LAUSR

Soil nutrients vary greatly and information regarding their spatial-temporal variability is critical for achieving site-specific nutrient management. However, the problem is that the high cost of soil sampling and laboratory analysis limits the number of samples current research budgets can afford. Besides, large sampling intervals sometimes fail to capture the variability of typical soil properties. Consequently, determining a reasonable yet feasible sampling interval to characterize the spatial variation of soil properties is greatly needed to delineate soil nutrient management zones (MZs). For this purpose, this study used four sampling intervals, 25 m × 25 m, 50 m × 50 m, 75 m × 75 m, and 100 m × 100 m by resampling the raw data acquired from seven adjacent fields located in Quzhou County to estimate the optimal design of soil sampling. Five soil properties (pH, electrical conductivity, extractable phosphorus, extractable potassium, and soil organic matter) were selected to delineate MZs using principal component analysis and fuzzy c-means clustering algorithm. The results indicated that indigenous P and K varied significantly, with the coefficients of variation of 34.33% and 24.01 %, respectively. The findings suggested that the optimal soil sampling interval was 50 m to accurately delineate MZs in medium- and small-scale farmlands. Besides, findings also demonstrated that site-specific P and K management could save approximately 21 kg P ha −1 and 30 kg K ha −1 as compared to conventional farming practices while ensuring target yield. The results can help make precision P and K management decisions in medium- and small-scale farming systems.