Using network analysis to determine the soil quality indexes for land degradation

Land degradation poses a serious threat to soil quality health. Our study aimed to assess soil quality more effectively by establishing a valid and accurate soil quality index (SQI) for four land degraded levels in northeast China.

The minimum data set (MDS) and different scoring techniques (linear and nonlinear scoring) were selected through network analysis (NA) and principal component analysis (PCA). As potential SQI indicators, 11 physical, 12 chemical and 6 biological indicators were measured at 0 – 20 cm depth.

Our results showed that the soil properties were degraded and SQI decreased significantly with increasing land degradation. In addition, maize yield was positively related to SQI. The number of MDS generated by NA was much lower than that generated by PCA but increased the contributions of the indicators. For validating the accuracy and sensitivity of SQI, we found SQI-NA had greater correlations with maize yields and higher sensitivity indexes than SQI-PCA, implying that NA performs better in terms of accuracy and sensitivity to variation in soil quality under land degradation. So NA not only screens fewer metrics but also is more efficient in differentiating among SQIs. In addition, the SQIs calculated using the nonlinear integral through NA (NA-NL) had larger sensitivity index and F values than the other SQIs and were thus better able to discriminate under land degradation.

We conclude that NA-NL was recommended as a sensitive and effective approach for assessing SQIs at different land degradation levels.