Soil quality index variation and MIR-based discrimination in a former lakebed and adjacent soils under cultivation

We studied the influence of land use and depth on soil quality indices across a former lakebed and adjacent soils under cultivation and evaluated the predictability of soil quality index (SQI) calculated using linear (SQI-L) and non-linear (SQI-NL) approaches from Mid-infrared (MIR) spectra. A total of 108 soil samples were collected, and soil samples were scanned using a mid-infrared (MIR) (4000–600 cm –1 ) spectrometer. Land use significantly (P < 0.05) affected o;rganic carbon (OC), CaCO₃, concentrations of P, Cu, Fe, Mn, and total nitrogen (TN), and SQI-L and SQI-NL. Soil depth significantly (P < 0.05) influenced silt, AWC, OC, bulk density (BD), CaCO₃, pH, EC, concentrations of K, P, Cu, Zn, Mn, TN, SQI-L, and SQI-NL, with both SQI-L and SQI-NL decreasing with increasing soil depth. Significant Land use × Depth interactions (P < 0.05) were observed for P, Cu, Zn, and Fe, indicating depth-dependent responses to land use. Significant Land use × depth interactions (P < 0.05) were observed for P, Cu, Zn, and Fe concentrations, indicating depth-dependent responses to land use. Across all depths, OC and SQI-L and SQI-NL were significantly lower in soils of the former lakebed, whereas CaCO₃ and pH were higher in soils of the croplands. SQI-NL showed slightly higher variability than SQI-L, particularly at 0–20 cm and 20–40 cm. The analyzed lakebed soils contained higher proportions of calcite, dolomite, smectite, and vermiculite than the analyzed soils under cultivation. MIR spectroscopy successfully predicted SQI-L (R 2 = 0.87) over SQI-NL (R 2 = 0.85) using the PLSR model. We conclude that MIR spectra can distinguish SQI classes.