Ity) . The AUC = 1 worth plot inside the upper left corner represents the 100 sensitivity vs. 100ISPRS Int. J. Geo-Inf. 2021, ten,eight ofo-Inf. 2021, 10, x FOR PEER REVIEWspecificity representative of your higher predictive functionality in the model. Following the 9 of 22 recommendations in , the prediction talent of the model was acceptable, superb or outstanding in the event the AUC values exceeded 0.70, 0.80 or 0.90, respectively. The original gully inventory dataset of Bosino et al. (2020)  was split by using the Mkhomazi catchment basin gully information and also the Lotheni River a considerable role in As shown in Figure four, the following environmental layers played catchment information that have been used as education data. The Mkhomazana River catchment gullies have been used for testing and also the formation of the gully erosion: catchment region, NDVI, Vertical Distance to Channel validating the model.Network, lithology, valley depth, TPI, land use and SPI. To know the precise contributions of the3. Results the individual variable response curves were derived, as illusvariables, The initial benefits from A study are represented trated in Figures five and 6 for gully typesthisand B, respectively.by the inventory map from the 122 gully Lydicamycin Autophagy erosion types and features situated within the study region: 85 options of gully form A and For every single of your two gully forms, the variable significance with the single independent 37 options of gully 4). B. Variety A significant variables about 0.7 km2 while form variables was evaluated (Figure typeThe mostgullies covered an area2ofthat describe, form A B gullies covered an location of about 1 km2 for a total of 1.7 km . gullies are represented by the (1) catchment location (33.1), (2) Vertical Distance to Channel As shown in Figure 4, the following environmental layers played a important part Network (12.3), the formation of your gully(four) valley depth (11.1) and Vertical Distance to Chanin (3) lithology (11.7), erosion: catchment area, NDVI, (five) NDVI (ten.eight). As an alternative, the formation of kind B gullies was primarily dominated by the (1) NDVI (18.eight), nel Network, lithology, valley depth, TPI, land use and SPI. To understand the certain contributions the variables, the Position Index response (4) valley derived, (two) catchment location (15.2),of(three) Topographicindividual variable (11.six), curves have been depth as illustrated in Figures 5 and 6 for Network (ten.7). (11.1) and (5) Vertical Distance to Channelgully varieties A and B, respectively.Figure four. Relative significance with the environmentalenvironmental variables. Figure four. Relative importance in the variables.Sort A gullies For each and every of the two gully sorts, the Sonidegib metabolite M48 MedChemExpress variable1140 m andof the single independent develop in an elevation variety in between significance 1170 m, too as variables was evaluated (Figure four). One of the most essential variables that describe type A above 1460 m, mainly where the valley depth is in between 65 m and 75 m. The response gullies are represented by the (1) catchment location (33.1), (two) Vertical Distance to Channel curve for the catchment (12.three), (three) lithology (11.7), parameter, showed low values. AddiNetwork area, one of the most influencing (4) valley depth (11.1) and (5) NDVI (10.8). tionally, they had been characterized byof type B amongst 60 m and 90 m forby the (1) NDVI (18.8), Rather, the formation values gullies was mainly dominated the Vertical Dis(2) catchment Based on Figure 5, Position Index (11.6), connected with tance to Channel Network. area (15.2), (3) Topographictype A gullies have been (4) valley depth (11.1) and and.