Genotype × environment interaction (GE) is a perennial problem in plant breeding and must be dealt with. Dealing with GE starts with differentiation of repeatable GE from unrepeatable GE in a target region. Repeatable GE can be used by dividing the target region into mega‐environments (MEs) and breeding ME‐specific cultivars, and unrepeatable GE must be accommodated by testing adequately within a ME. This study applied LG (location‐grouping) biplot analysis to several datasets from multiyear oat (Avena sativa L.) variety trials conducted at locations across Canada. Analysis showed that the oat growing regions in Canada can be divided into three MEs: the crown rust (Puccinia coronata Corda f. sp. avenae Eriks.) prone regions in southern and eastern Ontario (ME1), other regions in eastern Canada (ME2), and the Canadian Prairies (ME3). In addition, two sub‐MEs existed within ME2. Latitude was shown to be the main factor for the ME differentiation. The results suggest that oat variety trials should be conducted and cultivar recommendation be made according to MEs, as opposed to by administrative regions that are currently in place.

中文翻译：

---

加拿大的燕麦超级环境

基因型×环境相互作用（GE）是植物育种中的常年问题，必须加以解决。与GE的交易始于在目标区域中将可重复的GE与不可重复的GE进行区分。可重复使用的GE可以通过将目标区域划分为巨型环境（ME）并繁殖ME特定的品种来使用，不可重复的GE必须通过在ME中进行充分测试来适应。这项研究将LG（位置分组）双标图分析应用于在加拿大各地进行的多年燕麦（Avena sativa L.）品种试验的几个数据集。分析表明，加拿大的燕麦种植区可分为三个ME：冠锈病（Puccinia coronata Corda f。sp。avenae）安大略省南部和东部（ME1），加拿大东部的其他区域（ME2）和加拿大大草原（ME3）的易发地区。此外，ME2中还存在两个子ME。纬度被证明是ME分化的主要因素。结果表明，应根据MEs进行燕麦品种试验并推荐品种，而不是与当前的行政区域相对。