Fig 1.
Location map of the study area.
((a) the black soil region of northeast China and the location of the watershed; (b) digital elevation model of the watershed; (c) land use classification and photographs showing exposed soil sections and land use within the study area).
Table 1.
Dataset and resources.
Table 2.
Descriptive statistics of thickness of black soil layer of soil samples.
Fig 2.
Histogram of thickness of black soil layer of sample points.
Table 3.
Analysis of variance of thickness of black soil in different terrain.
Fig 3.
Relationships between soil thickness and terrain variables.
Table 4.
Out-of-bag data error of random forest.
Table 5.
Comparison of estimated performance of random forest and interpolation models.
Fig 4.
Comparison of observed value and estimated value from random forest and interpolation models.
Fig 5.
Variable importance and variable explanation of random forest.
Fig 6.
Random forest and kriging modal results ((a) random forest; (b) classical bayesian kriging; (c), ordinary kriging; (d)simple kriging).
Table 6.
Soil erosion classification results.
Fig 7.
RUSLE results grading and proportion.
Fig 8.
Risk assessment levels and proportions of soil and water erosion.