3.4 Selecting soil quality indicators
To evaluate the sensitivity of different soil chemical and physical quality indicators, we performed a principal component analysis. The first three principal components were selected because they accounted for approximately 50% of the total variation. Though the contribution of the three PC to the explanation of the total variation is relatively small, it still displays loadings above 0.4 with at least three soil properties (S. Material 3). CEC, total carbon, total Phosphorus, and Inorganic Phosphorus contributed more strongly to the variability of PC1. PC2 is positive related to available ammonium, available Olsen-P and Porosity, and negative related to nitrate. PC3, which explains the 11% of the total variation, have large negative loadings for the infiltration rate and K-RUSLE and positive with porosity.
The k-means clustering algorithm display that the CG site tends to differ from the OG sites even though both PC’s only accounts for approximately 40% of the total variation (Figure 4). The chemical properties allow a fair clustering of the sites where loadings of total and inorganic Phosphorus, CEC, and Kjeldahl Nitrogen contribute the most. All the differences described above and shown in Figure 5.