Original Investigation of the Role of Soil Moisture Memory in Seasonal Rainfall Predictability Over the Guinea Savanna Belt of Nigeria

Soil moisture memory, the persistence of antecedent land surface wetness anomalies and their modulation of subsequent rainfall through land-atmosphere coupling, represents a theoretically significant but observationally underexplored source of seasonal predictability in tropical West Africa. This study investigates the role of soil moisture memory in seasonal rainfall predictability across Nigeria's Guinea Savanna belt through a combination of observational analysis, reanalysis diagnostics, and controlled model experiments. Soil moisture data from the ESA Climate Change Initiative (ESA-CCI) merged passive-active microwave product were analyzed alongside MERRA-2 surface heat fluxes and NiMet seasonal rainfall records for 1993 to 2022 to characterize land-atmosphere coupling strength. Global Climate Model (GCM) ensemble experiments in which soil moisture was either freely evolving or restored to climatological values were conducted to isolate the soil moisture memory contribution to seasonal forecast skill. Decorrelation time scale analysis reveals that soil moisture memory over the Guinea Savanna extends to approximately eight weeks at the start of the rainy season, substantially longer than atmospheric memory alone. Coupling hotspot analysis identifies the northern Guinea Savanna fringe (8-11 degrees N) as a zone of strong positive soil moisture-rainfall feedback, where antecedent May soil moisture anomalies significantly predict June-August rainfall with correlation coefficients of 0.52 to 0.67. Theoretical advancement includes development of a locally calibrated soil moisture coupling parameter that outperforms global estimates in diagnosing feedback strength for Nigerian conditions. Keywords: soil moisture memory, land-atmosphere coupling, Guinea Savanna, seasonal predictability, original contribution.