Ref: 184/AUS/26, Kaveeta Wadia , 24th ICSD 26, Australia.

Ref: 184/AUS/26 Geospatial Assessment to analyse the Spatial and Temporal Dynamics of Vegetation cover in the Kaluganga Catchment in Sri Lanka: Special Reference to Landslide hazards H.M.B.S. Herath 1*, Pavithra Jayasinghe 2, Isini Thisara Hewa Kokawalage 3, Thotamune Patabendige 4 1,2,3 Department of Geography & International Center for Multidisciplinary Studies (ICMS),Faculty of Humanities and Social Sciences, University of Sri Jayewardenepura, Sri Lanka. 4 National Building Research Institute Sri Lanka. *Corresponding author: badrahearath@sjp.ac.lk Abstract Upper catchments in the central hills in Sri Lanka face multiple overlapping natural and human-induced hazards, primarily driven by intense monsoonal rainfall, steep terrain, and changing land use. Landslides are one of the major forms of slope instability in hilly mountainous areas, commonly occurring through processes such as slope failure, cutting failure, and mass movement triggered by the interaction of rainfall, topography, geology, land-use changes, and vegetation conditions. Vegetation cover plays an important role in maintaining slope stability through root reinforcement, interception of rainfall, evapotranspiration, and regulation of soil moisture. Consequently, the progressive loss of vegetation may increase surface runoff, soil erosion, and slope instability, thereby contributing to conditions favorable for landslide occurrence. The Kalu Ganga Catchment in Sri Lanka encompasses diverse mountainous, undulating, agricultural, forested, and settlement environments and is frequently affected by rainfall-induced landslides and related slope failures. Therefore, understanding the spatial relationship between long-term vegetation loss and landslide occurrence is important forsustainable land and disaster-risk management. This study assessed the spatial and temporal dynamics of vegetation cover in the Kalu Ganga Catchment, Sri Lanka, and its spatial association with landslide occurrence, using geospatial technologies and remotely sensed data. Vegetation dynamics was assessed using Normalized Difference Vegetation Index (NDVI) using Landsat images acquired in 1995, 2005, 2015, and 2025. Selected vegetation cover classes were dense forest cover, sparse forests and agricultural lands, while non-vegetation samples represented bare land, built-up areas, and water bodies. Random Forest classification framework was adopted with training samples to classify vegetation and non-vegetation areas and derive multi-temporal vegetation dynamic patterns by comparing the classified vegetation conditions between 1995–2005, 2005–2015, and 2015–2025. Pearson’s coefficient test was carried out to identify the correlation between vegetation changes and the occurrence of landslides statistically. The results indicate substantial spatial and temporal variations in vegetation loss across the catchment. Approximately 73,139 acres of vegetation loss were identified during 1995–2005, followed by 61,456 acres during 2005–2015 and 100,176 acres during 2015–2025. To investigate the relationship between vegetation loss and landslide occurrence, landslide locations recorded in 2017, 2019, 2020, 2021, and 2022 were analyzed using ArcGIS Pro. A 180 m spatial buffer was generated around each landslide location, and the buffered landslide areas were overlaid with the identified vegetation-loss zones. Of the 87 recorded landslides, 85 were spatially associated with vegetation-loss areas identified during 1995–2015, comprising 56 landslides in 2017, one in 2019, 14 in 2020, 12 in 2021, and two in 2022. According to Pearson’s correlation analysis, there is a strong positive correlation (r =0.71) between the vegetation changes and the frequency of the occurrence of landslides. These findings confirm the strong spatial association between historical vegetation loss and subsequent landslide occurrence in the Kalu Ganga Catchment. Based on empirical evidence, the study proposes a proactive set of strategies for vegetation-centered slope stabilization with the participation of multi-stakeholders to reduce the risk of landslides in Kalu Ganga Catchment. The framework of the study can be taken as a model to reduce the risk of landslides in other tropical monsoon regions Keywords: Geo-Spatial Assessment, Landslides, Normalized Difference Vegetation Index, Vegetation Dynamics, Disaster Risk Management