Why did South Sudan experience unprecedented flooding in 2022? The role of upstream storage and hydrological memory in the White Nile.

Between 2019 and 2024, South Sudan experienced prolonged and widespread flooding that existing early-warning systems did not anticipate, largely because flood forecasting models were conditioned on short-term rainfall variability and did not reflect the basin’s long hydrological memory. This study examines the physical processes underlying this forecasting mismatch and shows how limited representation of upstream storage, wetland dynamics, and multi-season antecedent conditions constrained the skill of anticipatory flood information. Using daily lake levels, river discharge, CHIRPS rainfall, and MODIS flood extent, we quantified floodwave travel time from Lake Victoria to the Sudd Wetland to identify the mechanisms shaping this multi-year flood response.

The analysis shows that the mean upstream-to-downstream floodwave transit time is approximately 16.8 months, rather than the often-assumed five months, revealing a fundamentally slow system controlled by lake storage, floodplain buffering, and wetland attenuation. This long delay explains why downstream hydrological signals evolved differently from local rainfall patterns. Flooding in the central and western Sudd was shaped by the gradual movement of stored water through the Victoria–Kyoga–Albert–Sudd corridor, where each lake and wetland unit progressively reshapes and delays the floodwave. In contrast, eastern sub-catchments such as the Baro–Akobo–Sobat responded more directly to local rainfall, reflecting weaker connectivity to the lake–wetland system. The extensive inundation observed in 2022, including around Bentiu, therefore resulted from cumulative multi-year storage initiated by the 2019 positive Indian Ocean Dipole and reinforced by successive anomalous rainy seasons both up- and downstream, rather than from local rainfall downstream alone.

These findings highlight the limitations of flood forecasting modelling approaches that emphasise short-term precipitation forcing while under-representing storage, routing, and long hydrological memory in large lake–river systems. By identifying system-scale transit times and the spatial structure of storage-driven response, this work provides a physical basis for improving the interpretation of flood forecasts and for extending effective lead times for anticipatory action. Explicit recognition of long-memory dynamics can help distinguish precipitation-driven from storage-driven flooding, supporting more timely and proportionate preparedness decisions along the White Nile corridor.