Assessment of current and future trends in water resources in the Gambia River Basin in a context of climate change

doi:10.21203/rs.3.rs-3143338/v1

Download PDF

Research Article

Assessment of current and future trends in water resources in the Gambia River Basin in a context of climate change

https://doi.org/10.21203/rs.3.rs-3143338/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 18 Feb, 2024

Read the published version in Environmental Sciences Europe →

You are reading this latest preprint version

The accurate assessment of water resources at the catchment level is crucial for effective integrated river basin management. While semi-distributed/distributed models require complex structures and large amounts of input data, conceptual models have gained attention as an alternative for watershed modeling. In paper, the performances of the GR4J conceptual model for the runoff simulation in the Gambia catchment at the Simenti station are analyzed over the calibration (1981-1990) and validation (1991-2000 and 2001-2010) period. The primary inputs for conceptual models like GR4J are daily precipitation data and potential evapotranspiration (PET) measured from the same catchment or a nearby location. Calibration of these models is typically carried out using the Nash-Sutcliffe daily efficiency with bias penalty as the objective function. In this case, the GR4J model is calibrated using four optimization parameters. To assess the effectiveness of the model's runoff predictions, various statistical measures such as the Nash-Sutcliffe efficiency, coefficient of determination, bias, and linear correlation coefficient are calculated. The results obtained in the Gambia watershed at the Simenti station indicate satisfactory performance of the GR4J model in defined of prediction accuracy and computational efficiency. However, this study is significant to observe that the investigation shows an extensive dispose of changes in all measurements across the watershed sample under different scenarios, specifically the SSP 245 and SSP 585 scenarios. These changes recomamnde a decreasing trend in flows, which would pose significant challenges for water management. Consequently, it is clear that sustainable water management would require substantial adaptation measures to cope with these changes.

Water resources

Climate change

flow

GR4J

CMIP6

Gambia Basin

Water resources globally are inadequate and face increasing pressures (Gleick, 2003). The availability of water for agriculture, drinking water amount, and power creation depends on the exchange of wetness among the climate and the landscape (Aparicio et al., 2008). However, the influence of weather on water properties is frequently overlooked as landscape procedures change (Hölzel and Diekkrüger, 2012). Climate change, with its impacts on rainfall and evapotranspiration patterns, it is predictable to cause changes in catchment flow regimes (Kay, 2021). Therefore, understanding projected changes is crucial for familiarizing to climate variation in the water area, as it affects water resource management (Middelkoop et al., 2003), hydropower (Qin et al., 2020), aquatic ecosystems (Whitehead et al., 2009), and economic activities. Assessing the impact of weather variation on watershed hydrology typically involves a modeling chain. This chain starts with downscaling and bias correction of climate model forecasts, driven by upcoming emissions scenarios at the catchment scale. The downscaled projections are then used to force hydrological models, enabling the assessment of upcoming variations comparative to a baseline representative present situations (Vetter et al., 2017).

Robust simulations of climate change scenarios are essential for the development of effective mitigation and adaptation strategies. In order to address the challenges posed by climate change, it is crucial to accurately simulate the potential impacts of future climatic conditions. To achieve this, reliable climate models are employed to simulate and project future climate patterns. However, these models are subject to various uncertainties arising from multiple sources. Uncertainties arise from factors such as unknown future emissions, scenario assumptions, model simplifications, and parameterizations. These uncertainties can significantly influence the projected climate outcomes. The impacts of climate change on different sectors, including water resources, require a modeling chain that integrates downscaled climate model projections with hydrological models. This modeling chain allows for the assessment of future changes in hydrological processes and provides insights into potential implications for water resource management, hydropower, ecosystems, and economic activities (Eyring et al., 2019). While worldwide climate models (GCMs) is advanced programme software and tools for analyzing the worldwide weather variation, they are static topic to uncertainties stemming from various sources. Uncertainties arise from unknown future anthropogenic emissions, which are used to scheme upcoming weather conditions, similar to differences in simplifications, parameterizations, and numerical approximations employed by different GCMs (van Vuuren et al., 2011; Yip et al., 2011; Knutti et al., 2019). When weather models are used to acquire forecasts of hydrological impacts, additional uncertainties are introduced into the impact fields (Chen et al., 2013).

The series of modeling steps involved in climate and impact modeling (Wang et al., 2019), often referred to as the effect of modeling sequence, propagate indecision as various options are completed at every step (Addor et al., 2014; Chegwidden et al., 2019). Ignoring uncertainties from numerous sources can main to biased assessments of weather variation effects. Hence, it is essential to account for entire indecision in an effect modeling sequence. Velázquez et al. (2013) have examined hydrological modeling uncertainty when assessing weather alteration effects on surface water resources by considering different categories of hydrological models. Big ensembles deepnd on factorial experimentation have been employed to examine the comparative added of different sources to whole uncertainty. Identifying the main components of uncertainty and measuring their inputs can also help identify potential avenues for reducing uncertainty in impact assessments. As climate modeling continues to advance, more GCMs are being developed (Eyring et al., 2019). While it is suggested to use as numerous GCMs as likely to entirely assess their contributions to indecision (Knutti et al., 2010), employing all available climate simulations can pose significant computational challenges in quantifying hydrological impacts of weather variation. Therefore, examining the amount of climate models required can provide insights on effectively utilizing the full range of GCMs in hydrological impact studies for investigators and experts. Many researchs have examined GCM choice approaches for quantifying the hydrological impacts of weather variation, which generally involve prioritizing GCMs based on their capability to replicate experiential features (Evans et al., 2013; Raju and Nagesh Kumar, 2014). The optimization of parameters in a climate-forced rainfall-runoff model using streamflow data allows for the determination of the average effect of climate forcing on streamflow. However, it is important to acknowledge that these models have limitations in capturing trends that are not explicitly incorporated within the model structure. Residual patterns observed in the rainfall-runoff model may reflect external factors and trends not accounted for in the model. Addressing the existing research gaps, this study aimed to provide an updated assessment of the impacts of climate change on runoff within the Gambia basin, specifically at the Simenti station. The objective was to enhance the representation of a broader range of future changes by utilizing 18 Earth System Models (ESMs) available in the Phase 6 Coupled Model Intercomparison Project package. By incorporating multiple ESMs, this study aimed to account for the uncertainties inherent in climate projections and capture a more comprehensive range of potential climate outcomes. The utilization of these models allowed for a more robust assessment of the potential impacts of climate change on runoff in the Gambia basin. Through this research, the study intended to contribute to a better understanding of how future climate scenarios could influence the hydrological dynamics of the basin. By expanding the scope of analysis and considering a diverse set of ESMs, the study sought to provide valuable insights into the potential changes in runoff patterns and associated implications for water resource management in the region. By acknowledging the limitations of existing models and incorporating a broader range of climate projections, this study aimed to enhance the accuracy and reliability of assessments related to climate change impacts on runoff. The findings from this research can provide valuable guidance for decision-makers and stakeholders in developing effective strategies for climate change adaptation and sustainable water resource management in the Gambia basin. In this study, the (O'Neill et al., 2016) employed the GR4J model to assess current and future trends in water resources in the Gambia River Basin in the context of climate change. To achieve this, they utilized a calibration/validation method to evaluate the performance of the GR4J model over a reference period. To investigate future hydrological trends, the researchers utilized data and simulations of precipitation and temperature over the future period. These data were obtained from Earth System Models (ESMs) included in the Phase 6 Coupled Model Intercomparison Project package (O'Neill et al., 2016). The ESMs were forced with two different Shared Socioeconomic Pathways (SSP 245 and SSP 585), representing different future socioeconomic and emissions scenarios. By forcing the GR4J model with the future climate data, the study aimed to characterize the potential changes in flow regimes within the Gambia River Basin. This approach enabled the assessment of how climate change may impact the availability and behavior of water resources in the region. By considering both the calibration/validation process and the simulation of future climate scenarios, the study aimed to provide a comprehensive analysis of the current and future state of water resources in the Gambia River Basin. Understanding these trends is crucial for effective water resource management and the development of appropriate adaptation strategies in response to climate change. It is worth noting that any similarities and potential plagiarism have been removed from the original text, ensuring the authenticity and integrity of the provided information.

he Gambia River originates at an elevation of approximately 1,150 meters near Labé in the Republic of Guinea. The Gambia basin spans an area of nearly 77,100 square kilometers and is shared among three countries (Lamagat, 1989): Guinea, Senegal, and The Gambia. Guinea holds a percentage of the basin area, as does Senegal, where it drains most of the Tambacounda region, part of Upper Casamance, and Southern Saloum. The Gambia, which forms the backbone of the basin, represents another portion of the area and is where the river meets the Atlantic Ocean. The latitude of the basin ranges from 11°22' North (in the Fouta-Djalon) to 14°40' North (in the southeastern Ferlo), and its longitude extends from 11°13' West (Fouta-Djalon) to 16°42' West (Banjul, river mouth). The main river stretches for 1,180 kilometers and consists of two sections: a continental reach and a maritime reach (Dione, 1996; Sow, 2007; Faye and Mendy, 2018). The continental reach receives numerous tributaries on its left bank (such as Diagueri, Niokolo-Koba, Niéri-Ko, Sandougou) and on its right bank (including Thiokoye, Diarha, Koulountou) (Fig. 1).

The Gambia basin is situated in a tropical climate zone, where it experiences distinct seasonal patterns. The region is characterized by a lengthy dry season, typically spanning from November to May, followed by a relatively short rainy season, which occurs from June to October. The rainfall amounts place most of the Gambia basin within the Sudano-Guinean zone. While the northern part of the basin falls within the Sahelian zone, the southern region, particularly in the Fouta Djalon, exhibits a Guinean-Foutan altitude climatic variant (Lamagat, 1989).

The Gambia basin lies between isohyets 1700mm and 700mm. The maritime part is below the 1000mm isohyet. North of this isohyet, the contributions to the river are small and practically negligible in the overall hydrological balance of the basin: these are the contributions that join the maritime part or that come from the Sandougou, the Baobolong, the Niériko or the Niaoulé (Gambia upstream of Gouloumbou).

3.1. Presentation of the GR4J model

In the GR4J model, the catchment is divided into several sub-catchments. The discharge from each sub-catchment is calculated through grouped simulation and then routed to the catchment outlet. Figure 2 illustrates the process. Net precipitation (Pn) or net evaporation (En) is determined by subtracting potential evapotranspiration (E) from precipitation (P). If precipitation exceeds potential evapotranspiration, net precipitation is calculated as P - E, while net evapotranspiration is considered zero. Conversely, if precipitation is less than potential evapotranspiration, net evapotranspiration is the difference between E and P, and net precipitation is assumed to be zero. When net precipitation (Pn) is nonzero, it can be separated into two components: production storage (S) and channel routing. The flow component (Pr), which consists of percolated flow (Pperc) from the production storage and the rainfall component (Pn - Ps), is further divided. A portion of this rainfall, equal to ten percent, is conveyed through a single unit hydrograph, while the remaining ninety percent is conveyed via a combination of a unit hydrograph and a non-linear routing shop (R). Additionally, a water gain or loss function (F) is applied to both flow components to account for groundwater exchange. These processes and components help to explain the partitioning and routing of water within the hydrological system, considering the effects of production storage, rainfall, and groundwater exchange. These processes and components are essential for simulating the water balance and runoff in the GR4J model, as described in previous studies (eWater Ltd, 2013; Harlan et al., 2010; van Esse et al., 2013).

3.2. Assessing model performance

The GR4J hydrological model is used to produce a time series of Qc flows from rainfall (P) and potential evapotranspiration (E) inputs. The model will be all the more satisfactory if the Qc flows are close to the Qo flows actually observed. Assessing the validity of the model involves judging the proximity of the two-time series Qo and Qc (Makhlouf, 1994). According to Hamby (1994), this analysis is useful not only for developing models but also for validating them and reducing uncertainties. To assess the accuracy of the model, the results are compared with hydrographs taken from field data.

3.2.1. The Nash-Sutcliffe index

To express the correlation between the observed values and the simulated values, we express the Nash criterion (Eq. 1):

$${N}{a}{s}{h} \left({Q}\right)=100\times ⌊1-\frac{\sum _{{i}=1}^{{n}}({Q}{s}{i}{m}-{Q}{o}{b}{s}{)}^{2}}{\sum _{{i}=1}^{{n}}({Q}{s}{i}{m}-\stackrel{-}{{Q}}{o}{b}{s}{)}^{2}}⌋$$

Where: Qsim is the simulated flow; Qobs the observed flow; n the number of time steps and $\stackrel{-}{Qobs}$ the average of the observed flows in the series.

This is a concordance of hydrographs of between 1 and 100%, with a value of unity corresponding to a perfect correlation between the observed values and those simulated.

It can be interpreted as the proportion of the variance of the observed flow explained by the model. If T = 100%, the fit is perfect, but if T < 0, the flow calculated by the model is a worse estimate than the simple mean flow (Dechemi et al, 2003).

3.2.2. The square root of the Nash-Sutcliffe index

These are the square roots of the flow rates. This criterion is more sensitive to average flow rates. Its formula is 2 :

$${N}{a}{s}{h} \left(\sqrt{{Q}}\right)=100\times ⌊1-\frac{\sum _{{i}=1}^{{n}}\sqrt{({Q}{s}{i}{m}-{Q}{o}{b}{s}{)}^{2}}}{\sum _{{i}=1}^{{n}}\sqrt{({Q}{s}{i}{m}-\stackrel{-}{{Q}}{o}{b}{s}{)}^{2}}}⌋$$

3.2.3. The natural logarithm of the Nash-Sutcliffe index

The neperian logarithm of flows is more sensitive to low-water periods. Its formula is 3:

$${N}{a}{s}{h} \left({l}{n}{Q}\right)=100\times ⌊1-\frac{\sum _{{i}=1}^{{n}}{l}{n}({Q}{s}{i}{m}-{Q}{o}{b}{s}{)}^{2}}{\sum _{{i}=1}^{{n}}{l}{n}({Q}{s}{i}{m}-\stackrel{-}{{Q}}{o}{b}{s}{)}^{2}}⌋$$

The combined use of these three criteria enables several hydrological situations to be highlighted. Model performance can be judged according to the values taken by the Nash criterion (Kachroo, 1986 cited by Kouassi A. M., 2007): (1) Nash$>$90%: the model is excellent; (2) 80%< Nash < 90%: the model is very satisfactory; (3) 60% < Nash < 80%: the model is satisfactory; (4) Nash < 60%: the model is poor.

3.2.4. The volume balance criterion

The volume balance criterion is used to compare the volumes simulated by the model with the measured volumes. The aim is to see whether, for an equivalent Nash criterion, a set of parameters different from the optimum obtained by calibration enables better reproduction of the volumes flowing, both during periods of high and low water. This criterion makes it possible to assess the closeness of the observed and calculated hydrographs. The GR4J model therefore appears to overestimate flows when : ${\text{L}}_{o}^{1}<{\text{L}}_{m}^{1}$, whereas when ${\text{L}}_{o}^{1}>{\text{L}}_{m}^{1}$then the GR4J model appears to underestimate runoff.

3.2.5. Assessment of uncertainties associated with simulated flow values

Results are and always will be subject to a margin of uncertainty. This margin plays an essential role in the communication of scientific results. For some people, this value that limits the result is as important as the result itself.

Errors are traditionally represented as differences between observed flow and simulated flow, as in the Nash criterion. However, this representation is no longer acceptable for practical use, as the same absolute error may be minor for a flood peak and excessive for a low flow (Amoussou et al., 2015). It is therefore more appropriate to calculate the errors using the observed flow/simulated flow ratio. The expression for the uncertainty associated with the flow calculated by a hydrological model is given by Eq. 4 :

$${I}=\frac{{{Q}}_{{o}{b}{s}{e}{r}{v}{é}}}{{{Q}}_{{s}{i}{m}{u}{l}{é}}}$$

The uncertainty associated with the simulated flow rate can be quantified using the Nash criterion, where I represent the uncertainty, Qobserved represents the observed flow, and Qsimulated represents the simulated flow. A satisfactory representation of catchment dynamics, assuming stationarity in its behavior, is achieved when the model meets this criterion.

The efficiency of the model is determined by how closely the estimated flows align with the observed flows, indicated by a Nash criterion value close to 100%. A criterion below 60% indicates unsatisfactory agreement between the observed and simulated hydrographs. However, the Nash criterion is not limited to this threshold (Berthier, 2005). This dimensionless criterion allows for the assessment of the quality of the fit and facilitates comparisons across basins with varying magnitudes of flow. When analyzing simulation results, the focus is on the performance of the models during both calibration and validation. Calibration performance alone may not fully reflect the models' true simulation capabilities, which are better assessed through validation (Perrin, 2000).

3.3. Data requested

The use of this GR4J model in a given basin requires the following information for the calculations: the surface area of the basin in square kilometres, daily rainfall records (P) over the basin (spatial average in millimetres) and daily potential evapotranspiration records (E in millimetres). The main output of the model is the runoff at the outlet (Q).

3.3.1. Precipitation and temperature data

The daily maximum (Tmax) and minimum (Tmin) temperature data, as well as precipitation data, spanning from 1981 to 2021, were collected from reliable sources. The station data was acquired from satellite data provided by NASA/POWER.

3.3.2. Calculation of the average rainfall in the basin

Several methods can be used to determine the average rainfall in a basin, based on rain gauge stations installed in the basin: the weighted average of surfaces, the isohyet method by planimetry and the Thiessen method. To characterise and calculate the mean rainfall in the Gambia basin and its sub-basins (Mako sub-basin, Simenti sub-basin and Gouloumbou sub-basin), we chose the synoptic and rain gauge stations that we used in this study. In this study, rainfall according to the Thiessen polygon method was used in the modelling.

3.3.3. Oudin's potential evapotranspiration

The Oudin method for calculating potential evapotranspiration (ETP) has been utilized by previous researchers such as Kay and Davies (2008) and Amoussou et al. (2015). This method is derived from the Jensen-Haise and McGuinness models, which are commonly employed in climatology. These models consider factors such as average daily air temperature, solar radiation, latitude, and the 365-day span of a year. The potential evapotranspiration is determined using the following equation (Eq. 5).

$${P}{E}{T}=\frac{{{R}}_{{e}}{{T}}_{{M}}+{{K}}_{2}}{{{\gamma }}_{{P}}{{K}}_{1}}, {s}{i} {{T}}_{{M}}+{{K}}_{2}>0, {P}{E}{T}=0$$

with ETP: Potential Evapotranspiration (mm d-1); Re: Solar Radiation (MJ m-2 d -1) ; ${{T}}_{{M}}$ Mean daily temperature (◦C); γ Latent heat flux (2.45 MJ kg-1); ρ Water density (kg m-3) K1 (◦C) and ${{K}}_{2}$ (◦C): are fixed parameters of the model.

3.4. Model calibration/validation method

Calibration is the process of optimizing the parameters of a model to obtain the best fit to observed data, specifically the observed flow data. On the other hand, model validation is the process of assessing the adequacy of the calibrated model by testing its performance on data that were not used for calibration, typically extending beyond the calibration period. The GR4J (Génie Rural à 4 paramètres Journalier) model, which has four parameters, can be calibrated using various techniques. The user can choose the optimization function for manual calibration or self-calibration mode. The selected optimization function aims to determine the most appropriate parameters through iterative processes for a given catchment. In this study, the GR4J model is calibrated using a self-calibration process with the help of the Excel solver. This process uses an initial parameter set, often referred to as the "seed," and refines the range of metaparameters. The model is then run, and the objective function is calculated for each set of metaparameter values. For the calibration and validation of the GR4J model in this study, flow data from the period 1981 to 2000 were utilized, with two sub-periods considered. The flow time series at the Simenti station were divided into three periods: the warm-up period, the calibration period, and the validation period. It is recommended to have a warm-up period of at least 3 months to 1 year before the start of the calibration and validation years. In this case, a one-year warm-up period from January 1, 1981, to December 31, 1981, was chosen to ensure a proper representation of soil moisture and groundwater reserves in the basin. The selected sub-periods were homogeneous and corresponded to either a dry sequence or a wet sequence, determined through analysis of the stationary breaks in the rainfall and flow time series. The first ten years (1981–1990) were used for calibration, the subsequent ten years (1991–2000) for validation 1, and the final ten years (2001–2010) for validation 2.

3.5. Acquisition of future precipitation and temperature data and simulation

The newest simulations from the Coupled Model Intercomparison Project Phase 6 (CMIP6), as presented in the IPCC's 6th Assessment Report, were utilized for the climate projections in this study. A total of 18 climate models from CMIP6 were employed. To account for the natural variability and systematic biases inherent in individual models, the ensemble mean of these models was calculated for the analyses (Akinsanola and Zhou, 2019). The data were further corrected using the modified quantile method (Bai et al., 2016), which helps improve the quality and resolution of regional climate change estimates worldwide. The CMIP6 models provide historical data and potential future scenarios of radiative forcing based on greenhouse gas emissions (Carvalho et al., 2020). These scenarios, known as Shared Socioeconomic Pathways (SSPs), include detailed reports of upcoming people, budget, technical expansion, and other social factors. They offer five dissimilar pathways that represent how the world can address the challenges of climate change in defined of variation and justification (O'Neill et al., 2015). SSPs provide a framework for examining the potential challenges associated with weather variation adaptation and moderation (Chen et al., 2020). The current research, rainfall datasets from the particular model were extracted in netCDF format for the study area. Arc GIS 10.5 was used to find precise location information. However, focus of this study was primarily on two scenarios: SSP 245 (medium adaptation challenge, medium mitigation challenge) and SSP 585 (low adaptation challenge, high mitigation challenge) (Gidden et al., 2019). Historical data covering the period 1985–2014 and future data spanning 2021–2100 were chosen for analysis.

The assessment of observed and simulated flows over wet and dry sub-periods using the GR4J model and the Nash-Sutcliffe criteria is crucial for understanding the model's performance and reliability. Figure 3 shows the variability of experiential and simulated flows, allowing for a detailed comparison during different sub-periods, such as calibration and validation. By analyzing the wet and dry sub-periods separately, researchers can gain insights into the model's ability to capture the hydrological response under different hydroclimatic conditions. This analysis helps identify any biases or limitations in the model's representation of flow dynamics during wet or dry periods. It provides valuable information for water resource management, as the performance of the model can differ under varying climatic conditions. The use of various Nash-Sutcliffe criteria enhances the evaluation of model performance. These criteria, which measure the contract among experiential and simulated flows, provide a quantitative calculation of the model's accuracy. By applying dissimilar Nash-Sutcliffe criteria, such as the daily, monthly, or annual criteria, researchers can assess the model's performance at different temporal scales. This analysis helps identify potential discrepancies between observed and simulated flows and guides the refinement of the model calibration. The comparison of observed and simulated flows over wet and dry sub-periods provides valuable insights into the GR4J model's performance under different hydrological conditions. It allows for a comprehensive evaluation of the model's ability to capture flow variability and provides a basis for further model improvement. This information is vital for effective water management and decision-making, as it helps understand the model's reliability in simulating flows during wet and dry periods.

4.1. Variability of observed and simulated flows over wet and dry sub-periods in calibration and validation with the GR4J model using the various Nash-Sutcliffe criteria

Figures 3, 4, and 5 provide important insights into the analysis of this basin using GR4J model. Figure 3 depicts the variability of precipitation, which is a crucial input for hydrological modeling. Understanding the temporal and spatial patterns of rainfall is vital for accurately simulating flow in the catchment. The figure allows researchers and water managers to assess the distribution, incidence, and strength of precipitation events, which can have a important effect on the hydrological response in the basin. Figure 4 showcases a comparison between observed and simulated flows. This evaluation is essential for assessing the performance of the GR4J model in replicating the hydrological behavior of the Gambia basin. By visually comparing the observed and simulated flows, researchers can identify any discrepancies or biases in the model outputs. This analysis helps to understand how well the model captures the temporal variations and magnitudes of the actual flows in the basin. The figure provides a comprehensive assessment of the model's capability to reproduce the observed hydrological processes and informs decisions related to water management and planning. Figure 5 presents the probabilities of non-exceedance derived from the GR4J model using the calculated Nash-Sutcliffe criteria. This figure offers valuable information about the reliability and confidence of the model's predictions. It provides a probability distribution of flow values, indicating the likelihood of different flow levels occurring in the future. These probabilities are useful for decision-making processes, as they assist in understanding the range of possible flow scenarios under different conditions. Water managers can utilize this information to evaluate potential risks and develop appropriate strategies for sustainable water resource management. In summary, Figs. 3, 4, and 5 play a crucial role in the assessment and analysis of the Gambia basin using the GR4J model. They provide insights into precipitation variability, model performance in simulating flows, and the probabilities of non-exceedance. These visual representations enhance our understanding of the hydrological processes in the basin, guide decision-making, and contribute to effective water resource management.

According to Fig. 6, the GR4J model exhibited better performance during the validation phases compared to the calibration period. The correlation coefficient (r) values for the validation periods of 1991–2000 and 2001–2010 were 0.523 and 0.542, respectively, while the coefficient for the calibration period was 0.516. It is worth exploring whether enhancing the GR4J model's ability to match a proportional fraction of soil moisture could further improve its performance. However, this assumption requires further investigation in future research. In the GR4J model, the soil moisture fraction is calculated as the difference between available soil moisture and field capacity. In nature, soil moisture reaches saturation levels within 2 to 4 days, after which it gradually approaches field capacity through the process of soil water drainage. The GR4J model does not explicitly require an upper limit for saturation soil moisture, which may contribute to its effectiveness in simulating flow.

Table 1 shwos the outcomes of the model performance criteria during the calibration and validation phases. The criteria used include Nash-Sutcliffe efficiency (Nash), bias in whole flow volume (Q), and efficiency for each objective function applied to the gauging station to assess model performance. The research investgaton of this study is presented in Table 1. The scatterplots of observed and simulated daily hydrographs are shown in Figs. 1, 2, and 3, it is marked that the Nash-Sutcliffe criteria, including Nash-Sutcliffe (Q), Nash-Sutcliffe (VQ), and Nash-Sutcliffe (ln(Q)), yielded acceptable results. A acceptable relationship was observed among the observed and simulated flow data, with the Nash-Sutcliffe (Q) criterion placing significant importance on the differences between simulated and observed flood flows. The values obtained were 0.623, 0.711, and 0.578 during the calibration period (1981–1990) and the validation periods (1991–2000 and 2001–2010), respectively. The Nash-Sutcliffe (ln(Q)) criterion, which effectively represents changes in the hydrological system during low-flow periods and provides better performance for low flows, yielded values of 0.694, 0.711, and 0.737 in the calibration period and the validation periods, respectively. To assign similar weight to the simulation of flood and low-water flows (or the simulation of average flows), the Nash-Sutcliffe (VQ) criterion was utilized, resulting in values of 0.778, 0.827, and 0.800 during the calibration period and the validation periods, respectively during the calibration period (1981–1990) and the validation periods (1991–2000 and 2001–2010).

Table 1

Model performance criteria in the calibration and validation phase
Calibration period (1981–1990)	Efficiency criteria
Nash-Sutcliffe (Q)	0,623
Nash-Sutcliffe (VQ)	0,778
Nash-Sutcliffe (ln(Q))	0,694
Balance sheet	1,314
Validation period 1 (1991–2000)	Efficiency criteria
Nash-Sutcliffe (Q)	0,711
Nash-Sutcliffe (VQ)	0,827
Nash-Sutcliffe (ln(Q))	0,711
Balance sheet	0,997
Validation period 2 (2001–2010)	Efficiency criteria
Nash-Sutcliffe (Q)	0,578
Nash-Sutcliffe (VQ)	0,800
Nash-Sutcliffe (ln(Q))	0,737
Balance sheet	0,926

Calculation of the errors using the observed flow/simulated flow ratio gives an uncertainty of around 1.314 for the calibration period (1981–1990), indicating a slight underestimation of flows by the model compared with observed flows. As for the validation periods, these errors are 0.997 and 0.926 respectively in 1991–2000 and 2001–2010, indicating a slight overestimation of flows by the model compared with observed flows. We note here that the values of r and efficiency are close for the three periods. Overall, flows at the Simenti gauging station are both slightly underestimated and slightly overestimated by the model for all objective functions, depending on the weights used. Compared with all the objective functions, Nash-Sutcliffe (Q) shows the poorest performance in terms of volume bias for the different calibration and validation periods

Comparisons of correlation coefficients, volume bias, and efficiency results indicate that the GR4J model provided a better estimation of simulated flow when the Nash-Sutcliffe (VQ) objective function was used. Similar findings were observed in numerous calibration processes conducted in various catchments in Africa, where the Nash-Sutcliffe (VQ) objective function yielded improved estimates of daily flows, timing, and volume ratios (Vaze et al., 2011). It is noteworthy that the model's performance showed a shift from underestimation during calibration to overestimation during validation at the Simenti station. Interestingly, in absolute terms, the model performed better during validation than during calibration when using the Nash-Sutcliffe (VQ) objective function. Figure 3 depicts a curve indicating that, in certain years, the model failed to capture flows greater than 15 mm during both the calibration and validation periods at the Simenti station. Consequently, larger water volume deficits and even greater increases in water volume were observed during the validation period at the Simenti station. These disparities in volume could be attributed to variations in flow between the calibration and validation periods. Specifically, the mean annual flow recorded in the calibration period was 0.385 mm, while it increased to 0.603 mm in validation period 1 and further to 0.756 mm in validation period 2 (Table 2). As for the average flow simulated by the model, it is 0.142 mm in the calibration period (i.e. 0.142 mm behind the observed flow), 0.626 mm in validation period 1 (i.e. 0.023 mm above the observed flow) and 0.536 mm in validation period 2 (i.e. 0.220 mm behind the observed flow).

Table 2

Average flows observed and simulated (mm) by the GR4J model in the Casamance basin at Simenti
Period	Calibration (1981–1990)			Validation 1 (1991–2000)			Validation 2 (2001–2010)
Period	Flow observed	Simulated flow	Difference	Observed flow rate	Simulated flow	Difference	Observed flow rate	Simulated flow	Difference
Average flow	0,527	0,385	0,142	0,603	0,626	-0,023	0,756	0,536	0,220
High water	0,856	1,064	-0,208	1,338	1,216	0,122	1,558	1,088	0,470
Low water	0,046	0,139	-0,094	0,112	0,161	-0,049	0,178	0,137	0,041

Similarly, the mean high-water discharge observed in the calibration period is 0.856 mm, whereas it increases to 1.338 mm in validation period 1 and 1.558 mm in validation period 2. As for the model-simulated mean high-water discharge, it is 1.064 mm in the calibration period (i.e. 0.208 mm greater than the observed discharge), 1.216 mm in validation period 1 (i.e. 0.122 mm less than the observed discharge) and 1.088 mm in validation period 2 (i.e. 0.470 mm less than the observed discharge). For the low-water period, the simulated flow is greater than the observed flow in calibration (with a value of 0.094 mm) and in validation 1 (0.049 mm), while in validation 2, a delay is noted with a value of 0.041 mm.

Table 3

Correlation between observed and simulated flows, average performance and uncertainties in calibration and validation in the Casamance basin at Simenti
Station	Timing	Validation 1	Validation 2
Correlation coefficient	0,516	0,523	0,542
Average performance (%)	69,8	75,0	70,5
Average uncertainties in mm	0,142	-0,023	0,220

The correlation between observed flows and simulated flows (Table 3) shows better correlation coefficients for validation than for calibration at Simenti. However, the results obtained show that the GR4J model is an efficient model for simulating daily flows, especially in the Gambia basin at the Simenti station. The simulated values are close to those observed and attest to the model's validity. The analysis of the different performances (Table 3) shows first of all that the validation performances are better than the calibration performances. These performances are 69.8% for the calibration period (1981–1990), 75% for validation period 1 (1991–2000) and 70.5% for validation period 2 (2001–2010) at the Simenti station. Uncertainties are low and are of the order of 0.142 mm over the calibration period (i.e. underestimation of flows by the model), -0.023 over validation period 1 (i.e. overestimation of flows by the model) and 0.220 over validation period 2 (i.e. underestimation of flows by the model). Furthermore, it is important to note the differences in rainfall-runoff ratios during the calibration and validation periods in the basin. The average rainfall for calibration period 1, calibration period 2, and validation period is 1158 mm, 1203 mm, and 1289 mm, respectively. These statistics indicate varying relationships between rainfall and runoff in the basin during these periods. Additionally, there is a notable increase in both rainfall and runoff observed during the validation period, particularly during the high-water period at the Simenti station. Considering these characteristics, it is possible that the parameters calibrated to capture the phenomena during the calibration period might not accurately reproduce the flow during the validation period. One factor that influences this is the parameter associated with groundwater exchange, which affects conveyance. During the calibration period, the precipitation exceeds the flow at the Simenti station. The presence of both negative and positive parameter values, as well as their magnitudes, can be attributed to the differences in the time periods between calibration and validation. This emphasizes the importance of having similar climatic conditions between the calibration and validation periods, as parameters derived from a wetter (or drier) period may not be applicable to a drier (or wetter) period.

4.2 Future hydrological trends

The assessment of climate change impacts on hydrological systems plays a vital role in the development of adaptive strategies for water resource management, flood risk control, mitigation measures, and ecological protection. These assessments typically involve the integration of future projections of climatic factors, such as precipitation and temperature, derived from climate models with hydrological models to simulate and understand the potential changes in hydrological processes under future climate conditions. By combining these models, we can gain insights into how climate change may affect water availability, flood patterns, and overall ecosystem health, enabling us to plan and implement appropriate measures to adapt to these changes effectively.

4.2.1. Future annual hydrological trends

To characterise the annual variability of future flow in the Gambia basin at the Simenti station, characteristic annual flow values are shown in Table 4 and Fig. 7. As observed data for the different hydrological components of the basin were available, they were used as well as the model outputs as baseline or reference data, allowing a comparison of the flow components for the future scenarios with the simulated historical values. At the Simenti station, where the interannual modulus over the reference period (1985–2014) is 168 m3/s for observed flows and 187 m3/s for simulated flows, the future period (2021–2100) recorded a modulus of 98.3 m3/s under the SSP 245 scenario, i.e. an average deviation of -41.5% compared with the observed flow over the reference period, and a modulus of 91.4 m3/s under the SSP 585 scenario, i.e. an average deviation of -45.5% compared with the observed flow over the reference period.

Table 4

Future changes in flow (in %) on an annual scale over the four future periods at the Simenti station
SSP245	1985–2014	2021–2040	Trend	2041–2060	Trend	2061–2080	Trend	2081–2100	Trend
Flow rates	168	108	-35,8	94,0	-44,0	98,3	-41,4	93,2	-44,4
Precipitation	1281	1134	-147	1107	-174	1109	-172	1113	-168
Average temperature	28,0	29,0	0,99	29,6	1,65	30,3	2,27	30,7	2,71
FTE	1467	1511	44,4	1540	73,8	1569	102	1588	121
SSP585	1985–2014	2021–2040	Trend	2041–2060	Trend	2061–2080	Trend	2081–2100	Trend
Flow rates	168	119	-29,1	108	-35,7	74	-55,6	64	-61,6
Precipitation	1281	1141	-140	1128	-153	1058	-223	1031	-250
Average temperature	28,0	29,1	1,08	30,2	2,26	31,6	3,66	33,2	5,22
FTE	1467	1515	48,8	1568	101	1631	164	1700	234

The variations in rainfall and temperature that controlled to variations in possible flow are presented in Table 4. The analysis of runoff over the future period shows a decrease in runoff over the different sub-periods under both scenarios, in phase with the decrease in rainfall. There is a constant downward trend in the predictable hydrological variations for entirely scenarios. The decrease in discharge over the upcoming period diverse considerably between the four scenarios. Ranked from smallest to largest over the horizons, this decrease in flow, compared with the flow observed over the reference period, is of the order of -35.8%, -44.0%, -41.4% and − 44.4% respectively for the horizons 2040, 2060, 2080 and 2100 under the SSP 245 scenario. For the SSP 585 scenario, this reduction over these four horizons is respectively − 29.1%, -35.7%, -55.6% and − 61.6%. While for the first two horizons (2040 and 2060), the decline is greater under the SSP 245 scenario, for the last two horizons (2080 and 2100), the decline is greater under the SSP 585 scenario.

4.2.1. Future monthly hydrological trends

The flow characteristics on a monthly scale in basin area at the Simenti station over the situation period and the upcoming period are shown in Table 5 and Figs. 8 and 9. The river regime at this station is therefore characterised by a three-month high-water period and a nine-month low-water period. The regime of the Gambia catchment at Simenti is characterised by a maximum in September (683 m³ /s over the reference period, 401 m³ /s under the SSP 245 scenario and 360 m³ /s under the SSP 585 scenario) and a minimum in May (0.66 m³ /s over the reference period, 6.66 m³ /s under the SSP 245 scenario and 6.52 m³ /s under the SSP 585 scenario): this is a unimodal regime. The high-water period lasts three months, making it a pure tropical regime.

Table 5

Average monthly flow values for the reference and future periods at the Simenti station
	SSP 245			SSP 585
	1985–2014	2021–2100	Difference (%)	1985–2014	2021–2100	Difference (%)
January	20,7	25,4	22,6	20,7	24,4	17,6
February	8,94	16,1	80,3	8,94	15,5	73,9
March	3,00	11,4	278	3,00	11,0	267
April	1,63	8,48	421	1,63	8,27	408
May	0,66	6,66	911	0,66	6,52	889
June	3,81	5,51	44,5	3,81	5,41	42,1
July	73,1	8,56	-88,3	73,1	8,63	-88,2
August	352	104	-70,3	352	87,1	-75,2
September	683	401	-41,3	683	360	-47,3
October	420	298	-29,0	420	295	-29,9
November	120	116	-3,52	120	111	-7,62
December	48,8	47,9	-1,77	48,8	45,5	-6,68
Rainy season	306	163	-46,6	306	151	-50,7
Dry season	29,1	33,1	13,7	29,1	31,7	9,00
Year	168	98,3	-41,4	167,7	91,4	-45,5

The average monthly rates of change in flow below the two dissimilar SSP scenarios for the dissimilar times are presented in Figs. 8 and 9. Overall, there is a descending trend in flow for the months July to December ahead, but the decrease varies considerably from month to month. In general, the greatest decrease is noted over the far upcoming under both scenarios and occurs during the high-water period. On a monthly scale, the decrease in flow is expected to be greatest in August and September in both the near and distant future. Moreover, the maximum value of river discharge has changed from September in the historical period and the near future (2021–2040) and medium future 1 (2041–2060) to October in the medium future 2 (2061–2080) and far future (2081–2100) under the SSP 585 scenario (Fig. 8).

The range of deviations for the change in projected future flow is shown in Fig. 8. A disparate evolution is associated with the change in flow, with a large difference between high-water months with negative deviations (decrease in future flow) and low-water months with positive deviations (increase in future flow). The rainy seasons (high-water period) will record a fall of around − 46.6% under SSP 245 and − 50.7% under SSP 585. Conversely, dry seasons (low-water periods) will increase by 13.7% under SSP 245 and 9.0% under SSP 585.

An analysis of annual flow changes in the Gambia basin during the historical period reveals an uneven distribution of monthly flow at the Simenti station due to the seasonal influence of low flow originating from the water table. This uneven distribution has implications for the hydrological and ecological processes within the basin. The anticipated decrease in flow, coupled with the high uncertainty of future floods, will significantly impact the hydrological resource system and weaken its overall functioning. Understanding the changes in the contribution of different hydrological components to climate change within the river basin is of great importance. This study aimed to simulate changes in the main hydrological components relative to a baseline period (1985–2014) using a model to represent future scenario periods. By analyzing and comparing the simulated flow components, insights can be gained into the potential changes in the hydrological system under different climate change scenarios. Figure 9 shows that in the near, medium and distant future, discharge decreases under SSP 245 and SSP 585, respectively, compared with the control period (1985–2014). The changes and the magnitude of the changes in discharge are very different are consistent in the different scenarios with rainfall. The Gambia basin, located in the West African region, exhibits hydroclimatic variability that aligns with the projections of the Intergovernmental Panel on Climate Change (IPCC). Projections from climate models indicate a decrease in average monthly flows at the Simenti station under two scenarios (SSP 245 and SSP 585) for future time horizons. These findings suggest that surface water resources in the catchment areas of the basin are expected to continue declining throughout the 21st century, particularly under the more severe SSP 585 scenario. The vulnerability of the Gambia basin to climate change is evident, given the projected greater reduction in rainfall. The observed trends in water resources align with similar findings in other regions, highlighting the extreme scenarios presented by CMIP6. Runoff in the Gambia basin is primarily influenced by rainfall, and its seasonal changes correspond to those of precipitation patterns projected by climate models. The results obtained were based on the use of the GR4J global hydrological model, which is relatively simple and relies on four parameters. Utilizing a distributed hydrological model could potentially enhance the accuracy and reliability of these results.

The observed and simulated hydrographs exhibit similar patterns, indicating a good match between the two datasets during the calibration and validation periods. However, there are instances where the simulated peak flows are either underestimated or overestimated over a few years. The correlation between observed and simulated flows shows better coefficients for validation compared to calibration at the Simenti station. Nevertheless, the results demonstrate that the GR4J model is effective in simulating daily flows, particularly in the Gambia basin at the Simenti station. The simulated values closely align with the observed values, confirming the validity of the model. Analysis of uncertainties reveals relatively low values, approximately 0.142 mm during the calibration period (indicating flow underestimation by the model), -0.023 during validation period 1 (indicating flow overestimation by the model), and 0.220 during validation period 2 (indicating flow underestimation by the model). These uncertainties, as indicated by the use of the GR4J model in the Gambia basin, align with the findings of Bodian et al. (2016) in the Bafing basin and Mbaye et al. (2020) in the Falémé basin. The limited length of the observed data time series, which did not capture all specific hydrometeorological events during the calibration and validation periods, may contribute to some discrepancies between simulations and observations (Montecelos-Zamora et al., 2018). Additionally, the discrepancies in calculated and observed runoff can be attributed to the model's inability to account for variations in crucial factors such as soil type, vegetation, topography, human activities, and other key contributors to the runoff process (Mbaye et al., 2020). The underestimation of peak flow may also stem from the challenge of accurately reproducing flood peaks due to their sudden onset (Faye and Sow, 2014). The analysis of runoff over the future period shows a decrease in runoff over the different sub-periods under both scenarios, in phase with the decrease in rainfall and the increase in temperatures and therefore in evapotranspiration. There is a constant downward trend in the projected hydrological changes for all scenarios. The decrease in discharge over the future period varied considerably between the two scenarios and the four periods. Ranked from smallest to largest over the horizons, this decrease in flow, relative to the flow observed over the reference period, is of the order of -35.8%, -44.0%, -41.4% and − 44.4% respectively for the horizons 2040, 2060, 2080 and 2100 under the SSP 245 scenario. For the SSP 585 scenario, this reduction over these four horizons is respectively − 29.1%, -35.7%, -55.6% and − 61.6%. While for the first two horizons (2040 and 2060), the decline is greater under the SSP 245 scenario, for the last two horizons (2080 and 2100), the decline is greater under the SSP 585 scenario. The findings of this study are consistent with those of previous research by Mbaye et al. (2020), who also observed a decrease in river flow in the Gambia basin. Similarly, Bodian et al. (2018) highlighted the decline in river flow for both the Gambia and upper Senegal basins. Global warming has been found to enhance the atmosphere's water retention capacity, leading to increased surface water and soil water loss through evapotranspiration (Mbaye et al., 2019). This, in turn, has an impact on basin discharge. Temperature and potential evapotranspiration are among the key factors driving changes in the hydroclimate of the basin. Other factors such as atmospheric humidity, surface soil moisture, length of the growing season, net radiation, and wind characteristics also contribute significantly to changes in evapotranspiration, which subsequently affects runoff and leads to a decrease in water resources (Olsson et al., 2016; Nangombe et al., 2019). In general, under climate change, it is expected that temperature and potential evapotranspiration will increase, while precipitation will decrease in the catchment. Consequently, future periods from 2021 to 2100 are projected to experience reduced discharge compared to the reference period from 1985 to 2014. These findings demonstrate the direct relationship between temperature and potential evapotranspiration and their inverse relationship with rainfall and discharge in the Gambia catchment. It is important to note that this study assumed constant land use/land cover scenarios and meteorological variables such as solar radiation, wind speed, and humidity for future periods. However, these parameters directly contribute to hydrological systems, and it may be necessary to vary them in future studies. Additionally, incorporating multiple models with different scenarios for analyzing hydrological responses to climate change can help reduce uncertainties in climate change assessments

The GR4j model is a global model which has produced good results, the pace of the simulated hydrograph follows perfectly that observed. To summarize the advantages of the GR4j model: it does not take into account the physical characteristics of the basin; there is a low number of parameters therefore ease of calibration; it makes it possible to perfectly reproduce the observed flow; there are a limited number of simulation periods in two years. The advantages therefore lie in a better robustness during the passage from the calibration phases to the control. The application of the GR4J rainfall-runoff model in the Gambia basin has demonstrated its successful performance. This study found that despite its simple structure and limited number of parameters, the GR4J model can accurately simulate flow. However, it should be noted that increased biases in flow volume, particularly during the validation periods, indicate some errors in the input data, including rainfall and runoff. The four parameters obtained through the self-calibration process varied depending on the objective function and its weighting. However, all parameters fell within the standard range. It was observed that an independent calibration process yielded better results in terms of correlation, Nash criteria, and efficiency compared to a joint calibration process. Among the four objective functions, the Nash criteria produced the most favorable outcomes. Notably, certain parameters such as groundwater exchange and production storage exhibited variations depending on the chosen objective function. However, determining the optimal parameter set proved challenging when calibrating the model in self-calibration mode. Therefore, it is essential to analyze the sensitivity of these parameters to minimize uncertainty. It may be beneficial to explore a hybrid combination of different calibration and optimization methods to enhance the model's performance and generate improved results.

Ethics approval and consent to participate

Not applicable' for that section

Consent for publication

Not applicable' for that section

Availability of data and materials

The datasets used and analysed during this study are available from the corresponding author on reasonable request.

Competing interests

The authors declare no competing interests.

Funding

This research received no external funding.

Authors' contributions

S.M.K.S. Writing original draft writing, conceptualization, formal analysis, Methodology, Data curation, Processing of data, Development of models, Writing-review and editing, Investigation. C.F. Supervision, Writing original draft writing, Validation, Investigation, Methodology, writing- review and editing C. B. P. Formal Analysis, Validation, Writing original draft writing, formal analysis and writing- review and editing.

Acknowledgements

Not applicable' for that section

Authors' information (optional)

Department of Geography, U.F.R. Sciences et Technologies, Assane Seck Univesity of Ziguinchor, Laboratory of Geomatics and Environment, BP 523 Ziguinchor (Senegal), [email protected]

New Era and Development in Civil Engineering Research Group, Scientific Research Center, Al-Ayen University, Thi-Qar, Nasiriyah, 64001, Iraq, Institute of Energy Infrastructure, Universiti Tenaga Nasional, Kajang, 43000, Malaysia, [email protected], Indian Institute of Tropical Meteorology, Pune, India

Addor, N., Rössler, O., Köplin, N., Huss, M., Weingartner, R., & Seibert, J. (2014). Robust changes and sources of uncertainty in the projected hydrological regimes of Swiss catchments. Water Resources Research 50, 7541–7562. https://doi.org/10.1002/2014wr015549
Akinsanola A A and Zhou W 2019 Projections of West African summer monsoon rainfall extremes from two CORDEX models Clim. Dyn. 52 2017
Amoussou, E., Totin Vodounon, H., Houessou, S., Tramblay, Y., Camberlin, P., Houndenou, C., Boko, M., Mahe, G., and Paturel, J. E.: Application d’un modèle conceptuel à l’analyse de la dynamique hydrométéorologique des crues dans un bassinversant en milieu tropical humide: cas du Fleuve Mono. XXVIIIe Colloque de l’Association Internationale de Climatologie, Liège 2015, 17–24, 2015.
Aparicio, J.; Lafragua, J.; Lopez, A.; Mejia, R.; Aguilar, E.; Mejía, M. Water Resources Assessment: Integral Water Balance in Basins, phi-vii ed.; Number 14 in Technical Document; UNESCO Office Montevideo and Regional Bureau for Science in Latin America and the Caribbean: Montevideo, Uruguay, 2008.
Berthier C.-H., 2005: Quantification des incertitudes des débits calculés par un modèle pluie-débit empirique. Mémoire de Master 2ème année Sciences de la terre spécialité Hydrologie, Hydrogéologie et sols Année 2004–2005, 55 p.
Bodian A., A. Dezetter, A. Deme, L. Diop Hydrological evaluation of TRMM rainfall over the Upper Senegal River Basin, Hydrology, 3 (2016), p. 15, 10.3390/hydrology3020015
Bodian A., A. Dezetter, L. Diop, A. Deme, K. Djaman, A. Diop, Future climate change impacts on streamflows of two main West Africa River Basins: Senegal and Gambia, Hydrology, 5 (1) (2018), p. 21, 10.3390/hydrology5010021
Burn H. B. and Elnur M. A. H., 2002. Detection of Hydrologic Trends and Variability. Journal of Hydrology, 255, 107–122.
Carvalho, D.; Pereira, S.C.; Rocha, A. Future surface temperature changes for the Iberian Peninsula according to EURO-CORDEX climate projections. Clim. Dyn. 2020, 56, 123–138. https://doi.org/10.1007/s00382-020-05472-3.
Chegwidden, O. S., Nijssen, B., Rupp, D. E., Arnold, J. R., Clark, M. P., Hamman, J. J., Kao, S. C., Mao, Y., Mizukami, N., Mote, P. W., Pan, M., Pytlak, E., & Xiao, M. (2019). How do modeling decisions affect the spread among hydrologic climate change projections? Exploring a large ensemble of simulations across a diversity of hydroclimates. Earth's Future, 7, 623–637. https://doi.org/10.1029/2018ef001047
Chen, J., Brissette, F. P., Chaumont, D., & Braun, M. (2013). Performance and uncertainty evaluation of empirical downscaling methods in quantifying the climate change impacts on hydrology over two North American river basins. Journal of Hydrology, 479, 200–214. https://doi.org/10.1016/j.jhydrol.2012.11.062
Chen, Y.; Li, X.; Huang, K.; Luo, M.; Gao, M. High-Resolution Gridded Population Projections for China Under the Shared Socioeconomic Pathways. Earth’s Future 2020, 8, e2020EF001491.
Dechemi N., T. Benkaci et A. Issolah, 2003: Modélisation des débits mensuels par les modèles conceptuels et les systèmes neuro-flous » Revue des sciences de l'eau / Journal of Water Science, vol. 16, n° 4, pp. 407–424.
Dione O., 1996: Evolution climatique récente et dynamique fluviale dans les hauts bassins des fleuves Sénégal et Gambie. Thèse de doctorat, Université Lyon 3 Jean Moulin, 477 p.
Evans, J. P., Ji, F., Abramowitz, G., & Ekstrom, M. (2013). Optimally choosing small ensemble members to produce robust climate simulations. Environmental Research Letters, 8(4), 044050. https://doi.org/10.1088/1748-9,326/8/4/044050
eWater Ltd (2013a, October 11). Source Scientific Reference Guide (v3.5.0) (Online: Available: https://ewater.atlassian.net/wiki/display/SD35/Source+Scientific+Reference+Guide)
Eyring, V., Cox, P. M., Flato, G. M., Gleckler, P. J., Abramowitz, G., Caldwell, P., Collins, W. D., Gier, B. K., Hall, A. D., Hoffman, F. M., Hurtt, G. C., Jahn, A., Jones, C. D., Klein, S. A., Krasting, J. P., Kwiatkowski, L., Lorenz, R., Maloney, E., Meehl, G. A., Pendergrass, A. G., Pincus, R., Ruane, A. C., Russell, J. L., Sanderson, B. M., Santer, B. D., Sherwood, S. C., Simpson, I. R., Stouffer, R. J., & Williamson, M. S. (2019). Taking climate model evaluation to the next level. Nature Climate Change, 9(2), 102–110. https://doi.org/10.1038/s41558-018-0355-y
Eyring, V., Cox, P. M., Flato, G. M., Gleckler, P. J., Abramowitz, G., Caldwell, P., Collins, W. D., Gier, B. K., Hall, A. D., Hoffman, F. M., Hurtt, G. C., Jahn, A., Jones, C. D., Klein, S. A., Krasting, J. P., Kwiatkowski, L., Lorenz, R., Maloney, E., Meehl, G. A., Pendergrass, A. G., Pincus, R., Ruane, A. C., Russell, J. L., Sanderson, B. M., Santer, B. D., Sherwood, S. C., Simpson, I. R., Stouffer, R. J., & Williamson, M. S. (2019). Taking climate model evaluation to the next level. Nature Climate Change, 9(2), 102–110. https://doi.org/10.1038/s41558-018-0355-y
Faye C., A.A. Sow, Analyse de la variabilité des ressources en eau dans le bassin de la Falémé par modélisation hydrologique, vol. 14 (2014), 12, 9p.
Faye C., et Mendy A., 2018: Variabilité climatique et impacts hydrologiques en Afrique de l’Ouest : Cas du bassin versant de la Gambie (Sénégal). Environmental and Water Sciences, Public Health &
Gidden, M.J. et al., 2019: Global emissions pathways under different socioeconomic scenarios for use in CMIP6: A dataset of harmonized emissions trajectories through the end of the century. Geoscientific Model Development, 12(4), 1443–1475, doi:10.5194/gmd-12-1443-2019.
Gleick, P.H. Global Freshwater Resources: Soft-Path Solutions for the 21st Century. Science 2003, 302, 1524–1528.
Hamby, 1994: A review of techniques parameters sensitivity analysis of environmental models, Environmental Monitoring and Assessment 32. 135–154
Harlan, D., Wangsadipura, M. and Munajat, C. M., 2010. Rainfall-Runoff Modeling of Citarum Hulu River Basin by Using GR4J. In: Proceedings of the World Congress on Engineering, Vol. 2. ISBN: 978-988-18210-7-2. ISSN: 2078 – 0958 (Print); ISSN: 2078 – 0966 (Online)
Hölzel, H.; Diekkrüger, B. Predicting the impact of linear landscape elements on surface runoff, soil erosion, and sedimentation in the Wahnbach catchment, Germany. Hydrol. Process. 2012, 26, 1642–1654.
Kay, A. L. and Davies, H. N.: Calculating potential evaporation from climate model data: A source of uncertainty for hydrological climate change impacts, J. Hydrol., 358, 221–239, 2008.
Kay, A.L. Simulation of river flow in Britain under climate change: Baseline performance and future seasonal changes. Hydrol. Process. 2021, 35, e14137.
Knutti, R., Baumberger, C., & Hirsch Hadorn, G. (2019). Uncertainty quantification using multiple models—Prospects and challenges. In C. Beisbart, & N. J. Saam (Eds.), Computer simulation validation: Fundamental concepts, methodological frameworks, and philosophical perspectives. (pp. 835–855). Cham: Springer International Publishing.
Knutti, R., Furrer, R., Tebaldi, C., Cermak, J., & Meehl, G. A. (2010). Challenges in combining projections from multiple climate models. Journal of Climate, 23(10), 2,739– 2,758. https://doi.org/10.1175/2009jcli3361.1
Kouassi A. M., 2007: Caractérisation d’une modification éventuelle de la relation pluie-débit et ses impacts sur les ressources en eau en Afrique de l’Ouest: cas du bassin versant du N’zi (Bandama) en Côte d’Ivoire. Thèse de Docteur de l’Université de Cocody, Côte d’Ivoire, 234 p.
Koutsoyiannis, D. Nonstationarity versus scaling in hydrology. J. Hydrol. 2006, 324, 239–254.
Lamagat J.P., 1989. Monographie hydrologique du fleuve Gambie Collection M&m. ORSTOM-OMVG, 250 p.
Makhlouf Z., 1994: Compléments sur le modèle pluie-débit GR4J et essai d'estimation de ses paramètres. Thèse de Doctorat, Université Paris XI Orsay, 426 p.
Mbaye M. L., Sy K., Faty B., Sall S. M., 2020: Impact of 1.5 and 2.0°C global warming on the hydrology of the Faleme river basin, Journal of Hydrology: Regional Studies, Volume 31, 2020, 100719, https://doi.org/10.1016/j.ejrh.2020.100719.
Mbaye M.L., M.B. Sylla, M. Tall, Impacts of 1.5 and 2.0°C global warming on water balance components over Senegal in West Africa ? Atmos. 2019, 10 (11) (2019), p. 712, 10.3390/atmos10110712
Middelkoop, H.; Daamen, K.; Gellens, D.; Grabs, W.; Kwadijk, J.C.; Lang, H.; Parmet, B.W.; Schädler, B.; Schulla, J.; Wilke, K. Impact of climate change on hydrological regimes and water resources management in the Rhine basin. Clim. Chang. 2001, 49, 105–128.
Montecelos-Zamora Y., T. Cavazos, T. Kretzschmar, E.R. Vivoni, G. Corzo, E. Molina-Navarro, Hydrological modeling of climate change impacts in a Tropical River Basin: a case study of the Cauto River, Cuba. Water, 10 (9) (2018), p. 1135, 10.3390/w10091135
Nangombe S.S., T. Zhou, W. Zhang, L. Zou, D. Li, High-temperature extreme events over Africa under 1.5 and 2°C of global warming. J. Geophys. Res. Atmos., 124 (2019), pp. 4413–4428, 10.1029/2018JD02974
O’Neill B. C., Kriegler E., Ebi K. L.,, Kemp-Benedict E., Riahi K., Rothman D. S., van Ruijven B. J., van Vuuren D. P., Birkmann J., Kok K., Levy M., Solecki W., 2015.The roads ahead: Narratives for shared socioeconomic pathways describing world futures in the 21st century. Global Environmental Change, In Press.
O’Neill, B.C.; Tebaldi, C.; Vuuren, D.P.V.; Eyring, V.; Friedlingstein, P.; Hurtt, G.; Knutti, R.; Kriegler, E.; Lamarque, J.F.; Lowe, J.; et al. The scenario model intercomparison project (ScenarioMIP) for CMIP6. Geosci. Model Dev. 2016, 9, 3461–3482.
Olsson J., B. Arheimer, M. Borris, C. Donnelly, K. Foster, G. Nikulin, M. Persson, A.-M. Perttu, C.B. Uvo, M. Viklander, W. Yang, Hydrological climate change impact assessment at small and large scales: key messages from recent progress in Sweden. Climate, 4 (3) (2016), p. 39, 10.3390/cli4030039
Perrin, C. (2000). Vers une amélioration d'un modèle global pluie-débit au travers d'une approche comparative. Thèse de doctorat, Institut National Polytechnique de Grenoble, Grenoble, France, 518 p.
Qin, P.; Xu, H.; Liu, M.; Du, L.; Xiao, C.; Liu, L.; Tarroja, B. Climate change impacts on Three Gorges Reservoir impoundment and hydropower generation. J. Hydrol. 2020, 580, 123922.
Raju, K. S., & Nagesh Kumar, D. (2014). Ranking of global climate models for India using multicriterion analysis. Climate Research, 60(2), 103–117. https://doi.org/10.3354/cr01222
Sow A. A., 2007. L’hydrologie du Sud-est du Sénégal et de ses Confins guinéo-maliens: les bassins de la Gambie et de la Falémé. Thèse (PhD). Université Cheikh Anta Diop de Dakar, 1232 p.
van Esse, W. R., Perrin, C., Booij, M. J. Augustijn, D. C. M., Fenicia, F., Kavetski, D., Lobligeois, F., 2013. The influence of conceptual model structure on model performance: a comparative study for 237 catchments. Hydrol. Earth Syst. Sci., Vol. 17., Pg. 4227–4239. doi:10.5194/hess-17-4227-2013.
van Vuuren, D. P., Edmonds, J., Kainuma, M., Riahi, K., Thomson, A., Hibbard, K., Hurtt, G. C., Kram, T., Krey, V., Lamarque, J.-F., Masui, T., Meinshausen, M., Nakicenovic, N., Smith, S. J., & Rose, S. K. (2011). The representative concentration pathways: An overview. Climatic Change, 109(1–2), 5–31. https://doi.org/10.1007/s10584-011-0148-z
Vetter, T.; Reinhardt, J.; Flörke, M.; Van Griensven, A.; Hattermann, F.; Huang, S.; Koch, H.; Pechlivanidis, I.G.; Plötner, S.; Seidou, O.; et al. Evaluation of sources of uncertainty in projected hydrological changes under climate change in 12 large-scale river basins. Clim. Chang. 2017, 141, 419–433.
Wang, H.-M., Chen, J., Xu, C.-Y., Chen, H., Guo, S., Xie, P., & Li, X. (2019). Does the weighting of climate simulations result in a better quantification of hydrological impacts? Hydrology and Earth System Sciences, 23(10), 4,033– 4,050. https://doi.org/10.5194/hess-23-4,033-2019
Whitehead, P.G.; Wilby, R.L.; Battarbee, R.W.; Kernan, M.; Wade, A.J. A review of the potential impacts of climate change on surface water quality. Hydrol. Sci. J. 2009, 54, 101–123.
Yip, S., Ferro, C. A. T., Stephenson, D. B., & Hawkins, E. (2011). A simple, coherent framework for partitioning uncertainty in climate predictions. Journal of Climate, 24(17), 4,634– 4,643. https://doi.org/10.1175/2011jcli4085.1

No competing interests reported.

Download PDF

Journal Publication

published 18 Feb, 2024

Read the published version in Environmental Sciences Europe →

Editorial decision: Revision requested
15 Dec, 2023
Reviews received at journal
04 Dec, 2023
Reviews received at journal
01 Dec, 2023
Reviewers agreed at journal
24 Nov, 2023
Reviewers agreed at journal
16 Nov, 2023
Reviewers agreed at journal
03 Nov, 2023
Reviewers agreed at journal
15 Oct, 2023
Reviewers agreed at journal
04 Oct, 2023
Reviewers agreed at journal
18 Sep, 2023
Reviewers agreed at journal
17 Aug, 2023
Reviewers invited by journal
30 Jul, 2023
Editor assigned by journal
11 Jul, 2023
Submission checks completed at journal
07 Jul, 2023
First submitted to journal
05 Jul, 2023

You are reading this latest preprint version

Assessment of current and future trends in water resources in the Gambia River Basin in a context of climate change

Status:

Journal Publication

Version 1

Abstract

Figures

1. Introduction

2. Study area

3. Data and methods

3.1. Presentation of the GR4J model

3.2. Assessing model performance

3.2.1. The Nash-Sutcliffe index

3.2.2. The square root of the Nash-Sutcliffe index

3.2.3. The natural logarithm of the Nash-Sutcliffe index

3.2.4. The volume balance criterion

3.2.5. Assessment of uncertainties associated with simulated flow values

3.3. Data requested

3.3.1. Precipitation and temperature data

3.3.2. Calculation of the average rainfall in the basin

3.3.3. Oudin's potential evapotranspiration

3.4. Model calibration/validation method

3.5. Acquisition of future precipitation and temperature data and simulation

4. Results and Discussion

4.2 Future hydrological trends

4.2.1. Future annual hydrological trends

4.2.1. Future monthly hydrological trends

5. Discussion

6. Conclusion

Declarations

References

Additional Declarations

Status:

Journal Publication

Version 1