The paper describes a geostatistical approach for combining multi-source data with different support for field delineation into homogeneous soil zones. It takes into account change of support explicitly given the critical influence of spatial resolution on the statistical characteristics of estimates. Geophysical and hyperspectral data were used in combination with soil chemical properties measured in the laboratory on 50 samples collected in a field cropped with Tomato (Solanum lycopersicum, L. cv San Marzano). The approach consisted in performing Gaussian anamorphosis with support correction, multi-collocated block cokriging and factorial block cokriging to jointly analyse all data. Two regionalised factors at different spatial scales were retained to split the field into homogeneous zones for site-specific management. The results emphasize the impact of spatial scale on site-specific management.

本文介绍了一种地统计学方法，用于将多源数据与具有不同字段支持力的字段组合到均匀土壤区域中。考虑到空间分辨率对估计的统计特征的关键影响，它明确考虑了支持的变化。将地球物理和高光谱数据与在实验室中测量的在种植番茄（茄子）的田间收集的50个样品的土壤化学性质结合起来使用，L. cv San Marzano）。该方法包括通过支持校正，多并置块协同克里格和阶乘块协同克里格执行高斯变形，以共同分析所有数据。保留了两个在不同空间尺度上的区域化因素，以将田野划分为同质区域，以进行特定地点的管理。结果强调了空间规模对特定地点管理的影响。