Forecasting Deforestation Trends in the Sequoia National Park

Abstract

This study examines several models for predicting deforestation trends in a specific region of Sequoia National Park using the Normalized Difference Vegetation Index (NDVI) and the Leaf Area Index (LAI). For the study, a comprehensive extracted dataset from Google Earth Engine was used, containing images from MODIS sensors of the Terra and Aqua satellites from 2000 to 2022. These indices, which indicate the vitality of vegetation and are not constant over time, serve as a basis for predicting deforestation trends. Advanced forecasting models, including XGBoost, SARIMAX, and Prophet, are used to estimate deforestation rates for future intervals of one to ten years, based on the dataset and model specifics. The accuracy of these predictions is primarily assessed using the root mean square error (RMSE) to ensure the reliability of the model assessments. This study contributes to the understanding of deforestation dynamics and provides a methodological framework for future conservation planning and management in forested national parks.