Revealing Short-Term Behavioral Patterns of Storage Dynamics in African Natural Lakes and Reservoirs Using High-Frequency Satellite Observations

This dataset provides 3-day resolution time series of surface area, water level, active storage (AS), and normalized active storage (NAS) for 510 lakes and reservoirs across Africa from 2017 to 2025. The surface area time series were derived by integrating Harmonized Landsat and Sentinel-2 (HLS) and Sentinel-1 data through a workflow including mapping, filtering, gap filling, and interpolation. Water level time series were obtained by combining ICESat-2, DAHITI, and SWOT observations and deriving elevations using four area–elevation (A–E) relationship approaches. The AS time series were calculated by integrating the A–E relationships, while NAS was obtained by normalizing the AS series. After extraction, the dataset consists of 510 CSV files, with filenames corresponding to IDs in the HydroLAKES dataset. Each CSV file contains the time series of surface area and related variables for an individual water body. Some uncertainties remain in the dataset. For example, LOWESS interpolation may introduce extreme values, A–E relationships are subject to fitting errors, and water level estimates have been constrained by observational limits.