Application of machine learning and remote sensing in monitoring land use dynamics in tourism area

Abstract

Background: This study uses remote sensing and machine learning techniques to investigate the spatial-temporal changes in land use and land cover (LULC) within the Lake Toba tourism area over the past 35 years. Increasing tourism activities have significantly altered the region's landscape, particularly leading to a reduction in forest cover and an expansion of built-up areas. Method: By applying the Random Forest algorithm to satellite imagery data from Landsat 5, 8, and 9, and integrating Geographic Information System (GIS) technology, we analyzed and accurately predicted these changes. Additionally, indices such as NDVI and SAVI were used to monitor ecosystem health in detail, particularly for tracking the growth of invasive species like water hyacinths. Findings: LULC analysis of the Lake Toba tourism area reveals significant changes, including an increase in built-up areas, a decrease in vegetation, and the potential growth of water hyacinths. Surface temperature analysis indicates higher temperatures in built-up areas and cooler temperatures in natural vegetation. Using NDVI, SAVI, and MDWI indices also helped in monitoring water hyacinth growth, supporting improved ecosystem management for sustainability. Conclusion: This study highlights the environmental impacts of tourism and emphasizes the need for sustainable land management practices to balance development with ecological preservation. Novelty/Originality of this Research: This research demonstrates the effectiveness of combining machine learning with spatial technologies to support informed decision-making in land use planning.