Assessing the effects of deforestation and intensive agriculture on the soil quality through digital soil mapping

[Display omitted] •Different soil quality indices were applied to evaluate soil quality.•Different data set and scoring approaches used to calculate soil quality indices.•SQIn showed better estimation of soil quality comparing to SQIa and SQIw.•Deforestation led to a decrease in soil quality.•RF technique through digital soil mapping is a robust method to monitor soil quality. This study was designed to evaluate soil quality (SQ) in deforested and intensively cultured lands in Mazandaran Province, Iran. For this purpose, three soil quality indices (SQIs: additive soil quality index (SQIa), nemoro soil quality index (SQIn) and weighted additive soil quality index (SQIw)) and two datasets (the total data set (TDS) and minimum data set (MDS)) were determined. The linear (L) and non-linear (NL) scoring systems were also used to calculate each SQI. Eight soil properties, including pH, cation exchange capacity (CEC), electrical conductivity (EC), bulk density (BD), soil microbial respiration (SMR), total nitrogen (TN), soil organic carbon (SOC), and calcium carbonate equivalent (CCE), were measured in 108 locations (0–30 cm depth). The MDS was determined by the principal component analysis. A digital soil mapping (DSM) method, more specifically the random forest technique, was applied to produce the SQI maps. The maximum and minimum values of the SQIs were obtained in the natural forest (NF) and the dry farming (DR) land uses, respectively. The results indicated that both methods (i.e., TDS and MDS) properly could describe SQ in this area but MDS can be recommended as an appropriate method because it was able to classify the SQ using a lower number of the soil properties without losing information to SQ assessment. In addition, NF and the more-than-ten-year-old commercial garden (G10 + ) land uses had the highest proportions of SQ grades I and II, respectively. Conversely, DR and the less-than-ten-year-old commercial gardens (G10) land uses had the highest proportions of the SQ grade V (very low quality) among the six SQIs. The findings indicated that the L scoring system had higher agreement values for all SQIs compared to the NL scoring system, and also SQIn (R2 = 0.874) could provide a better estimation compared to the SQIa (R2 = 0.864) and SQIw (R2 = 0.865). The spatial distribution of the SQ grades using DSM indicated that the land use conversion could decrease SQIs, suggesting that close attention should be paid to the sustainable use of the agricultural l