Ravazzolo et al. (2026) Towards integrated short-term Rain-on-Grid modeling and long-term RUSLE estimates for improved erosion susceptibility assessment in the Oltrepò Pavese hills of Northern Italy

Identification

• Journal: Journal of Hydrology Regional Studies
• Year: 2026
• Date: 2026-03-03
• Authors: D. Ravazzolo, Diego Ravazzolo, Andrea Fenocchi, G. Petaccia, C. Costanzo, Wafae Ennouini, M. Bordoni, C. Meisina, Pierfranco Costabile, Stefano Sibilla
• DOI: 10.1016/j.ejrh.2026.103290

Research Groups

• Department of Civil Engineering and Architecture, University of Pavia, Italy
• Department of Environmental Engineering, University of Calabria, Italy
• Department of Earth and Environmental Sciences, University of Pavia, Italy

Short Summary

This study evaluates the complementary use of the empirical Revised Universal Soil Loss Equation (RUSLE) and a two-dimensional Rain-on-Grid (RoG) hydrodynamic model for erosion susceptibility assessment in Northern Italy. The models showed over 50% spatial overlap in identifying erosion-prone areas, with RoG better reproducing event-based erosion zones and RUSLE capturing land-cover effects, offering a practical integrated framework for data-scarce catchments.

Objective

• To evaluate the complementary use of the empirical Revised Universal Soil Loss Equation (RUSLE) and a two-dimensional Rain-on-Grid (RoG) hydrodynamic model to assess their spatial agreement in identifying erosion-prone areas and compare them with an erosion inventory.

Study Configuration

• Spatial Scale: Three catchments in the Oltrepò Pavese area of Northern Italy: Scuropasso (40.66 km²), Versa (56.17 km²), and Ardivestra (47.23 km²). Digital Terrain Model (DTM) resolution: 20 m. RoG model grid: triangular elements with 3 m edge lengths. Rainfall Intensity-Duration-Frequency (IDF) data spatial resolution: approximately 1.5 km × 1.5 km.
• Temporal Scale: RUSLE provides long-term average annual erosion estimates. RoG simulates event-based runoff dynamics and sediment transport capacity for a synthetic storm event with a 2-year return period (4.5 hours duration, 15-minute discretization). Model predictions were compared with an erosion inventory mapped after a major storm event in 2009 (estimated return period 5–50 years).

Methodology and Data

• Models used:
  • Revised Universal Soil Loss Equation (RUSLE)
  • Two-dimensional Rain-on-Grid (RoG) hydrodynamic model (solving 2D Shallow Water Equations) coupled with an empirical sediment transport formulation (Zhang et al., 2009).
  • Soil Conservation Service Curve Number (SCS-CN) method for runoff estimation in RoG.
  • Chicago hyetograph for rainfall temporal distribution in RoG.
• Data sources:
  • Digital Terrain Model (DTM) 20 m resolution (Regione Lombardia).
  • Erosion zone inventory from the 2009 storm event (digitized from field survey data and 0.15 m resolution orthophotos).
  • Intensity-Duration-Frequency (IDF) curves from ARPA Lombardia (based on 1987–2021 rainfall data).
  • Daily precipitation data from the ARCIS database (2001–2021) for RUSLE R-factor calculation.
  • Regional soil database of Regione Lombardia for RUSLE K-factor determination.
  • Land use classification maps (DUSAF 7, 2023) from Regione Lombardia for RUSLE C-factor and RoG Manning's roughness coefficients and Curve Numbers.

Main Results

• The RoG-based and RUSLE approaches showed significant spatial agreement, with over 50% overlap in identifying erosion-prone areas (Ardivestra: 55%, Scuropasso: 66%, Versa: 54%). These overlapping zones were primarily located on steep hillslopes and areas of high flow accumulation.
• The RoG-based approach reproduced the mapped 2009 erosion zones more effectively than RUSLE, particularly in the Versa catchment (93.3% RoG vs 42% RUSLE overlap), highlighting the dominant role of short-duration, high-intensity runoff-driven processes. In Scuropasso, both models showed high agreement (86.6% RoG vs 86.7% RUSLE).
• RUSLE better captured land-cover effects due to its explicit C-factor.
• Most observed erosion zones fell within the lowest erosion classes predicted by both models, suggesting their primary utility is in identifying spatial susceptibility patterns rather than precise quantitative erosion magnitudes.
• Mean annual soil loss estimated by RUSLE (after removing stream channels): Ardivestra 0.83 t ha⁻¹ a⁻¹, Scuropasso 1.31 t ha⁻¹ a⁻¹, Versa 1.09 t ha⁻¹ a⁻¹.
• Mean event-scale erosion indicators from RoG (for a 2-year return period event, after removing stream channels): Ardivestra 0.07 t ha⁻¹ event⁻¹, Scuropasso 0.10 t ha⁻¹ event⁻¹, Versa 0.06 t ha⁻¹ event⁻¹.

Contributions

• Proposes and evaluates a comparative, integrated framework for erosion susceptibility assessment by leveraging the complementary strengths of empirical (RUSLE) and process-based (RoG) models.
• Demonstrates that spatial overlap between models operating at different temporal scales (long-term average vs. event-based) can enhance confidence in identifying high-risk erosion areas, particularly in data-scarce environments.
• Provides insights into the distinct sensitivities of each model: RoG effectively captures runoff-driven erosion along concentrated flow paths, while RUSLE integrates long-term factors including land cover.
• Offers a practical and transferable framework for prioritizing areas for field validation, monitoring, and targeted soil conservation and land management interventions.
• Contributes to understanding the spatial consistency between model-predicted erosion-prone areas and observed geomorphic responses during extreme rainfall events.

Funding

• Project Nord Ovest Digitale E Sostenibile (NODES) funded by the European Union - NextGenerationEU, Mission 4 Component 1.5 - ECS00000036 - CUP F17G22000190007.

Citation

@article{Ravazzolo2026Towards,
  author = {Ravazzolo, D. and Ravazzolo, Diego and Fenocchi, Andrea and Petaccia, G. and Costanzo, C. and Ennouini, Wafae and Bordoni, M. and Meisina, C. and Costabile, Pierfranco and Sibilla, Stefano},
  title = {Towards integrated short-term Rain-on-Grid modeling and long-term RUSLE estimates for improved erosion susceptibility assessment in the Oltrepò Pavese hills of Northern Italy},
  journal = {Journal of Hydrology Regional Studies},
  year = {2026},
  doi = {10.1016/j.ejrh.2026.103290},
  url = {https://doi.org/10.1016/j.ejrh.2026.103290}
}