Spatiotemporal Dynamic Analysis of Tropical Wetlands using Sentinel-1 SAR Images: Case of Study in Colombia

Abstract

SAR-Synthetic Aperture Radar-Sentinel-1 data complements the current possibilities offered by multispectral information, and it is mainly useful in areas of high cloudiness. This study presents the spatiotemporal dynamics of a tropical wetland system named Ciénaga Colombia from 2014 to 2021 through Sentinel-1 SAR data. The analysis is mainly done using Google Earth Engine-GEE-and boxplots were used to analyze SAR backscattering. Thresholds and SWIM-Surface Water Index Extraction Model-methods were used to calculate the surface water of the wetland system. This study also explored some possible relationships between the system surface water with precipitation and the ENSO –El Niño Southern Oscillation-. The results showed that 2015 and 2017 present the greater-lesser backscattering values (respectively), indicating mean surface water sizes of 110 ha and 334 ha. SWIM and thresholds values calculated for the lowest backscattering moment in 2017 differ almost double (384 ha and 657 ha). The consistency of the SAR data results was validated through NDWI values from Sentinel 2-1C images. Finally, this study recognizes the need to delve into different analyses to identify changes in the wetland system as a result of direct interventions on its drainage network.