The coupled evolution of rainfall (P) and annual flow (Q) over the upstream basin of the Ubangi River at Mobaye is studied in order to discuss the role of the forest compared to the savannah on the hydropluviometric behavior of the Kotto, Mbomu, and Uele basins, and to assess the behavior of aquifers in this hydropluviometric deficit context recorded since 1970 in the region. We approach this analysis by looking for homogeneous hydroclimatic periods of the 1938–2015 sequence on the Ubangi River at Mobaye, and over the 1951–1995 period for its three major tributaries: the Kotto, Mbomu, and Uele rivers.
6.2. STUDY AREA: UPPER BASIN OF THE UBANGI RIVER AT MOBAYE
The Mobaye outlet on the Ubangi River is located 151 km from the confluence of the Mbomu River (CAR) (Central African Republic) and the Uele River (DRC) (Democratic Republic of the Congo), the confluence from which the Ubangi River originates at Kemba (Figure 6.1). The Ubangi is the largest right‐bank tributary of the Congo and a natural border between the CAR to the north and the DRC to the south. The Ubangi at Mobaye has a watershed of 403,800 km2. The SE (Uele Highlands) and N (Kotto Highlands) ends of the basin are dominated by mountains with respective altitudes of 1,700 m and 1,300 m, nested between two plateaus (the Uele Sandstone Plateau in the south and the Central African Surface in the north), then the plain of the Ubangi foothills, which is quite narrow, confines the downstream valleys of the Mbomu, Bili, and Uele rivers up to Mobaye, including the downstream valley of the Kotto. In our study, the Bili River of only 21,400 km2 is integrated into the Uele basin (i.e. Orange, 1996).
Figure 6.1 Topographic map of the Ubangi basin at Mobaye, and of the hydropluviometric network studied, with the limits of the five studied sub‐watersheds (at the outlets of Bondo for the Uele River, Zemio and Bangassou for the Mbomu River, and Bria and Kembe for the Kotto River).
Source: Based on Orange et al., 1994.
Figure 6.2 Vegetation map of the Ubangi basin at Mobaye, the geological bedrock and the hydropluviometric network studied, with the limits of the five studied sub‐watersheds (at the outlets of Bondo for the Uele River, Zemio and Bangassou for the Mbomu River, and Bria and Kembe for the Kotto River).
Source: Based on Orange et al., 1996.
The rainfall gradient decreases from south to north, from 1,700 mm on the Uele to 1,000 mm on average over the upper Kotto basin (Nguimalet & Ndjendolé, 2008). The basin of the Ubangi at Mobaye is a well‐marked latitudinal stratification of the edaphic stages (Figure 6.2). From south to north, one observes the tropical rainforest domain of the equatorial forest with 15% primary forest (Global Forest Watch, 2019), followed by the wooded savannah zone, successively wooded, shrubby, and grassy in the extreme north (Nguimalet, 2017). The latest 2018 surveys by Global Forest Watch confirm that in the Central African Republic and Northern Congo the forest area has decreased very little (–1.5% between 2001 and 2018), with a slightly greater loss of primary forest (–1.9%). This study considers that the forest area has not changed. Thus, according to Orange & Ghiloufi (1996), the Uele basin (including Bili) is 70% covered by tropical forest and 30% by wooded and tree‐covered savannah, the Mbomu basin has only 10% tropical forest and 30% wooded savannah (according to Orange, 1996), and the Kotto basin is representative of wooded and tree‐covered savannah (Figure 6.2). The humid forest thus represents 30% of the Ubangi basin at Mobaye.
In the region, population growth is low: the population in Bangassou on the Mbomu River, the regional capital, increased from 902,205 inhabitants in 2006 to 1,126,730 in 2015, an increase of only 2.5% per year. The entire Ubangi basin at Mobaye thus remains a sparsely populated area, with a population density varying between 10 inhab/km2 and 3.3 inhab/km2. In the Uele basin, the population density was 7 inhab/km2 in 2008 (Haut‐Uélé, 2020). Human activities are essentially primary (slash‐and‐burn agriculture, artisanal mining, production of firewood and charcoal, etc.), concentrated around the Ubangi, Uele, and Mbomu valleys; these activities have changed little since 1970 (Haut‐Uélé, 2020; Nguimalet & Orange, 2019; Orange & Ghiloufi, 1996).
6.3. DATA AND METHODS
6.3.1. Hydropluviometric Data
The hydrological stations studied are the Ubangi River at Mobaye (403,800 km2), the Kotto River at Kembe (78,400 km2, downstream of the basin) and at Bria (61,500 km2, upstream of the basin), the Mbomu River at Bangassou (115,000 km2, downstream of the basin) and at Zemio (29,300 km2, upstream of the basin), and the Uele + Bili hydrographic complex (161,100 km2), whose flows were estimated by deducting the confluences, according to the work of Orange (1995).
Daily flow data were used for the Bangui and Mobaye stations on the Ubangi River from 1938 to 2015. These data come from the IRD database published in the Monography of the Ubangi (Callède et al., 2009) and supplemented by data from Wesselink et al. (1995) over the period 1986–1993, and from the Direction de la Météorologie Nationale from 1994 to 2015. On the one hand, there are no missing data on Ubangi at Bangui. On the other hand, between 1975 and 2014, a total of 21 years is completely missing from Mobaye (Table 6.1). For the four other Central African hydrological stations (Kembe and Bria on the Kotto; Bangassou and Zemio on the Mbomu), the hydrological data used are monthly and cover only the period 1948–1995, taken entirely from the IRD database, published in part (1986–1993) by Wesselink et al (1995). On average, over 45 years, about 15 full years are missing, i.e., a third of the period studied. For the other years, there may be occasional monthly shortfalls (Table 6.1), with the result that hydrological data gaps represent between 43 and 52% of the monthly bank.
When monthly data are functionally missing, the monthly reconstitution of month i is done by the method of averaging the two years of month i surrounding the missing data. For the periods 1977–1985 and 1995–2015, the missing daily flow data at the Mobaye outlet were reconstructed from the Bangui station, considering that the specific flow is unchanged between Mobaye and Bangui. Finally, for the complete missing years, data for the Kotto and Mbomu rivers were reconstructed from the nearby hydrological station, according to the formula (Jacob et al., 2006; Azouka, 2011):
(6.1)
where q is the estimated stream discharge; Q is the discharge from the nearest hydrological station; A is the surface area of the basin at the nearest hydrological station; and a