Yield performance and adaptation of some Australian-grown rice varieties through multivariate analysis

Abstract

Rice breeders are involved in developing new varieties for a diverse range of production environments to increase average yields. Interpretation of performance of several varieties evaluated in a broad range of environments is usually affected by variety × environment interactions. Application of multivariate analyses, especially multi-criteria decision making (MCDM) methods in variety × environment interaction studies can benefit interpretation of yield performance and adaptation of varieties in response to different environments. The case study discussed in this paper highlights the significant advantages of using MCDM methods to overcome constraints imposed by having to investigate a large number of variables inherent in such analysis. The grain yield of 13 rice varieties with varying characteristics, grown in 4 different regions in southern New South Wales (NSW), Australia, during the 2000–06 growing seasons was analysed using classification techniques and preference ranking organization method for enrichment evaluation (PROMETHEE) analysis. Significant variety × environment interaction was found to influence the yield performance of individual varieties across a range of environmental conditions. Classification of environments and geometrical analysis for interactive aid (GAIA) plot of PROMETHEE analysis identified the Murrumbidgee Irrigation Area (MIA) and Coleambally Irrigation Area (CIA) as higher yielding regions compared with the Eastern Murray Valley (EMV) and Western Murray Valley (WMV). Turbid water and cold weather conditions are suggested to explain the lower yield obtained in the WMV environment. In terms of varieties, Amaroo and Opus were identified as widely adaptable to most of the environments, while Jarrah was the least adaptable. Illabong can be considered as best adapted to the EMV or WMV regions, while Paragon and Reiziq can be regarded as best adapted to the MIA and CIA regions. Partial and complete ranking showed the interrelationships between the varieties for their yield performance and adaptation across all environments. Amaroo, Illabong, and Opus were ranked as the most preferred varieties, while Koshihikari, Kyeema, and Jarrah were ranked as the least preferred ones. Partial pre-order with 13 classes from PROMETHEE I analysis identified varieties with similar characteristics and aided the selection of suitable alternative varieties. The outcomes from the analyses reported here allow rice varieties or genotypes to be rated for yield stability for a specific or a range of different environments.