Cotton Megaenvironment Investigation And Test Environment Evaluation Based On Gge Model

The formation and development of cotton lint yield and fiber quality traits is strongly affected by multiple ecological environment factors, thus leading to the universal existence of significant genotype by environment interaction. Insight exploration to genotype by environment interaction effect is the basis of megaenvironment investigation and sufficient test environment evaluation and selection for representative test environment as a breeding environment within the target planting region, and thus to greatly improve the cotton genetic improvement and new cultivar selection efficiency. To select and recommend particular varieties to its special adaptive subregions will give rise to full play of the positive interactive effects of ecological resources and varieties’production potential. In this study, the heritability adjusted GGE biplot was adopted to intensively explore the genotype by environment interaction patterns in the datasets of27independent cotton multi-environment cultivar trials in the Yangtze River Valley during2000-2010. The possible existence of megaenvironment in the Yangtze River Valley was investigated based on the selection for single breeding traits (lint yield, fiber length, fiber strength, micronaire), and for multi-traits simultaneous selection index of the fiber quality selection index (FSI) and lint yield plus fiber quality integrated selection index (ISI). In the same time the desirability of test environments was comprehensively evaluated in terms of the discriminating ability, representativeness of the target region and also the desirability index for ideal test environment selection and recommendation. On the basis of the construction of information ratio adjusted GGE model and the responding goodness-of-fit test method, and the establishment of the mathematical formula for ideal distance calculation based on GGE principal component scores, the goodness of fit of GGE model was tested and the test environment evaluation result was revised, and hence after the desirability index was adjusted with the Euclidean distance to the ideal environment to enhance the accuracy, reliability and efficiency of the megaenvironment investigation and test environment evaluation. The conclusion of the ideal test environment selection and megaenvironment investigation scheme in the Yangtze River Valley was proposed out based on single trait or multiple yield and fiber quality trait integrated selection index, in order to provide the theoretical implications and decision-making support for broad adaptive and specific adaptive cultivar selection and application targeting at the whole cotton planting region in the Yangtze River Valley and the megaenvironment within the region as well. The main findings were listed as follows.1. The readjustment and application effect of the HA-GGE modelIn accordance with the effective information ratio judgment criteria of principal component analysis, the principal component scores with IR≥1were entered the GGE model simulation procedure, and thus revised the GGE model simulation goodness of fit. Based on the selection of cotton lint yield, fiber length, fiber strength, micronaire, fiber combined selection index, lint yield plus fiber quality integrated selection index, the goodness of fit of IR-GGE model was improved8.2%,3.1%,6%,6.7%,5.4%and5.4%respectively, in this way the megaenvironment investigation and text environment evaluation efficiency were effectively enhanced.2. Cotton megaenvironment identification and investigation based on HA-GGE modelBased on GGE model and its information ratio revised version, the possible existence of megaenvironments in the Yangtze River Valley was intensively investigated for the selection of cotton lint yield, fiber length, fiber strength, micronaire, fiber combined selection index(FSI) and lint yield plus fiber quality integrated selection index(ISI).(1) Based on the fiber quality combined selection, the cotton planting region in the Yangtze River Valley as target region were divided into three distinct megaenvironments: A major complex megaenvironment covering the majority test environments including Anqing, Wuhan, Xiangfan, Yueyang, Jiujiang, Nanyang, Huanggang, Changde, Jinzhou and Nanjing, and two minority megaenvironments covering the test environment Shehong, Yancheng, Nantong Cixi and Jianyang representing for the cotton planting region in Sichuan basin and the coastal region in Jiangsu and Zhejiang province.(2) Based on the cotton lint yield and fiber quality integrated selection index, the whole cotton growing region in the Yangtze River Valley were divided into two minor megaenvironments, one is located at the inland Nan-Xiang basin bordering with the Yellow River Valley in the north including two locations, the other is at the mountainous basin in west area including two locations too, and a major complex megaenvironment covering the majority eleven locations in the target region, where are of the most traditional climatotype and agrotype around the western Pacific temperate monsoon climate regions.3. Test environment comprehensive evaluation based on GGE modelBased on GGE biplot and the selection of cotton lint yield, fiber length, fiber strength, micronaire, fiber combined selection index(FSI) and lint yield plus fiber quality integrated selection index(ISI), the test environment was elaborately evaluated as far as its discriminating ability, representativeness, desirability index and the distance to the ideal environment were concerned.(1) Globally speaking, Huanggang, Jingzhou, Nanjing and Changde were ideal test environments and the most effective breeding environments for cultivar selection targeting at the whole cotton planting region in the Yangtze River Valley, while Xiangfan and Nanyang in Nan-Xiang basin were ineffective test environment in terms of cotton lint yield and lint yield plus fiber quality integrated selection, Nantong, Yancheng and Cixi in Jiangsu and Zhejiang provincial coastal cotton region were less effective for fiber quality selection, Shehong and Jiangyan in Sichuan basin were poor test environments for cotton cultivar selection in neither of lint yield nor fiber quality traits.(2) Based on the selection of yield plus fiber quality integrated selection index (ISI), the ranking by desirability index of the test environments was relevant to the cultivar comprehensive selection efficiency implementing in the locations, which might be linked to the special geographical position and the climate characteristics. Shehong and Jianyang were located at the mountainous basin in the west area which might implicated its lower efficiency in cultivar selection, while Nanyang and Xiangfan test environment representing for the Nan-Xiang basin region were located in the northern border of the Yangtze River Valley, where frost descents earlier and temperature declines fast in the later autumn might be the reason explaining its poor representativeness to the whole region and less efficiency of cotton lint yield selection.