Climate change and ongoing human activity have been threatening Ethiopia's Lake Ziway's water balance. However, few studies have been conducted to investigate the combined effects of climate change and water withdrawal on the lake's water balance using climate change and water withdrawal for irrigation. We used high-resolution multiple climate models and Representative Concentration Pathways (RCP) scenarios to assess the impact of climatic variables for two future periods: 2021–2050 and 2051–2080. Rainfall and temperature data biases were corrected using power transformation and variance scaling methods, respectively. The Hydrologiska Byråns Vattenbalansavdelning (HBV) hydrological model was employed to simulate surface inflow into Lake Ziway from the Meki and Katar sub-catchments. The FAO CROPWAT model was used to estimate the irrigation water demand of major crops grown in the study area. The results indicate that future temperatures and wet season runoff levels are expected to rise. Under the worst climate change scenario, climate change and water withdrawal from the lake for agriculture may cause the lake level to drop by 25 cm per year, resulting in a 10 km2 surface area and 101 Mm3 volume reductions. Therefore, implementing preventive measures, proper planning and careful monitoring of lake water use is advised.

  • Integration of multiple climate model, RCP scenarios, water abstraction, HBV hydrological and water balance model were used to evaluate Lake Ziway's water balance for impact analysis.

  • Bias-corrected climate model output resulted in improved rainfall and streamflow simulation.

  • Our findings indicate that the combined effect of climate change and water withdrawal affected both lake level and volume.

Graphical Abstract

Graphical Abstract
Graphical Abstract

Greenhouse gases have increased the heat stored in the earth's atmosphere, contributing to global warming and climate change (IPCC 2014). According to preliminary observations and climate projections, climate change significantly affects freshwater resources, with far-reaching consequences for human societies and ecosystems (Arnell 2004). Climate change affects temperature trends, evapotranspiration, precipitation, runoff, and streamflow. Temperatures will rise because of climate change, affecting catchment evapotranspiration and lake evaporation (Nigatu et al. 2016), and consequently, lake levels fluctuate due to climate change (Seyoum et al. 2015). The IPCC (2014) projected a general increase in precipitation in East Africa in the future. Simultaneously, the IPCC's Fifth Assessment Report revealed that increased global precipitation in the long term is associated with significant seasonal variations in precipitation (Collins et al. 2013). Arnell (2004) also reported an increase in annual streamflow in East Africa based on climate change projections, which is consistent with the findings of other researchers (IPCC 2014). The changes/variability in hydro-climatic variables yield frequent floods, droughts, and reduced food production, which eventually influences social well-being. Because water is used for a variety of purposes, increased human interference harms freshwater resources. As a result, it is critical to investigate the combined effect of climate change and anthropogenic influences on a catchment's water resources. The Lake Ziway catchment is one of the most important catchments in Ethiopia's rift valley lakes basin. Climate change and human influences may jeopardize the availability of water resources in the future. However, the magnitude of change caused by these factors has yet to be quantified.

Lake Ziway and its environs are vital to the livelihoods of about 2 million people, according to Central Statistical Authority (CSA) report in 2014. According to Hayal et al. (2015), Lake Ziway serves as a source of drinking water for local towns, irrigation supplies, and the supply of fish to the country's largest market center. However, climate change/variability and the intensification of agricultural operations in the catchment are putting a strain on the lake. Water withdrawal from the lake and Feeder Rivers for irrigated farms is rising (Ayenew 2004). Rivers feeding into Lake Ziway have been diverted for irrigation recently. Furthermore, farmers and investors around the lake regularly pump water from the lake for irrigation resulting in a decreased outflow into the Bulbula River (Herco et al. 2007; Desta et al. 2020). Climate change and anthropogenic factors may have a folded effect on the lake's hydrological and ecological integrity, which could be harmful (Ayenew 2004; Seyoum et al. 2015; Desta et al. 2020; Goshime et al. 2019a). However, the combined effects of future climate change and water withdrawal from the lake and tributary rivers are poorly understood.

Variations in main climate variables such as temperature, precipitation, and evapotranspiration are one of the repercussions of climate change. This, in turn, could cause changes in the hydrological cycle, reducing lake level and ultimately affecting the catchment's water availability. Several researchers revealed the presence of both climate variability/change and a strong anthropogenic disturbance on Lake Ziway's water levels (Zeray et al. 2006; Herco et al. 2007; Hayal et al. 2015; Desta et al. 2017, 2020; Damtew et al. 2021). However, most of these studies focused solely on the water balance under natural conditions, water abstraction, or land-use changes. Therefore, this study uses copious Global and Regional Climate Models as input to the Hydrologiska Byråns Vattenbalansavdelning (HBV) hydrological model to evaluate and quantify the implications of climate change and future planned water abstraction for irrigation scenarios from Lake Ziway.

The Lake Ziway catchment is situated approximately between 7°25′29.9″ and 8°34′30.5″N latitudes and 38°12′00″ and 39°15′01.5″E longitudes in the rift valley basin, Ethiopia (Figure 1). The Meki and Katar sub-catchments contribute most of the water volume entering Lake Ziway. The surface area of the Meki and Katar sub-catchments is approximately 6,569 km2. The open water surface area of the lake is 423 km2. Lake Ziway is Ethiopia's shallowest lake, and its outflow flows into the Bulbula River before merging with Lake Abiyata (Dagnachew et al. 2005). The elevation of the Lake Ziway catchment ranges from 1,600 to 4,200 meters above mean sea level, with a mean elevation of 2,900 meters. The highlands have a higher drainage density than the escarpments due to rock permeability, climate, and slope changes (Dagnachew et al. 2004).
Figure 1

The location of Lake Ziway including major feeding rivers (Meki and Katar) and outflow river (Bulbula) with hydro-meteorological stations.

Figure 1

The location of Lake Ziway including major feeding rivers (Meki and Katar) and outflow river (Bulbula) with hydro-meteorological stations.

Close modal

The climate in the Lake Ziway catchment ranges from humid to sub-humid. The lowland catchment areas surrounding the lake have arid or semi-arid climates, while the mountains have dry-humid to humid climates (Tenalem et al. 2007). The annual temperature ranges between 12.5 and 25.8 °C. The yearly oscillation of the inter-tropical convergence zone has a significant impact on rainfall patterns in the study area, resulting in warm and wet summers and dry, cold, and windy winters. According to time series data from 1986 to 2016, the catchment's mean annual rainfall ranges from 713 mm near the lake floor to 1,146 mm on the plateaus at the Ziway and Butajira stations. Agriculture/cultivated area is the most common land use/cover, accounting for 81% of the total land use area. The remaining areas are covered by forests, bare surfaces, and water bodies.

Observed data

We collected meteorological data such as temperature, rainfall, sunlight hour, wind speed, and relative humidity from the Ethiopian Meteorology Institute (EMI). The period of time-series data spans from the years 1980 to 2018. Data quality measures such as homogeneity, outlier, and consistency tests were used to find complete and consistent datasets. The streamflow and lake level data were obtained from the Hydrology Department of the Ministry of Water and Energy (MoWE) and used to calibrate and validate the hydrological model. MoWE Ethiopia also provided the bathymetric survey data of Lake Ziway conducted in 2013.

The Digital Elevation Model (DEM), having a resolution of 30 m × 30 m from ASTER GDEM V2, was downloaded from https://lpdaac.usg.gov.data access and was used to delineate the catchment. The land use/land cover (LULC) map of Ethiopia for 2016 was used to acquire land use and land cover statistics. The research area's soil map was also received from Ethiopia's Ministry of Water and Energy. The dominant crops grown in the area and cropping pattern data were collected using a field survey in the irrigation area. Throughout the year, the study area has two cropping seasons. A total of 4,099 hectares of the Lake Ziway sub-catchment was surveyed for data on commercial and large-scale farms managed by the government and private investors. Moreover, the Meki-Ziway large-scale irrigation project-II, having the potential of a 2,000 ha irrigable area, not yet functional, was considered in this study for future scenarios. Earlier studies did not consider this scheme and new development plans in this study area.

Future data

We used dynamically downscaled regional climate model (RCM) output data from the Coordinated Regional Climate Downscaling Experiment for the Regional Climate Change Assessment Project (CORDEX) from https://esgf-node.llnl.gov/projects/esgf-llnl/ in the current study. Using downscaled climate data with a resolution of 0.44° × 0.44°, CORDEX Africa was used to investigate future precipitation and temperature projections (Musie et al. 2020). This study used data from six (6) GCM and four (4) RCM models (Table 1). With a temporal resolution of one day, the climate model output includes precipitation, maximum and minimum temperature, humidity, wind speed, sunshine hours, and solar radiation for the reference and future periods from RCP 4.5 and RCP 8.5.

Table 1

GCM-driven CORDEX RCM models selected for this study

GCM ModelsHorizontal resolutionRCMInstitute (GCM)Country
CNRM 1.4° × 1.4° RCA4 CNRM–CERFACS: Centre National de Recherches Météorologiques Groupe d'études de l'Atmosphère Météorologique and Centre Européen de Recherche et de Formation Avancée France 
ICHEC 1.125° × 1.12° CCLM4.8, ReMo2009 ICHEC: Consortium of European research institutions and researchers Europe 
MPI 1.9° × 1.9° ReMo2009 MPI-M: Max-Planck-Institute Germany 
GFDL 2.5° × 2° RCA4, ReMo2009 NOAA GFDL: Geophysical Fluid Dynamics Laboratory (GFDL) United States 
MOHC 1.875° × 1.25° CCLM4-8 MOHC: Met Office Hadley Centre United Kingdom 
IPSL 3.75° × 2.5° ReMo2009, RCA4 Institute Pierre Simon Laplace France 
GCM ModelsHorizontal resolutionRCMInstitute (GCM)Country
CNRM 1.4° × 1.4° RCA4 CNRM–CERFACS: Centre National de Recherches Météorologiques Groupe d'études de l'Atmosphère Météorologique and Centre Européen de Recherche et de Formation Avancée France 
ICHEC 1.125° × 1.12° CCLM4.8, ReMo2009 ICHEC: Consortium of European research institutions and researchers Europe 
MPI 1.9° × 1.9° ReMo2009 MPI-M: Max-Planck-Institute Germany 
GFDL 2.5° × 2° RCA4, ReMo2009 NOAA GFDL: Geophysical Fluid Dynamics Laboratory (GFDL) United States 
MOHC 1.875° × 1.25° CCLM4-8 MOHC: Met Office Hadley Centre United Kingdom 
IPSL 3.75° × 2.5° ReMo2009, RCA4 Institute Pierre Simon Laplace France 

Evaluation of climate models

Several climate models were used to forecast future climate variables. Climate model data for hydrologic modeling (CMhyd) was used to extract data from global and regional climate models (Rathjens et al. 2016). We compared climate models before bias correction to identify high-performing models capable of capturing basin climatic aspects. The climate model that meets a set of statistical criteria (Musie et al. 2020) was chosen. Statistical metrics such as the root mean square error (RMSE), percentage bias, correlation coefficient (CC), and coefficient of variation (CV) were used to evaluate the effectiveness of the climate models. Table 2 shows the statistical measures used to evaluate model performance.

Table 2

Statistical measures used for performance evaluation of the GCM/RCM products

Statistical MeasureEquationsUnitBest fit value
Bias  
¥RMSE  mm 
CV  
CC  – 
Statistical MeasureEquationsUnitBest fit value
Bias  
¥RMSE  mm 
CV  
CC  – 

¥RMSE, Root mean square error; CV, Coefficient of variation; CC, Correlation Coefficient; RR and Rob is RCM simulated rainfall and observed rainfall, respectively, over bar symbol denotes the mean of the statistical value over the analysis period (n); σR refers to the standard deviation of either RCM or observed rainfall.

Bias correction of climate model data

The downscaled RCM products, including daily precipitation and minimum and maximum temperature values, may have biases that must be corrected before they are used. We selected the power transformation method for precipitation bias correction based on the recommendations of Lafon et al. (2013) and Pratama et al. (2018). The power transformation method is given as follows:
formula
(1)
where Pcorr = corrected precipitation on the dth day of mth month, Praw = raw precipitation on the dth day of mth month, a & b are the bias parameters.
The variance scaling technique adjusts the parameter of temperature, especially the mean and variance of the normally distributed variable (Teutschbein & Seibert 2012). The equation is given as (Equation (2)):
formula
(2)
where = corrected temperature; = observed temperature; = raw temperature all are on the dth day of mth month; = mean and standard deviation, respectively.

HBV hydrological model

In this study, the Integrated Hydrological Modeling System (IHMS) version 6.3 HBV rainfall-runoff model was selected to simulate the streamflow. The HBV hydrologic model was used in this study because of its efficacy in Ethiopian catchments (Wale et al. 2009; Goshime et al. 2019b). The governing water balance equation used by the model is:
formula
(3)
where SP is the snowpack (mm); SM is the soil moisture (mm); SUZ is the upper groundwater zone storage (mm); SLZ is the lower groundwater zone storage (mm); P is the precipitation (mm/day); E is the evapotranspiration (mm/day); Q is the runoff (mm/day) and t refers to the time scale (day).

The contents of the model, i.e., the routines for each variable, the procedures of simulations for mass exchange between the upper and lower boundaries, and other detailed descriptions of the model are documented in recent publications such as in Habib et al. (2014), Dessie et al. (2015) and Goshime et al. (2019b). Daily rainfall, temperature, potential evapotranspiration, river flow, land use/land cover, and the research area's Digital Elevation Model (DEM) were used as input variables for the HBV hydrological model. The measured river discharge at Meki and Katar was used to calibrate and validate the HBV hydrologic model. Eight model parameters, K4, Khq, Alfa, CFLUX, FC, BETA, LP, and PREC, from the earlier studies were selected to calibrate the model (Wale et al. 2009; Goshime et al. 2019b, 2020).

Table 3 presents the descriptions of the selected calibrated model parameters and their initial, minimum, and maximum values.

Table 3

Selected model parameters for calibration and their value range

ParameterDescriptionUnitMinimumMaximumInitial value
Alfa The coefficient for non-linearity of flow – 1.5 0.6 
BETA The exponent in drainage from the soil layer – 2.5 
CFLUX Maximum capillary flow mm 0.5 
FC Field capacity mm 100 1,500 200 
K4 Recession coefficient for the lower zone d−1 0.001 0.1 0.01 
Khq Recession coefficient for the upper zone d−1 0.005 0.5 0.1 
LP Limit for potential evaporation – 0.1 0.9 
PERC Percolation capacity mmd−1 0.01 0.5 
ParameterDescriptionUnitMinimumMaximumInitial value
Alfa The coefficient for non-linearity of flow – 1.5 0.6 
BETA The exponent in drainage from the soil layer – 2.5 
CFLUX Maximum capillary flow mm 0.5 
FC Field capacity mm 100 1,500 200 
K4 Recession coefficient for the lower zone d−1 0.001 0.1 0.01 
Khq Recession coefficient for the upper zone d−1 0.005 0.5 0.1 
LP Limit for potential evaporation – 0.1 0.9 
PERC Percolation capacity mmd−1 0.01 0.5 

Model calibration and evaluation

The model was initialized for a 1-year warm-up period to conduct the model calibration process. Then, model sensitivity analysis was carried out to obtain model-sensitive parameters so that attention is given to these parameters during model calibration. The default values of model parameters were used as a benchmark in deciding sensitive parameters. The sensitivity analysis was conducted by varying a model parameter and investigating the overall model run-up performance using statistical indices such as relative volume error (RVE) and Nash-Sutcliffe efficiency (NSE). The parameters that produce significant variation in the statistical indices for varying model runs were selected. Parameter sensitivity is measured by varying one parameter at a time and fixing all other parameters during the model run. A parameter whose rate of change is high compared to others during model runs is considered a sensitive parameter. Finally, the model was calibrated using streamflow data from 1985 to 1991 and verified using daily flow data from 2000 to 2007.

The model was calibrated and validated using streamflow data from the Katar and Meki sub-catchments. Model calibration was carried out by minimizing the difference between the observed and simulated streamflow at each of the sub-catchments using statistical indices. The model performance was evaluated using graphical observation and statistical indices. Error statistics such as the coefficient of determination (R2), Nash Sutcliffe Efficiency (NSE), and relative volume error (RVE) were utilized to evaluate the performance of the hydrologic model. The coefficient of determination (R2) measures how well the model replicates the observed results. NSE is a model efficiency metric that compares simulated values to observed values. The relative volume error (RVE) measures the volume difference between the computed and observed streamflow. The equations of the performance measure are given by:
formula
(4)
formula
(5)
formula
(6)
where Qo is the observed streamflow (ms−3), Qs is the computed streamflow (ms−3), the over-bar symbol represents the average values, and n is the number of data points.

Lake water balance

Changes in water balance are managed through the difference between the input, and output fluxes, which are commonly influenced by the local hydrological processes (Benn & Lehmkuhl 2000). The inflow and outflow components of the water balance were approximated monthly to create the Lake Ziway water balance model. The water balance equations using various components were used to compute the water volume, area, and depth. Monthly inflow, precipitation, outflow, lake evaporation, water abstractions, and lake-level data were used to assess water balance. The recursive continuity equation was formulated and described as follows (Ayenew 2007):
formula
(7)
where St and St−1 are storage at the current month and end of the preceding month (mm), respectively; Qit and Qot are inflow to and outflow from the lake at the current month, Pt is the mean areal rainfall on the lake, Abt is an abstraction of water and Et is evaporation loss from the lake (all dimensions for the length variables are given in mm).

Estimation of lake precipitation and evaporation

The areal precipitation over the lake's catchment area was estimated using data from four stations (Meki, Ziway, Adami Tulu, and Ogolcho). The areal precipitation was calculated using the Thiessen polygon method. After calculating the depth of rainfall over the surface area, the volume of rainfall was calculated by multiplying the average lake area by the rainfall depth. Finally, the potential evapotranspiration (PET) was calculated using a combination of the Penman-Monteith (Allen et al. 1998) and Hargreaves (Hargreaves & Allen 1985) methods.

Estimation of surface inflows and outflows

The streamflow from the two gauged main rivers (Meki and Katar) was simulated using the calibrated HBV model. Then, the rainfall in the ungauged catchment was translated into runoff using the area-ratio approach for the two sub-catchments located in similar rainfall regime areas (Nruthya & Srinivas 2015). The area-ratio method was chosen due to similarities in the hydrological, geological, and land cover of the gauged and ungauged catchments.

Water demand estimations

The water demand for irrigation was estimated using the CROPWAT 8.0 method (Allen et al. 1998) by applying site-specific crop data. The CROPWAT model was used to compute the crop's water requirements (CWR) using the study area's climatic, soil, and crop data. Following the FAO protocol, crop and soil data were utilized to assess consumptive usage. Field data were collected to identify common crops grown and cropping patterns of the study region. Accordingly, onion, tomato, green beans, cabbage, maize, pepper, papaya, and wheat are the most commonly grown crops in the study area. We also used crop coefficients, growth stage, depletion, and soil data from FAO (Allen et al. 1998) and other documents, including the basin master plan and irrigation system design report to compute the irrigation water demand. The water demand was estimated by considering the following:

  • i.

    Irrigable area expansion with the current precipitation and temperature condition

  • ii.

    Area expansion with climate change in the future climatic variables

  • iii.

    The potential irrigable area is reaching 30,090 ha in future periods

Lake water level

After determining all water balance components, a spreadsheet model was created to mimic lake water levels using Equation (7) (Goshime et al. 2020). Lake bathymetric data was utilized to obtain the observed lake-level data from Ethiopia's Ministry of Water and Energy in 2013. The simulated lake level was then compared to the observed lake level to investigate feasible differences. Using the bathymetric data, the lake volume was converted to lake level. The simulated lake level's accuracy was then compared to the observed lake level.

Evaluation of climate models

The mean annual value comparison of observed and climate model outputs of precipitation revealed seasonal variability and shifts compared to baseline observed climate records (Figure 2). In general, it can be noticed that most of the climate models attempt to capture the seasonal pattern of rainfall in the region. However, one of the climate models, MOHC-CCLM4.8, ultimately failed to capture the general trend of the monthly rainfall over the catchment area. A significant disparity between the observed and model-estimated rainfall is observed during the rainy season compared to the dry seasons. This may be attributed to the inherent nature of climate models, the assumptions included in the models, regional variations in rainfall, temporal and spatial scales, the behavior of the climatic variable, and others (Räisänen 2007; Randall et al. 2007; Sun et al. 2018). Climate models tend to more accurately estimate temperature than precipitation (Räisänen 2007; Randall et al. 2007).
Figure 2

Comparison of monthly precipitation between observed and nine (9) climate model data for 1984–2000.

Figure 2

Comparison of monthly precipitation between observed and nine (9) climate model data for 1984–2000.

Close modal

When comparing the observed mean annual rainfall with model counterparts, there exists an overestimate and underestimation for the respective RCMs. Seven models (IPSL-RCA4, ICHEC-CCLM4.8, ICHEC-ReMo2009, CNRM-RCA4, MPI-ReMo2009, GFDLG-RCA4, and GFDL2M-ReMo2009) show higher mean annual precipitation value than the observed mean annual values. GFDLM-RCA4, IPSL-ReMo2009, and CNRM-RCA4 climate models highly overestimated the rainfall in the wet season, while GFDL2G-ReMo2009 and ICHEC-ReMo2009 slightly overestimated the rainfall during the same season. The ICHEC-CCLM4.8 model underestimated wet-season rainfall over the catchment area. However, MPI-ReMo2009 and IPSL-RCA4 models revealed a more consistent and closer pattern with observed precipitation values (Figure 2 and Table 4). As a result, these two models were chosen for this study's intended purpose.

Table 4

Comparison of observed and climate models annual average precipitation over Lake Ziway sub-catchment for the period 1984–2000

Observation/Climate ModelCV (%)Bias (%)RMSE (mm)CC (–)Annual Rainfall (mm)
Observed 1.91 – – – 837 
IPSL-RCA4 1.98 15.13 5.37 0.88 845 
ICHEC-CCLM4-8 1.59 28.10 6.77 0.65 1,082 
ICHEC-ReMo2009 1.37 40.63 5.80 0.67 850 
CNRM-RCA4 1.53 27.79 5.77 0.64 1,126 
MPI-ReMO2009 1.83 16.48 5.22 0.78 840 
GFDLM-RCA4 1.63 55.78 6.72 0.73 1,346 
IPSL -ReMO2009 1.77 −29.40 5.38 0.69 758 
MOHC -CCLM4.8 2.79 −31.16 6.41 −0.67 566 
GFDLG-ReMO2009 1.48 16.58 5.68 0.57 1,083 
Observation/Climate ModelCV (%)Bias (%)RMSE (mm)CC (–)Annual Rainfall (mm)
Observed 1.91 – – – 837 
IPSL-RCA4 1.98 15.13 5.37 0.88 845 
ICHEC-CCLM4-8 1.59 28.10 6.77 0.65 1,082 
ICHEC-ReMo2009 1.37 40.63 5.80 0.67 850 
CNRM-RCA4 1.53 27.79 5.77 0.64 1,126 
MPI-ReMO2009 1.83 16.48 5.22 0.78 840 
GFDLM-RCA4 1.63 55.78 6.72 0.73 1,346 
IPSL -ReMO2009 1.77 −29.40 5.38 0.69 758 
MOHC -CCLM4.8 2.79 −31.16 6.41 −0.67 566 
GFDLG-ReMO2009 1.48 16.58 5.68 0.57 1,083 

Figure 3 depicts a graph of the observed and simulated GCM/RCM models' mean maximum (top) and minimum (bottom) monthly temperatures. Most models did not capture the observed mean maximum and minimum monthly temperature values. Four models completely underestimate the catchment area's mean monthly temperature. The ICHEC-CCLM4.8 and MOHC -CCLM4.8 climate models revealed the worst trends in the average monthly minimum and maximum temperature simulations. The MPI-ReMo2009 and IPSL-RCA4 models were more efficient than the other models in capturing the observed values for both minimum and maximum temperatures. As a result, the two models could be used to study climate effects in the catchment area.
Figure 3

Comparison of monthly mean (a) maximum temperature (top) and (b) minimum temperature (bottom) between observed and climate model data for 1984–2000.

Figure 3

Comparison of monthly mean (a) maximum temperature (top) and (b) minimum temperature (bottom) between observed and climate model data for 1984–2000.

Close modal

Table 4 shows the statistical indices used to evaluate the performance of the climate models. The mean annual value comparison of precipitation variables between observed and climate model outputs revealed that MPI-ReMO2009 (bias = 16.48%) and IPSL-RCA4 (bias = 15.13%) performed well. In contrast, GFDLM-RCA4 (bias = 55.78%) showed a significant deviation from the observed value based on the bias (Table 4). In addition, MOHC -CCLM4.8 shows a negative association with observed precipitation in terms of correlation coefficient (CC). The MPI-ReMO2009 and IPSL-RCA4 climate models estimated precipitation values closer to those observed values. Therefore, these models were used to assess the effects of climate change on the water balance of the study area.

This study's result agrees with the study by Kumar et al. (2020) that revealed the RCM (REMO2009) and its derivative GCM (MPI) model portrayed the rainfall value near the Global Precipitation Climatology Centre (GPCC) observed rainfall. The authors reported that the biases in REMO2009 and GCM (MPI) are comparable in amplitude, making them suitable for future projection of climate variables under Representative Concentration Pathways (RCPs) 4.5 and 8.5 at 99 and 95% confidence levels. According to Ogega et al. (2020), when compared to the other 24 multiple climate models runs from five Coordinated Regional Climates Downscaling Experiments, the RCM (ReMo2009) forced by MPI has the best performance in simulating East Africa's spatio-temporal precipitation characteristics (CORDEX).

We used a bias correction approach for precipitation and temperature for the IPSL-RCA4 and MPI-ReMo2009 climate models for the RCP 4.5 and RCP 8.5 scenarios (Figures 4(a) and 4(b)). After bias correction, the models accurately captured the pattern and magnitude of precipitation as well as the mean maximum and minimum temperature values. The corrected minimum and maximum temperatures, as well as precipitation values, corresponded closely to the observed values. Following bias correction, the selected climate models outperformed the uncorrected climatic variables in terms of performance and improvement. Biases in the original uncorrected climate variables observed across the study area have been significantly reduced and used to simulate the water balance and water level of Lake Ziway.
Figure 4

Monthly observed, corrected, and uncorrected (a) precipitation (top) and (b) maximum temperature and minimum temperature (bottom) for the period from 1984–2000.

Figure 4

Monthly observed, corrected, and uncorrected (a) precipitation (top) and (b) maximum temperature and minimum temperature (bottom) for the period from 1984–2000.

Close modal

Climate change impact on precipitation

The mean annual precipitation change in the Lake Ziway sub-catchment is expected to range from 2% (2021–2050) to 9.2% (2051–2080) for the RCP 4.5 scenarios. For the RCP 8.5 scenario, the annual range of precipitation change is between 3.5% (2051–2080) to 18% (2051–2080). The change in precipitation on a monthly period does not exhibit any consistent rises or declines. Climate estimates imply that precipitation will likely fall between January and December in the future, with a few exceptions where the drop will vary from 7.8 to 32 percent. Furthermore, due to the rainy season, precipitation is projected to increase from June to August in both RCP scenarios (Figure 5). Previous research has also shown that climate change has an effect on the expected annual mean precipitation pattern in East Africa (Zeray et al. 2006; Tekleab et al. 2013). Because global warming is expected to increase evaporation and intensify water cycling, climate model projections show that global mean precipitation will rise (Meehl et al. 2007). However, precipitation increases in east Africa have been less noticeable than temperature trends over the last 50 years, and the frequency and magnitude of changes vary significantly across the region.
Figure 5

Projected changes in monthly, seasonal, and annual precipitation of the Meki (bottom row) and Katar (top row) catchments compared to the baseline period (1971–2000) under the RCP 4.5 & 8.5 scenarios.

Figure 5

Projected changes in monthly, seasonal, and annual precipitation of the Meki (bottom row) and Katar (top row) catchments compared to the baseline period (1971–2000) under the RCP 4.5 & 8.5 scenarios.

Close modal

Figure 5 shows the projected changes in monthly, seasonal, and annual precipitation of the Meki and Katar sub-catchments compared to the baseline period (1971–2000) under the RCP 4.5 & 8.5 scenarios.

Climate change impact on the minimum and maximum temperature

The mean annual maximum temperature over the Lake Ziway catchment will rise by 1.6 degrees Celsius for all models under the RCP 4.5 emission scenario and 2.5 degrees Celsius for all models under the RCP 8.5 scenario between 2021 and 2080, compared to the baseline period (1971–2000). All models show that the monthly maximum temperature will rise in all months except June and August, which showed the smallest increase compared to the other months under both scenarios. However, due to increased CO2 emissions, the RCP 8.5 scenario (2051–2080) has shown the greatest increase of all models.

According to the Ethiopian Meteorological Institute (EMI), the Ethiopian climate is divided into three seasons: Belg (humid, moist season from February – May), Kiremt (wet summer season from June – September), and Bega (dry winter from October – January) (Degefu 1987). In comparison to the other seasons (Bega and Belg), the maximum temperature in Kiremt (June-September) showed the smallest increase (Figure 6). The results revealed that the maximum temperature in the study area is likely to rise. Previous investigations in the study area (Zeray et al. 2006; Abraham et al. 2018) also found similar results. Under both RCP (RCP 4.5 & RCP8.5) scenarios, the lowest temperature on an annual, monthly, and seasonal basis is anticipated to rise in the future timeframe for all models. The mean annual minimum temperature will rise from 2.5 to 3.2 °C for all models under the RCP 4.5 scenario and from 3.8 to 4.2 °C under the RCP 8.5 scenario. The minimum temperature rises at a faster rate than the maximum temperature. This signifies that the nights are warmer than the days (Tekleab et al. 2013; Osima et al. 2018). Because RCP 8.5 is a high-emission scenario that produces more greenhouse gas than RCP 4.5, the maximum and lowest temperature changes are more remarkable in RCP 8.5 than in RCP 4.5 (Riahi et al. 2011). Regional investigations across East Africa have revealed substantial warming (Anyah & Qiu 2012). According to the authors, warm/cold extremes are becoming more common, with warm days and nights becoming common. Daron (2014) also indicates that temperatures in East Africa's central regions have risen by 1.5–2 °C. By 2050, average annual temperature increases are expected to range from no change to 4 °C, while model forecasts are susceptible to significant uncertainty. In a situation with higher/lower greenhouse gas emissions, relatively high/low increases are more plausible (Daron 2014).
Figure 6

Annual, monthly, and seasonal minimum temperature changes for 2021–2050 (a) and for 2051–2080 (b) and maximum temperature changes for 2021–2050 (c) for 2051–2080 (d) compared to the baseline period.

Figure 6

Annual, monthly, and seasonal minimum temperature changes for 2021–2050 (a) and for 2051–2080 (b) and maximum temperature changes for 2021–2050 (c) for 2051–2080 (d) compared to the baseline period.

Close modal

Model sensitivity, calibration and validation

The results of model sensitivity analysis of the HBV model parameters in the Katar and Meki catchments revealed that the BETA, FC, and LP parameters were the most sensitive, determining the catchment storage. The response of the NSE statistical index indicated that the parameters that determine the shape of the hydrograph, such as K4, PERC, and Khq, were moderately sensitive. When compared to other parameters, the PERC and Alfa parameters were found to be less sensitive. Goshime et al. (2019b; 2020) reported similar findings over the Lake Ziway area, as did Dessie et al. (2015) over the Lake Tana sub-basin in Ethiopia. Hence, the calibration process should take caution while calibrating volume and shape-controlling parameters in the study area.

Figure 7 shows the model calibration (1985–1999) and validation (2000–2007) hydrographs for the Katar sub-catchment, comparing the simulated and observed streamflow hydrographs. The model reasonably captured the pattern of the observed hydrograph in both the recession and rising limbs of the hydrograph. The peak and base flow hydrographs were also reasonably caught, except for a few peak discharges. We note that the pattern and magnitude of some peaks were not captured which is mainly attributed to the uncertainty in observed data, the inherent nature and structure of hydrological models, and inaccuracy in the rating curve of the gauging site (Renard et al. 2010; Kahsay et al. 2018).
Figure 7

Observed and simulated hydrograph of Katar catchment for calibration (1985–1999) and validation (2000–2007) periods for daily time scale (the solid black line represents the period between calibration and validation).

Figure 7

Observed and simulated hydrograph of Katar catchment for calibration (1985–1999) and validation (2000–2007) periods for daily time scale (the solid black line represents the period between calibration and validation).

Close modal
Figure 8 depicts the Meki sub-catchments model calibration (1985–1999) and validation (2000–2007) hydrographs, comparing simulated and observed streamflow. The HBV model captured the pattern and magnitude of the observed hydrographs better. However, the model failed to capture the rapidly varying portion of the observed hydrograph and some of the observed low streamflow magnitudes. Except for a few peak discharge values, the model captured the observed stream flow better in the Katar sub-catchment than in the Meki sub-catchment.
Figure 8

Observed and simulated hydrograph of Meki catchment for calibration (1985–1999) and validation (2000–2007) periods for daily time scale (the pink line represents the period between calibration and validation).

Figure 8

Observed and simulated hydrograph of Meki catchment for calibration (1985–1999) and validation (2000–2007) periods for daily time scale (the pink line represents the period between calibration and validation).

Close modal

The HBV model performs well when measured quantitatively using the NSE and RVE statistical indices (Table 5). Table 5 shows that the model-calibrated parameter values are within the acceptable range. Using data from the Katar sub-catchment, the model's performance was found to be very good, with R2 = 0.8 and NSE = 0.63 during the model calibration period from 1985 to 1999, and R2 = 0.78 and NSE = 0.62 during the validation period from 2000 to 2007. Similarly, when the Meki catchment data was used, the model performed very well, with R2 = 0.81 and NSE = 0.64 during calibration and R2 = 0.79 and NSE = 0.54 during validation for the same period as the Katar sub-catchment. The model's performance was slightly lower during the validation period than during the calibration period. However, its performance during the validation period is still very good, indicating that the model can be used to project future streamflow into Lake Ziway.

Table 5

Calibrated model parameters and their performance statistical measures

Sub-catchmentParameters
Calibration
Validation
AlfaBetaCfluxFcK4KhqLpPercR2NSERVER2NSERVE
Katar 0.7 0.8 860 0.045 0.07 0.52 0.54 0.8 0.63 2.94 0.78 0.62 2.40 
Meki 1.05 1.5 0.01 750 0.115 0.3 0.42 1.5 0.82 0.64 5.57 0.79 0.54 0.67 
Range 0–1.5 0–2 0–2 100–1,500 0.001–0.1 0.005–0.5 ≤1 0.01–6  
Sub-catchmentParameters
Calibration
Validation
AlfaBetaCfluxFcK4KhqLpPercR2NSERVER2NSERVE
Katar 0.7 0.8 860 0.045 0.07 0.52 0.54 0.8 0.63 2.94 0.78 0.62 2.40 
Meki 1.05 1.5 0.01 750 0.115 0.3 0.42 1.5 0.82 0.64 5.57 0.79 0.54 0.67 
Range 0–1.5 0–2 0–2 100–1,500 0.001–0.1 0.005–0.5 ≤1 0.01–6  

Impact of climate change on streamflow

In the RCP 8.5 scenario, the annual estimated streamflow volume for the basin increases from 11% (2021–2050) to 22% (2051–2080). Similarly, the catchment's total average annual inflow volume change is anticipated to grow by 13% in 2021–2050 and 18% in 2051–2080 under RCP 4.5 scenarios. The average total volume of flow in the catchment is expected to increase every month, except in April, May, and November in the future. The flow volume is also expected to increase in all three seasons in the future anticipated periods (2021–2050 and 2051–2080). However, under the two RCP scenarios, the highest increase will be recorded in the Ziway catchment during the Kiremit season (June-September) (Figure 9). A rise in average total yearly runoff volume is observed during the Kiremt season and for periods with comparable increases in mean annual precipitation.
Figure 9

Percentage change of ensemble annual, monthly, and seasonal surface runoff volume of Meki (first row) and Katar (second row) gauged and ungauged catchments under both RCP scenarios compared with the baseline period.

Figure 9

Percentage change of ensemble annual, monthly, and seasonal surface runoff volume of Meki (first row) and Katar (second row) gauged and ungauged catchments under both RCP scenarios compared with the baseline period.

Close modal

Irrigation water demand around Lake Ziway

According to CROPWAT model results, the total annual irrigation water demand is 39.4 Mm3 (Table 6). The current irrigation water requirement was calculated using average climate data from 1980 to 2009. Because the streamflow of the Meki and Katar rivers has decreased and there is a scarcity of water for irrigation using these rivers, the Meki and Katar catchments have only one cropping season. Therefore, water withdrawals for other purposes or sectors, aside from agricultural water use, have not been considered in this study. According to Ayenew (2004), increased irrigated farming in the sub-catchments has contributed to a significant decrease in the flow of the feeder rivers and the water level of Lake Ziway.

Table 6

Monthly existing irrigation water requirement using CROPWAT

MonthsMeki sub-catchment (Mm3)Katar sub-catchment (Mm3) Lake Ziway catchment (Mm3)Total existing irrigation water requirement (Mm3)
Jan 0.79 1.66 3.45 
Feb 1.95 1.95 
Mar 3.5 3.5 
Apr 3.41 3.41 
May 2.72 2.72 
Jun 2.52 2.52 
Jul 1.18 1.18 
Aug 0.07 0.19 1.06 1.32 
Sep 0.4 0.49 2.89 
Oct 1.69 1.13 3.02 5.84 
Nov 1.71 1.5 3.15 6.36 
Dec 0.52 0.56 3.14 4.22 
Annual 4.39 4.87 29.3 39.4 
MonthsMeki sub-catchment (Mm3)Katar sub-catchment (Mm3) Lake Ziway catchment (Mm3)Total existing irrigation water requirement (Mm3)
Jan 0.79 1.66 3.45 
Feb 1.95 1.95 
Mar 3.5 3.5 
Apr 3.41 3.41 
May 2.72 2.72 
Jun 2.52 2.52 
Jul 1.18 1.18 
Aug 0.07 0.19 1.06 1.32 
Sep 0.4 0.49 2.89 
Oct 1.69 1.13 3.02 5.84 
Nov 1.71 1.5 3.15 6.36 
Dec 0.52 0.56 3.14 4.22 
Annual 4.39 4.87 29.3 39.4 

From 2021 to 2050, if 40% of the total irrigable area is used, the annual irrigation water requirement will be 81.1 Mm3, doubling from the baseline. Between 2051 and 2080, the yearly irrigation water needs for 60% of the total irrigable area will increase to 121 Mm3, a threefold increase over the baseline. The RCP 4.5 scenario projects annual irrigation demand of 85.6 Mm3 (2021–2050) and 124.4 vMm3 (2051–2080), respectively, whereas the RCP 8.5 scenario projects 102.8 Mm3 (2021–2050) and 125.6 Mm3 (2051–2080). As a result, if all of the intended irrigated areas are constructed, the annual water requirement is expected to be 150 Mm3. Such irrigation practices may cause a 3 m drop in Lake Ziway's level, dramatically dropping Lake Abiyata's level and the Bulbula River drying up (Ayenew 2007). Previous research found that abnormally dry years and extensive irrigation near Lake Ziway caused the lake to drop 1.5 meters below the long-term average (Tenalem 2002).

We considered the existing yearly total water abstraction of 42 Mm3 in this analysis. Earlier studies put the water abstraction at 38 Mm3 (Goshime et al. 2021), 28 Mm3 (Ayenew 2004), and 41 Mm3 (Desta & Lemma 2017). The dataset and methodology used in each of these studies may differ slightly. For example, under the RCP 4.5 scenario, the lake's annual total inflow and outflow are estimated to be 1,471.6 Mm3 and 1,268.4 Mm3 for 2021–2050, respectively, and 1,751.1 Mm3 and 1,321.9 Mm3 for the period 2051–2080. In contrast, the lake's inflow and outflow volume under RCP8.5 are 1,384.7 Mm3 and 1,256.8 Mm3 for 2021–2050 and 1,972.8 Mm3 and 1,352.8 Mm3 for 2051–2080, respectively (Table 7).

Table 7

Lake level simulation results for natural, existing, and future developments

ScenariosPeriodLake Water level (m.a.s.l)Surface area (km2)Volume (Mm3)Level variation (m)Area variation (km2)Volume variation (Mm3)
BS 1980–2009 1,636.33 443.31 1,492.36 – – – 
ED 1980–2009 1,635.91 422.89 1,313.77 −0.42 −20.42 −178.59 
FD 2021–2050 1,635.729 420.71 1,249.47 −0.601 −22.6 −242.89 
2051–2080 1,635.555 418.29 1,235.17 −0.775 −25.02 −257.19 
ScenariosPeriodLake Water level (m.a.s.l)Surface area (km2)Volume (Mm3)Level variation (m)Area variation (km2)Volume variation (Mm3)
BS 1980–2009 1,636.33 443.31 1,492.36 – – – 
ED 1980–2009 1,635.91 422.89 1,313.77 −0.42 −20.42 −178.59 
FD 2021–2050 1,635.729 420.71 1,249.47 −0.601 −22.6 −242.89 
2051–2080 1,635.555 418.29 1,235.17 −0.775 −25.02 −257.19 

Note: BS, Baseline natural; ED, Existing development; FD, Future development.

Lake Ziway water level simulation

The comparison of observed, modeled (for the natural state, BS), and existing lake level (ED) demonstrates that water abstraction has a significant impact on Lake Ziway's water level (Figure 10). This result is also in agreement with Goshime et al. (2021). The effects became more noticeable after 1990, which can be attributed to the cumulative impact of water abstraction and climatic variability across the research area. The lake's water level had been reduced by 42 cm, and its surface area had fallen by 20.42 km2, resulting in a volume reduction of 178.6 Mm3. According to Goshime et al. (2021), the lake's depth dropped by 36 cm, and its surface area shrank. The most significant drop in lake level is expected between 2051 and 2080, with a drop of up to 77.5 cm possible, considering future developments. As a result, the surface area could drop by 25.02 km2, and the volume could shrink by 257.2 Mm3. Abraham et al. (2018) found that due to climate change, the overall mean annual inflow volume into Lake Ziway will considerably decrease, resulting in a fall in the lake's water level and surface area of 0.62 m and 25 km2, respectively.
Figure 10

Observed, simulated, and existing Ziway Lake water level for the period 1980–2009.

Figure 10

Observed, simulated, and existing Ziway Lake water level for the period 1980–2009.

Close modal

Under the RCP 8.5 scenario, the lake water level will rise by 62.4 cm, while the surface area and storage will grow by 25.4 km2 and 293.3 Mm3, respectively, for the period 2051–2080. On the other hand, the lake's water level might rise by 56.7 cm above the average baseline lake level under the RCP 4.5 scenario, with a surface area of 23 km2. This occurrence could be linked to the use of GCM models and climate model uncertainties, the amount of data used and the methodology used in the study. We believe that the accuracy of this study's findings was enhanced by using two climate models and an ensemble.

Finally, the combined effect of irrigation expansion and climatic change was assessed, and the results revealed a decrease in lake water level, volume, and surface area (Table 8). Under the RCP 4.5 scenario, the lake water level, surface area, and storage loss might be as much as 25 cm, 10.3 km2, and 101.7 Mm3, respectively, between 2021 and 2050. The highest lake level reduction under RCP 8.5 might be up to 23 cm. This suggests that water abstraction, rather than climate change, will have the most detrimental impact on Lake Ziway's water level. Seyoum et al. (2015) found similar results for Lakes Abiyata and Ziway, informing that lake level rise is attributed to increased river inflow and over-lake precipitation. However, extensive irrigation water abstraction reduced the volume of both lakes.

Table 8

Lake level simulation results for natural (baseline), current, and future developments

ScenariosPeriodLake Water level (m.a.s.l)Surface Area (km2)Volume (Mm3)Level variation (m)Area variation (km2)Volume variation (Mm3)
BS 1980–2009 1,636.33 443.31 1,492.36 – – – 
ED 1980–2009 1,635.91 422.89 1,313.77 −0.42 −20.42 −178.59 
FD 2021–2050 1,635.729 420.71 1,249.47 −0.601 −22.6 −242.89 
2051–2080 1,635.555 418.29 1,235.17 −0.775 −25.02 −257.19 
RCP 4.5 2021–2050 1,636.659 459.08 1,643.7 0.329 15.77 151.34 
2051–2080 1,636.898 462.54 1,729.81 0.567 23.23 237.45 
RCP 8.5 2021–2050 1,636.557 455.25 1,594.33 0.226 11.94 101.97 
2051–2080 1,636.954 463.68 1,735.7 0.624 25.37 243.34 
RCP 4.5 + FD 2021–2050 1,636.16 433.96 1,410.7 −0.17 −9.35 −81.7 
2051–2080 1,636.08 432.99 1,390.67 −0.25 −10.32 −101.73 
RCP 8.5 + FD 2021–2050 1,636.17 435.57 1,417.2 −0.16 −7.74 −75.2 
2051–2080 1,636.1 433.9 1,396.7 −0.23 −9.41 −95.7 
ScenariosPeriodLake Water level (m.a.s.l)Surface Area (km2)Volume (Mm3)Level variation (m)Area variation (km2)Volume variation (Mm3)
BS 1980–2009 1,636.33 443.31 1,492.36 – – – 
ED 1980–2009 1,635.91 422.89 1,313.77 −0.42 −20.42 −178.59 
FD 2021–2050 1,635.729 420.71 1,249.47 −0.601 −22.6 −242.89 
2051–2080 1,635.555 418.29 1,235.17 −0.775 −25.02 −257.19 
RCP 4.5 2021–2050 1,636.659 459.08 1,643.7 0.329 15.77 151.34 
2051–2080 1,636.898 462.54 1,729.81 0.567 23.23 237.45 
RCP 8.5 2021–2050 1,636.557 455.25 1,594.33 0.226 11.94 101.97 
2051–2080 1,636.954 463.68 1,735.7 0.624 25.37 243.34 
RCP 4.5 + FD 2021–2050 1,636.16 433.96 1,410.7 −0.17 −9.35 −81.7 
2051–2080 1,636.08 432.99 1,390.67 −0.25 −10.32 −101.73 
RCP 8.5 + FD 2021–2050 1,636.17 435.57 1,417.2 −0.16 −7.74 −75.2 
2051–2080 1,636.1 433.9 1,396.7 −0.23 −9.41 −95.7 

Note: BS, Baseline natural; ED, Existing development; FD, Future development.

In contrast to earlier studies, this analysis combined rainfall-runoff and water balance models with observed and projected climate data from different model output datasets. The HBV conceptual rainfall-runoff model was used to simulate the streamflow of the Lake Ziway gauged catchment. The amount of streamflow from ungauged catchments was calculated using area-ratio approaches. Unlike the earlier studies, which used only one station's data, evaporation was calculated using four stations' air temperature data. The use of various climate model outputs has aided in the estimation of lake area rainfall. This study is also relatively comprehensive because it employed the CROPWAT model to incorporate the irrigation water demand considering site-specific crops to calculate the water balance study area. Prior studies, however, relied on few specific crops with the highest water requirements.

It should be noted that the calculation of the lake's water balance components may be subject to much uncertainty. The first source of uncertainty is groundwater's contribution to the lake's water balance, which was overlooked in this study. Estimating lake evaporation and lake areal precipitation is another source of uncertainty. Finally, runoff simulation from gauged and ungauged catchments has inaccuracies. As a result, future studies should consider uncertainties in the water balance by combining the advantages of multiple methodologies in assessing each water balance component.

The influence of climate change on water balance and the water level was explored in this study utilizing outputs from different climate models for currently feasible scenarios and water resource development near Lake Ziway. Future research should consider water balance uncertainties by employing improved techniques, enhanced model calibration procedures, and more advanced bias correction algorithms.

The combined effect of climate change and irrigation water abstraction has been explored in this study under the RCP 4.5 and RCP 8.5 emission scenarios for 2021–2050 and 2051–2080 periods. Annual precipitation, mean annual maximum, and minimum temperatures over the Katar and Meki catchments are projected to increase in the two future periods. As a result, the average yearly potential evapotranspiration in the sub-catchments and across the lake will rise. Despite climate change-induced increases in lake level and volume, both RCP scenarios project a net decline in lake level and concomitant lake volume reduction in the future. From 2051 to 2080, the consequences of water abstraction may result in a 257.2 Mm3 loss in annual lake volume, a 77.5 cm drop in lake level, and a 25 km2 surface area decline. The combined effects of climate change and water use could result in a 25 cm drop in lake level, with annual reductions of 10 km2 surface area and 101 Mm3 volume. Therefore, strict monitoring procedures for water abstraction and proper lake ecosystem management policies should be in place for the Lake Ziway catchment. We suggest future studies to assess and evaluate various water management scenarios to mitigate the adverse effects of water withdrawal and inform integrated water resource management among all stakeholders.

We acknowledge the Ministry of Water and Energy, MoWE, and the National Meteorology Institute, NMI, for providing this study's necessary hydrology and climatic data. In addition, we recognize the central Rift Valley lakes basin development office for providing information during field data collection. Finally, the authors also gratefully acknowledge the Swedish Meteorological and Hydrological Institute (SMHI) for allowing the HBV model free of charge.

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

The authors declare there is no conflict.

Abraham
T.
,
Woldemicheal
A.
,
Muluneh
A.
&
Abate
B.
2018
Hydrological responses of climate change on Lake Ziway catchment, Central Rift Valley of Ethiopia
.
J. Earth Sci. Clim. Change
9
,
474
.
Allen
R. G.
,
Pereira
L. S.
,
Raes
D.
&
Smith
M.
1998
Crop Evapotranspiration: Guidelines for Computing Crop Water Requirements
.
Irrigation and Drainage Paper 56, Food and Agriculture Organization (FAO)
,
Rome
.
Ayenew T. 2007 Water management problems in the Ethiopian rift: Challenges for development. J. African Earth Sci. 48(2-3), 222–236. doi:10.1016/j.jafrearsci.2006.05.010
Collins
M.
,
Knutti
R.
,
Arblaster
J.
,
Dufresne
J. L.
,
Fichefet
T.
,
Friedlingstein
P.
,
Gao
X.
,
Gutowski
W.J.
,
Johns
T.
,
Krinner
G.
,
Shongwe
M.
,
Tebaldi
C.
,
Weaver
A.J.
&
Wehner
M.
2013
Long-term Climate Change: Projections, Commitments and Irreversibility
. In:
Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change
(T.F. Stocker, D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex & P.M. Midgley, eds.).
Cambridge University Press
,
Cambridge, UK and New York, NY
.
Dagnachew
L.
,
Zerihun
W.
,
Assefa
M.
,
Seyoum
M.
&
Tenalem
A.
2005
A Review of the Current Status and an Outline of a Future Management Plan for Lakes Abijata and Ziway, Report, Addis Ababa
.
The FDRE, Oromia Environmental Protection Authority
,
Ethiopia
.
Damtew
Y. T.
,
Verbeiren
B.
,
Awoke
A.
&
Triest
L.
2021
Satellite imageries and field data of macrophytes reveal a regime shift of a tropical lake (Lake Ziway, Ethiopia)
.
Water
13
,
396
.
https://doi.org/10.3390/w13040396
.
Daron
J. D.
2014
Regional Climate Messages: Southern Africa. Scientific Report From the CARIIA Adaptation at Scale in Semi-Arid Regions (ASSAR) Project
.
Ottawa
,
Canada
.
Degefu
W.
1987
Some Aspects of Meteorological Drought in Ethiopia
. In:
Ethiopia, drought and hunger in Africa: denying famine a future
(M.H. Glantz, ed.)
Cambridge University Press
,
Cambridge, UK
.
Dessie
M.
,
Verhoest
N. E.
,
Pauwels
V. R.
,
Adgo
E.
,
Deckers
J.
,
Poesen
J.
&
Nyssen
J.
2015
Water balance of a lake with floodplain buffering: Lake Tana, Blue Nile Basin, Ethiopia
.
J. Hydrol.
522
,
174
186
.
doi:10.1016/j.jhydrol.2014.12.049
.
Desta
H.
&
Lemma
B.
2017
SWAT-based hydrological assessment and characterization of Lake Ziway sub-catchments, Ethiopia
.
J. Hydrol.: Reg. Stud.
13
,
122
137
.
doi.10.1016/j.ejrh.2017.08.002
.
Desta
H.
,
Lemma
B.
&
Gebremariam
E.
2017
Identifying sustainability challenges on land and water uses: the case of Lake Ziway catchment, Ethiopia
.
Appl. Geogr.
88
,
130
143
.
doi:10.1016/j.apgeog.2017.09.005
.
Desta
H.
,
Lemma
B.
,
Stellmacher
T.
&
Gebremariam
E.
2020
Water use and management of Lake Ziway and its catchment, Ethiopia: the perception of experts vis-à-vis the latest state of research
.
Environ. Dev. Sustainable
,
0123456789
.
doi:10.1007/s10668-019-00359-8
.
Goshime
D. W.
,
Rafik
A.
,
Béatrice
L.
,
Florence
D.
&
Haile
A. T.
2019a
Impact of water abstraction on the water level of Lake Ziway, Ethiopia
.
Water Soc.
239
,
67
78
.
doi:10.2495/WS190071
.
Goshime
D. W.
,
Absi
R.
,
Haile
A. T.
,
Ledésert
B.
&
Rientjes
T.
2020
Bias-Corrected CHIRP satellite rainfall for water level simulation, Lake Ziway, Ethiopia
.
J. Hydrol. Eng.
doi:10.1061/(ASCE)HE.1943-5584.0001965
.
Goshime
D. W.
,
Haile
A. T.
,
Absi
R.
&
Ledésert
B.
2021
Impact of water resource development plan on water abstraction and water balance of Lake Ziway, Ethiopia
.
Sustainable Water Resour. Manage.
36
(
7
).
doi:10.1007/s40899-021-00516-w
.
Habib
E.
,
Haile
A. T.
,
Sazib
N.
,
Zhang
Y.
&
Rientjes
T.
2014
Effect of bias correction of satellite-rainfall estimates on runoff simulations at the source of the Upper Blue Nile
.
Remote Sens.
6
,
6688
6708
.
doi:10.3390/rs6076688
.
Hargreaves
G. H.
&
Allen
R. G.
1985
History and evaluation of Hargreaves evapotranspiration equation
.
J. Irrig. Drain. Eng.
,
53
63
.
doi:10.1061/(ASCE)0733-9437
.
Hayal
D.
,
Lemma
B.
,
Albert
G.
&
Stellmacher
T.
2015
Degradation of Lake Ziway, Ethiopia: a study of the environmental perceptions of school students
.
Lakes Reservoirs Res. Manage.
20
(
4
),
243
255
.
Herco
J.
,
Hengsdijk
H.
,
Dagnachew
L.
,
Tenalem
A.
,
Hellegers
P.
&
Spliethoff
P.
2007
Land and water resources assessment in the Ethiopian Central Rift Valley. Alterra-rapport
,
Alterra Wageningen
,
Wageningen
, p.
1587
.
IPCC
2014
Climate change 2014: synthesis report
. In:
Contribution of Working Groups I, II and III to the AR5 of the IPCC Core Writing Team
(
Pachauri
R. K.
&
Meyer
L. A.
, eds).
IPCC
,
Geneva
,
Switzerland
, pp.
151
.
Kahsay
K. D.
,
Pingale
S. M.
&
Hatiye
S. D.
2018
Impact of climate change on groundwater recharge and base flow in the sub-catchment of Tekeze basin, Ethiopia
.
Groundwater Sustainable Dev.
6
,
121
133
.
https://doi.org/10.1016/j.gsd.2017.12.002
.
Kumar
P.
,
Sarthi
P. P.
,
Kumar
S.
,
Barat
A.
&
Sinha
A. K.
2020
Evaluation of CORDEX-RCMS and their driving GCMs of CMIP5 in the simulation of Indian summer monsoon rainfall and its future projections
.
Arab. J. Geosci.
13
(
5
).
doi:10.1007/s12517-020
.
Lafon
T.
,
Dadson
S.
,
Buys
G.
&
Prudhomme
C.
2013
Bias correction of daily precipitation simulated by a regional climate model: a comparison of methods
.
Int. J. Climatol.
33
(
6
),
1367
1381
.
doi: 10.1002/joc.3518
.
Meehl
G. A.
,
Covey
C.
,
Delworth
T.
,
Latif
M.
,
McAvaney
B.
,
Mitchell
J. F.
,
Stouffer
R. J.
&
Taylor
K. E.
2007
The WCRP CMIP3 multimodel dataset: a new era in climate change research
.
Bull. Am. Meteorol. Soc.
88
(
9
),
1383
1394
.
Nigatu
Z. M.
,
Rientjes
T.
&
Haile
A. T.
2016
Hydrological impact assessment of climate change on Lake Tana's water balance, Ethiopia
.
Am. J. Clim. Change
5
,
27
37
.
Ogega
O. M.
,
Koske
J.
,
Kung'u
J. B.
,
Scoccimarro
E.
,
Endris
H. S.
&
Mistry
M. N.
2020
Heavy precipitation events over east Africa in a changing climate:results from CORDEX RCMs
.
Clim. Dyn.
55
(
3–4
),
993
1009
.
doi.10.1007/s00382-020-05309-z
.
Osima
S.
,
Indasi
V. S.
,
Zaroug
M.
,
Endris
H. S.
,
Gudoshava
M.
,
Misiani
H. O.
,
Anyah
R. O.
,
Otieno
G.
,
Ogwang
B. A.
,
Jain
S.
,
Kondowe
A. L.
,
Mwangi
E.
,
Lennard
C.
,
Nikulin
G.
&
Alessandro
D.
2018
The projected climate over the Greater Horn of Africa under 1.5 C and 2 C global warming
.
Environ. Res. Lett.
13
(
6
),
065004
.
Pratama
A. W.
,
Buono
A.
,
Hidayat
R.
&
Harsa
H.
2018
Bias correction of daily satellite rainfall data using genetic algorithm
. In:
IOP Conf. Ser.: Earth. Environ. Sci.
Vol.
149
.
Räisänen
J.
2007
How reliable are climate models?
Tellus A: Dyn. Meteorol. Oceanogr.
59
(
1
),
2
29
.
http://doi.org/10.1111/j.1600-0870.2006.00211.x
.
Randall
D. A.
,
Wood
R. A.
,
Bony
S.
,
Colman
R.
,
Fichefet
T.
,
Fyfe
J.
,
Kattsov
V.
,
Pitman
A.
,
Shukla
J.
,
Srinivasan
J.
,
Stouffer
R.
,
Sumi
A.
&
Taylor
K. E.
2007
Climate models and their evaluation
. In:
Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change
(
Solomon
S.
,
Qin
D.
,
Manning
M.
,
Chen
Z.
,
Marquis
M.
,
Averyt
K. B.
,
Tignor
M.
&
Miller
H. L.
, eds).
Cambridge University Press
,
Cambridge
,
United Kingdom
;
Rathjens
H.
,
Bieger
K.
,
Srinivasan
R.
,
Chaubey
I.
&
Arnold
J. G.
2016
CMhyd User Manual
.
Renard
B.
,
Kavetski
B.
,
Kucezera
G.
,
Thyer
M.
&
Franks
S. W.
2010
Understanding predictive uncertainty in hydrologic modeling: the challenge of identifying input and structural errors
.
Water Resour. Res.
46
,
W05521
.
Riahi
K.
,
Rao
S.
,
Krey
V.
,
Cho
C.
,
Chirkov
V.
,
Fischer
G.
,
Kindermann
G.
,
Nakicenovic
N.
&
Rafaj
P.
2011
RCP 8.5-A scenario of comparatively high greenhouse gas emissions
.
Clim. Change
109
(
1
),
33
57
.
Seyoum
W. M.
,
Milewski
A. M.
&
Durham
M. C.
2015
Understanding the relative impacts of natural processes and human activities on the hydrology of the Central Rift Valley lakes, Eastern Africa
.
Hydrol. Processes
29
(
19
),
4312
4324
.
doi.10.1002/hyp.10490
.
Sun
Q.
,
Miao
C.
,
Duan
Q.
,
Ashouri
H.
,
Sorooshian
S.
&
Hsu
K.-L.
2018
A review of global precipitation data sets: data sources, estimation, and intercomparisons
.
Rev. Geophys.
56
,
79
107
.
https://doi.org/10.1002/2017RG000574
.
Tekleab
S.
,
Mohamed
Y.
&
Uhlenbrook
S.
2013
Hydro-climatic trends in the Abay Upper Blue Nile basin, Ethiopia
.
Phys. Chem. Earth Parts A/B/C
61
,
32
42
.
Tenalem
A.
,
Becht
R.
,
van-Lieshour
A.
,
Gebreegziabher
Y.
,
Legesse
D.
&
Onyando
J.
2007
Hydrodynamics of topographically closed lakes in the Ethio-Kenyan Rift: the case of lakes Awassa and Naivasha
.
J. Spatial Hydrol.
7
(
1
),
81
100
.
Wale
A.
,
Rientjes
T. H. M.
,
Gieske
A. S. M.
&
Getachew
H. A.
2009
Ungauged catchment contributions to Lake Tana's water balance
.
Hydrol. Processes
23
(
26
),
3682
3693
.
Zeray
L.
,
Roehrig
J.
&
Chekol
D. A.
2006
Climate change impact on Lake Ziway water availability, Ethiopia
. In:
Conference on Inter. Agricultural Research for Development
. pp.
1
4
.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY 4.0), which permits copying, adaptation and redistribution, provided the original work is properly cited (http://creativecommons.org/licenses/by/4.0/).