Congo Basin Hydrology, Climate, and Biogeochemistry. Группа авторов

that a time series is homogeneous between two given dates and according to two laws of different distributions. On the one hand, these tests are known for their robustness but do not allow the detection of more than one break in a time series. On the other hand, Hubert’s test gives the year of the different breaks observed in the series with the means and standard deviations of the different segments. This method has been the subject of numerous publications in different regions of Africa (Ardoin et al., 2003; Bodian et al., 2011; Servat et al., 1999) and in Central Africa (Nguimamet & Orange, 2020). These tests have not been applied to the Uele time series, which are series inferred from the surrounding basins.

      6.3.3. Calculation of River Depletion Coefficients

      The river depletion coefficients of the Ubangi at Mobaye and the volume mobilized by the basin aquifer were calculated from daily data (according to Cosandey et al., 2003; Goula et al., 2006; Nguimalet & Orange, 2019; Savané et al., 2001). However, numerous gaps in the 1960s with exceptional rainfall and after 1976 shortened the series. This approach makes it possible to assess the contribution of surface runoff on aquifer recharge, by homogeneous hydropluviometric period in this deficit context.

      The mathematical expression for the river depletion coefficient is written:

      (6.3)

      where Q_t = discharge at time t, Q₀ = initial discharge (discharge at the beginning of the depletion), and k = Maillet depletion coefficient. Maillet’s law is modeled on a single reservoir, which corresponds to all the aquifers contributing to flows in the watershed under study (Goula et al., 2006; Nguimalet & Orange, 2019). The integration of this equation over the interval [0, +] gives an estimate of the volume mobilized from the aquifers in the watershed, i.e.:

      (6.4)

6.4. RESULTS

      6.4.1. Interannual Evolution of Rainfall and Annual Flows of the Ubangi River at Mobaye

The average interannual rainfall on the Ubangi basin at Mobaye is 1,506 mm over the period 1938–2015. On the one hand, the search for breaks with Hubert’s segmentation reveals two breaks (1968 and 2006) and thus three homogeneous periods (1938–1968, 1969–2006, and 2007–2013). On the other hand, a single break appears in 1969 with the Pettitt and Lee‐Heghinian tests (Table 6.3). This time series is thus characterized by a definite rainfall break in 1968/1969 (Figure 6.3), marking the passage from a previous wet period to a dry period, and a probable second break in 2006 to be confirmed in the future, and which would characterize the return to a wetter period. The average rainfall amounts for the periods 1938–1968, 1969–2006, and 2007–2015 are 1,568 mm, 1,436 mm, and 1,583 mm respectively. The dry period of 1969–2006 corresponds to a reduction of 8% in rainfall compared to the previous period. And the current wet period (2007–2015) describes a 10% increase in rainfall compared to the previous period.

Table 6.3 Statistical segmentation of annual rainfall in the Ubangi basin at Mobaye (1938–2015), and in the upstream basins from 1951–1995 (in the Mbomu basin at Bangassou and Zemio, and in the Kotto basin at Kembe and Bria).

Hydrological basins	Interannual average (mm)	Pettitt	Lee and Heghinian	Segmentation of Hubert (mm)	Standard Deviation	Hydro‐climatic trends	P divided by interannual average (%)
Ubangi at Mobaye (403,800 km2)	1,506	1969	1969	1938–1968: 1,568 1969–2006: 1,436 2007–2013: 1,583	109 88 81	Wet Dry Wet	+4 –5 +5
Mbomu at Bangassou (115,000 km2)	1,491	1970		1951–1994: 1,851
Mbomu at Zemio (29,300 km2)	1,421	1970		1951–1994: 1,748
Kotto at Kembe (77,500 km2)	1,348	1969	1969	1948–1980: 1,458 1981–1991: 1,306 1992–1993: 1,577	94 89 32	Wet Dry Wet	+2 –8 +11
Kotto at Bria (61,500 km2)	1,348	1967	1967	1948–1966: 1,436 1967–1991: 1,267 1992–1993: 1,515	94 105 83	Wet Dry Wet	+7 –7 +12

Figure 6.3 Rainfall index and flow index of the Ubangi basin at Mobaye, 1938–2015.

The 1938–2015 time series of annual flows from the Ubangi at Mobaye is marked by a sharp decline in flows in 1970 (Figure 6.2). Statistical tests indicate only one break with the Pettitt and Lee‐Heghinian tests: it is noted in 1981 on the 1938–2015 series, and in 1970 on the 1948–1995 series. And Hubert’s segmentation describes five homogeneous periods for the two series tested (1938–2015 and 1948–1995) but only the years 1968 and 1980 appear for the two hydrological series tested, attesting to the importance of the statistical break at these two dates. Three hydroclimatic periods are thus retained for the Ubangi at Mobaye (Table 6.4): 1954–1968, 1969–1980, 1981–2013, with interannual mean flows of 3,919 m³/s, 2,986 m³/s (–24%) and 2,372 m³/s (–21%), respectively.

      6.4.2. Interannual Evolution of Rainfall and Annual Flows in Sub‐Basins of the Ubangi at Mobaye

The rainfall series of the Uele basin and the Mbomu basin at Bangassou have a similar evolution to that of the Ubangi basin at Mobaye, and to a lesser extent the Kotto basin at Kembe. On the other hand, the rainfall series of the Mbomu at Zemio and the Kotto at Bria do not show a clear separation between a wet and a dry phase (Figure 6.4). The last two basins drain the wooded and shrubby savannah, while the first two are of East–West extension and drain a significant part of the