The determination of cultivated soil quality within a semi-humid climate using integrated quality index model and prediction with artificial neural network

The quality of the soil is greatly influenced by soil management practices that either raise or decrease the criteria of soil quality. Therefore, it is crucial to monitor our soil to determine whether agricultural practices have a major positive or negative impact on it. The Havza district in Samsun, Turkey, which is a semi-humid climate region, experiences intensive agricultural land use, diverse soil-forming conditions, and increasing pressure from management practices, making it a vulnerable agroecosystem necessitating monitoring. Thus, this research was carried out to assess and predict the soil quality of this region. 217 soil samples were collected from the study area, and 33 soil quality parameters were selected and analyzed. The data were subjected to the Integrated Quality Index (IQI) and the Artificial Neural Network (ANN). This is to support sustainable land management, productivity, and long-term soil conservation under increasing human and climatic pressure. In addition, the Total Dataset (TDS) of soil parameters was subjected to Principal Component Analysis (PCA), and 13 soil quality parameters were chosen for the creation of a Minimum Dataset (MDS), and spatial distribution maps of the study area were created. The result showed that the Soil Quality Index (SQI) determined by IQI was similar to that predicted by ANN, with R2 values of 0.999, 0.970, and 0.987 for training, validation, and testing, respectively. The distribution maps show sporadic low-quality areas within the interior, with the lowest quality in the northern middle part of the study area. The overall soil quality was classified as medium quality. Also, the distribution maps provide valuable information for land management, ecosystem management, and the sustainability of agricultural farmlands.