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.

土壤养分差异很大，有关其时空变异性的信息对于实现特定地点的养分管理至关重要。然而，问题在于土壤采样和实验室分析的高成本限制了当前研究预算能够承受的样本数量。此外，较大的采样间隔有时无法捕捉典型土壤特性的变异性。因此，非常需要确定一个合理但可行的采样间隔来表征土壤特性的空间变化，以划定土壤养分管理区 (MZ)。为此，本研究使用 25 m × 25 m、50 m × 50 m、75 m × 75 m 和 100 m × 100 m 四个采样间隔，通过对从位于衢州县的 7 个相邻油田获得的原始数据进行重新采样，以获得估计土壤取样的最佳设计。使用主成分分析和模糊 c 均值聚类算法选择了五种土壤特性（pH、电导率、可提取磷、可提取钾和土壤有机质）来描绘 MZ。结果表明，土生磷、钾变异显着，变异系数分别为34.33%和24.01%。研究结果表明，最佳土壤采样间隔为 50 m，以准确划定中小型农田的 MZ。此外，研究结果还表明，特定地点的磷钾管理可以节省大约 21 公斤磷公顷 结果表明，土生磷、钾变异显着，变异系数分别为34.33%和24.01%。研究结果表明，最佳土壤采样间隔为 50 m，以准确划定中小型农田的 MZ。此外，研究结果还表明，特定地点的磷钾管理可以节省大约 21 公斤磷公顷 结果表明，土生磷、钾变异显着，变异系数分别为34.33%和24.01%。研究结果表明，最佳土壤采样间隔为 50 m，以准确划定中小型农田的 MZ。此外，研究结果还表明，特定地点的磷钾管理可以节省大约 21 公斤磷公顷^-1和 30 kg K ha ^-1与传统耕作方法相比，同时确保目标产量。结果有助于在中小型农业系统中做出精确的磷和钾管理决策。