MONITORING WATER BODIES TO DETERMINE THEIR CONNECTIVITY, SEASONAL VARIATIONAL AND PERMANENCE IN THE GRIJALVA-USUMACINTA REGION

摘要

The article focuses on the detection and delineation of water bodies’ surfaces for the processing of digital images for monthly inter-annual variability, using Sentinel 1A data. The study area is a sub-region of lowland of the Grijalva-Usumacinta basin. Results are shown for different algorithms used to recognize patterns for the detection, delineation and segmentation of water bodies in an automated manner, with methods such as Support Vector Machine, and the Bayesian classifier. The study shows the spatial-temporal behavior of the form, size and distribution of water bodies, and allows for the description of rivers, their connectivity and analysis of superficial hydrological connectivity. It also shows a cartography of the annual flooding frequency, where vulnerable areas can be seen.