Assessment Of Soil Quality For Sustainable Land Management Using Machine Learning And Digital Soil Mapping Techniques In Obudu Cattle Ranch, Nigeria

Abstract

Soil quality assessment is essential to know variation in nutrient concentrations within landscape for sustainable soil management. This study assessed soil quality in Obudu cattle ranch using machine learning and digital techniques. A total of 60 composite soil samples (0–30 cm depth) were collected at intervals of 200–500 m and selected soil physicochemical properties were determined. Digital elevation model (DEM) and Sentinel-2 satellite imageries were obtained, processed and applied for modelling. Soil quality was measured using total dataset (TDS) and minimum dataset (MDS). Linear (L) and non-linear (NL) scoring functions were applied, yielding four indices, MDS_L, MDS_NL, TDS_L and TDS_NL. Sixteen soil quality indicators (SQI) were used as TDS and were further screened for MDS using principal component analysis (PCA). Multiple linear regression was used to predict soil quality index in unsampled locations. The result showed that the soils were sandy loam, loam and sandy clay loam in texture.  pH ranged from strongly acidic to slightly acidic.  Soil organic carbon, CEC and base saturation were high while available P, exchangeable cations, exchangeable acidity as well as ECEC were low. The mean estimated soil quality for MDS_L, MDS_NL, TDS_L and TDS_NL were 0.415, 0.51, 0.42, and 0.49 respectively. MDS_NL model was the best model in predicting soil quality index in the area. All the models showed almost similar spatial distribution, with a high soil quality region mostly found in the southwestern part while low soil quality areas were located mostly in the central part and northwestern part of Obudu Mountain Resort.  The soil quality prediction class showed moderate class (class III) to be the dominant class covering greater part of the area with MDS_NL model. The predictive maps derived from this study should serve as a guide in the establishment of regionalized soil nutrient management programmes.