Mapping Wetland Dynamics With SAR-Based Change Detection in the Cloud

Abstract

The wetlands are often coupled with anthropogenic systems. This makes them highly dynamic, and thus difficult to map and monitor from space. The Sentinel-1 constellation allows monitoring land changes regardless of cloud conditions and with a high frequency of pass. We use a time series of 33 Sentinel-1 dual-polarized images from 2016 to map seasonal changes at the artificial Lake of Kerkini, in Greece. The images are accessed and processed in Google Earth Engine via a Flask web application. Points of change within the time series and frequency of change were determined according to an omnibus-test statistic. The results were compared to an optical-based land cover map and gaging measurements. Frequency of change patterns matched different land cover types, indicating that frequency of change is related to the physical and structural properties of the land cover. This methodology can be used operationally to improve our understanding of ecosystems, enhance land cover maps, and to monitor wetland dynamics as well as cover changes caused by human activities.