Understanding the spatiotemporal dynamics of vegetation cover change (VCC) in the Teesta basin: a geospatial and statistical modelling–based investigation of environmental and human factors

Abstract

This study presents an insightful analysis of vegetation cover change (VCC) within the Teesta Basin, utilizing various statistical models and geospatial techniques. The Ordinary Least Squares (OSL) regression model reveals a modest explanatory power with an adjusted R² of 0.1398, indicating its ability to account for approximately 13.98% of the variance in the data. This model, however, hints at potential heteroscedasticity and non-normal distribution of errors. In contrast, the geographically weighted regression (GWR) model, accounting for 84.556% of the variance, demonstrates a robust spatial heterogeneity in the relationships between the variables, offering a more nuanced understanding of the regional disparities. The study further incorporates a hot spot analysis using the Getis-Ord Gi* statistic, which exposes significant spatial clustering patterns in VCC, emphasizing the influence of both environmental and anthropogenic factors. The Boosted Regression Tree (BRT) model, with a substantial relative influence of 44.11% from ‘Population Proximity’, highlights the critical role of human-driven factors in vegetation dynamics. This model shows a moderate to strong correlation in predicting NDVI values. Analysis of seasonal trends reveals a cyclic pattern in NDVI values, indicating pronounced seasonal variations and negative trends in vegetation activity over time, particularly in the lower basin area. The Mann–Kendall time series analysis further confirms this declining vegetation trend. The study’s findings are crucial for understanding the spatial and temporal dynamics of vegetation cover in the Teesta Basin. They underscore the importance of considering both environmental and human-driven factors in conservation strategies, especially in protected forest regions.

Graphical Abstract