玉米品种多环境测试站点的空间布局方法研究

Multi-environment trial (MET) is a key point for crop varieties to test their productivity and adaptability. A suitable test station distribution is a prerequisite to obtain reliable and accurate result. There are three challenges for the distribution of MET test station. First, the sampling data available for test station distribution is in low density. Second, there are significant effects of genotype by environment interaction (GEI) during crop growth. Third, little test stations with high stress and high representation were included in the MET. In this study, a series of methods for spatial data processing, such as spatial analyzing, sampling, and modeling, will be utilized to solve these problems. Based on high density meteorology, soil, and geographic data, and the mechanism of agronomic and stress characters, spatial distribution and high density grid data of agronomic and stress characters for Maize will be simulated. And then, the typical effect and spatial distribution of GEI of Maize will be revealed. Finally, the model suitable for test stations sampling, selection, and spatial distribution of MET will be built, and a MET station distributing program for Huang-Huai-Hai Plain will be proposed to improve the model through actual multi-environment tests. These studies will bring new methods supplement for spatial sampling and distribution and the MET technology, as well as provide supporting methods and data for the test station distribution of field crops. In addition, these studies will contribute to increase the test efficiency of MET and lower the planting risk of new varieties.

多环境测试(MET)是获知作物品种丰产性和适应性的关键环节，科学合理的测试站点布局是确保测试结果准确可靠的基础。针对MET站点布局研究缺少高密度农艺或胁迫采样数据，测试结果受基因-环境互作效应影响显著，高代表性测试环境和逆境偏少等问题。利用空间分析、抽样与建模等空间数据处理分析技术，研究建立玉米关键农艺或胁迫指标的机理模型，基于高密度气象、土壤等基础环境数据和机理模型拟合构建高密度农艺和胁迫数据集，分析玉米关键农艺和胁迫指标的时空分布规律；研究并揭示玉米典型基因-环境互作效应及其空间分布规律，基于上述结果研究MET站点数量计算方法、测试环境选取方法及站点空间布局优化方法，以黄淮海夏播玉米区为例进行MET站点布局优化方案研究和验证。有助于完善空间抽样与布局优化方法和MET技术，可为大田作物MET站点布局优化提供方法和数据支持，提高测试效率、降低品种推广风险。

作物品种多环境测试通过多年多点大田测试来检测新品种的丰产性和适应性等表现；科学合理的测试站点布局是确保测试结果准确可靠的基础。本项目针对当前玉米品种多环境测试站点的布局依靠主观经验，缺少量化分析方法，导致品种试验结果的代表性和可靠性较差、效率低等问题。完成了玉米品种测试站点布局研究所需的基础种植环境与品种测试数据整理、环境模型构建与生育期内环境指标值计算；探究了典型玉米品种的基因-环境互作效应空间分布规律。建立了基于10公里的种植环境基础数据分析单元。.提出了玉米种植环境空间型精细区划与测试站点布局方法,以10公里地理网格为分析单元，以其环境指标的多年平均值作为聚类指标，将东北三省春玉米区分为8个生态亚区，使用属性聚类方法对生态亚区进一步细分，结合实际玉米种植面积，确定82个测试点；根据环境代表度对测试站点的布点优先级进行排序。提出了玉米种植环境时空型区划与测试点布局方法。以10公里地理网格为分析单元，以其每年的环境指标值作为聚类指标，利用属性聚类方法，将东北三省春玉米区分为7类生态亚区，分析了7类生态亚区从1993-2013年间的空间分布；将出现次数超过50%的生态亚区类别作为该分析单元的典型生态环境类型，其余分析单元归为非典型生态环境。利用空间分层抽样方法，计算各类生态亚区所需的最少测试次数及测试站点数，得出东北三省共需要布设112个测试点；典型和非典型生态环境分别需要60个和52个测试点。以环境代表度、测试充分度计算了每个分析单元的概率栅格；再利用空间分层抽样方法确定测试站点位置布局。 .基于空间型区划的试点布局代表了种植环境的空间分布，基于时空型区划的试点代表了种植环境的空间分布和年际间的变化趋势。为科研合理布设品种试验站点提供了方法和工具,能够提高布设品种测试点的代表性，能够量化评估不同试点布局下试验结果的可靠程度，并开始用于管理部门、科研单位和种子企业的品种试验站点布局优化工作。