Karen Escalona *,1, Rodrigo Abarca-del-Rio 2, Maria Pedreros 2, Ignacio García 2
1 Doctoral Program in Physical Sciences, Faculty of Physical Sciences and Mathematics, University of Concepción, Chile.
2 Department of Geophysics, Faculty of Physical Sciences and Mathematics, University of Concepción, Chile.
Contact: Karen Escalona. Email: kescalona@udec.cl
This study seeks to identify and monitor long-term aquatic vegetation encroachment. We are using Maracaibo lake (Venezuela), which is highly anthropogenic, as a test of a robust and fast remote sensing methodology. The strategy was created and implemented in Google Earth Engine (GEE) using Landsat 7 and 8 satellite images and combining cloud masking, spectral indices, and machine learning classification techniques. Thus, monthly coverages were routinely calculated from 2013 to 2021. Subsequently, annual, seasonal and climatological frequency products were derived and analyzed. The results indicate that the average yearly coverage has been increasing significantly (slope from 2013 to 2021).
Also, it was identified that the slope is dependent on the season of the year (dry and/or rainy), being more important during the latter. The climatological distribution shows that proliferations predominate between October and January, weaken from February to May, and then have a secondary maximum between June and August. From a spatial perspective, the areas of most significant accumulation and concern are on the northwest shores and the lake strait to the north. Our routine use of freely available imagery and a fast cloud computing environment could be used as a complement to monitor the performance of eradication and control plans by local governments in Maracaibo or in other regions of the world beset by similar invasions, especially when climate change threatens water resources.