Time to get moving: assisted gene flow of forest trees

Abstract

Geographic variation in trees has been investigated since the mid-18th century. Similar patterns of clinal variation have been observed along latitudinal and elevational gradients in common garden experiments for many temperate and boreal species. These studies convinced forest managers that a 'local is best' seed source policy was usually safest for reforestation. In recent decades, experimental design, phenotyping methods, climatic data and statistical analyses have improved greatly and refined but not radically changed knowledge of clines. The maintenance of local adaptation despite high gene flow suggests selection for local adaptation to climate is strong. Concerns over maladaptation resulting from climate change have motivated many new genecological and population genomics studies; however, few jurisdictions have implemented assisted gene flow (AGF), the translocation of pre-adapted individuals to facilitate adaptation of planted forests to climate change. Here, we provide evidence that temperate tree species show clines along climatic gradients sufficiently similar for average patterns or climate models to guide AGF in the absence of species-specific knowledge. Composite provenancing of multiple seed sources can be used to increase diversity and buffer against future climate uncertainty. New knowledge will continue to refine and improve AGF as climates warm further.

Keywords: assisted migration; climate change; ecological genetics; forest policy; genetic clines; local adaptation; provenance; temperate species.

Figure 1

Schematic diagram of management options for reforestation and restoration in a changing climate. While this illustrates the northward movement of individuals, assisted gene flow may also occur along elevational or longitudinal climatic gradients.

Figure 2

(A) A schematic diagram of the annual developmental cycle of temperate and boreal trees for 1 year illustrating the timing of growth, dormancy and bud phenology, modelled after the degree growth stage model of Fuchigami et al. (1982). (B) The timing of physiological changes, environmental cues, and biotic and abiotic stresses. Orange arrows illustrate the expected environmental effects of climate change (timing, magnitude and direction); blue arrows indicate the expected genetic changes in phenotypic traits and vulnerability resulting from assisted gene flow from populations matching projected new conditions.

Figure 3

Map of Picea sitchensis showing the proportion of 17 putatively adaptive SNPs found to be associated with fall cold hardiness by Holliday et al. (2010a) that were polymorphic based on genotyping of 122 to 164 mature trees in each population, from Lobo (2011).

Figure 4

Adaptive clines in height growth potential along a temperature gradient are similar among many temperate tree species from western North America. (A) Regression slopes of standardized height growth potential versus provenance mean annual temperature (MAT), for populations grown in a common test environment. Each line represents an independent provenance trial, and line type corresponds to the species and reference indicated in the legend. Only statistically significant regression slopes are shown (P < 0.05). (B) The climatic scope of provenance trial data sets compared to the species realized niche. For each species, the bottom line represents the core climatic niche (solid black line) and full climatic niche (dashed extension) in MAT (see text; Thompson et al. 2000). The upper lines represent the range of provenance MAT values among populations within each provenance trial for that species. Open circles represent the MAT at the test site(s) for each provenance trial. Provenance trials missing an open circle were grown in a controlled environment. Note that the y‐axis scale in (A) and x‐axis scale in (B) do not cover the full range of data for some species; this scaling was applied to increase visibility for the majority of the data sets.

Figure 5

Provenance variation in 8‐year‐old Picea sitchensis from across the species range grown in a common garden in Vancouver, BC, Canada. For illustrative purposes, trees were selected from each population that were closest to the provenance mean. The state or province of origin and the provenance mean annual temperature are indicated. This species shows the steepest cline of all species for height at age 2 in Figure 4(A). The population from Kodiak Island, AK (Fig. 3) has poorer growth than expected for an MAT of 5°C, likely due to inbreeding (Mimura and Aitken 2007b).