Wehliye et al. (2026) Nano clay mitigation of wind induced soil loss across aggregate sizes through wind tunnel experiments and explainable machine learning

Identification

• Journal: Scientific Reports
• Year: 2026
• Date: 2026-09-27
• Authors: Abdirashid Ali Wehliye, Sema Kaplan
• DOI: 10.1038/s41598-026-73412-y

Research Groups

• Graduate School of Natural and Applied Sciences, Soil Science and Plant Nutrition, Erciyes University, Kayseri, Türkiye
• Department of Soil Science and Plant Nutrition, Faculty of Agriculture, Erciyes University, Kayseri, Türkiye

Short Summary

This study investigated the effectiveness of nano-clay in mitigating wind-induced soil loss across various aggregate sizes and wind speeds using wind tunnel experiments and explainable machine learning, demonstrating that nano-clay significantly reduced soil loss, particularly for finer aggregates and at higher wind speeds.

Objective

• To investigate the effects of different nano-clay application rates (0, 0.94, 1.88, and 3.75 g m⁻²) on wind-induced soil loss across different aggregate sizes (0.5, 1.0, and 2.0 mm) and wind speeds (8, 10, and 12 m s⁻¹) under wind-tunnel conditions.
• To model wind-induced soil loss processes using machine learning and explainable artificial intelligence methods to characterize predictive relationships within the experimental domain.

Study Configuration

• Spatial Scale: Wind tunnel experiments using soil trays with a surface area of 0.16 m² (0.75 m × 0.21 m × 0.025 m). The wind tunnel's test section was 5 m long with an inlet cross-section of 0.80 m × 0.90 m.
• Temporal Scale: Nano-clay treated trays were kept under laboratory conditions for 48 hours before experiments. Each tray was exposed to wind for 5 minutes during the experiments.

Methodology and Data

• Models used: Linear Regression, Random Forest, Extreme Gradient Boosting (XGBoost). Explainable AI methods included Shapley Additive Explanations (SHAP), Partial Dependence Plots (PDPs), and Accumulated Local Effects (ALE) analyses.
• Data sources: Wind tunnel experiments (108 experimental units from 3 aggregate sizes × 4 nano-clay doses × 3 wind speeds × 3 replicates) measuring soil loss (%). Soil material was collected from a wind-erosion-prone sandy agricultural area in İncesu district of Kayseri province, Türkiye. Field Emission Scanning Electron Microscopy (FE-SEM) was used for qualitative surface morphology analysis.

Main Results

• Nano-clay applications significantly reduced wind-induced soil loss across all aggregate sizes and wind-speed conditions.
• The highest protective effect was observed at the 3.75 g m⁻² nano-clay rate, reducing soil loss by 80.8% in 2 mm aggregates, 93.2% in 1 mm aggregates, and 95.9% in 0.5 mm aggregates compared to the control.
• The protective effect of nano-clay was particularly pronounced at high wind speeds and in fine aggregates.
• XGBoost demonstrated the strongest development-stage predictive performance with a pooled out-of-fold R² of 0.835 ± 0.064, RMSE of 11.896 ± 2.229, and MAE of 6.686 ± 0.854. On the independent holdout test set, it achieved R² = 0.882, RMSE = 13.298, and MAE = 9.443.
• SHAP analysis indicated that nano-clay dose had the largest mean absolute contribution to model predictions (22.661), followed by wind speed (13.866) and aggregate size (2.717). Higher nano-clay doses were associated with lower predicted soil loss, while higher wind speeds were associated with higher predicted soil loss.
• FE-SEM images qualitatively showed apparent particle bridging and clustering in nano-clay-treated samples.

Contributions

• First study to integrate a controlled factorial wind-tunnel experiment with treatment-combination-aware machine-learning validation for wind erosion mitigation.
• Experimentally quantified nano-clay effects across a range of wind speeds and aggregate sizes, providing insights into its protective mechanisms.
• Utilized explainable artificial intelligence (SHAP, PDP, ALE) to characterize and interpret the predictive relationships of nano-clay dose, wind speed, and aggregate size on soil loss.
• Demonstrated the potential of nano-clay as a novel and sustainable surface-stabilization approach for mitigating wind erosion in highly erodible sandy soils.

Funding

This research received no specific grant, financial support, or funding from any public, commercial, or not-for-profit funding agency.

Citation

@article{Wehliye2026Nano,
  author = {Wehliye, Abdirashid Ali and Kaplan, Sema},
  title = {Nano clay mitigation of wind induced soil loss across aggregate sizes through wind tunnel experiments and explainable machine learning},
  journal = {Scientific Reports},
  year = {2026},
  doi = {10.1038/s41598-026-73412-y},
  url = {https://doi.org/10.1038/s41598-026-73412-y}
}