A framework based on mechanistic modelling and machine learning for soil moisture estimation
Abstract Soil moisture has many important applications in subsurface water movement. Methods for measuring soil moisture are either destructive or require prior calibration. The aim of this study was to provide a framework for estimating soil water content with limited data (rainfall and evaporation...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Springer
2025-07-01
|
| Series: | Discover Soil |
| Subjects: | |
| Online Access: | https://doi.org/10.1007/s44378-025-00086-9 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract Soil moisture has many important applications in subsurface water movement. Methods for measuring soil moisture are either destructive or require prior calibration. The aim of this study was to provide a framework for estimating soil water content with limited data (rainfall and evaporation). The hydraulic properties of four layers of soil from Tunisia were measured to a depth of 80 cm. These values were used as inputs to the Hydrus 1D model to generate soil water content profiles over a 27-year period. 109 profiles were calculated using machine learning algorithms (regression trees; random forest; support vector machine) to predict soil moisture content from daily values of rainfall and evaporation. The statistical performance indices of prediction, root mean square error and correlation coefficient indicate the superiority of regression trees over other methods for all soil layers and during both calibration and validation processes and reproducing the seasonal variation of soil moisture. |
|---|---|
| ISSN: | 3005-1223 |