Impact of climate and anthropogenic changes on current and future variability in flows in the Nyong River Basin (equatorial Central Africa)

Variations in river flow generally result from interactions between climate change and/or anthropogenic changes (Diem et al. 2018; Oudin et al. 2018; Ebodé et al. 2021a; Ebodé 2022a, 2022b). However, the study of the impact of these environmental forcings is done in different ways. In some works, this impact has been studied using the historical/contemporary period (Cissé et al. 2014; Bodian et al. 2020; Descroix et al. 2020). Meanwhile, other studies have addressed it over future periods to plan how available water resources will be distributed among different sectors (Abe et al. 2018; Dibaba et al. 2020). Until now, there are few studies (Ardoin-Bardin 2004; Sighomnou 2004) that have addressed both the study of the impact of environmental forcings on flows using current and future climate projections in some regions like Central Africa, due to the scarcity of data. The active network of this part of the continent now only includes 35 stations in the Democratic Republic of Congo (DRC) (i.e. one station for 67,000 km²), 22 in Cameroon and less than 20 for all of Gabon, Congo and Central African Republic (Bigot et al. 2016). Such studies are urgently needed to assess the current and future evolution of water resources in a given region, in order to put in place different water management strategies.

In equatorial Central Africa, recent studies focusing on the impact of environmental forcings on flows during the historical period have highlighted a drop in average and extreme flows linked to a drop in rainfall for some basins such as the Ntem and the Ogooue (Conway et al. 2009; Ebodé et al. 2020, 2021b). In other basins such as the Nyong and the Mefou, it has been demonstrated that maximum flows may stay the same or increase, as a result of changes in land use patterns. These changes, which are essentially marked by an increase in impervious areas (buildings, roads and bare soil) to the detriment of the forest, would then compensate for the rainfall deficit by an increase in runoff, hence maximum flows will stay the same in the larger basins and increase in the smaller ones (Ebodé et al. 2022a). In forested equatorial Central Africa, there are few studies addressing the impact of environmental forcings on flows during future periods using the state-of-the-art methodology and new data (Akoko et al. 2021).

In other regions of the world, predictive hydrological modelling is done using a global or distributed/semi-distributed approach. The global approach considers the watershed as a single entity. The GR2J (Rural Engineering model with two daily Parameters) and GR4J (Rural Engineering model with four daily Parameters) models are some of the reference models generally used in this type of approach (Perrin et al. 2003; Bodian et al. 2012; Amoussou 2014). These models take into account parameters such as precipitation, evapotranspiration and soil water capacity. In the distributed/semi-distributed approach, on the other hand, the watershed is considered as a complex entity and flow modelling requires a subdivision into homogeneous elementary surfaces (Legesse et al. 2003; Githui et al. 2009; Taleb et al. 2019). Distributed/semi-distributed models require a wide range of input data, ranging from physical characteristics of the basin (slopes, land cover, soils, etc.) to meteorological data (precipitation, maximum and minimum temperatures, wind speed, relative humidity, insolation, etc.).

In terms of performance, the comparison of global and distributed/semi-distributed approaches in modelling is a problem that has been strongly developed for a long time (Wending 1992; Ague et al. 2014). Results from such model assessments provide a relatively complex picture as some authors clearly state the benefits of using the distributed/semi-distributed approach (Michaud & Soroshian 1994; Krysanova et al. 1999; Boyle et al. 2001) while other studies (Diermanse 1999; Kokkonen & Jakeman 2001) present opposite results. Some studies also show that the global approach gives better results in small watersheds, while the distributed/semi-distributed approach performs better in the case of large watersheds. These findings suggest that distributed/semi-distributed models are particularly applicable to complex watersheds due to their physical heterogeneity (Tegegne et al. 2017). Due to the complex nature of our study area, this study adopted a semi-distributed modelling approach considering that a previous study in the area using the lump conceptual modelling approach produced poor results (Sighomnou 2004). One of the complexities of the study area is the existence of two rainy seasons in the region.

In recent decades, several distributed/semi-distributed hydrological models have been developed to simulate the hydrological processes of watersheds and predict flows. Typical examples of distributed hydrological models include TOPMODEL (Topography-based Hydrological Model) (Beven & Kirkby 1979), SHE (European Hydrological System) (Abbott et al. 1986), SWAT (Arnold et al. 1998), MGB-IPH (Large Basin Hydrological Model) (Collischonn & Tucci 2001), etc. Since the others only allow an approximate characterization of the physical environment of the watershed through the use of data and parameters in a point-grid network (Cao et al. 2006; Wang et al. 2012), SWAT appears to be the most efficient model in the wide range of applications.

The SWAT model has been widely used around the world for hydrological simulations of basins with different environmental conditions (climate, topography, geology, soils and vegetation) (Von Stackelberg et al. 2007; Pereira et al. 2010; Andrade et al. 2013; Ebodé et al. 2022b) with satisfactory results obtained. For SWAT to become a universal hydrological model, more studies are needed in basins with climatic regimes and soils typical of equatorial conditions, such as that of the Nyong basin in southern Cameroon.

The objectives of this study were to (1) evaluate the capacity of the SWAT model to simulate flows in a relatively large complex equatorial river basin with two rainy season regimes and (2) use the model to simulate future flows in the basin under different climate change scenarios. One of the biggest challenges in hydrological modelling is the absence of sufficient flow gauging stations and the ability of SWAT to simulate flows in this poorly gauged basin will be of great importance for socioeconomic development considering the number of construction projects including dams and bridges ongoing in the basin. The results obtained could help to improve the management of water resources in this basin and the region.

DEM data used in this study have a spatial resolution of 30 × 30 m. It was obtained from the United States Geological Survey website (https://earthexplorer.usgs.gov/). The DEM was used to delineate the watershed, slope classification and to generate the hydrological response units (HRUs). In total, 21 HRUs were delineated in the NRB, with slopes varying from 0 to 69% (Table 1). The dominant slope class (14–69%) which covers 1.81% of the total area of the basin is mostly found in its North West part (Figure 2). The slopes of 0–3%, 4–7% and 8–13%, respectively, occupy 41.09, 38.28 and 18.82% of the watershed. These three slope classes are observed more in the East and middle parts of the NRB (Figure 2).

The FAO world digital soil map downloaded from the site: http://www.fao.org/geonetwork/srv/en/metadata.show?id=14116 was used as soil data in this study. The soil classification used is based on the FAO classification system and was customized as required by the SWAT model (Figure 2). Five types of soils characteristic of the forest zone, namely the Orthic Ferralsols (Fo) have been identified in the basin (Ao7-437, Fo1-ab-481, Fo14-2a-1162, Fo26-ab-490 and Fo35-494). Types Fo1-ab-481 and Fo35-494 are the most widely represented with 64.20 and 31.05% of the whole basin.

Apart from topographic and soil information which remain unchanged on the human time scale, the simulation of flows from the SWAT model requires other dynamic information such as land use. Three Landsat satellite images (Table 2) were used to produce land cover maps. These are the Landsat Multispectral Scanner System (MSS) of 1978, Landsat Thematic Mapper (TM) of 2005 and Landsat 8 Operational Land Imager (OLI) of 2020. The maps produced from the images of 1978 and 2005 were used in the model to simulate the flows during the first (1984–1986 for calibration and 1981–1982 for validation) and the second (2008–2011 for calibration and 2010–2014 for validation) historical period retained. The choice of two relatively distant periods (distinct from the point of view of land cover) aimed to assess the model's performance in simulating the region flows in a situation of land cover change. Taking this aspect into account is currently a major challenge in modelling the flows of tropical rivers, given the strong anthropization recently observed in their catchment areas, the hydrological impacts of which are well known. Recent modelling work in West Africa has already integrated these aspects (Paturel et al. 2017). It would also be interesting to continue this work in Central Africa on larger basins. The choice of these periods also considered the availability of observed hydrological data (to calibrate the model) and usable satellite images. The maps made from these 2 years (1978 and 2005) also served as a reference in the QGIS software, to predict the land cover in 2040 and 2080. The 2020 map was used to validate the one simulated for the same year from the two previous ones (1978 and 2005 maps) since any prediction from a model requires calibration and validation of the said model. Similarly, the reliability of a model's outputs depends on the validation results of the model in question.

Rainfall data used to study the current period were obtained from the Climate Research Unit (CRU) of the University of East Anglia in the United Kingdom. The data are available from 1901, via the site https://climexp.knmi.nl/selectfield_obs2.cgi?id=2833fad3fef1bedc6761d5cba64775f0/ in NetCDF format, on a monthly time step and at a spatial resolution of 0.25° × 0, 25°. The flow series used is that of ORE-BVET (Environmental Research Observatory/Tropical Experimental Watersheds).

The analysis of average rainfall and flow was carried out using Pettitt (Pettitt 1979) and modified Mann–Kendall (Hirsch & Slack 1984; Araghi et al. 2014) statistical tests at the 95% significance level. Following the application of autocorrelation (from the calculation of the R statistic) and seasonality (from the employment of the correlogram) tests to the rainfall and flow series used, it turned out that there is a seasonality in the latter. That is why we chose to use the modified Mann–Kendall test of Hirsch & Slack (1984) to the detriment of the classic Mann–Kendall test and other modified Mann–Kendall tests.

If the null hypothesis is rejected, an estimate of the break date is given by the instant defining the maximum in absolute value of the variable Ut, N.

There is no significant trend in the series analyzed when the calculated p-value is above the chosen significance level. More details on the modified version of the Mann–Kendall test can be found in the relevant literature (Hirsch & Slack 1984).

To assess the behaviour of extreme flows, the Indicators of Hydrologic Alteration (IHA) tool, version 7.1, developed by The Nature Conservancy was used. This tool offers the possibility of comparing the parameters characterizing the flow regimes under different conditions. It uses daily discharge values and produces several important statistics. Only four of them were considered essential for this study, among which were the average, the coefficient of variation (CV) of extreme discharge and the Julian date of the annual minimum and maximum. By dividing the series of values in the period before and after the discontinuity, the tool calculates the change that occurred in the evolution of each of these parameters after the discontinuity. We can thus analyze not only the sign of change between the two periods but also the magnitude of this difference.

The response of discharges from the Nyong watershed to climate change and anthropization was assessed over two future periods (2022–2060 and 2061–2100). The interval 2009–2014 served as the reference/historical period. Rainfall data of the station used are without gaps over this interval. The 2005 land cover map was used to simulate runoff for the reference period (2009–2014). Those simulated for the years 2040 and 2080 were respectively used to simulate basin flows in the near (2022–2060) and distant (2061–2100) future.

CA-Markov is the procedure used for land cover prediction in this work. It combines Markov chains (quantity), multi-criteria evaluation (location) and filtering. This procedure is described as ‘cellular automata’ (CA) (Halmy et al. 2015). Markovian chains analyze two images of land cover at different dates and produce two transition matrices (probability and affected area in pixels for persistence and transition), and a set of conditional probability images. They make it possible to calculate a future state from the known present state, based on the observation of past evolutions and their probability. This makes this method one of the best for modelling the temporal and spatial dimensions of land use patterns (Halmy et al. 2015; Yang et al. 2019).

In this equation, Po is the proportion of correctly simulated cells; Pc is the expected proportion correction by chance between the observed and simulated map. If the Kappa index ≤ 0.5, this indicates poor proximity between the two compared cards (simulated and observed). If 0.5 ≤ Kappa ≤ 0.75, the proximity between the two maps is acceptable. When 0.75 ≤ Kappa ≤ 1, the proximity between the two maps is good. When Kappa = 1, the two cards are identical. The Kappa coefficient obtained for the comparison between the observed and simulated maps for the year 2020 in this study is 0.87, which gives credibility to the simulated maps for the years 2040 and 2080.

In this study, two regional climate models (RCMs) [RCA4 and CCCma] from the CORDEX project that have proven to be effective in simulating precipitation and temperature in Africa (Gadissa et al. 2018; Dibaba et al. 2019) were retained. However, despite their reliability and the degree of confidence that can be granted to them, the outputs of the models sometimes present considerable biases, which require corrections before using them to study the impact of climate change. For each of the RCMs, data from two scenarios (representative concentration pathway (RCP)4.5 and RCP8.5) were collected. The first and second scenarios are respectively representative of high greenhouse gas emissions and moderate emissions. The other meteorological variables (solar radiation, relative humidity and wind speed) considered for the historical period have been taken over for the two future periods without making any changes, given that their modifications have no significant impact on the modelling result (Gadissa et al. 2018).

Climate Model Data for Hydrological Modeling (CMhyd) software (Rathjens et al. 2016) obtained from: https://swat.tamu.edu/software/ was used to correct for precipitation and temperature biases. Teutschbein & Seibert (2012) provided a comprehensive review of bias correction techniques based on this tool. According to the authors, all the correction techniques improved the simulations of precipitation and temperature. However, they noted differences between the correction methods. Based on the proximity between the corrected datasets and the observed datasets, distribution mapping (DM) was considered to be the best correction method, both for temperature and precipitation. According to the authors, DM uses a transfer function to adjust the cumulative distribution of the corrected data to that of the observed data, which makes the results significantly better. Zhang et al. (2018) compared five bias correction methods using the CMhyd tool. They demonstrated that DM was the most efficient correction method for studying the impact of climate change on the flow dynamics of two rivers in the northern basin of Lake Erie (Canada). Based on these results, DM was retained for the precipitation and temperature corrections of the model outputs used in this study.

SWAT is a physically based semi-distributed hydrological model, designed and developed by researchers at the USDA (United States Department of Agriculture) (Arnold et al. 1998). The physical aspect of the model makes it possible to reproduce the processes that take place in the environment, using a different set of equations (Neitsch et al. 2005; Arnold et al. 2012). This model is continuous over time and is designed to run simulations over long periods (Payraudeau 2002). The SWAT model analyzes the watershed as a whole by subdividing it into sub-watersheds containing homogeneous portions called HRUs. Each HRU is characterized by unique land use, soil type and topography. SWAT provides access to the different water balance components at the HRU scale for each time step (daily, monthly and annual) over the simulation period (Neitsch et al. 2005).

The validity of a hydrological model was checked by comparing the model simulated (Q_sim) (Q_cal) and observed (Q_obs) flows through subjective and quantitative criteria. Initially, a good match between the observed and simulated flow hydrographs will attest to good calibration. In the second step, we used three of the most widely used criteria for the validation of hydrological models, correlation coefficient (r), the Nash index (NSE) and the bias (Akoko et al. 2020).

Previous studies in the Nyong basin at Mbalmayo have reported the absence of a rupture in the time series of annual precipitation in this basin (Liénou et al. 2008; Ebodé et al. 2020). This result differs from that obtained in the present study. A rupture marking a decrease in precipitation was highlighted in the series of precipitation in this same basin in 1975–1976. This difference in results could be attributed to the different sources of precipitation used in the two studies. This study used rainfall data obtained from the CRU while Ebodé et al. (2020), for its part, used rainfall data from SIEREM (1950–1999) and TRMM 3B42 V7 (2000–2016) data. Liénou et al. (2008) used data obtained from SIEREM precipitation only. This could also be due to the difference between the lengths of the precipitation series used. This study covers the period 1950–1951 to 2017–2018, while that of Liénou et al. (2008) covers the period 1950–2000.

Precipitation during the rainy season decreased in the studied basin (Table 4). This decrease is statistically significant in all cases, according to the Pettitt test (Figure 4). Spring precipitation decreased statistically in 1975–1976. The rate of change recorded following this rupture is –2.3% (Figure 4). Despite this decrease since the 1970s, it should be noted that Spring precipitation increased slightly over the past two decades (Table 5).

In the case of the Autumn precipitation, it also decreased significantly. The deficits caused by the ruptures identified in 1974–1975 is –2% (Figure 4). The analysis of the decadal deviations from the interannual average shows an evolution in the Autumn precipitation which is in line with that described for the annual precipitation (Table 5).

The opposite trends have been highlighted in the dry season's rainfall (Table 4). There is an increase in Summer precipitation and a decrease in Winter precipitation. The increase in summer precipitation is statistically significant according to the Pettitt test. A rupture was identified in their series in 1964–1965 (+5.3%) (Figure 4). Despite a general increase observed during the 1960s, the analysis of the decadal deviations from the interannual mean nevertheless reveals a significant drop in summer precipitations which started during the 2000s and intensifies considerably in the following decade (Table 5).

The decrease in winter precipitation is significant according to the Pettitt test. Compared to the whole period mean, the rate of change recorded after the rupture detected in their series in 1975–1976 is –15% (Figure 4). Although having experienced a general decrease during the 1970s, Winter precipitations seem to increase considerably since the 2000s (Table 5).

Like the previous studies relating to the studied region (Liénou et al. 2008), this study highlights a cross-trend evolution of Summer and Winter precipitation between the 1960 and 1990 decades. However, the present work reveals a reversal of these previously known trends from the 2000s. These new trends have only been demonstrated by a small number of studies so far (Ebodé et al. 2020, 2021a).

The analysis results of the interannual evolution of rainfall show that the 1970s represent the pivotal period of rainfall evolution in the study area. This led us to investigate the spatial distribution of rainfall in the Nyong basin during the period after the breaks which occurred during this decade. The question therefore arises as to whether the changes in annual and seasonal rainfall observed after the breaks were accompanied by a modification in their spatial distribution.

As for the rainy season's rainfall, we also observe, as with the annual precipitation, a decrease in rainfall in the basin from East to West between the periods compared. For spring, the limit of the weakest precipitation class (<660 mm) was around Nyabong during the first period but reached Dzeng during the second period (Figure 5). Also, the 700 mm isohyet located near Dzeng during the first period disappeared in the basin during the second. The autumn precipitation evolved almost identically to spring in the basin studied (Figure 5).

In the case of the dry season precipitation, we note an increase in summer precipitation from North to South in the basin (Figure 5). Summer precipitation did not exceed 215 mm during the first period, which is not the case during the second. Winter precipitation decreased in the same direction (North–South). The lowest precipitation class (<100 mm) was confined to the North of the basin during the first period and covered the whole catchment during the second (Figure 5).

The average discharges of Nyong are analyzed through the annual and seasonal modules. According to the Pettitt test, the annual modules of this river evolve statistically downwards (Figure 4). This decrease is accompanied by a reduction in Cv compared to the post-discontinuity period which seems negligible (−1%) (Table 6). The Pettitt test shows discontinuities in the series of this river in 1973–1974 (Figure 4). The homogeneous sequence following this discontinuity recorded a deficit compared to the long-term average estimated at −5% (Figure 4). However timidly started in the 1970s, the decrease in the annual modules of Nyong increased considerably during the 2000s (Table 5). This decade is the driest of the entire observation period. The decadal deviations from the whole period average for this decade is −7.4% (Table 5).

Regarding the discharges of the rainy seasons, a statistically significant decrease was noted in spring modules that the Pettitt test situated in the same years as that of the annual modules. As in the annual timescale, the change in Cv of discharges for this season is slight (−1%). The deficit of the period occurring after the discontinuity −12.4% (Figure 4). The decrease started in the 1970s seems to persist (Table 5). In autumn, Nyong discharges decrease statistically insignificantly (Figure 4). However, there has been an important decline in them during the 2000 and 2010s, following the increase that began in the 1960s (Table 5).

In dry seasons, during the summer, the discharges do not change statistically according to the Pettitt test, although a slight decrease is observed (Figure 4). However, a general decrease since the 1970s can be observed through the decadal deviations, which increased considerably during the 2010s (Table 5). In winter, the Pettitt test does not show any discontinuity in the Nyong discharges, although an increase in them is noted in the 2010s after the decrease started in the 1970s (Table 5).

The Nyong extreme flows are analyzed through maximum and minimum over consecutive days (1, 3, 7, 30 and 90 days).

The evolution of minimum flows differs from that of maximum (Figure 6). All of the different ranges taken into account have decreased considerably after the rupture, regardless of the flow duration. The rates of change recorded after the discontinuity are quite close from one range to another, varying between −29.3 and −36.2% (Table 7). The 90-day minimum flow rate is the only one where there is a slight reduction in variability after the rupture. For the other four ranges, an increase was observed, which seems greater for the short-term minimum (1-day and 3-day). Their coefficients of variation increased from 30% before the rupture to 41% after (Table 7). The minimums are slightly late on the Nyong after the rupture. Their average Julian date increased from 61 before the rupture to 67 after (Table 7).

The changes in the flow regimes in the studied basin require explanations which are provided herein by examining the flow regimes using a combination of environmental forcings (land use modes and precipitation).

Significant hydrological changes observed after the rupture occurred in the studied basins seem not to have any links with changes in precipitation. All this prompts us to investigate the evolution of land use in the two basins.

A change detection analysis performed with SNAP, through the diachronic comparison of the results of supervised classifications of Landsat images, shows significant changes in land occupation and use patterns in the Nyong basins at Mbalmayo outlet (Figure 2). In general, there is an increase in impervious areas in this basin (made up of buildings, roads, bare soil, young fallows and cultivation areas) to the detriment of forest areas and water bodies. Between 1978 and 2020, the increases are between +414.62% (bare soil, savannah and cultivated area) and +261.62% (buildings, roads) (Table 2).

The stability observed for the maximums after the rupture could be a consequence of the changes in surface conditions. The most logical process would have been to see the maximum decrease because of the evolution of the precipitation of the two rainy seasons that generate them. The current rate of urbanization of this basin, although modest, seems to be the main factor that can justify this lack of trend. The increase in runoff has offset the decrease in precipitation. It is quite possible to wonder about the impact of such a modest imperviousness rate on runoff. Given that the changes in extremes do not agree with other processes having a proven impact on runoff, in this case, the decrease in precipitation highlighted in this study and the increase in temperatures postulated by Sighomnou (2004) in the region, which act rather in the direction of flows reduction and a late observation of the maximums, it is quite possible to attribute this flows trend to the changes in land cover, however modest they may be. Their combined action on the flows could have been considerably amplified by their location in the basin (to the west, near the outlet of the basin), which would favour the rapid delivery of runoff with the least possible losses towards the outlet, thus erasing the effect of size and forest cover of the basin. If this situation persists, there is every reason to believe that an increase in the Nyong's flood will occur within a few years, despite the drop in precipitation.

Thresholds of impervious surfaces beyond which urbanization is supposed to have, at the basin scale, a statistically significant influence on river flows are proposed in the literature, although the figures put forward by the various authors are somewhat contradictory. Some authors place this threshold at 10% of impervious surfaces (Schueler 1994; Booth & Jackson 1997), while others place it at 20% (Brun & Band 2000). The threshold proposed by Yang et al. (2010) is much lower (3–5%). In all cases, by remaining strictly within the limits of the NRB at the Mbalmayo outlet on the one hand, and by considering the impervious surfaces (buildings, roads, crops and bare soils) analyzed using Landsat 8 image of 2020 on the other hand, it appears that this imperviousness rate, which exceeds 10%, is exceeded in the case studied. Under these conditions, it is logical that hydrological alterations such as those observed in this study occur. Studies in sub-Saharan Africa (Amogu et al. 2010; Ebodé et al. 2021b) and elsewhere (Getachew & Melesse 2013; Lee et al. 2018) have already reported the impact of urbanization on runoff.

At the annual time scale, the impact of precipitation is perceptible in the flows of the studied basin. There is a significant synchronous decrease in precipitation and flow in the Nyong basin at the Mbalmayo outlet during the 1970s. Although the dates of ruptures highlighted for the precipitation and the flows do not strictly coincide, the fact that they occur during the same decade, marked by the beginning of an evolution in the same direction (decrease) led us to postulate a possible link between these two variables (precipitation and flows).

In the rainy season, the autumn flows and runoff coefficients increase in the Nyong basin at the Mbalmayo outlet between the 1970 and 1990 decades, while the precipitation does not vary much in the region. This increase seems to be the result of the summer precipitation increasing. The summer precipitation impact in autumn discharges is more noticeable during the decades 2000 and 2010. For these last, there is a joint decrease in summer precipitation, autumn flows and runoff coefficients despite a negligible variation in autumn precipitation (Table 1). Because of these developments, the increase in precipitation during the summer dry season during the decades from 1970 to 1990, considerably reduced the deficit of evaporation and water reserves in the soil at the start of the autumn rainy season, favouring runoff (Liénou et al. 2008). The portion of precipitation that is converted to runoff increases, resulting in an increase in the autumn runoff coefficient during these decades (Table 5). The decrease in summer precipitation during the decades 2000 and 2010 had the opposite effect, which is why we see during the latter a reduction in runoff coefficients, responsible for a reduction in flow.

In the spring, the precipitation and the flows decrease in a statistically significant way during the decade 1970. As with the annual time step, one could expect a possible impact of the precipitation of this season on its discharges. However, the winter precipitation impact also appears noticeable in the spring runoff, between the 1970 and 1990 decades only. The winter precipitation decline over this interval appears to be closely related to that of spring runoff. The rates of change in these two variables during this interval are consistent (decrease in both cases) (Figure 4). Liénou et al. (2008) already suggested this influence of winter precipitation over the same period. The authors argue that from the 1970s, the decline in winter precipitation created a water deficit (evaporation and soil water reserve) in the basin at the start of the first spring rainy season. This winter deficit means that a part of the precipitation received during the spring first fills this water deficit, and therefore the fraction that generates the runoff is reduced. Then, this results in low runoff for the same average precipitation levels during the spring, which explains the decrease in the runoff coefficient (Figure 4). In the decades 2000–2010, on the other hand, the impact of winter precipitation in spring runoff is not visible (Table 5). Over this interval, the winter precipitation impacts the winter runoff itself. It is visible during the 2010s, during which a lesser drop in winter precipitation is concomitant with an increase in the runoff for the same season. The lesser impact of winter precipitation on the spring flows of the Nyong between the 2000 and 2010 decades, has been already suggested in the literature (Ebodé et al. 2020). However, it should be noted that, unlike the Nyong basin, the increase in Winter precipitation observed in the region between the decades 2000 and 2010 is visible in the Spring flows from basins located further south, such as the Ntem and the Ogooue (Ebodé et al. 2021b).

To identify the parameters having a major influence on the outputs of the model, a sensitivity analysis was carried out from the daily flows observed. Initially, four parameters relating globally to runoff and groundwater were identified as being the most sensitive in calibration (Table 8).

The second simulation (2009–2011) shows relatively better results than the first in calibration, which is observed by a clearer synchronism between the flow curves compared to the daily and monthly time steps (Figures 7–9). For daily flows, the model performance is confirmed by a correlation coefficient and a Nash index of 0.89 and 0.81 (Figure 7). As for the monthly flows, they show a correlation coefficient and a Nash index of 0.94 and 0.86 (Figure 9). The overestimation of flows by the model is slightly greater than in the case of the first simulation (Figures 7 and 9).

In validation, the model is slightly more efficient in the daily and monthly flows simulation during the first period (Figures 7–9). The correlation coefficient and the Nash index calculated for the daily flows are 0.86 and 0.81 (Figure 7). Those obtained for the monthly flows are 0.9 and 0.79 (Figure 9).

For the second period, on the other hand, the agreement between the observed and simulated flow curves is less clear than during the calibration period (Figures 7–9). The criteria functions retained attest to satisfactory performance, but slightly less improved than for the first period. The correlation coefficient and the Nash index show values of 0.81 and 0.68 for the daily flows (Figure 7). Their values are 0.86 and 0.73 for monthly flows (Figure 9).

This work confirms the ability of the SWAT model to reproduce flows satisfactorily in the equatorial region in a situation of land cover change. Good results were already obtained from the SWAT model elsewhere. For example, Arnold & Allen (1996) have used observed data from three watersheds, ranging in size from 122 to 246 km², to successfully validate flows simulated from SWAT. Other authors (Arnold et al. 1999) have also successfully evaluated the ability of the model to reproduce flows in the Gulf of Texas over basin sizes between 2,253 and 304,260 km².

To highlight the evolution of the future climate, the averages of the corrected precipitation and temperature series for each of the two RCMs (CCCma and RCA4), for each of the two emission scenarios (RCP4.5 and RCP8.5) and for each future period (2022–260 and 2061–2100) were compared with those of the historical period (2009–2014).

In their study of the Finchaa watershed in Ethiopia, Dibaba et al. (2020) highlighted a decrease in rainfall from the RCA4 model in the near (2021–2050) and distant (2051–2080) future. This study, however, notes a decrease in precipitation for the same model in the near future only (2022–2060). An increase is projected by the model in the distant future (2061–2100).

The CCCma model predicts a decrease in precipitation in the NRB for both periods and under the two scenarios considered (Table 9). For the first period, the decreases projected to be, respectively, –28.7 and –28.4% under the RCP4.5 and RCP8.5 scenarios. For the second, respective decreases of –25.5 and –15.2% are forecast under the RCP4.5 and RCP8.5 scenarios (Table 9). This decrease in precipitation will mainly result from that of the summer and autumn precipitation (Figure 10). Under the RCP8.5 scenario, the 1940s is the only decade for which the deficit seems more severe compared to the others, three decades present an apparent deficit under the RCP4.5 scenario including the 2020s, 2040 and 2070s (Figure 11).

The two models (RCA4 and CCCma) project an increase in maximum and minimum temperatures, regardless of the period and the scenario considered (Figure 10 and Table 10). In general, this temperature rise will gradually increase over the decades, and only peak towards the end of the century (Figure 11). The RCP8.5 scenario also appears to be the one for which the increase is greater, especially after the 1960s (Table 10 and Figure 11). Under the RCP8.5 scenario, the average maximum temperature of the 2100 decade exceeds 35°, regardless of the model considered. However, it does not exceed 32° under the RCP4.5 scenario (Figure 11). The observation is practically the same for minimum temperatures. Those predicted by the two models under the RCP8.5 scenario are generally about 2° higher than those predicted under the RCP4.5 scenario during the last two decades of the century. Seasonally, both models predict a larger increase in winter and spring minimum temperatures. For maximum temperatures, on the other hand, the expected increase is almost regular throughout the year (Figure 10).

As is the case in this work, several studies dealing with the impact of climate change on resources (Kingston & Taylor 2010; Basheer et al. 2015) have already noted in the hydrological units studied, a gradual increase in future maximum and minimum temperatures under the different scenarios retained.

Four classes of land cover having a direct link with the runoff have been identified in the Nyong basin (forest; buildings and roads; bare soils, savannahs and crops; water bodies) (Figure 2). Land cover forecasts for the periods 2040 and 2080 were made based on the changes observed between 1978 and 2005. 2005 is the reference year in relation to which the future land cover changes are assessed.

Changes in surface conditions are noted over the 2005–2040 and 2005–2080 intervals (Figure 2 and Table 2). Between 2005 and 2040, there will be a decrease in forest areas of –912.72 km² (Table 2). Conversely, we will note an increase in buildings and bare soil of +113.67 and +801.92 km², respectively (Table 2). Between 2005 and 2080, similar developments to the previous ones are supposed to occur. A decrease in forests (–925.16 km²) and an increase in buildings (+122.17 km²) and bare soils (+805.85 km²) are expected (Table 2).

Predictions from climate models (precipitation and temperature) were integrated into the SWAT model to get an idea of possible future changes in flows in the NRB.

With regard firstly to the RCA4 model, the results show that a change in precipitation in line with its forecasts will cause an increase in runoff, whatever the period and the scenario taken into account. In general, this increase will be greater during the spring, and the decades at the end of the century will be the wettest (Figures 10 and 11). The changes (increases) noted during the second period (2061–2100) are consistent with those observed for rainfall under the two scenarios (RCP4.5 and RCP8.5). We could also see an impact of climate change on flows during this period and under these two scenarios. However, the climate forecasts of this model do not explain the increase noted during the first period (2022–2060) under the two scenarios. This increase is indeed concomitant with a decrease in precipitation and an increase in temperature. The analysis of the impact of anthropization and its corollaries which will be done later in this part of the paper will allow us to see more clearly on this subject.

About the CCCma model, a change in precipitation in line with its forecasts will also cause an increase in runoff, regardless of the period and scenario considered (Table 9). This increase will mainly concern the autumn and spring months (Figure 10). The last two centuries decades will be the most affected (Figure 11). The impact of climate change is not perceptible on the flows simulated from the outputs of this model for all the periods and scenarios taken into account. This again raises questions about the possible impact of anthropization and its corollaries on the future flows of the NRB.

To ensure better management of water resources, the impact of climate change on them has already been addressed in several studies around the world (Notter et al. 2013; Wagena et al. 2016; Danvi et al. 2018; Duku et al. 2018). In these various studies, the outputs of the climate models were integrated into the SWAT model to simulate future flows and get an idea of their possible evolution. As is the case in this study, some of them also predict identical trends in the evolution of future precipitation and runoff (Beyene et al. 2010; Basheer et al. 2015; Dibaba et al. 2020).

Explanations for certain changes in future flows in the NRB could not be found in those of precipitation and temperature. In this case, the evolution of the flows simulated for the first period (2022–2060) from the outputs of the RCA4 models under the two scenarios, and those simulated for all periods and all scenarios from the outputs of the CCCma model. In these different cases, an increase in flows has been noted, while rainfall decreases and temperature increases, which should rather cause a decrease in flows, due to a reduced supply of groundwater and increased evapotranspiration. This raises questions about the evolution of surface conditions, knowing that an increase in impervious areas, for example, increases runoff and therefore flows.

Let us first recall that in this study, the land cover maps used to simulate flows for the historical period (used to assess changes in future periods) and future periods 2022–2060 and 2061–2100 are respectively those of 2005, 2040 and 2080. Compared to the 2005 situation, impervious areas (buildings, roads, bare soil, crops, etc.) are expected to increase in 2040 and 2080 (Table 2). Because of the increase in runoff that it will cause, this increase in impervious areas is the only possible explanation for the increase in future flows in the various cases where a drop in precipitation is expected.

The objective of this article was to study the impact of climate change and anthropization on flows in the Nyong watershed during the contemporary (1950–2017) and future (2022–2100) periods. During the contemporary period, the Nyong basin experienced a decrease in discharges of around −5% during the 1970s, which would result from a decrease in precipitation that occurred during the same decade. Model RCA4 is the only one of the two retained that predicts an increase in precipitation. This increase is assumed to occur during the second future period (2061–2100) under both scenarios (RCP4.5 and RCP8.5). However, in this future deficit context, flows will experience an increase regardless of the climate model and the period considered. This increase in runoff will be greater in the case of the RCA4 model, which forecasts smaller decreases in precipitation during the first future period (–6 and –9.6% under the RCP4.5 and RCP8.5 scenarios), and an increase during the second. Because of the increase in runoff they will cause, the changes in land cover expected in the basin (reduction of forest and increases in impervious areas) constitute the only possible explanation for the increase in flows in cases where a decrease in rainfall is forecast.

Data cannot be made publicly available; readers should contact the corresponding author for details.

The authors declare there is no conflict.