Mapping suspended sediment dynamics in the Pantanal wetland using remote sensing and ANN-based models

<p>The Pantanal is the largest tropical wetland on earth, covering an area of 158000 km&#178; between Brazil (~70%), Bolivia (~20%) and Paraguay (~10%). The regular flood pulse of this region produces unique ecological and geomorphological processes in the floodplains. Due to the extensive areas with flat topography, the water velocity of the rivers gets reduced, and large sediment deposition processes begin to take place in the Pantanal floodplain. Despite its unique characteristics and great environmental importance, the rivers from this region have very scarce in situ monitoring of suspended sediment concentration (SSC), with approximately one gauge per 8000 km², and experiences low collection frequency (average of four measurements per year). Therefore, the characterization of sediment dynamics in this region remains challenging, and the spatial-temporal variation of suspended sediments in the Pantanal rivers is still poorly understood.</p><p>Remote sensing techniques offer enormous advantages by providing cost-effective systematic observations of large water systems, allowing spatial-temporal mapping of wild areas such as the Pantanal. The suspended matter in water bodies increases the reflectance in the green, red, and near-infrared (NIR) bands, i.e., the backscatter radiation increases as the SSC in water increases. Therefore, reflectance from the visible bands and NIR band can be used as proxies of SSC in water bodies.</p><p>The focus of this study is to assess the spatial-temporal variations of SSC in rivers that drain to and through the Pantanal wetland, by using surface reflectance (SR) from satellite images and artificial neural network (ANN)-based models. We used atmospherically corrected SR from Sentinel-2, Landsat 8, and Landsat 9 (bands of blue, green, red and NIR) as input variables, and in situ data on SSC from 23 gauges along the Pantanal rivers as output variables in the models.</p><p>Through this methodology, we expect to obtain time series of SSC estimated by the ANN-based model and reflectance data from satellite images for different parts of the Pantanal hydrographic basin.</p><p>The resulting time series allows us to:</p><ul><li>Characterize the spatial variations of suspended sediments along different rivers in the Pantanal wetland.</li> <li>Identify the main drivers of these spatial variations by comparing these differences with land use, vegetation cover, topography, and types of soil within the drainage watershed of the rivers.</li> <li>Characterize the influence of the seasonal hydrological regime on SSC transport.</li> <li>Identify the influence of anthropic activities on the amount of SSC transported to the wetland.</li> </ul>