Open Conference Systems, CLADAG2023

Font Size: 
Spatial modelling of pyroclastic cover deposit thickness with remote sensing data and ground measurements: a forecasting combination approach
Raffaele Mattera, Germana Scepi, Pooria Ebrahimi, Fabio Matano

Last modified: 2023-06-19


Thickness of pyroclastic deposits governs various geomorphological and hydrological processes, but studies on the areas characterized by pyroclastic soil coverage are limited in the literature worldwide and the existing models predict thickness mainly based on morphological features of the slope. In this paper, additional variables are also derived from Digital Elevation Model (DEM) and satellite multispectral images to propose a spatial model for forecasting the thickness of pyroclastic deposits. For the prediction model, a two-step procedure is adopted: (1) the best subset of variables is selected; and (2) the predictions from different schemes are combined for deriving the final model. Predictive accuracy tests verify that the combination procedure provides a statistically significant improvement in predictions.