MAPPING THE HYDROLOGICAL CONNECTIVITY OF FLOODPLAIN LAKES USING PLANET MOSAICS: AN EXPLORATORY ANALYSIS

Hydrological Connectivity (HC) of small floodplain lakes is rarely investigated using remote sensing techniques. This paper explores the application of planet mosaics to estimate the HC between the middle Juruá River and its floodplain lakes. Remote Sensing features from 41 floodplain lakes and related main river reaches were provided by a time series (2020-2022) of Planet mosaics. These features and river water level were input to a machine learning algorithm. The trained HC-model had a total accuracy of 82% but showed a trend to overestimate connected lakes. Model performance was compromised by planet mosaic spectral information and the quality of the connectivity data used for the training. Finally, this paper encourages new applications related to the HC of floodplain lakes using water color data as a proxy.