Robust Statistical Processing of Long-Time Data Series to Estimate Soil Water Content

The research presented in this paper aims at providing a statistical model that is capable of estimating soil water content based on weather data. The model was tested using a long-time series of field experimental data from continuous monitoring at a test site in Oltrepo Pavese (northern Italy). An innovative statistical function was developed in order to predict the evolution of soil-water content from precipitation and air temperature. The data were analysed in a framework of robust statistics by using a combination of robust parametric and non-parametric models. Specifically, a statistical model, which includes the typical seasonal trend of field data, has been set up. The proposed model showed that relevant features present in the field of experimental data can be obtained and correctly described for predictive purposes.

Automated summary

This research presents a statistical model for estimating soil water content based on weather data. The model was tested using field experimental data, and a combination of robust parametric and non-parametric models was used for data analysis. The proposed model successfully captures the relevant features of the field data for predictive purposes.