Impacts of climate change on crop growing season characteristics in Northern Ethiopia

Developing countries are the most vulnerable countries to the impacts of present and future climate change and variability. Climate change and variability are the most determinant constraints to agricultural production and food security in areas that rely on rainfed production systems (Gitz et al. 2016).

The impacts of climate change and variability are predominantly noticed through changes in the two most important climate parameters, which are temperature and rainfall. The trends of these parameters are crucial for crop production, as they have changed in Ethiopia for the last several decades. For instance, Ethiopia's annual average temperature has increased by 1 °C since 1960 at an average rate of 0.25 °C per decade (World Bank Group 2021). Moreover, Ethiopia has experienced a very high degree of rainfall variability annually and seasonally (Mekasha et al. 2014; Zeray & Demie 2016; Gebremicael et al. 2017). The impacts of increased temperatures and high variability of rainfall patterns are expected to reduce crop production and water availability for irrigation and other water-consuming sectors, especially in the north, northeast, and eastern lowlands of the country (Aragie 2013). Moreover, projections of climate change indicated a decrease in the length of the growing period through the shifting of onset and cessation dates over different regions in Ethiopia (Kassie 2014; Gebrekiros et al. 2016; Jima et al. 2019). This is very critical to the majority of the population of the country who are dependent on rainfed agriculture for their livelihood, such as in the case of the Tigray region.

Climate change and variability today are serious threats to crop production in Ethiopia in general and in the Tigray region in particular. For example, the climate of the study area (the eastern part of the Tigray region in northern Ethiopia) indicates that it is characterized by frequent droughts. The region has encountered frequent meteorological droughts (e.g., in 1982, 1983, 1984, 1985, 1987, 1991, 1999, 2000, 2002, 2004, and 2009) (Gebrehiwot et al. 2011). Especially in the Tigray region, the majority of the droughts had a drastic impact on agricultural outputs, with total crop failure and massive livestock deaths (Asheber 2010).

Even though rainfall and temperature variability and associated localized droughts in the study area have been of the greatest concern, many studies have emphasized only a single parameter, i.e., rainfall variability and its distribution. Despite the fact that temperature is a crucial variable in crop production, its impact analysis has been ignored by many studies (e.g., Meze-Hausken 2004; Gebrehiwot et al. 2011; Gebrehiwot & van der Veen 2013; Abrha & Simhadri 2015; Hayelom et al. 2017). Yet, some studies have pointed out that temperature increases have a much stronger impact on crop production than rainfall (Schlenker & Lobell 2010; Ochieng et al. 2016). Similarly, some other studies indicated that even with sufficient rainfall, increasing temperatures contribute to reduced yields (Cooper et al. 2009; Luhunga et al. 2017). In addition, the shortening of the growing period with increasing temperatures has been identified as the main yield-reducing factor (Gardi et al. 2022). In an arid and semi-arid climate with a dry environment, such as the study area, a slight increase in temperature will have a significant impact. Hence, in a drier environment, temperature changes, particularly the minimum temperature, have a drastic impact on crop production (Mupangwa et al. 2023).

Moreover, previous studies regarding rainfall analyses were restricted to trends in annual, monthly, or seasonal total values. Rainfall variability based on crop growing season characteristics such as onset and cessation at an interval of days, length of the growing period (LGP), and dry spells (DSs) has not been included in many of the studies. In addition, the analysis of rainfall and rainy season characteristics have not been covered within the context of climate change impacts, with the only exception of Gebrekiros et al. (2016) who determined the LGP of sorghum (Sorghum bicolor) crop under climate change in southern Tigray. This implies that there are no previous studies that consider the impact of climate model projections on growing season characteristics (onset and cessation, LGP, and DSs) over the eastern zone of the Tigray region. Yet, the characterization of these systematic variations is very important for practitioners to improve water resources and agricultural planning. This study was therefore conducted to investigate the impact of climate change on growing season characteristics in the eastern zone of the Tigray region in northern Ethiopia.

In the present study, a series of daily rainfall and temperature (maximum and minimum) data were utilized. The weather datasets were collected from seven stations within and nearby the eastern zone of the Tigray region for the years covering 1980–2009. These were obtained from the Enhancing National Climate Services Initiative (ENACTS), recently implemented at the Ethiopian National Meteorological Agency (NMA). The quality of the climate data collected from the ENACTS is improved by combining careful quality control of data from weather stations with that of satellite estimates. The available dataset at the NMA is the best, most homogeneous, and recommended for climate analysis (Dinku et al. 2014). More information about the ENACTS is available in Dinku et al. (2014) and Dinku et al. (2018). The selected stations, along with their geographical locations and elevations, are shown in Table 1.

Daily temperatures and rainfall data were projected from the Fifth Phase Coupled Model Inter-Comparison Project (CMIP5) global climate modes (GCMs) using a 30-year baseline daily weather dataset (1980–2009). As presented in Table 2, five GCMs were used for climate change projections. These GCMs were selected based on their consistency and resolution performance for East and sub-Saharan Africa (Sillmann & Roeckner 2008; Msongaleli et al. 2015; Rosenzweig et al. 2015). The five selected GCMs were used in many climate change impact studies in East Africa in general and in northern Ethiopia in particular. Moreover, two RCPs (RCP4.5 and RCP8.5) were used for future climate change scenario analysis. RCP4.5 is an intermediate forcing level, and RCP8.5 is a very high emission forcing level. Each RCP defines a specific trajectory and radiative forcing level. The radiative forcing values are 4.5 and 8.5 W/m², respectively (Wayne 2014).

GCMs have course resolution. Hence, the delta statistical downscaling method of GCMs was applied to represent local conditions. The delta statistical method is one of the simplest methods of downscaling. In this method, historical data and future projections are expressed and interpolated from a common reference period. R script in R statistical software was used to prepare delta-based future climate change scenarios in the mid-term (2050:2040–2069) and end-term (2080:2070–2099) periods. The process of the delta downscaling method comprises the following steps: (i) Gathering of baseline data (current climates corresponding to WorldClim). WorldClim contains files including latitude, altitude, elevation, precipitation, and temperatures for each sub-region. (ii) Gathering of the full GCM time series. WorldClim and the full GCM time series are freely available on the internet at www.worldclim.org. (iii) Calculation of 30-year running averages for the present-day simulations (1980–2009) and two future periods in this study. (iv) Calculation of anomalies as the absolute difference between future values in temperatures and proportional differences in total precipitation. (v) Interpolation of these anomalies using centroids of GCM cells as points for interpolation. (vi) Addition of the interpolated gridded data to the current climates from WorldClim, using absolute sum for temperatures, and addition of relative changes for precipitation (Ruiter 2012).

In the present study, crop risks associated with extreme events and crop growing season characteristics, such as DS length, onset, cessation, and LGP, were analyzed using R-Instat Statistical Program (Version 0.6.2), http://r-instat.org/Download (AMI 2018). The definitions of the crop growing season characteristics adopted from Berhe et al. (2023) are presented on p. 5. Berhe et al. (2023) examined long-term trends and variabilities in temperature, rainfall, and crop growth season characteristics based on years from 1980 to 2009. The years are the historical period and served as a reference for several studies on climate change. Hence, these 30-year daily data were used as input to the current study as a baseline period in order to predict the future periods under two RCPs and five selected GCMs.

The mean projections of all GCMs under RCP8.5 are higher and more variable compared to RCP4.5 scenarios. Mid-term projections under RCP8.5 are higher than those predicted under RCP4.5 and varied from +30 to +50%. Similarly, end-term projections under RCP8.5 are about +84 to +120% higher than projected results under RCP4.5 at all stations. On average, end-term period projections are higher by about 75% than those predicted for the mid-term under RCP8.5, while in the case of RCP4.5, the end-century projections are higher by about 30%. Across all scenarios, temperatures will continue to increase in the study area throughout the end of the century. Temperature increases are expected to result in more intense heat waves and higher rates of evapotranspiration. This will have significant implications for human and animal health, agriculture, water resources, and ecosystems. Results on projected temperature increases will help to develop effective adaptation measures to reduce the impacts of climate change and draw up long-term crop and water resource management plans in the study area.

Because of differences in the time periods, GCMs, RCPs, and downscaling methods, it is difficult to make a direct comparison between the results of previous studies and the current ones. Yet, most previous studies on the impact of climate change on future temperatures in northern Ethiopia indicated an overall rise. The same studies, however, reported the absence of a significant and clear trend in the annual rainfall pattern. In line with the findings in this study, some previous studies indicated an increase in annual rainfall (Niguse & Aleme 2015; Gebrekiros et al. 2016; Berhe et al. 2023; Shiferaw et al. 2018; Thomas et al. 2020; Kidanemariam et al. 2021), while other studies indicated a slight decrease in rainfall (Gebrehiwot & van der Veen 2013; Kahsay et al. 2018; Takele et al. 2022). Some other studies, however, indicated that the rainfall did not show any systematic increase or decrease in their study locations (Tesfaye et al. 2014; Gardi et al. 2022). But it is worthwhile to note that most of the studies found that the increasing or decreasing trend of rainfall is not significant. In spite of the fact that the high variability of the rainfall associated with climate change damages crop and livestock interventions, the impact of temperatures is found to be a main limiting factor for crop production as compared to rainfall (Niguse & Aleme 2015; Gebrekiros et al. 2016; Gardi et al. 2022).

Different crop species all over the world have a given set of temperature thresholds for growth and reproduction. The increase in temperatures in the next years in the study area will thus lead to an alteration in growing days and yield. Furthermore, high temperatures cause crops to mature quickly, thereby reducing grain and forage production. One critical period of exposure to temperatures is the pollination stage. High-temperature exposure during the pollination stage can greatly reduce crop yield and increase the risk of crop failure (Walthall et al. 2012). Similarly, rising temperatures create warm environments. This would favor rapid insect development, with more insect generations and a higher population of pests. In addition, it could also create favorable environmental conditions for some secondary insects to become key pests (El Bouhssini et al. 2011). Moreover, rising temperatures and changes in moisture are expected to cause changes in the distribution of crop diseases, the development of epidemic diseases, and the appearance of new pathogens, as well as new or minor pathogens that are likely to become key diseases under climate change (Ahmed et al. 2011). On the other hand, some areas will likely experience increased rainfall over a short period, leading to increased soil erosion and land degradation due to increased rainfall intensities and an increased incidence of floods in flood-prone areas. All of them undermine the production capability of the area. In general, by reducing production capabilities and increasing production risks, the area becomes unsuitable for crop production. Hence, as a coping mechanism, specific crop production will gradually shift to another suitable area (Girmay et al. 2022).

Moreover, several literatures indicate that there are major challenges to be faced concerning climate change and variability in crop production in the Tigray region (Niguse & Aleme 2015; Gebrekiros et al. 2016; Berhe et al. 2018; Araya et al. 2021). Model predictions have shown that climate change will affect grain yield significantly in different locations of the region. For example, Araya et al. (2021) predicted the impact of climate change from three global climate models using the Decision Support System for Agrotechnology Transfer (DSSAT) crop model in the Adigudom area. They found that an increase in the temperature of 2–8 °C is expected to significantly decrease the grain yield of barley (Hordeum vulgare L.) by 6–11% in mid-century (2040–2069) under a higher emission scenario (RCP8.5). Similarly, projections done by Gebrekiros et al. (2016), also using the DSSAT model, have found that a reduction in sorghum (S. bicolor L.) yields by 5–24%. Likewise, Niguse & Aleme (2015) projected the yield of sesame (Sesamum indicum) using the Aqua crop model in the western part of the Tigray region and found that the yield of sesame is expected to increase by about 33.1% using the GFDL-ESM2M model in the mid-century (2050) under RCP4.5, while it will decline by about −5.88% and by −23.31% at the end of the century (2100) under RCP8.5 using the GFDL-ESM2M and HadGEM2-EM climate models, respectively.

Overall, most of the studies in the region projected that crop production under future climate change will be very challenging as the yield of different crops is expected to decline significantly. In the past, farmers in the region have been practicing locally adopted coping mechanisms in response to changes; however, climate variability and change are eroding the coping mechanisms by causing climatic extremes with an increasing rate of recurrence and intensity that local people have never experienced before. Therefore, future climate change adaptation mechanisms should correspond with the magnitude of the problem. This requires an in-depth, multidimensional understanding of the problems and research-based solutions all over the world.

Future changes in ETo are of increasing importance in assessing the potential impacts on hydrology and water resources. Similar to the findings of this study, several previous studies indicated an increase in evapotranspiration in their corresponding study locations (Gebremeskel & Kebede 2018; Kahsay et al. 2018; Shiferaw et al. 2018; Takele et al. 2022). The studies pointed out that the change in evapotranspiration is associated with changes in temperatures and rainfall variability, which will influence the hydrological responses of river basins over northern Ethiopia in the future. For example, Shiferaw et al. (2018) reported that increasing trends in temperature and evaporation are predicted in the future in the Illala watershed. They found that the surface runoff will also decline from 1.74% under RCP4.5 in the near term (2010–2039) to 0.36% under RCP8.5 in the end-term periods (2070–2099). Likewise, Takele et al. (2022) pointed out a significant increasing trend of temperatures and a decreasing trend of precipitation in the upper Blue Nile basin in the mid-century (2040–2069). This leads to an increase in the annual evapotranspiration of about 10.4%, while it causes a reduction in the streamflow of up to 54%, surface runoff of up to 31%, and water yield of up to 31%. Similarly, Kahsay et al. (2018) found an increase in the evapotranspiration of 0.4% under RCP2.6 and 8.1% under RCP4.5 in the sub-catchment of the Tekeze River basin throughout 2020–2079. As a result, they found a decrease in the groundwater recharge of 3.4 and 1.3% and a decrease in the base flow of 1.5 and 0.55% under RCP2.6 and RCP4.5, respectively. Gebremeskel & Kebede (2018) also investigated the hydrological responses in the Werii watershed based on A1B and B1 special reports for emission scenarios (SRES) and found an increase in rainfall as well as temperatures in mid-century (2015–2050). This causes an increase in the evapotranspiration of 15% under A1B and 18% under B1, and a decrease in the runoff of 13 and 14% under the respective scenarios. Many of the studies mentioned above projected that the changes in evapotranspiration, in association with changes in temperatures and rainfall, are expected to bring changes in river flows and runoff in the future at different river basins in northern Ethiopia. And the changes in these hydrological variables are expected to affect the availability of water for irrigation and other water-consuming sectors in the region. Hence, to meet the food and potable water demands of the ever-growing population, water resources need to be managed efficiently.

The cessation of the season ranged from 255 to 269 (DOY) in the baseline period. Contrary to the onset, there is a tendency toward early cessation at most of the stations under future climate scenarios, ranging from no change at the Edagahamus station to 8 days at the Adigrat station. The tendency of late onset and early cessation will make the LGP decline over the study area in the future. The LGP of the season shortens by 4 days at the Wukro station to a maximum of 19 days at the Hawzen station in all periods compared to the baseline period. The results indicated that the LGP will decline at all stations under all RCPs and all time periods. Consequently, the short nature of the LGP of the study area will be further shortened, which is expected to affect crop production in the study area. Our results corroborate the findings of the previous study that indicated a decrease in the LGP in the future in southern Tigray (Gebrekiros et al. 2016).

The average DS length determined based on a 1 mm threshold rainfall in the Kiremt season varied from 11 to 15 days in the baseline period (Figures 5 and 6). Comparing the future scenarios with the baseline, the future DS length showed decreasing with a maximum of 5 days in the Edagahamus and Sinkata stations, but there is no significant difference under the two RCPs as well as across the stations. The small decreasing tendency of the DS length during the Kiremt season in the future time period could be a result of the contribution of the increasing rainfall projections obtained by the multiple GCMs at most of the stations over the study area.

The effects of climate change are being felt in many sectors all over the world. Hence, quantifying its effect on crop growing season characteristics will allow better preparation for small-scale farmers who are dependent on rainfed agriculture to make decisions on selecting crop varieties, where and when to plant, and when to harvest. In this study, daily temperatures and rainfall variables were projected for a future time period using selected GCMs under two RCPs (RCP4.5 and RCP8.5) to investigate the impact of climate change on crop growing season characteristics.

Results from this analysis indicate that the length of the growing period (LGP) shortened from 5.5 to 19% from mid-to-end of the century compared to the baseline period (1980–2009). In addition, the shortening of the growing season over the study area is expected toward the end of the century under RCP8.5. This shortening is associated with both a later onset and an early cessation. These changes are attributed to modifications in both precipitation and potential evapotranspiration. Hence, farmers need more caution in selecting early maturing and drought-tolerant crop varieties. Crop varieties with LGP greater than 90 days cannot be produced currently or in future climates unless appropriate adaptation strategies are implemented.

Overall, the results revealed that the study area may experience more warming in the future with the high and increasing trend of temperatures, which will exacerbate moisture stress and affect yield. Furthermore, higher variability and an increasing trend in rainfall may increase the risk of flooding and waterlogging conditions, eventually damaging crops and infrastructure. The implications of high temperatures as well as changes in moisture conditions in the future time period will possibly create favorable conditions and increase the risk of disease and pest outbreaks. The negative outcomes as a result of climate change in association with the decreasing trend of the LGP in the future pose a clear message that the continuation of crop production in the study area depending solely on rainfed conditions will be unthinkable. Hence, crop production in the area demands adaptation strategies that encourage supplemental irrigation and water harvesting technologies, as well as introducing tolerant varieties for high temperatures and new diseases while formulating appropriate pesticides for the adapted ones. It is very important to note that the results may have practical implications for formulating long-term adaptation strategies for sustainable crop production as well as the sustainable management of water resources in the region in particular and in other parts of the world in general. Finally, this research recommends that further research is needed that compares the latest CMIP6 with CMIP5 for the study region and examines the similarities and differences to be used for future climate change projections considering CMIPs.

The first author would like to express his sincere appreciation to the Open Society Foundation (OSF)-Africa Climate Change Adaptation Initiative (ACCAI) (Grant No. OR2016-30576) project based at Mekelle University for fully financing the research study. We would also like to acknowledge the NMA of Ethiopia for providing the weather data.

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

The authors declare there is no conflict.