University of Tehran Press Iranian Journal of Soil and Water Research 2008-479X 57 3 2026 05 22 Digital Mapping of Soil Quality Using Satellite Imagery and Machine Learning Algorithms (A Case Study of Lushan, Guilan Province, Iran) Digital Mapping of Soil Quality Using Satellite Imagery and Machine Learning Algorithms (A Case Study of Lushan, Guilan Province, Iran) 611 630 107248 10.22059/ijswr.2026.410504.670095 FA Samira Hemmati . Department of Soil Science, Faculty of Agriculture, University of Zanjan, Zanjan, Iran 0009-0000-5452-0907 Kamran Moravej Department of Soil Science, Faculty of Agriculture, University of Zanjan, Zanjan, Iran 0000-0002-7929-742X Ahmad Golchin Department of Soil Science, Faculty of Agriculture, University of Zanjan, Zanjan, Iran 0000-0002-7193-6821 Mir Naser Navidi Soil and Water Research Institute, Agricultural Research Education and Extension Organization (AREEO), Karaj, Iran, 0000-0001-5087-9760 Mohammad Sadegh Askari Department of Soil Science, Faculty of Agriculture, University of Zanjan, Zanjan, Iran 0000-0002-2110-0217 Journal Article 2026 02 04 Soil quality is a fundamental indicator for assessing ecosystem sustainability and land productivity, and it is influenced by a combination of natural and anthropogenic factors. This study aimed to analyze the spatial variability of the soil quality index (SQI) at the regional scale using the random forest (RF) machine learning algorithm and multiple linear regression (MLR) based on environmental variables in the lands of the Loshan region in Guilan Province. For this purpose, 76 soil samples were collected from the 0–30 cm soil layer, and soil physical, chemical, and biological properties were measured using standard laboratory methods. In addition, remote sensing-based indices, including normalised difference vegetaion index (NDVI), normalised difference water index (NDWI), normalised difference moisture index (NDMI), normalized difference built-up index (NDBI), the bare soil index (BSI), and land surface temperature (LST), were derived as environmental auxiliary variables. The SQI was calculated using both the total data set (TDS) and minimum data set (MDS) approaches, combined with fuzzy scoring functions. The results showed that the random forest model predicted the spatial variability of SQI with higher accuracy (R² = 0.75) than multiple linear regression (R² = 0.53). Moreover, spectral indices particularly NDVI, NDWI and BSI played the most important roles in explaining the spatial variation of soil quality. This study demonstrates that the proposed digital soil quality mapping framework can serve as an effective tool for sustainable land management, soil conservation, and supporting decision-making in precision agriculture. Soil quality is a fundamental indicator for assessing ecosystem sustainability and land productivity, and it is influenced by a combination of natural and anthropogenic factors. This study aimed to analyze the spatial variability of the soil quality index (SQI) at the regional scale using the random forest (RF) machine learning algorithm and multiple linear regression (MLR) based on environmental variables in the lands of the Loshan region in Guilan Province. For this purpose, 76 soil samples were collected from the 0–30 cm soil layer, and soil physical, chemical, and biological properties were measured using standard laboratory methods. In addition, remote sensing-based indices, including normalised difference vegetaion index (NDVI), normalised difference water index (NDWI), normalised difference moisture index (NDMI), normalized difference built-up index (NDBI), the bare soil index (BSI), and land surface temperature (LST), were derived as environmental auxiliary variables. The SQI was calculated using both the total data set (TDS) and minimum data set (MDS) approaches, combined with fuzzy scoring functions. The results showed that the random forest model predicted the spatial variability of SQI with higher accuracy (R² = 0.75) than multiple linear regression (R² = 0.53). Moreover, spectral indices particularly NDVI, NDWI and BSI played the most important roles in explaining the spatial variation of soil quality. This study demonstrates that the proposed digital soil quality mapping framework can serve as an effective tool for sustainable land management, soil conservation, and supporting decision-making in precision agriculture. https://ijswr.ut.ac.ir/article_107248_cc9d11aeace9b0b3eb57de576ce3c92b.pdf