Climatic drivers of genotype–environment interactions in lodgepole pine based on multi-environment trial data and a factor analytic model of additive covariance

Abstract

The optimum deployment of select material from tree breeding programs is affected by the presence of genotype–environment interactions (G × E) and further complicated by future climate change. Here, we analyzed tree height data from 28 progeny test sites in a multi-environment trial (MET) dataset for two testing cycles of the lodgepole pine (Pinus contorta Douglas ex Loudon) breeding program in British Columbia to characterize one approach to investigating the climatic variables influencing G × E and the potential impacts of climate change. Linear mixed model analysis was conducted using an approximate reduced animal model with a factor analytic (FA) variance model to estimate the complex additive (co)variance structure. Test sites were grouped according to patterns of G × E, and climate modelling was employed to project historical and future deployment zones for each group. Based on these findings, it appears that breeding groups with historically wide deployment zones from northern environments will become less important as the climate warms, and therefore investment should be directed toward southern breeding groups, which will be useful across a very wide geographic range in the near to mid-term future.