Genetic Variation and Climatic Impacts on Survival and Growth of White Spruce in Alberta, Canada

Abstract

Because climate has the greatest effect in determining the genetic structure of forest tree species, climatic variables with large effects on growth and survival need to be identified. This would enable proper matching of tree populations to planting sites in the present and future climates. We analysed 24-year survival (S24), height (H24) and diameter (D24) from a series of white spruce provenance trials with 46 populations and 8 test sites in Alberta, Canada. We determined: (1) the amount and pattern of genetic variation, (2) the response of populations to climatic transfer and (3) the potential effects of climate change (2030-2039) on H24 and S24 of the species in Alberta. We found that: (1) using the intraclass correlation, the between-population genetic variance was 10.6% (H24) and 6.6% (D24) of the betweenpopulation phenotypic variance across sites, (2) three climatic white spruce regions exist in Alberta within which variation in growth potential is strongly clinal, (3) the annual moisture index (AMI) expressed as a ratio of degree days above 5°C (GDD) and mean annual precipitation (MAP) was the major determinant of survival and growth at the test sites, (4) we found that at the level of AMI predicted for the 2030-2039 period, survival and growth would decline substantially in the continental part (northern and central) of Alberta where drought already exists. However, during the same period, survival and growth would increase substantially in the foothills and Rocky Mountains region where growth is currently limited by low GDD due to a short growing season.

Population Differentiation and Climatic Adaptation for Growth Potential of White Spruce (Picea glauca) in Alberta, Canada

“…The data were measurements of height and diameter of 19 populations planted at 8 locations in Alberta, Canada (Table 1; Figure 1), which are a subset of the 46 populations described by RWEYONGEZA et al (2007). Because some of the analyses in this study used data derived from population response functions, the 27 populations planted on fewer sites were omitted because their response functions could not be reliably developed.…”

“…Climatic variables were selected by fitting population response functions of individual variables to identify variables that fit the data well and also based on knowledge from published work (e.g., REHFELDT et al, 1999, 2002, RWEYONGEZA et al, 2007. To simplify the interpretation of PCA axes, we avoided using the ratio variables AMI and SMI in conjunction with GDD, MAP and MSP from which they were derived.…”

“…A quadratic model (Eq. 2) was selected based on preliminary analyses and our previous study on these populations (RWEYONGEZA et al, 2007). Regression coefficients were then used to obtain predicted optimum climate (elsewhere simply called "optimum climate"), which is a P I1 or P I2 score corresponding to the maximum predicted height or diameter (P O = -β 1 / 2β 2 , where P O = optimum climate) on the response function.…”

“…Thus, characterization of populations based on their growth response to test site climate was done only for P I1 . As previously discussed (RWEYONGEZA et al, 2007), site C is located on an abandoned sawmill with significant soil compaction. Consequently, trees on this site were much smaller than trees on sites with comparable climate and soil fertility.…”

“…Many studies (e.g., REHFELDT et al, 1999;RWEYONGEZA et al, 2007;THOMSON et al, 2009) have used individual climatic variables to model the response of height or diameter to the climate of the test sites. However, climate is a complex variable of moisture and precipitation and their seasonal variation.…”