Building on the availability of high revisit frequency Earth Observation satellites at medium spatial resolution (250 m), this study investigates the feasibility of temporal monitoring of water bodies at a continental scale with MODIS. A 2004–2010 time series of twice-daily observations covering the whole African continent was systematically processed using a surface water detection method to derive 10-day indicators describing the location, the intra- and inter-annual variability as well as the temporal characterization of water bodies (i.e. seasonal or permanent water and maximum extent). The multispectral surface reflectance transformation in the HSV color space allows a per-pixel identification of surface water. The water aggregation time indicator provides the water occurrence for each 10-day period built from the seven years of observations. The cartographic products were successfully cross-validated with already existing maps and water products. The validation of the water body maximum extent map estimates the commission error at less than 6% and the seasonality information was also found to be consistent with the Köppen climatic classification.
d’Andrimont, R., & Defourny, P. (2017). Monitoring African water bodies from twice-daily MODIS observation. GIScience and Remote Sensing, 1-24. https://doi.org/10.1080/15481603.2017.1366677 (Original work published 2017)