Current availability and distribution of Congo Basin's freshwater resources

Freshwater resources on land are vital for societies and ecosystems, yet their quantity, spatial distribution, and variability remain poorly known in many regions. This knowledge gap is acute in the Congo Basin, the world’s second largest river basin (3.7 × 10⁶ km²) and a major outlet of freshwater to the oceans, which sustains about 120 million inhabitants and harbors globally significant tropical forests and peatlands that act as a vast carbon sink. Despite its importance, the basin is understudied and sparsely gauged today, hindering robust water management. Existing basin-scale water resource estimates (e.g., FAO AQUASTAT) rely on survey-based hydro-climatic parameters but do not quantify Total Drainable Water Storage (TDWS)—the long-term average amount of freshwater in soils and near-surface that can drain to rivers. Spaceborne observations, notably GRACE/GRACE-FO gravimetry and satellite altimetry/imagery, have advanced understanding of storage variability but TDWS remains a “known unknown” in hydrology. This study addresses two objectives: (1) provide the first quantitative estimate of the contemporary TDWS over the entire Congo Basin and its major sub-basins (Kasaï, Middle Congo, Ubangui, Sangha, and Lualaba and its sub-divisions), and (2) estimate the hydraulic time constant (basin resistance) governing how quickly stored water is discharged. The results aim to inform sustainable water management under growing climate and anthropogenic pressures.

The paper situates its contribution within several strands: (a) limited in-situ hydrometric coverage in the Congo since the 1960s and reliance on legacy datasets; (b) satellite-based advances—GRACE for basin-scale total water storage anomalies (TWSA), satellite altimetry (TOPEX/Poseidon, Jason series), and optical/radar-derived surface water extent (e.g., MODIS, Landsat, GIEMS-2)—that have improved characterization of hydrology in large tropical basins; (c) prior GRACE-based studies in the Congo that quantified storage dynamics and groundwater stress but not TDWS; (d) prior TDWS estimates exist for few basins (Amazon and Mississippi) showing feasibility of storage–discharge methods; (e) climate projections indicating rising temperatures (+2 to +4 °C by late century) and uncertain precipitation changes (−9% to +27%), with likely increases in mean annual runoff, raising water management challenges amid anthropogenic pressures (groundwater stress, damming, deforestation). This context underscores the need for a robust, observation-based TDWS estimate and hydraulically meaningful timescales for the Congo.

Data and preprocessing: (1) GRACE TWSA (2002–2017) from ITSG-Grace2018 spherical harmonics with degree-1 added, degree-2/3 zonals replaced by SLR; primary/secondary tidal aliasing (S1, S2, P1, K1, K2, M2, O2, O1, Q1) removed via least-squares Fourier analysis; Gaussian (350 km) and destriping filters applied; GIA corrected (ICE6G-D/viscoelastic model); basin-wise leakage mitigated using a data-driven method; anomalies referenced to 2004–2009. (2) Lake water storage anomaly (LWSA): Monthly lake volume anomalies for Mai-Ndombe, Mweru, Upemba, Bangwelu, Kivu, Tanganyika computed by combining satellite-derived lake area (HydroSat/Hydroweb) and water level (altimetry) using a pyramidal geometry approach between successive observations; anomalies referenced to 2004–2009 and converted to equivalent water height per sub-basin. (3) Wetland water storage anomaly (WWSA): Derived via hypsometric curve approach combining GIEMS-2 monthly surface water extent (0.25° equal-area pixels, 1992–2015) with FABDEM 30 m topography; DEM-based area–elevation relationships corrected to avoid overestimation, converted to area–storage relationships, and merged with SWE time series to obtain surface water storage for targeted wetlands (Mai-Ndombe; Bangwelu, Mweru, Mweru Wantipa, Upemba). Anomalies referenced to 2004–2009 and scaled by sub-basin area. (4) River discharge: Used available in-situ discharge at sub-basin outlets where possible; for Kasaï, Lualaba-North, Lualaba-South, Lualaba-Lukuga (with only legacy discharge up to 1959), estimated discharge from satellite altimetry water levels using a quantile-matching rating curve method that does not require synchronous datasets; virtual stations selected to avoid >1-month transit times. Computation of DWSA and lag removal: For each sub-basin, Drainable Water Storage Anomaly (DWSA) = TWSA − LWSA − WWSA, treating major lakes and permanent wetlands as loosely coupled storages except where floodplain connectivity implies coupling (e.g., parts of Cuvette Centrale). For basins without large lakes/wetlands (e.g., Ubangui, Sangha, Middle Congo, Lualaba-North), TWSA approximates DWSA. Mean monthly time series used to filter fast runoff events. The storage–discharge phase lag was removed by harmonic decomposition of both series (up to six annual harmonics) and phase alignment of storage to discharge to yield lag-corrected storage anomalies. Estimation of TDWS and hydraulic time constant: A linear, time-independent relationship Q(t) = (1/τ)(S0 + ΔS(t)) was fit between mean monthly discharge and lag-corrected storage anomaly using a Gauss-Helmert Model that accounts for errors in both variables, estimating the hydraulic time constant τ and the x-intercept S0 defining Total Drainable Water Storage (TDWS). Two cases were analyzed: (i) DWSA–Q yielding a lower bound for TDWS (excluding decoupled lakes/wetlands), and (ii) TWSA–Q yielding an upper bound (including all storages). Analyses were performed for the full Congo and seven sub-basins (Kasaï, Middle Congo, Ubangui, Sangha, Lualaba-North, Lualaba-South, Lualaba-Lukuga).

• Basin-wide TDWS: 476 ± 10 km³ (lower bound from DWSA–Q), corresponding to 133 ± 3 mm equivalent drainable water height. Upper bound from TWSA–Q: 502 ± 22 km³ (139 ± 6 mm). Hence, Congo’s TDWS lies between ~476 and ~502 km³.
• Seasonal storage range: Minimum drainable storage in August ~40 mm (137 km³); maximum total volume in December ~613 km³; annual fluctuation amplitude ~137 km³ around the mean TDWS.
• Spatial distribution: Weighted sub-basin analysis yields ~503 ± 10 km³ total, consistent with full-basin estimate. Kasaï stores 220 ± 4 km³ (upper bound 228 ± 18 km³), the largest share (~43% of lower-bound total). Lualaba totals 109 ± 4 km³ (North 22 ± 3; South 67 ± 2; Lukuga 20 ± 1). Northern sub-basins (Middle Congo 90 ± 6; Ubangui 63 ± 5; Sangha 20 ± 2) together store 173 ± 8 km³. Overall, approximately 63–65% of Congo’s drainable water is in the southern Kasaï and Lualaba sub-basins.
• Hydraulic time constant (basin resistance): Congo-wide τ = 4.3 ± 0.1 months. Sub-basin τ (DWSA–Q): Middle Congo 2.6 ± 0.2 months; Lualaba-North 1.7 ± 0.3; Sangha 5.7 ± 0.5; Ubangui 8.6 ± 0.8; Kasaï 9.3 ± 0.2; Lualaba-South 7.7 ± 0.2; Lualaba-Lukuga 105.8 ± 3.0 months. Including fully coupled surface water (TWSA–Q) increases τ notably where large lakes/wetlands are present (e.g., Lualaba-Lukuga τ = 174.2 ± 4.7 months; Lualaba-South τ = 13.9 ± 0.9; Kasaï τ = 10.1 ± 0.8), quantifying added resistance from lakes/wetlands.
• Storage components and seasonality: Over the Congo, LWSA contributes about 10% of TWSA amplitude (dominated by Lake Tanganyika), and WWSA about 20% (largely from southern wetlands). Basin exhibits generally clockwise storage–discharge hysteresis, with regional differences; Lualaba-Lukuga shows rapid storage loss in dry season driven by strong evaporative losses from Lake Tanganyika (~82% of its annual water loss).
• Congo vs Amazon benchmarking: Congo TDWS (476 ± 10 km³) ≈ 25% of Amazon’s (1766 ± 47 km³); similar ratios are seen for mean discharge, TWSA amplitude, and recharge (P−ET). Congo τ (4.3 ± 0.1 months) matches Amazon τ (~4.4 months).

The study addresses the key unknown of Congo Basin freshwater availability by operationally defining and estimating TDWS via a robust storage–discharge framework using satellite observations. The estimated TDWS (476–502 km³) and τ (4.3 months) quantify both volume and dynamical response time of the basin. Sub-basin contrasts highlight the dominant role of southern sub-basins (Kasaï and Lualaba) in storing drainable water and the substantial buffering (resistance) effects of large lakes and permanent wetlands (e.g., extreme τ in Lualaba-Lukuga due to Lakes Tanganyika and Kivu). Internal consistency checks (agreement between basin-wide and aggregated sub-basin estimates) and physical plausibility (larger τ with more extensive open water) support the methodology. Additional credibility comes from independent GRACE-based driest-state mapping (all-time lowest anomaly sum −316 km³, below TDWS) and congruent Congo–Amazon ratios across multiple hydrological metrics. The findings are relevant for assessing water security under climate variability and increasing anthropogenic pressure, informing allocation, drought resilience, and infrastructure planning by providing both storage capacity and depletion timescales.

This work delivers the first quantitative estimate of Total Drainable Water Storage for the Congo Basin and its distribution among major sub-basins, together with basin resistance (hydraulic time constants). The basin holds approximately 476–502 km³ of drainable freshwater, with a characteristic depletion timescale of about 4.3 months, and a majority of storage located in the southern Kasaï and Lualaba sub-basins. Large lakes and wetlands substantially increase regional resistance, modulating hydrograph timing and low flows. These observation-based benchmarks provide a foundation for sustainable water supply management, transboundary coordination, and climate risk assessments. Future research should refine coupling between surface waters and drainage, improve discharge estimation with expanded in-situ and altimetry coverage, reduce GRACE leakage uncertainties, and extend analyses with GRACE-FO to assess non-stationarity and trends under ongoing climate and land-use change.

• Cyclostationarity assumption: Mean monthly relationships are constructed over differing time periods for datasets and sub-basins, assuming stationary seasonal behavior despite known interannual to decadal variability and recent extremes.
• Discharge estimation: For several sub-basins, discharge after 2002 is inferred from altimetry–legacy rating curves under stationarity assumptions; spatial separation between gauges and virtual stations and potential non-stationary hydraulics add uncertainty.
• GRACE limitations: Coarse resolution and filtering induce signal leakage, especially for smaller sub-basins; uncertainties persist despite leakage mitigation. LWSA and WWSA estimates also inherit uncertainties from satellite imagery, altimetry, and hypsometric reconstructions.
• Decoupling assumption: Treating lakes/wetlands as loosely coupled storage to derive lower-bound TDWS is an approximation; partial coupling exists (e.g., Lake Tanganyika outflow), motivating bracketed (lower/upper bound) estimates.