Water level monitoring using radar remote sensing data: Application to Lake Kivu, central Africa

Satellite radar altimetry measures the time required for a pulse to travel from the satellite antenna to the earth’s surface and back to the satellite receiver. Altimetry on inland lakes generally shows some deviation from in situ level measurements. The deviation is attributed to the geographically varying corrections applied to account for atmospheric effects on radar waves. This study was focused on verification of altimetry data for Lake Kivu (2400 km2), a large inland lake between Rwanda and the Democratic Republic of Congo (DRC) and estimating the lake water levels using bathymetric data combined with satellite images. Altimetry data obtained from ENVISAT and ERS-2 satellite missions were compared with water level data from gauging stations for Lake Kivu. Gauge data for Lake Kivu were collected from the stations ELECTROGAZ and Rusizi. ENVISAT and ERS-2 data sets for Lake Kivu are in good agreement with gauge data having R2 of 0.86 and 0.77, respectively. A combination of the two data sets improved the coefficient of determination to 95% due to the improved temporal resolution of the data sets. The calculated standard deviation for Lake Kivu water levels was 0.642 m and 0.701 m, for ENVISAT and ERS-2 measurements, respectively. The elevation-surface area characteristics derived from bathymetric data in combination with satellite images were used to estimate the lake level gauge. Consequently, the water level of Lake Kivu could be estimated with an RMSE of 0.294 m and an accuracy of ±0.58 m. In situations where gauges become malfunctioning or inaccessible due to damage or extreme meteorological events, the method can be used to ensure data continuity.