Development of optimal breeding zones for white spruce in Ontario under current, past, and anticipated future climate change / by Ashley Marie Thomson.

Abstract

Tree breeding zones are delimited areas where trees are selected and bred, and their progeny are planted. Within these zones, the level of adaptation between seedlings and planting sites is controlled to ensure adequate survival and growth. Delineation of adaptively-based zones requires detailed knowledge of patterns of genetic variation within species. Five-year height, root-collar diameter, current increment and survival were measured at each of six Ontario white spruce provenance trials containing 127 sources from across Ontario and western Quebec. Survival generally did not demonstrate significant variation amongst provenances, except at the Englehart test site, while the remaining three traits were significant at most sites. Intraclass correlation coefficients for significant traits ranged from 0.7% for elongation at the Englehart site to 19.2% for survival at the Englehart site. Growth data from 2007 were combined with 2002-2004 growth and phonological data. Principal components analysis was used to summarize the main components of variation, explaining a total of 53.8% of the total variation with the first three axes. The first principal component represented mainly growth potential, while the second and third represented field trial phenology and greenhouse timing, respectively. Multiple linear regressions revealed significant correlations to both temperature and precipitation variables, explaining 28%, 53%, and 27% o f the total variation in the first three principal components axes. Regression equations were used to map predicted performance of white spruce across the study area of Ontario. The focal point seed zone procedure (FPSZ) was used to determine areas of adaptive similarity for each of 618 gridpoints across Ontario based on current climate. These areas were used to represent candidate breeding zones constructed at levels of +/- 1.0 and +/- 0.5 least significance difference values. A maximal-covering model was then used to determine the optimal combination of a specified number of zones that could be used to maximize the area covered by breeding zones. Thirteen breeding zones were required to cover the entire range of white spruce in Ontario at the 1.0 LSD level of adaptive similarity, while more than 40 zones were required at the 0.5 LSD level. The northwest and northeast forest management regions could be covered with 6 breeding zones at the 1.0 LSD level and the southern region with just 5 zones. The analysis was repeated using simulated past climate data from the CGCM2, HADCM3, and CSIRO climate models as input to the procedure. The HADCM3 model produced optimal breeding zones that were most similar to those developed for the current climate and is probably most reliable for predicting climate in Ontario. Future climate data for the same 3 models based on the A1 and B2 emissions scenarios was used to predict the location of future breeding zones for each of the six future climate scenarios. The HADCM3 model predicted the location of future zones to be relatively similar to those based on current climate, while CSIRO predicted wide shifts in the location of future zone boundaries and CGCM2 predicted moderate shifts.