Features of Forecasting the State of Arid Territories Based on the SARIMA Model Using Remote Sensing Data

In the context of climate change, forecasting the state and growth of vegetation has become especially important. Software tools for forecasting the state of arid territories based on NDVI (Normalized Difference Vegetation Index) time series processing created. SARIMA (Seasonal Auto-Regressive Integrated Moving Average) model is used for NDVI time series forecasting and different scenarios of climatic changes for the period up to the end of the century for forecasting changes in the arid zone. Forecasts of spatial distributions of the aridity index for the territory of Volgograd Oblast made. For the averaged NDVI time series at the study area over the test period, R
²
was equal to 0.908, RMSE was equal to 0.052. To date, this is the largest public study conducted with the ARIMA family models for forecasting vegetation growth.