Influence geometric anisotropy in management zones delineation

ABSTRACT The proper handling soil allows the reduction of contaminants, maximize agricultural productivity, and is directly related the knowledge spatial variability of soil attributes. This spatial variability can express isotropic and anisotropic form. The latter being neglected in research related to management zones delineation. In this context, the present study aimed to evaluate the effect of the geometric anisotropy correction on the management zone delineation. The methodology was applied under database of soybean productivity and apparent electrical conductivity (CEa) of a rural property in Ponta Porã - MS. By means of this georeferenced database, maps was interpolated with ordinary kriging. For each combination, attribute (productivity and CEa) and number of classes, were produced two maps management zones, one without and one with anisotropy correction, the same were compared through the kappa index, with significance tested by the Z-test. The management zones number was also evaluated by Fuzziness Performance Index (FPI) and the Modified Partition Entropy (MPE). The area subdivision in two management zones, without and with anisotropy correction, presented higher Kappa index, with values of 0.89 and 0.91 respectively, but not presented significant differences with each other.

摘要 合理的土壤管控可降低污染物含量、最大化农业生产力，且与土壤属性空间变异的认知直接相关。此类空间变异可表现为各向同性（isotropic）与各向异性（anisotropic）两种形式，其中各向异性在管理区划分（management zones delineation）相关研究中常被忽视。在此背景下，本研究旨在评估几何各向异性校正（geometric anisotropy correction）对管理区划分的影响。
本研究以巴西马托格罗索州蓬塔波朗（Ponta Porã - MS）某农场的大豆产量与表观电导率（apparent electrical conductivity, CEa）地理参考数据库为依托开展方法应用。基于该数据集，采用普通克里金（ordinary kriging）法对空间信息进行插值并绘制专题图。针对「属性（产量与CEa）+分级数」的每一种组合，分别生成未校正各向异性与经各向异性校正的两类管理区图，并通过Kappa指数（Kappa index）对二者进行对比，同时采用Z检验（Z-test）验证差异显著性。本研究同时借助模糊性能指数（Fuzziness Performance Index, FPI）与修正划分熵（Modified Partition Entropy, MPE）对管理区的最优分级数量进行评估。当将研究区划分为2个管理区时，未校正与经各向异性校正的方案对应的Kappa指数分别为0.89与0.91，整体表现最优，但二者之间未呈现显著差异。