Mapping the Spatial Extent and Dynamics of Sago Palm (Metroxylon sagu Rottb.) in Butuan City, Philippines Using Sentinel-2 and Google Earth Engine

Sago palm (Metroxylon sagu Rottb.) is an important wetland plant that holds significant ecological, economic, and cultural value. However, updated information on its spatial distribution and change dynamics remains limited. In this study, we mapped the spatial extent of sago palm in Butuan City and assessed changes in its distribution from 2012 to 2024 using advanced geospatial and remote sensing technologies. Sentinel-2 imagery acquired in 2024 was processed using Google Earth Engine (GEE). Five supervised machine-learning algorithms—Linear Support Vector Machine (SVM), Naïve Bayes (NB), Random Forest (RF), Classification and Regression Tree (CART), and Gradient Tree Boost (GTB)—were evaluated for land-cover classification. Eight land-cover classes were considered, including Sago Palm. Classification accuracy was rigorously assessed using an independent validation dataset, evaluating overall accuracy, producer’s accuracy, and user’s accuracy. Of the algorithms tested, Random Forest achieved the highest overall accuracy (93.65%) and was thus adopted for the initial 2024 sago palm mapping. The resulting sago palm classification was further refined using high-resolution imagery in Google Earth Pro and validated with field-based reference information. The finalized 2024 map delineated 26.37 ha of sago palm distributed across 13 barangays in Butuan City, with Baan KM 3 encompassing the largest mapped area at 6.04 ha. Comparative analysis with the 2012 sago palm map revealed 9.59 ha of retained areas, 17.04 ha of diminished or lost areas, and 16.78 ha of expanded areas. These findings indicate notable spatial turnover in sago palm distribution, despite a relatively modest net change in overall mapped area between the two periods. This study demonstrates the effectiveness of integrating Sentinel-2 imagery, Google Earth Engine, and high-resolution image refinement for generating reliable spatial information on sago palm distribution and change. The resulting maps offer a valuable baseline for ongoing monitoring, conservation planning, and land-use management of sago palm ecosystems in Butuan City, supporting both scientific inquiry and policy formulation.