Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

< Back to Article

Fig 1.

Study area.

Geomorphological regions of Shaanxi: Northern, Guanzhong, and Southern. Right panels show DEM data for the study areas in Northern and Southern Shaanxi.

More »

Fig 1 Expand

Table 1.

Evaluation factors and data sources affecting landslides in the study area.

More »

Table 1 Expand

Fig 2.

Evaluation factor map of Northern Shaanxi.

The map reveals the spatial distribution characteristics of the aspect factor in the loess hilly-gully regions of Northern Shaanxi, with spatial attributes extracted from a high-resolution Digital Elevation Model.

More »

Fig 2 Expand

Fig 3.

Evaluation Factor Map of Southern Shaanxi.

The map reveals the spatial distribution characteristics of the aspect factor in the mountainous landscapes of Southern Shaanxi, with spatial attributes extracted from a high-resolution Digital Elevation Model.

More »

Fig 3 Expand

Fig 4.

Factor Correlation and Multicollinearity Analysis in Northern Shaanxi Region.

(a) The sub-figure combines a Pearson correlation matrix with a topological network to quantitatively reveal the interconnections among evaluation factors and their nature of influence on the target variable (red lines for positive and blue lines for negative correlation). (b) The radial plot further evaluates the multicollinearity of input features, ensuring the robustness of the predictive model by calculating the VIF values of each factor.

More »

Fig 4 Expand

Fig 5.

Factor Correlation and Multicollinearity Analysis in Southern Shaanxi Region.

(a) The sub-figure combines a Pearson correlation matrix with a topological network to quantitatively reveal the interconnections among evaluation factors and their nature of influence on the target variable (red lines for positive and blue lines for negative correlation). (b) The radial plot further evaluates the multicollinearity of input features, ensuring the robustness of the predictive model by calculating the VIF values of each factor.

More »

Fig 5 Expand

Table 2.

Comparison of Sensitivity Model Accuracy for Landslides in Northern Shaanxi.

More »

Table 2 Expand

Fig 6.

Augmentation values of ROC curves in Northern Shaanxi.

(a) training set; (b) test set.

More »

Fig 6 Expand

Table 3.

Comparison of Sensitivity Model Accuracy for Landslides in Southern Shaanxi.

More »

Table 3 Expand

Fig 7.

Augmentation values of ROC curves in Southern Shaanxi.

(a) training set; (b) test set.

More »

Fig 7 Expand

Fig 8.

Prediction map of landslide-prone areas in Northern Shaanxi.

The prediction map further categorizes the susceptibility into five levels (from low to high risk), revealing the landslide distribution patterns in the Northern Shaanxi loess hilly-gully regions.

More »

Fig 8 Expand

Fig 9.

Prediction map of landslide-prone areas in Southern Shaanxi.

The prediction map further categorizes the susceptibility into five levels (from low to high risk), revealing the landslide distribution patterns in the Southern Shaanxi mountainous regions.

More »

Fig 9 Expand

Fig 10.

SHAP Analysis Map of Northern Shaanxi Region.

(a) The sub-figure illustrates the global feature importance for landslide prediction in Northern Shaanxi, with population density (POP) and Topographic Wetness Index (TWI) identified as the most influential factors. (b) The summary plot reveals the specific impact of feature values: the color represents the feature value (red for high, blue for low), and the horizontal axis indicates the contribution to the landslide probability.

More »

Fig 10 Expand

Fig 11.

SHAP Analysis Map of Southern Shaanxi Region.

(a) The sub-figure displays the global importance ranking of evaluation factors based on the mean absolute SHAP values, identifying DEM, LULC, and Lithology as the dominant factors. (b) The summary plot further reveals the direction of influence: colors represent feature values (red for high, blue for low), and SHAP values greater than 0 indicate an increased probability of landslide occurrence.

More »

Fig 11 Expand

Fig 12.

Single-factor dependency map of the Northern Shaanxi region.

(a) Dependence Polt: POP; (b) Dependence Polt: TWI; (c) Dependence Polt: DEM; (d) Dependence Polt: Slope; (e) Dependence Polt: NDVI; (f) Dependence Polt: LULC. The red solid lines represent the marginal effects of evaluation factors on landslide susceptibility, while the pink dots indicate sample distribution. Green dashed lines and shaded areas denote the critical thresholds for each factor.

More »

Fig 12 Expand

Fig 13.

Single-factor dependency map of the Southern Shaanxi region.

(a) Dependence Polt: DEM; (b) Dependence Polt: LULC; (c) Dependence Polt: Litholo; (d) Dependence Polt: Profile_curve; (e) Dependence Polt: NDVI; (f) Dependence Polt: DTR. The red solid lines represent the marginal effects of evaluation factors on landslide susceptibility, while the pink dots indicate sample distribution. Green dashed lines and shaded areas denote the critical thresholds for each factor.

More »

Fig 13 Expand

Fig 14.

Correlation Analysis of Evaluation Factors in Northern Shaanxi Region.

Node size represents the feature importance, and the thickness of the lines indicates the interaction strength between factors. Line colors denote the interaction type (green for synergistic and pink for antagonistic), while node colors reflect the impact direction of each factor on landslide susceptibility (red for increasing risk and blue for decreasing risk).

More »

Fig 14 Expand

Fig 15.

Correlation Analysis of Evaluation Factors in Southern Shaanxi Region.

Node size represents the feature importance, and the thickness of the lines indicates the interaction strength between factors. Line colors denote the interaction type (green for synergistic and pink for antagonistic), while node colors reflect the impact direction of each factor on landslide susceptibility (red for increasing risk and blue for decreasing risk).

More »

Fig 15 Expand