An assessment of mangrove forest in northwestern Mexico using the Google Earth Engine cloud computing platform

Mangrove forests are commonly mapped using spaceborne remote sensing data due to the challenges of field endeavors in such harsh environments. However, these methods usually require a substantial level of manual processing for each image. Hence, conservation practitioners prioritize using cloud computing platforms to obtain accurate canopy classifications of large extensions of mangrove forests. The objective of this study was to analyze the spatial distribution and rate of change (area gain and loss) of the red mangrove (Rhizophora mangle) and other dominant mangrove species, mainly Avicennia germinans and Laguncularia racemosa, between 2015 and 2020 throughout the northwestern coast of Mexico. Bimonthly data of the Combined Mangrove Recognition Index (CMRI) from all available Sentinel-2 data were processed with the Google Earth Engine cloud computing platform. The results indicated an extension of 42865 ha of red mangrove and 139602 ha of other dominant mangrove species in the Gulf of California and the Pacific northwestern coast of Mexico for 2020. The mangrove extension experienced a notable decline of 1817 ha from 2015 to 2020, largely attributed to the expansion of aquaculture ponds and the destructive effects of hurricanes. Considering the two mangrove classes, the overall classification accuracies were 90% and 92% for the 2015 and 2020 maps, respectively. The advantages of the method compared to supervised classifications and traditional vegetation indices are discussed, as are the disadvantages concerning the spatial resolution and the minimum detection area. The work is a national effort to assist in decision-making to prioritize resource allocations for blue carbon, rehabilitation, and climate change mitigation programs.