Geospatial Mapping of Soil Properties of Forest Types Using the k-Means Fuzzy Clustering Approach to Delineate Site-Specific Nutrient Management Zones in Goa, India

Numerous human activities have impacted the soils of Goa, a region inside the ecologically sensitive Western Ghats. The erosion of topsoil and diminishing soil fertility have jeopardized ecological security. The objective of the present study was to delineate management zones to safeguard the integrity of the forest soil ecosystem and to spatially differentiate the soil qualities within forested regions. The research was conducted in the Indian state of Goa. Fourteen soil parameters-pH, EC, SOC, accessible nitrogen, phosphorus, potassium, exchangeable calcium, exchangeable magnesium, sulfur, and DTPA-extractable micronutrients-zinc, copper, iron, manganese, and boron-were assessed in thirty-six georeferenced soil samples. The soil, with a pH of 6.31 +/- 0.45, was classified as acidic. The soil's electrical conductivity was 0.34 +/- 0.16 dSm(-1), indicating neutrality. The mean values for available nitrogen (AN), available phosphorus (AP), and available potassium (AK) were 299.25 +/- 76.9 kg ha(-1), 37.62 +/- 16.20 kg ha(-1), and 264.51 kg ha(-1), respectively, with ranges of 200.1 to 458.6 kg ha(-1), 14.3 to 68.1 kg ha(-1), and 107.1 to 481.6 kg ha(-1). The concentrations of accessible boron (AB) and available sulfur (AS) varied from 0.15 to 1.19 kg ha(-1) and 1.6 to 79.2 kg ha(-1), respectively, whereas the average availability of cationic DTPA-extractable AZn, ACu, AFe, and AMn was documented as 1.8 +/- 1.3, 1.2 +/- 1.1, 13.5 +/- 8.5, and 8.3 +/- 4.9 mg kg(-1). Soil pH had a substantial positive correlation with organic carbon, available phosphorus, available potassium, available zinc, available iron, and available boron. The EC maintained a direct association with OC, AMn, AFe, and AB, while exhibiting a negative relationship with AN, AP, AK, ACu, AZn, and AS. A positive link exists between SOC and AN, AZn, AMn, AFe, AB, and AS. Available nitrogen had a negative association with AK, ACu, AMn, AFe, and AB, while demonstrating a positive correlation with AP, AZn, and AS. The AP had a negative association with AFe, AB, and AS, while demonstrating a positive correlation with AK, ACu, AZn, and AMn. Geostatistical analysis was used to generate best fitted semivariogram models and spatial distribution maps for soil properties. In the present study, the Hole effect model for AN, ACu, Ex.Ca, and AFe and the J-Bessel model for EC, OC, AP, and Zn were developed. A Stable semivariogram model suitable for pH, AMn, and AB. For AS and AK, the Gaussian and K-Bessel models were generated, respectively. The rational quadratic model was best for ex.Mg. The majority of soil parameters exhibited substantial spatial dependency (ACu and ex.Mg.), moderate spatial dependence (pH, AN, AK, ex.Ca, AZn, and Amn.), and weak spatial dependence (OC and AP). Six principal components were selected based on their eigenvalues (> 1) using principal component analysis. About 74.46% of the variability in the dataset was accounted for by the six principal components. Kriged spatial distribution maps of soil variables, including pH, EC, SOC, AN, AP, AK, Ex. Ca, Ex. Mg, AZn, ACu, AFe, AMn, AS, and AB, were generated to illustrate the current status of these parameters in the forests of Goa. The FPI and NCE variables were utilized to delineate three MZs employing the k-means clustering methodology. The results indicates that soil fertility of Goa's forests is good, based on the average values of the soil parameters of developed MZs. The study's findings can be effectively utilized by forest managers and policymakers to improve forest soil fertility, refine precision silviculture management and advance afforestation initiatives.