Spatial-Temporal analysis of remotely sensed data in the Italian Alpine pastures

Exploring the spatial and temporal dynamic characteristics of regional forest net primary productivity (NPP) in the context of global climate change can not only provide a theoretical basis for terrestrial carbon cycle studies, but also provide data support for medium- and long-term sustainable pastures management planning of mountain regions. This study focuses on the Alpine Community of Giudicarie, located in the Province of Trento, Italy, through utilizing NASADEM (NASA Digital Elevation Model) 30m data from 2001 to 2020 as the main data sources. By integrating diverse variables: precipitation, Aspect, and Slopes, a statistical analysis was used in our study to clarify their relationship and interactions to Net Primary Productivity (NPP) in 20 years, then we examined five distinct Machine Learning algorithms: Linear Regression, Lasso Regression, KNN Regressor, Random Forest, and Gradient Boosting to identify the most effective model for estimating NPP in both spatial and temporal dimensions. The results show that the performance of the KNN model is better than the other models, by its adaptability to our datasets and its superior performance in terms of R² and RMSE values. The developed model is highly relevant for estimating Net Primary Productivity (NPP) in mountain pasture fields in both temporal and spatial dimensions, which serves as a valuable tool for informed decision-making in managing mountainous pastures, ensuring sustainable utilization and preservation of these vital ecosystems.

Spatial-Temporal analysis of remotely sensed data in the Italian alpine pastures