP4.5 and RCP8.5 future climate scenarios. The climate data of Babile meteorological station from 1980–2009 was used as a baseline for future climate projection using three global climate models, namely CanESM2, HadGEM2-ES and GGFDL-ESM2M used by the Intergovernmental Panel on Climate Change (IPCC) of the Fifth Assessment Report CMIP5) under two Representative Concentration Pathways (RCP4.5 and RCP8.5) for near, mid and end of the 21st century. The characteristics and trend of rainfall and temperature of the study area were analyzed. Drought characteristic indices of SPEI were evaluated based on temperature and rainfall while the SPI was based on rainfall. Drought characteristics such as duration, severity and intensity, which is the longest drought, the strongest severity of drought and the highest intensity of drought in the study area, have been identified within SPEI 3-, SPEI 4-, and SPEI 6-month scales related to agricultural crop production at the Babile District east Hararghe zone under RCP4.5 and RCP8.5 climate scenarios.

The results revealed that the annual maximum temperature would increase by 0.31 °C every 10 years and the annual minimum temperature increased by 0.22 °C every 10 years under the RCP4.5 scenario, whereas the annual maximum and minimum temperature increases by 0.48 and 0.46 °C in every 10 years under the RCP8.5 scenario. However, the rainfall trend varies with months and season's time scale. Consequently, the drought intensity and severity increase toward 21st century mainly under RCP4.5 in observed and mean across the three GCM climate model's output. The average drought durations (D) were 2.8, 3.5, 3.6 and 3.7 months in SPEI-3-month and 3.85, 4.54, 4.71 and 4.33 months in SPEI-4-month and 4.64, 5.64, 6.2 and 5 months in SPEI-6-month under RCP4.5 during observed, near, mid and end of 21st century respectively. Similarly, the average drought durations (D) were 3.21, 3.2 and 3.37 months in SPEI-3-month 4.17, 4 and 4.1 months in SPEI-4-month and 5, 6.7 and 4.8 months in SPEI-6-month under RCP8.5 during near, mid and end of 21st century. Increase in drought intensity and severity will affect sorghum production in in the future mainly under RCP4.5 scenarios. To tackle the impacts of drought, soil water conservation practices and drought resistance variety and soil water conservation cropping system are highly recommended to adapt crop production under future climate scenarios over the study area.

The authors would like to acknowledge the Ethiopia Meteorological Institute in the acquisition of necessary data and technical support for this work, who provided us climate data from 1980–2009 which was used as a baseline for future projection. We are also grateful to the African Center of Excellence in Climate Smart Agriculture and Biodiversity Conservation of Haramaya University for hosting a PhD study to undertake this research working facility and management support. The Authors also forward great thanks to those friends who stand by our side during the research work, share information as well as extended their technical support during the needed period. Momentous thank also goes to all anonymous reviewers for their valuable comments and remarks during the review process for further improvement of the manuscript.

A.A.T. was responsible for all activities of the research process such as the design, data collection from all sources, data compilation, and entry, data analysis, interpretation of results, editing as well as writing up of the manuscript. The four authors D.K.D., L.W.M., Z.B.E., and F.M.L. were involved in supervising from design to data collection, contributed to framing the manuscript, made valuable inputs, edited, and commented on improving the quality of the manuscript during the write-up of the manuscript. All authors contributed to the manuscript, and read and approved the final submitted version.

This research did not receive any specific grant from funding agencies in the public, commercial, or non-profit sectors.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.