Impact of anthropogenic changes and rainfall variability on river discharge in tropical central Africa

doi:10.21203/rs.3.rs-2195653/v1

Download PDF

Research Article

Impact of anthropogenic changes and rainfall variability on river discharge in tropical central Africa

https://doi.org/10.21203/rs.3.rs-2195653/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 02 Apr, 2024

Read the published version in Water Supply →

Version 1

posted

You are reading this latest preprint version

Climate variability and anthropogenic changes are key factors that influence river flow variability. However, understanding the simultaneous impact of these factors on flow variability remains limited in most parts of the world including central Africa. To shed light on this issue, the objective of this study is to investigate the effects of rainfall variability and anthropogenic changes on river discharge in the Benoue and the Logone river basins over the last 7 decades (1950–2018). To achieve this goal, hydro-meteorological data from these basins were analyzed using the Pettitt test. Likewise, land use change in these basins was also analysed using supervised classifications of Landsat satellite images over different periods (1987 and 2018). Our results show that the Sahel drought that occurred towards the end of the 1960s affected the Benoue and Logone basins, with remarkable persistence in streamflows. Negative rupture was observed in the hydro-meteorological timeseries of these basins at annual time step in 1970-71. The deficits associated with this rupture are estimated at -7% for rainfall, and − 28% for river flows. The wet season shows similar developments. However, from the 1990s onward, there has been a significant increase in the mean annual flows of the Benoue river, which coincides with that of the rainfall during the same decade. This increase over the recent decades could also be expected in response to an increase in impervious surface areas (ISA) in the catchment area, which could compensate the deficit generated by the post-1990s rainfall deficit through increasing in runoff. Since the filling of the Lagdo dam in 1983, an increase in all ranges of minimum flow, as well as an increase in the variability of extreme flows, has been detected. These results could be used for long-term planning of water demand and use in these basins, as well as for improving future simulations of the flow of these rivers. It also shows that due to the non-linearity of the rainfall-runoff relationships, it is very important to maintain observation networks operational, as the current situation might still evolve in the near future, requiring new adaptations for social response.

Hydrology

anthropization

Benoue

climate variability

dam

Land use modes

Logone

The impact of climate change and variability on surface water resources has been highlighted by several studies in tropical Africa (Liénou et al. 2008; Panthou et al. 2018; Getahun et al. al., 2020; Ebodé et al. 2021a; Kpoumié et al. 2022; Ebodé 2022a; Ebodé 2022b; Ebodé et al. 2022b). Most of these works are based on the detection of ruptures in the hydroclimatic timeseries. Their results confirm in the case of West Africa that the 1970s appears to be the main period of discontinuity marking the onset of Sahel drought in the region (Mahé et al. 2001; Nka et al. 2015; Bodian et al. 2020). In forested Central Africa, the most obvious fluctuations in runoff and rainfall have been observed at seasonal time steps (Buisson 1985; Ebodé et al. 2021b; Ebodé et al. 2022a). Liénou et al. (2008) demonstrated in the case of three equatorial rivers (Ntem, Nyong and Kienke), that the most important climatic variations leading to changes in flow result from variations in precipitation in the dry seasons. Most of the studies attributed the sensitivity of watersheds to changes in precipitation which in turn induces a significant deficit in soil moisture storage leading to a drop in runoff. Conversely, an increase in precipitation increases soil moisture storage during the dry season and therefore enhances runoff generation at the onset of the rainy season. It, therefore, appears that the variability of the flow regimes of equatorial rivers can not be better appreciated without considering precipitation variability.

In addition to climate variability, understanding the hydrological dynamics of hydro-systems in the Central Africa region, has to also consider anthropogenic factors such as dam construction and changes in land use which have the potential to influence runoff. For example, several studies in West Africa have highlighted the impact of land use change on runoff (Kouassi 2007; D'Orgeval and Polcher 2008; Amogu et al. 2010). This impact most often results in an increase in runoff in a context of decreasing rainfall. In forested Central Africa, only a few studies (Ebodé et al. 2020, Ewane et al. 2020) have attempted to investigate this issue. There is therefore an urgent need for more studies in the sub-region to investigate the impact of climate variability and anthropogenic changes on runoff variability. This is especially so considering the rapid increase in population observed in the in the sub-region (BUCREP 2011). Similarly, the Lagdo dam built on the Benoue river between 1977 and 1982, and filling since 1983, may also have contributed to modify the flow regime of this river. In other regions of the world, the construction of a dam is generally followed by an increase in minimum flows, an increase in the variability of extreme flows (maximum and minimum) and a modification of the regime (Crepet 2000; Assani et al. 2007; Amoussou et al. 2012; Nzango 2018). The impact of dam construction on runoff variability is still poorly understood in the Central Africa sub-region despite an increasing of dams for power generation.

This study focused on the analysis of rainfall/flow relationships over the recent period in the Benoue and Logone basins, using updated hydrometeorological timeseries. This study is crucial considering that the two basins are vulnerable to climate change where water scarcity could become a source conflict. Furthermore, insufficient hydroelectric production in the Benoue basin has been attributed to recent hydroclimatic changes mainly in the dry season, leading to frequent power cuts and load shedding. It is hoped that the results from this study will enhance the debate on these issues. In addition, the results from this study could help in the long-term planning of water resources, as well as improving future simulations of the flow in these rivers. The main objective of this article is therefore to document the type and extent of regime changes in the Benoue and Logone due to climate change and anthropization.

2.1. Study framework

Covering areas of 61810 km² in Garoua and 67012 km² in Bongor respectively (Table 1), the Benoue and Logone basins are located in the northern tropical zone, between shared by Cameroon, Chad and the Central African Republic (roughly between 6 °5'–11°N and 13°–17°E). The largest parts of the Benoue and Logone basins are located in Cameroon and Chad respectively (Fig. 1). These basins are subject to a tropical Sudanese climate recording an average of precipitation of above 1000 mm/year (Table 2). This climate is characterized by a dry and hot season extending on average from November to April, and a wet and less hot season from May to October during which most of the precipitation occurs. The relief of these basins are rugged. There are mountain ranges including the Adamaoua and Poli with altitudes varying between 1000 m and 1500 m, but also plateaus and several areas of undulations of altitudes between 200 m and 800 m. These reliefs are essentially underlain by granites and crystalline schists, in vertical beds and generally oriented South-West, North-East (Rodier 1953). Savannas and gallery forests are the main forms of vegetation in these basins (Rodier 1953).

Table 1

Gauging stations and corresponding characteristics in the studied basins
Station	Watercourse	Geographical coordinates		Drainage area (km²)	Data series lenght
		Latitude N	Longitude E		Rainfall	Discharge
Garoua	Benoue	9.29	13.4	61810	1950-51 to 2015-16	1950-51 to 2015-16
Bongor	Logone	10.26	15.41	67012	1950-51 to 2017-18	1952-53-2017-18

Table 2

Statistics of precipitation and average flows (annual and seasonal) of the basins over their respective study periods, but also on both sides of the identified ruptures
Periods	Interannual		Year of rupture	Mean		CV (%)
	Mean	CV (%)		Before rupture	After rupture	Before rupture	After rupture
Precipitation
Benoue
Annual	1050	0.1	1970-71	1114	1020	0.06	0.1
Wet season (May-October)	989	0.09	1970-71	1040	966	0.06	0.1
Dry season (November-April)	61	0.43	1970-71	74	54	0.33	0.47
Logone
Annual	1221	0.09	1969-70	1298	1185	0.06	0.08
Wet season (May-October)	1117	0.09	1970-71	1175	1090	0.08	0.08
Dry season (November-April)	104	0.31	1969-70	125	92	0.25	0.3
Discharges
Benoue
Annual	320.4	0.39	1970-71	388	275	0.22	0.46
Wet season (May-October)	563	0.4	1970-71	724	454	0.22	0.45
Dry season (November-April)	77	0.63	1984-85	41	151	0.43	0.3
Logone
Annual	496	0.33	1970-71	616	441	0.1	0.38
Wet season (May-October)	786	0.32	1970-71	932	720	0.08	0.37
Dry season (November-April)	197	0.47	1970-71	269	163	0.23	0.57
CV : Coefficient of variation

2.2. Data source

Several datasets (hydrometeorological and land cover) were used in this work. Available for the periods 1950–2015 and 1952–2017, the flow series for Benoue and Logone come from the SIEREM database (Environmental Information System on Water Resources and Modeling). That of Benoue has significant gaps between 1998 and 2009. The gaps contained in that of Logone are mainly located during the intervals 1980–1983, 1994–1995 and 2008–2012.

The rainfall data used in this study are from CRU (Climate Research Unit) of the University of East Anglia in the United Kingdom. These data have been available since 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° x 0.25 °. Precipitation and temperature data from CRU have been used to validate CMIP models in Logone basin which is a sub-basin of the Lake Chad basin (Nkiaka et al. 2018).

The spatial data used for the study of land cover in the Benoue basin are Landsat 8 satellite images from January to March 2018 and Landsat TM from January to March 1987, corresponding to path/row 185/53, 185 /54, 185/55, 184/53, 184/54, 184/55, 183/53 and 183/54. All of these images are made available to the general public free of charge by the National Aeronautics and Space Administration (NASA), via the US Geological Survey website (https://earthexplorer.usgs.gov/), in GeoTIFF format. The downloaded images taken during the dry season were preferred to those of the rainy seasons since they are less affected by cloud disturbances.

The demographic data used in this work are those of the population censuses in Cameroon (1976, 1987, 2005 and 2018) and Chad (1993 and 2009). These data were collected from the Central Bureau of Censuses and Population Studies.

2.3. Data analysis

The analysis of precipitation, average river discharge and runoff coefficients (at annual and seasonal time steps) was carried out using Pettitt (Pettitt 1979) and Mann-Kendall (Mann 1945; Kendall 1975; Yue et al. 2002) tests, at the 95% significance level.

The Pettitt test seems to be the most suitable for the analysis of incomplete series such as ours because it can separate the series into two periods with an overall distinct behavior, which avoids the detection of false discontinuities which can sometimes be observed with other tests such as Hubert segmentation. Its principle consists of dividing the studied series (of N size) into two sub-samples of sizes m and n, respectively. We then calculated the sum of the ranks of the elements of each sub-sample in the total sample. A statistical study is then carried out based on the two sums thus determined, then it is tested according to the hypothesis that the two subsamples do not belong to the same population. The Pettitt test is non-parametric and derives from that of Mann Whitney. The absence of a discontinuity in the series (Xi) of size N constitutes the null hypothesis. Its implementation supposes that for any instant T between 1 and N, the time series (Xi) 1 to t and t + 1 to N belong to the same population. The variable to be tested is the maximum in the absolute value of the variable Ut, N defined by:

$$\mathbf{U}\mathbf{t},\mathbf{N}= {\mathbf{Ʃ}}_{\mathbf{i}=1 }^{\mathbf{t}}{\mathbf{Ʃ}}_{\mathbf{i}=\mathbf{t}+1 }^{\mathbf{N}}\mathbf{D}\mathbf{i}\mathbf{j}$$

where Dij = Sign (Xi - Xj) with: sign (x) = 1 if x > 0, 0 if x = 0 and − 1 if x < 0. If the null hypothesis is rejected, an estimate of the date of discontinuity is given by defining the maximum in the absolute value of the variable Ut, N.

In addition to the Pettitt test, the Mann-Kendall test was also used to analyze precipitation, average river discharge and runoff coefficients (at annual and seasonal time steps). This test is based on the test statistic “S” defined as follow:

$$\mathbf{S}=\sum _{\mathbf{i}=1}^{\mathbf{n}-1}\sum _{\mathbf{j}=\mathbf{i}+1}^{\mathbf{n}}\mathbf{s}\mathbf{g}\mathbf{n}(\mathbf{x}\mathbf{j}-\mathbf{x}\mathbf{i})$$

Where the xj are the sequential data values, n is the length of the data set, and sgn = (θ) if θ > 1, 0 if θ = 0 and − 1 if θ < 0. There is no significant trend in the series analyzed when the calculated p-value is above the chosen significance level.

To assess the behavior 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 (Richter et al. 1998). It uses daily discharge values and produces several important statistics. Only four of them were considered essential for this study, among which are 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 the difference.

Landsat images were classified using the supervised maximum likelihood classification, using SNAP software (open access). This enabled us to perform a diachronic analysis of the evolution of land use in the basins studied. This operation was preceded by operations of preprocessing and recognition of objects in the field by photography and GPS (Global Positioning System). Satellite image preprocessing refers to all the process applied to raw data to correct geometric and radiometric errors that characterize certain satellite images. These errors are generally due to technical problems with the satellites and interactions between outgoing electromagnetic radiation and atmospheric aerosols, also called “atmospheric noise”. The atmospheric disturbances are influenced by various factors that are present on the day of acquisition, including weather, fires, and other human activities. They affect all the images acquired by passive satellites including Landsat 4-5-7 and 8. The downloaded Landsat images being orthorectified, the preprocessing involved atmospheric correction of the images and reprojection into the local system (WGS_84_UTM_Zone_32N). For this, neo-channels were created, to increase the readability of the data by enhancing certain properties less obvious in the original image, thus showing more clearly the elements of the scene. Three indices are therefore created, namely: the Normalized Difference Vegetation Index (NDVI, Eq. 1), the brightness index (BI, Eq. 2) and the Normalized Difference Water index (NDWI, Eq. 3). These indices respectively highlight vegetated surfaces, sterile (non-chlorophyllin) elements such as urban areas and water bodies. The formulae used in creating these indices are:

$\mathbf{N}\mathbf{D}\mathbf{V}\mathbf{I}=\frac{\mathbf{N}\mathbf{I}\mathbf{R}-\mathbf{R}}{\mathbf{N}\mathbf{I}\mathbf{R}+\mathbf{R}}$ Eq. 1

$\mathbf{B}\mathbf{I}={({\mathbf{R}}^{2}+{\mathbf{N}\mathbf{I}\mathbf{R}}^{2})}^{0.5}$ Eq. 2

$\mathbf{N}\mathbf{D}\mathbf{W}\mathbf{I}=\frac{\mathbf{N}\mathbf{I}\mathbf{R}-\mathbf{M}\mathbf{W}\mathbf{I}\mathbf{R}}{\mathbf{N}\mathbf{I}\mathbf{R}+\mathbf{M}\mathbf{W}\mathbf{I}\mathbf{R}}$ Eq. 3

where NIR: ground reflectance of the surface in the near-infrared channel; R: ground reflectance of the surface in the red channel and MWIR: ground reflectance of the surface in the mid-wave infrared channel. Due to the fact that the study area extends over several scenes, the enhancement operations were followed by the mosaic of the different scenes used on each date. The use of Google Earth, as well as the spaces sampled from the GPS, made it possible to identify with certainty the impervious areas (buildings, savannas, bare soils, and crops), water bodies (large rivers, lakes and ponds) and forest (secondary, degraded, non-degraded and swampy) of each mosaic. Before the classification, the separability of the spectral signatures of the sampled objects to avoid interclass confusion was assessed by calculating the “transformed divergence” index. The value of this index is between 0 and 2. A value > 1.8 indicates a good separability between two given classes. The different classes used in this study show good separability between them, irrespective of the image considered, with indices > 1.9. The validation of the classifications obtained was carried out using the confusion matrix, making it possible to obtain treatment details to validate the choice of training plots. After validating the land use/land cover maps, the statistical and spatial differences of each class between 1973 and 2018 were evaluated.

3.1. Rainfall evolution

3.1.1. Evolution of annual and seasonal rainfall

Annual rainfall decreases statistically significantly in the Benoue and Logone basins (Fig. 2; Table 3). The Pettitt test detected a rupture in both timeseries in 1970-71. The deficits associated with this rupture are − 8.4% (Benoue) and − 8.7% (Logone) (Fig. 2). The analysis of decadal deviations reveals a drop since the 1970s, which was only interrupted during the 1990s for Benoue. For Logone, in addition to the 1990s, the 2010s also had a slight surplus after 1970 (Table 4).

Rainfall during the wet season essentially follow the same trends as annual rainfall (Fig. 2). Notably, rainfall also show a downward rupture in 1970-71 in both basins, according to the Pettitt test. The deficits recorded following this rupture fluctuate around − 7% (Fig. 2). In the Benoue basin, the 1990s appears to be the only wet decade after the 1970s (Table 4). In the Logone, the 1990s and 2010s are also the only wet decades after the rupture.

As in the wet season, rainfall decreased significantly during the dry season in both basins according to the Pettitt test. A rupture was identified in the series of dry season rainfall in these basins in 1970-71 (Benoue) and 1969-70 (Logone) (Fig. 2). The deficits recorded following this rupture are approximate − 27% (Fig. 2). The drop in dry season rainfall observed during the 1970s was interrupted only during the 1990s in the two basins (Table 4).

A return to wetter conditions during the 1990s has also been reported in West Africa (Alhassane et al. 2013; Descroix et al. 2020). This is consistent with the results of our study, which also show an increase in annual precipitation during the 1990s in the Benoue, and during the 1990s and 2010s in the Logone. The results in the Logone basin are also consistent with previous studies in the basin (Nkiaka et al. 2017). This increase is consecutive to the drought of the 1970s and 1980s. It should however be noted that the increase in rainfall reported in West Africa has been uninterrupted since the 1990s (Descroix et al. 2020).

Table 3

Results of annual and seasonal rainfall, discharges and runoff coefficient analysis using Mann–Kendall’s test. The "+" and "-" signs indicate positive and negative trends respectively. Calculated p-values below the significance level (α = 0.05) indicate that there is a significant trend in the series (in bold).
Variables	Annual		Wet season		Dry season
	p-value	Evolution	p-value	Evolution	p-value	Evolution
Benoue
Rainfall	0.003733	–	0.01143	–	0.004764	–
Discharges	0.05128	–	0.002509	–	0.0002357	+
Ke	0.1577	–	0.00285	–	0.0003201	+
Logone
Rainfall	0.02447	–	0.1261	–	0.01579	–
Discharges	0.007317	–	0.01658	–	0.01152	–
Ke	0.00788	–	0.0577	–	0.4976	–

Table 4

Deviations (%) of the decadal annual and seasonal averages of precipitation, discharges and runoff coefficients compared to their inter-annual means in the basins studied
	Rainfall			Discharges			Ke
Decades	Annual	Rainy season	Dry season	Annual	Rainy season	Dry season	Annual	Rainy season	Dry season
Benoue
1950	6	5,7	11,9	11,3	16,6	-34,2	7	15,3	-39,2
1960	6,3	5,1	26,2	29,2	38,5	-39,3	23	36,5	-48,5
1970	0	-0,4	5,2	-6,5	0,6	-58,5	-4,8	4,7	-41,5
1980	-7,3	-6,9	-13,4	-43,7	-46	-27,3	-40,5	-42,2	2,3
1990	4	3,5	11,9	9,9	-6,4	128,8	8	-6,5	20,5
2000	-5,2	-3,7	-29,7	-	-	-	-	-	-
2010	-6,4	-5,1	-26,6	7,1	-10,6	136,7	16,7	-2,4	242,7
Logone
1950	6.5	5.5	18.2	30	15.6	93.3	25.1	13.2	25.4
1960	6.4	4.9	23.7	19.5	19.3	25.9	14	15.7	0.6
1970	-1.4	-1.1	-4.1	-7.3	-1.7	-25.4	-5.1	0.3	-16.6
1980	-7.8	-6.7	-18.8	-47.3	-43.6	-59.5	-41.5	-38.2	-46
1990	0.7	0.6	2.2	-3.4	-3.2	0.1	-7	-7.7	16.7
2000	-5.4	-4.6	-12.8	-13.6	-12.3	-15.1	-6.1	-5.6	0.5
2010	1.1	2.4	-13	17.7	21	9.9	17.3	19.6	15.1

3.1.2. Spatial evolution of precipitation

The results of 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 Benoue and Logone during the period after this decade. The question therefore arises as to whether the changes in annual and seasonal rainfall observed around the 1970s were accompanied by a modification in its spatial distribution.

Figure 3 highlights the spatial distribution of average precipitation over the periods 1950–1979 and 1980–2017. As regards annual rainfall, it accounts for a decrease in rainfall in the two basins from North to South between the two periods compared. The limits of the rainfall classes retained indeed progress towards the south of the two basins (Fig. 3). Similarly, if we stick to the 1000 mm isohyet, for example, it is located towards Sekande (to the north of the basin) during the first period defined, but it shifted to the middle zone of the basin during the second period (Fig. 3).

About rainfall during the rainy season, we also observe, as for annual rainfall, a decrease in rainfall in the basins, from North to South between the periods compared. The limit of the lowest rainfall class (< 950 mm) was located in the north of the basin during the first period, but it shifted to middle part during the second (Fig. 3). Also, rainfall > 1250 mm was recorded mostly in the southern zone of both basins during the first period, but which almost disappeared during the second period and mostly confined to the extreme southern part of the Logone basin (Fig. 3).

In the case of dry season rainfall, we note the same evolution as in the previous cases. Located in the middle zone of the basin during the first period, the lowest rainfall class (< 70 mm) evolved towards its southern part during the second (Fig. 3). Similarly, the rainfall > 190 mm observed in the south of the Logone basin during the first period completely disappeared in the area investigated during the second (Fig. 3).

3.2. Evolution of flows

3.2.1. Evolutions of mean flows

Average flows are analyzed through annual and seasonal modules. The average annual flows of the Benoue and the Logone recorded a downward rupture in 1970-71. The deficits observed following this break are respectively − 29.1% and − 28.4% (Fig. 2). Despite the general decrease of flows observed during the 1970s, an increase has been noted from the 1990s on the Benoue basin, and during the 2010s on the Logone basin (Table 4).

Average wet season flows also decrease statistically significantly according to the Pettitt test in 1970-71 (Fig. 2). The rates of variation recorded following this drop are − 37.3% (Benoue) and − 22.7% (Logone). The decadal deviations reveal for the flows of this season an uninterrupted decrease since the 1970s on the Benoue basin. For the Logone basin, on the other hand, the trend was reversed during the 2010 decade. This decade recorded a positive deviation from the climatological mean, which is + 21% (Table 4).

The average flows of the dry season evolved in opposite trends on the two basins studied. Flow in Benoue basin has increased since 1984-85 (+ 263.8%), while Logone flow has decreased since 1970-71 (-39.4%) (Fig. 2). In the Benoue basin, the increase in flow during the dry season has gradually amplified since the 1990s (Table 4).

The Logone regime has undergone a change which results in a significant drop in the August-November flood. On the Benoue, beyond the decrease of this same flood, its regime also recorded a significant increase in the low water levels that follow (Fig. 4).

This study highlights a significant decrease in flows in tropical Central Africa at the beginning of the 1970s. A large number of studies have obtained a similar result in West Africa (Cisse et al. 2014; Nka et al. 2015; Descroix et al. 2018) and central (Conway et al. 2009; Mahé et al. 2013; Ebodé et al. 2020).

3.2.2. Extreme flows

Extreme flows were studied through maximums and minimums on consecutive days (1, 3, 7, 30 and 90 days).

3.2.2.1. The maximums

First of all, it is worth noting that the period after the break corresponds to a deficit phase in both basins. The dry nature of this period is also observed in the maximum flows of the basins. There is a decrease in all the ranges of maximum flow rates evaluated (Table 5). The decreases observed in the Benoue are however much greater than those observed in the Logone. In the case of Benoue, the decreases are between − 40% (90-day maximum) and − 46.2% (1-day maximum). In the Logone, the decreases observed are between − 14.8% (1-day maximum) and 23.9% (90-day maximum) (Table 5). The maximums are earlier on average by 4 days on the Benoue and 8 days in the Logone after the rupture (Table 5).

3.2.2.2. The minimums

The minimum flows evolve in the same direction as the maximum in the Logone basin. The decreases recorded are between − 67.7% (1-day minimum) and − 70.1% (90-day minimum) (Table 5). Contrary to the Logone, the minimums increase on the Benoue at rates ranging from + 1149% (90-day minimum) to + 9633% (1-day minimum) (Table 5). The dates of occurrence of the minimums are slightly later on the Benoue after the rupture (+ 5 days on average). On the Logone, the average occurrence date of the minimums changed significantly, going from 306 before the rupture to 111 after (Table 5).

Table 5

Statistics relating to the maximum and minimum flows of Benoue and Logone basins before and after the break identified in their respective series
IHA statistics	Means (m³/s)		Cv (%)		Change
IHA statistics	Before break	After break	Before break	After break	m³/s	%
Benoue
Minimum flows
1-day minimum	0.6	58.4	1.13	0.96	57.8	9633
3-day minimum	0.7	59.7	0.9	0.96	59	8428
7-day minimum	0.8	61.6	0.85	0.95	60.8	7600
30-day minimum	1.4	66.6	0.84	0.92	65.1	4340
90-day minimum	6.1	76.2	0.6	0.89	70.1	1149
Maximum flows
1-day maximum	2957	1589	0.34	0.58	-1368	-46.2
3-day maximum	2851	1540	0.33	0.58	-1311	-46
7-day maximum	2608	1436	0.32	0.57	-1172	-45
30-day maximum	1979	1143	0.32	0.57	-836	-44.2
90-day maximum	1274	763.6	0.32	0.54	-510.4	-40
Average Julian dates
of minimum	113	118
of maximum	249	245
Logone
Minimum flows
1-day minimum	147.5	47.6	0.46	0.59	-99	-67.7
3-day minimum	148.6	48	0.46	0.59	-100.6	-67.7
7-day minimum	153.8	48.4	0.47	0.59	-105	-68.2
30-day minimum	182.4	53.3	0.49	0.58	-129.1	-70.8
90-day minimum	258	77	0.53	0.72	-181	-70.1
Maximum flows
1-day maximum	1956	1667	0.28	0.52	-289	-14.8
3-day maximum	1948	1654	0.28	0.51	-294	-15.1
7-day maximum	1930	1628	0.28	0.50	-302	-15.6
30-day maximum	1833	1496	0.27	0.49	-337	-18.4
90-day maximum	1534	1168	0.27	0.48	-366	-23.9
Average Julian dates
of minimum	306	111
of maximum	262	254

The evolutions of the flows in both basins studied require explanations which can be found by examining them in the light of natural and anthropogenic changes (rainfall variability and anthropogenic changes).

4.1. The role of rainfall in the flows of the basins studied

The average annual rainfall and flow rates in Benoue and Logone showed a downward rupture in 1970-71 (Fig. 2). The impact of rainfall in the evolution of the flows of these basins, therefore, seems quite clear at this time step. We note that the two variables recorded statistically significant decreases over the study period (Fig. 2). At this time step, the stationarity of the hydrological regime, and that of precipitation, follows the general logic which admits that hydrological ruptures occur either after or at the same time as rainfall ruptures (Fig. 2).

As with the annual time step, the influence of rainfall in the wet season is also perceptible in the runoff of the same season in the basins investigated (Fig. 2). We then note, in the respective timeseries of rainfall and discharges, a downward rupture in 1970-71, making it possible to consider a possible role of the decrease in rainfall on flows.

4.2. The impact of the dam on the flows of the Benoue basin

In the Benoue basin, since the impoundment of the Lagdo dam (1983), there has been a significant increase in all the minimum flow ranges on consecutive days (Table 5). These increases range from + 1149% (90-day minimum) to + 9633% (1-day minimum). Beyond the average delay of five days noted in the appearance of the minimums (Table 5), an increase in the interannual variability is also noted for the two categories of extreme flows evaluated (maximums and minimums), which seems much greater for the maximums (Table 5). All of these changes seem to be linked to the operation of the dam. This work makes it possible to store water in the wet season and release it in the dry season, in order to stabilise hydroelectric production and irrigation downstream during this season, thus modifying the natural behaviour of the river. The increase in the volumes released (also perceptible through the minimums) constantly increasing over time (Fig. 4), reflects an increasingly significant increase in the population and its growing energy needs which must be met permanently, including in the dry season. The Lagdo dam considerably affected the Benoue regime. Since its establishment, we note the development of a permanent character of the flow throughout the hydrological year, as well as the decline of the flood from August to October (Fig. 4). The increase in low flows is identified as a major consequence of hydraulic developments (dams), as previously proposed by Amoussou et al. (2012).

4.3. The contribution of changes in land-use patterns

It is certainly true that runoff in the Benoue has decreased statistically since the beginning of the 1970s, due to the drop in precipitation that began during the same decade. However, the analysis of decadal deviations confirms a significant increase in flows since the 1990s. The deviations noted are + 9.9% for the 1990s and + 7.1% for the 2010s (Table 4). It is quite possible to associate the increase in runoff in the 1990s with that of rainfall. On the other hand, nothing explains the maintenance of this increase over recent decades, since the rainfall decrease during this period. The calculated deficits are − 5.2% for the 2000s and − 6.4% for the 2010s (Table 4). All this leads us to question the evolution of land-use in the basin, which could explain such evolutions. In the same vein, we could also wonder whether the increase in runoff from the Logone during the 2010s (+ 17.7%) is solely the consequence of the modest increase in rainfall (+ 1%) (Table 4).

Classifications carried out using Landsat images from 1987 and 2018 show significant changes in the land use classes with a direct link to runoff in the Benoue and Logone basins (Fig. 5). In 31 years, a -25.7% and − 20% decrease in forest areas have been recorded, respectively for the Benoue and Logone basins (Table 6). Conversely, impervious areas (buildings, roads, bare soil and cultivated areas) and water bodies (lakes, ponds and large rivers) increased (Table 6). These results are consistent with population statistics. Significant population growth has been noted for some agglomerations in these basins over the intervals generally between 1970s-1990s and 2000s-2010s (Table 7). The highest average annual growth rates were recorded for Mokolo and Guider, while the lowest was for Rey Bouba and Tignere (Table 7). Some authors in Central (Ebodé et al. 2021b; Ewane et al. 2020), West Africa (Kergoat et al. 2007; Leblanc et al. 2008) and elsewhere (Gebrelibanos et al. 2015; Belay et al. 2019) have made similar observations relating to changes in land-use.

Table 6

Evolution of the land use modes in the studied basins between 1987 and 2018
Land Use Modes	Benoue				Logone
			Change				Change
	1987	2018	km	%	1987	2018	km	%
Built and Road	65	873	808	1243	43	705	662	1539.5
Cultivated area	449	2356	1907	424.7	2786	3048	262	9.4
Forest	51664	38362	-13302	-25.7	51712	41319	-10393	-20
Water	324	514	190	58.6	554	570	16	2.8
Bare soil and Savannah	14515	24806	10291	70.8	11912	21370	9458	79.3

Table 7

Evolution of the population of some main agglomerations of the studied basins
Agglomerations	Date 1 : Number of inhabitants	Date 2 : Number of inhabitants	Evolution (%)
Benoue
Garoua	1976 : 63900	2018 : 900000	6.5
Rey Bouba	1987 : 46547	2005: 109438	4.9
Mokolo	1976 : 2069	2005 : 242274	17.8
Tignère	1987 : 19165	2005 : 34167	3.2
Guider	1976 : 17197	2005 : 235996	9.4
Logone
Ngaoundere	1976 : 38800	2005 : 189800	5.6
Bongor	1993 : 20357	2009 : 30518	2.5
Moundou	1993 : 99530	2009 : 142462	2.2

We can suggest that the increase in flows, associated with the increase in rainfall in the 1990s, was immediately relayed by the increase in impervious areas in the basin. This increase in impervious areas is the most relevant factor that can justify the increase in flows during the 2000s and 2010s, since the rainfall decreased again during these decades. These formations (impervious areas) act in the direction of an increase in runoff and, consequently, flows. Studies carried out elsewhere in the world have reached similar conclusions (Coe et al. 2009; Dias et al. 2015). 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 et Jackson 1997), while others place it at 20% (Brun et 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 Benoue basin at Garoua outlet on the one hand, and by considering the urban surfaces (buildings, roads and bare soils) analysed using Landsat 8 image of 2018 on the other hand, it appears that this imperviousness rate, which exceeds 20%, 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 et al. 2013; Lee et al. 2018) have already reported the impact of urbanization on runoff. This work has shown an increase in average flows, following an increase in impervious areas in the studied basins.

It is also possible to review causes of the recent increase in flows observed in the Benoue basin, and in particular its absence during the wet season and in the maximum. This could be attributed to the significant increase in the volume of water stored during the rainy season, and released in the dry season (Fig. 4). This increase in rainy season storage in Lagdo implies that the excess water caused by runoff is not noticeable in the rainy season flows and the maximums at Garoua outlet.

In the case of Logone, it seems that changes in land-use patterns are partly linked to the increase in runoff during the 2010s. The + 1% increase in precipitation recorded during this decade cannot be the only cause of the significant increase in discharges (+ 17.7%) noted.

From the 1970s, the Benoue and Logone basins experienced a general decrease in rainfall which was only rarely interrupted. This change in the evolution of the rainfall has led to a modification of the flows these rivers. Thus, since 1970-71, there has been a decrease in average runoff during the rainy season and annual, with deficits of more than − 7%. Although annual flow decreased since the 1970s, average annual flows have witnessed an increase since the 1990s in the Benoue and 2010 in the Logone. The increase in runoff in the 1990s is attributable to abundant rainfall in the Benoue basin, but its stability in recent decades seems to be the consequence of changes in land-use patterns, since, during this period, the rainfall decreased. The rise in the minimums observed on this river, as well as the accentuation of the variability of extreme flows, seems to be associated with the filling of the Lagdo dam since 1983. This structure allows to store water in the rainy season and release during the dry season, to continue to produce hydroelectricity and irrigate the fields located downstream. The increase in storage to meet the needs of a growing population could also be the cause of the reduction in floods in Garoua.

Funding: This research received no external funding.

Conflicts of interest: the author declares no conflict of interest.

Alhassane, A., Salack, S., Ly, M., Lona, I., Traore, S.B. and Sarr, B. (2013) Evolution des risques agroclimatiques associés aux tendances récentes du régime pluviométrique en Afrique de l’Ouest soudano-sahélienne. Sécheresse 24: 282–93. http://dx.doi.org/10.1684/sec.2013.0400
Amogu, O., Descroix, L., Yéro, K.S., Le Breton, E., Mamadou, I., Ali, A., Vischel, T., Bader, J.C., Moussa, I.B., Gautier, E., et al. (2010) Increasing river flows in Sahel? Water 2: 170–199.
Amoussou, E., Camberlin, P. and Mahé, G. (2012) Impact de la variabilité climatique et du barrage Nangbéto sur l'hydrologie du système Mono-Couffo (Ouest Afrique). Hydrological Sciences Journal 57 (4): 805–817. https://doi.org/10.1080/02626667.2011.643799
Assani, A.A., Lajoie, F. and Laliberté, C. (2007) Impacts des barrages sur les caractéristiques des débits moyens annuels en fonction du mode de gestion et de la taille des bassins versants au Québec. Revue des sciences de l'eau 20 (1): 127–146. https://doi.org/10.7202/015741ar
Belay, T. and Mengistu, D.A. (2019) Land use and land cover dynamics and drivers in the Muga watershed, Upper Blue Nile basin, Ethiopia. Remote Sensing Applications: Society and Environment 15 (100249). https://doi.org/10.1016/j.rsase.2019.100249
Bodian, A., Diop, L., Panthou, G., Dacosta, H., Deme, A., Dezetter, A., Ndiaye, P.M., Diouf, I. and Vischel, T. (2020) Recent Trend in Hydroclimatic Conditions in the Senegal River Basin. Water 12 (436). https://doi.org/10.3390/w12020436
Booth, D.B. and Jackson, C.R. (1997) Urbanization of Aquatic Systems: Degradation Thresholds, Stormwater Detection, and the Limits of Mitigation. Journal of the American Water Resources Association 33 (5): 1077–1090. https://doi.org/10.1111/j.1752–1688.1997.tb04126.x
Brun, S.E. and Band, L.E. (2000) Simulating runoff behavior in an urbanizing watershed. Computers, Environment and Urban Systems 24: 5–22. https://doi.org/10.1016/S0198-9715%2899%2900040-X
BUCREP (2011) Rapport de présentation des résultats définitifs du recensement de la population en 2005, Yaoundé, Cameroun.
Buisson, A. (1985) La grande saison sèche 1985 au Gabon. Situation climatique en Afrique intertropicale. La Météorologie 15: 5–13.
Cisse, M.T., Sambou, S., Dieme, Y., Diatta, C. and Bop, M. (2014) Analyse des écoulements dans le bassin du fleuve Sénégal de 1960 à 2008. Revue des sciences de l’eau 27 (2): 167–187. https://doi.org/10.7202/1025566ar
Coe, M.T., Costa, M.H. and Soares-Filho, B.S. (2009) The influence of historical and potential future deforestation on the streamflow of the Amazon river–land surface processes and atmospheric feedbacks. Journal of Hydrology 369: 165–1774. doi:10.1016/j.jhydrol.2009.02.043
Conway, D.P., Persechino, A., Ardoin–Bardin, S., Hamandawana, H., Dieulin, C. and Mahé, G. (2009) Rainfall and river flow variability in sub-saharan Africa during the 20th century. J. Hydrom. 10: 41–59.
Crepet, F. (2000) Impact des aménagements hydrauliques sur le régime et la dynamique de la Loire amont. Implications pour la gestion du fleuve. Géocarrefour 75 (4): 365–374. doi: https://doi.org/10.3406/geoca.2000.2487
D’Orgeval, T. and Polcher, J. (2008) Impacts of precipitation events and land-use changes on West African river discharges during the years 1951–2000. Clim. Dyn. 31: 249–262.
Descroix, L., Guichard, F., Grippa, M., Lambert, L.A., Panthou, G., Mahé, G., et al. (2018) Evolution of Surface Hydrology in the Sahelo-Sudanian Strip: An Updated Review. Water 10 (748). https://doi.org/10.3390/w10060748
Descroix, L., Sané, Y., Thior, M., Manga, S.P., Ba, B.D., Mingou, J., et al. (2020) Inverse Estuaries in West Africa: Evidence of the Rainfall Recovery? Water 12 (647). https://doi.org/10.3390/w12030647
Dias, L.C., Macedo, M.N., Costa, M.H., Coe, M.T. and Neil, C. (2015) Effects of land cover change on evapotranspiration and streamflow of small catchments in the Upper Xingu river basin, Central Brazil. Journal of Hydrology 4: 108–122. doi:10.1016/j.ejrh.2015.05.010
Ebodé, VB., Dzana, J.G., Nkiaka, E. Nka, N.B., Braun, J.J. and Riotte, J. (2022b) Effects of climate and anthropogenic changes on current and future variability in flows in the So'o River Basin (south of Cameroon). Hydrology Research 53 (9): 1203–1220. https://doi.org/10.2166/nh.2022.047.
Ebodé, V.B. (2022b) Impact of rainfall variability and land-use changes on river discharge in Sanaga catchment (forest–savannah transition zone in Central Africa). Hydrology Research 53(7), 1017. https://doi.org/10.2166/nh.2022.046
Ebodé, V.B., Braun, J.J., Nnomo, B.N., Mahé, G., Nkiaka, E. and Riotte, J. (2022a) Impact of Rainfall Variability and Land Use Change on River Discharge in South Cameroon. Water 14, 941. https://doi.org/10.3390/w14060941.
Ebodé, V.B. (2022a) Variabilité hydroclimatique en Afrique centrale occidentale forestière: entre analyse des fluctuations observées, modélisation prédictive et recherche des facteurs explicatifs. Thèse de doctorat, Université de Yaoundé 1.
Ebodé, V.B., Mahé, G. and Amoussou, E. (2021a) Changement climatique dans le bassin versant de l'Ogooué: évolution récente et impact sur les écoulements. Proceedings of the International Association of Hydrological Sciences 384: 247–253. https://doi.org/10.5194/piahs-384-247-2021, 2021
Ebodé, V.B., Mahé, G., Dzana, J.G., Amougou, J.A. (2020) Anthropization and Climate Change: Impact on the Discharges of Forest Watersheds in Central Africa. Water 12 (2718). https://doi.org/10.3390/w12102718
Ebodé, V.B., Mahé, G. and Amoussou, E. (2021b) Impact de la variabilité climatique et de l'anthropisation sur les écoulements de la Bénoué (nord Cameroun). Proceedings of the International Association of Hydrological Sciences 384: 261–267. https://doi.org/10.5194/piahs-384-261-2021, 2021
Ewane, E.B. and Lee, H.H. (2020) Assessing land use/land cover change impacts on the hydrology of Nyong River Basin, Cameroon. J. Mt. Sci. 17 (1): 50–67. https://doi.org/10.1007/s11629-019-5611-8
Gebrelibanos, T. and Assen, M. (2015) Land use/land cover dynamics and their driving forces in the Hirmi watershed and its adjacent agro-ecosystem, highlands of Northern Ethiopia. J. Land Use Sci. 10 (1): 81–94. https://doi.org/10.1080/1747423X.2013.845614.
Getachew, H.E. and Melesse, A. (2013) The impact of land use change on the hydrology of the Angereb watershed. Int. J. Water Sci. 1 (4): 1–7.
Getahun, Y.S., Li, M.H. and Chen, P.Y. (2020) Assessing impact of climate change on hydrology of Melka Kuntrie Subbasin, Ethiopia with Ar4 and Ar5 projections. Water 12 (1308). doi:10.3390/w12051308
Kendall, M.G. (1975) Rank Correlation Methods, Griffin, London.
Kergoat, L., Hiernaux, P., Baup, F., Boulain, N., et al. (2007) Land surface in AMMA: Extending Ecosystem, Energy and Water Balance Studies in Space and Time is Some-Times Surprising. In Proceedings of the 2nd International Conference of AMMA Program, Karlsruhe, Germany.
Kouassi, A. (2007) Caractérisation d’une Modification Eventuelle 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 Doctorat Ph.D, Université de Cocody, Abidjan, Côte d’ivoire, 2007.
Kpoumié, A., Ngouh, A.N., Mfonka, Z., Nsangou, D., Bustillo, V., et al. (2022) Spatio–temporal assessing rainfall and dam impacts on surface runoff in the Sanaga river basin (transition tropical zone in central part of Cameroon). Sustainable Water Resources Management 8 (26). https://doi.org/10.1007/s40899-022-00624-1
Leblanc, M., Favreau, G., Massuel, S., Tweed, S., Loireau, M. and Cappelaere, B. (2008) Land clearance and hydrological change in the Sahel: South-west Niger. Glob. Planet. Chang. 61: 49–62.
Lee, E., Livino, A., Han, S.C., Zhang, K., Briscoe, J., Kelman, J. and Moorcroft, P. (2018) Land cover change explains the increasing discharge of the Paraná river. Reg. Env. Chang. 18: 1871–1881. doi:10.1007/s10113-018-1321-y
Liénou, G., Mahé, G., Paturel, J.E., Servat, E., Sighomnou, D., et al. (2008) Evolution des régimes hydrologiques en région équatoriale camerounaise: Un impact de la variabilité climatique en zone équatoriale? Hydrol. Sci. J. 53: 789–801. doi:10.1623/hysj.53.4.789
Mahé, G., L’Hôte, Y., Olivry, J.C. and Wotling, G. (2001) Trends and discontinuities in regional rainfall of West and Central Africa – 1951–1989. Hydrol. Sci. J. 46: 211–226. https://doi.org/10.1080/02626660109492817
Mahé, G., Liénou, G., Descroix, L., Bamba, F., et al. (2013) The rivers of Africa: Witness of climate change and the human impact on the environment. Hydrol. Process. 27: 2105–2114. doi:10.1002/hyp.9813
Mann, H.B. (1945) Nonparametric tests against trend. Econometrica 13: 245–259. doi:10.2307/1907187
Nka, N.B., Oudin, L., Karambiri, H., Paturel, J.E. and Ribstein, P. (2015) Trends in floods in West Africa: analysis based on 11 catchments in the region. Hydrology and Earth System Sciences 19: 4707–4719. doi:10.5194/hess-19-4707-2015
Nkiaka, E., Nawaz, R. and Lovett, J.C. (2016) Analysis of rainfall variability in the Logone catchment, Lake Chad basin. International Journal of Climatology 37 (9): 3553–3564. DOI: 10.1002/joc.4936
Nkiaka, E., Nawaz, R. and Lovett, J.C. (2018) Assessing the reliability and uncertainties of projected changes in precipitation and temperature in Coupled Model Intercomparison Project phase 5 models over the Lake Chad basin. International Journal of Climatology 38 (14): 5136–5152. https://doi.org/10.1002/joc.5717
Nzango, J. (2018) Les barrages de l'Oubangui: de l'impact hydraulique actuel à la prospective environnementale. Thèse de Doctorat, Université d'Orléans, France.
Panthou, G., Lebel, T., Vishel, T., Quantin, G., Sane, Y., Ba, A., et al. (2018) Rainfall intensification in tropical semi-arid regions: the Sahelian case. Environ. Res. Lett. 13 (064013). https://doi.org/10.1088/1748-9326/aac334
Pettitt, A.N. (1979) A non-parametric approach to the change-point problem. Applied Statistics 28 (2): 126–135.
Richter, B.D., Baumgartner, J.V., Braun, D.P. and Powell, J. (1998) A spatial assessment of hydrologic alteration within river network. Regul. Rivers Res. Mgmt. 39: 329–340.
Rodier, J. (1953) Monographie hydrologique du Logone supérieur. Ministère de la France d’Outre-mer, commission Logone-Tchad.
Schueler, T. (1994) The importance of imperviousness. Watershed Protection Techniques 3: 1–12.
Yang, G.X., Bowling, L.C., Cherkauer, K.A., Pijanowski, B.C. and Niyogi, D. (2010) Hydroclimatic Response of Watersheds to Urban Intensity: An Observational and Modeling-Based Analysis for the White River Basin, Indiana. Journal of Hydrometeorology 11 (1): 122–138. https://doi.org/10.1175/2009JHM1143.1
Yue, S., Pilon, P., Phinney, B. and Cavadias, G. (2002) The influence of autocorrelation on the ability to detect trend in hydrological series. Hydrol. Process. 16: 1807–1829. https://doi.org/10.1002/hyp.1095

Download PDF

Journal Publication

published 02 Apr, 2024

Read the published version in Water Supply →

Version 1

posted

You are reading this latest preprint version

Impact of anthropogenic changes and rainfall variability on river discharge in tropical central Africa

Status:

Journal Publication

Version 1

Abstract

Figures

1. Introduction

2. Data And Methods

2.1. Study framework

2.2. Data source

2.3. Data analysis

3. Results

3.1. Rainfall evolution

3.1.1. Evolution of annual and seasonal rainfall

3.1.2. Spatial evolution of precipitation

3.2. Evolution of flows

3.2.1. Evolutions of mean flows

3.2.2. Extreme flows

3.2.2.1. The maximums

3.2.2.2. The minimums

4. Discussion

4.1. The role of rainfall in the flows of the basins studied

4.2. The impact of the dam on the flows of the Benoue basin

4.3. The contribution of changes in land-use patterns

5. Conclusion

Declarations

References

Status:

Journal Publication

Version 1