Fuzzy-Clustering Driven Soil Nutrients Management Zone Delineation in Mandarin (Citrus reticulata Blanco) Growing Areas of Malwa Plateau Region of India
摘要
Purpose. Mandarin (Citrus reticulata Blanco), contributes 26% to global citrus fruits production; however, in India, yield remain suboptimal and spatially inconsistent owing to inefficient nutrient management. Therefore, understanding spatial variability of soil nutrients and delineating soil management zones (MZs) is essential for adopting site-specific nutrient management strategies (SSNMS) in order to enhance crop productivity and maintain long-term soil health. Methods. We investigated the spatial variability of soil nutrients and related soil attributes and delineated MZs in mandarin growing areas of Rajgarh district of Madhya Pradesh, located in the Malwa Plateau region of central India. A total of 104 geo-referenced surface (0 to 20 cm depth) soil samples were collected from mandarin orchards of the study area and were analyzed for soil pH, electrical conductivity (EC), soil organic carbon (SOC), available nitrogen (AN), available phosphorus (AP), available potassium (AK), exchangeable calcium (Ex. Ca), exchangeable magnesium (Ex. Mg), available sulphur (AS), available zinc (AZn), available iron (AFe), available copper(ACu), available manganese (AMn) and available boron (AB). Results. The values of soil parameters varied widely with low (4.92%) to moderate (20.4 to 54.6%) coefficient of variation. The mean values were 7.80 for soil pH, 0.17 dS m-1, 4.70 g kg-1 for SOC, 187 kg ha-1 for AN, 28.5 kg ha-1 for AP, 468 kg ha-1 for AK, 2864 mg kg-1 for Ex. Ca, 983 mg kg-1 for Ex. Mg, 15.6 mg kg-1 for AS, 1.07 mg kg-1 for AZn, 0.90 mg kg-1 for ACu, 10.9 mg kg-1 for AMn, 16.1 mg kg-1 for AFe and 0.75 mg kg-1 for AB. The semivariogram analysis revealed exponential best fitted model with moderate to strong spatial dependence for soil attributes. The soil attributes displayed varied spatial distribution patterns. Principal component analysis and fuzzy c-means clustering were used to delineate MZs. Six principal components (with eigen values > 1) explained 75.14% of the total variance and were the input for clustering. Based on fuzzy performance index and normalized classification entropy, 03 distinct MZs (MZ 3 (50.7% area) > MZ 2 (46.3% area) > MZ 1 (0.3% area)) were identified in the study area. The MZs differed significantly in soil attributes except SOC, AN, and AB. Conclusions. The results highlighted the potential of geostatistics and multivariate techniques for delineation of MZs in mandarin growing areas for adoption of site-specific nutrient management strategies (SSNMS).