Academia.eduAcademia.edu

Figure 1 – uploaded by Cesar Chagas

See full PDFdownloadDownload figure
Figure 1. Study area and location of soil sampling points.

Figure 1 Study area and location of soil sampling points.

Related Figures (10)

Table 1 Number of samples used per soil property and the depth of the layer after removal of “outliers”
Table 1 Number of samples used per soil property and the depth of the layer after removal of “outliers”
Table 2. Environmental covariates Table 3. Decision rules used to select the appropriate model for the prediction of soil properties
Table 2. Environmental covariates Table 3. Decision rules used to select the appropriate model for the prediction of soil properties
Source: Adapted from Ciampalini et al. (2012).
Source: Adapted from Ciampalini et al. (2012).
Figure 3. Variances of soil properties expressed as percentage of the overall variance. Sources: Batjes (2008) and Gray et al. (2009). Ordinary kriging was indicated as prediction model for sand and clay in the four surface layers to the depth of 60 cm (Table 4) because there were no significant correlations with the covariates,but
Figure 3. Variances of soil properties expressed as percentage of the overall variance. Sources: Batjes (2008) and Gray et al. (2009). Ordinary kriging was indicated as prediction model for sand and clay in the four surface layers to the depth of 60 cm (Table 4) because there were no significant correlations with the covariates,but
© Covariate correlated with p-value<0.005; spatial structure with p-value<0.10; “ M-medium; OK-ordinary kriging; CK-co- kriging; MLR-multiple linear regression; “ cross-validated r”. Table 4. Results of exploratory analysis based on the selected prediction model
© Covariate correlated with p-value<0.005; spatial structure with p-value<0.10; “ M-medium; OK-ordinary kriging; CK-co- kriging; MLR-multiple linear regression; “ cross-validated r”. Table 4. Results of exploratory analysis based on the selected prediction model
Table 5. Evaluation results of the MLR models by the introduction of categorical covariates for the functions “update” and “stepwise” of R Models defined as shown in figure 2; (*) interaction of the categorical covariate with all other continuous covariates; (:) specific interaction of categorical covariate with the continuous covariate. ® SWI: Soil Wetness Index; MSP: Mean Position of Declivity: TRI: Terrain Ruggedness Index; MRVBF: Multi-Resolution Valley Bottom Flatness; NDVI: Normalized Difference Vegetation Index); UCT: Land use and soil cover; SMU: Soil Mapping Units.
Table 5. Evaluation results of the MLR models by the introduction of categorical covariates for the functions “update” and “stepwise” of R Models defined as shown in figure 2; (*) interaction of the categorical covariate with all other continuous covariates; (:) specific interaction of categorical covariate with the continuous covariate. ® SWI: Soil Wetness Index; MSP: Mean Position of Declivity: TRI: Terrain Ruggedness Index; MRVBF: Multi-Resolution Valley Bottom Flatness; NDVI: Normalized Difference Vegetation Index); UCT: Land use and soil cover; SMU: Soil Mapping Units.
‘igure 5. Graphical representation of the comparison between predicted and estimated values for the variables OC and clay in the 15-30 cm layer and sand in the 60-100 cm layer, with the values of the coefficient of determination.
‘igure 5. Graphical representation of the comparison between predicted and estimated values for the variables OC and clay in the 15-30 cm layer and sand in the 60-100 cm layer, with the values of the coefficient of determination.
Figure 4.Adjustment of variograms for clay and sand.
Figure 4.Adjustment of variograms for clay and sand.
§oil layers defined by the natural break algorithm of Jenks (Jenks & Caspall, 1971). Table 6. Percentage distribution of area for OC, clay and sand
§oil layers defined by the natural break algorithm of Jenks (Jenks & Caspall, 1971). Table 6. Percentage distribution of area for OC, clay and sand
Related topics:
Environmental ScienceSoil ScienceDigital soil mappingNormalized Difference Vegetation Index
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved