WHEN LOCAL MAY NOT BE BEST
The risks associated with maintaining the default local provenancing strategy may be greater from an eco-evolutionary perspective than using alternative strategies that include both local and non-local provenances (summarised in Table 1). While this is not an exhaustive list, we present four case studies, with the aim of highlighting the transition from a state of known risks for non-local provenances to a state of known risks for both non-local and local provenances.
Climate change shifting local advantage
As climates change, there is a risk of local adaptations becoming decoupled from environmental selective pressures, resulting in the maladaptation of local genotypes. Evidence of such decoupling was demonstrated in Eucalyptus gunnii , a subalpine foundation tree species endemic to Tasmania, Australia. Consistent with local adaptation, local provenances initially showed higher fitness in reciprocal common garden trials (Potts 1985). However, following two-decades of declining precipitation and increasing temperatures (Sanger et al. 2011), local provenance fitness declined (<30% survival), with the lower elevation, non-local provenance showing higher fitness than the local provenance (>70% survival; Prober et al. 2016). Similar impacts of changing environments have been observed in other forest species (Jump et al. 2009) and natural systems more broadly (Parmesan 2006; Scheffers et al. 2016).
While future change may decouple local adaptations from home environments, climate change to date may already be outpacing adaptation – an outcome known as adaptation lag. Evidence for adaptation lag was shown in Quercus lobata (valley oak), where two, early-age provenance trials showed greater relative growth of provenances when planted into cooler climates than their current homesite climate (~10% greater growth; Browne et al. 2019). Using a genomics-informed selection model, Browne et al. (2019) predicted gains of up to 25% in relative growth of individuals pre-adapted to current environments of the test site (Browne et al. 2019). Optimal growth associated with historic rather than current climate is consistent with adaptation lag. While growth is a plastic trait and species may persist in suboptimal environments, similar evidence of adaptation lag has been shown in several tree species (Carter 1996; Aitken et al. 2008; Fréjaville et al. 2020).
Demographic and evolutionary history
Demographic history plays an important role in structuring the genetic and adaptive variation within and among populations (Hewitt 2000). The historical expansion from refugia or genetic bottlenecks can, for example, often leave evolutionary fingerprints that have fitness effects on modern-day populations. Such demographic processes have been argued to explain the differential survival and fecundity of two natural populations of Arabidopsis thaliana (Thale cress) from opposing latitudinal ends of its native European range (Sweden and Italy). While local provenances showed higher fitness on average in reciprocal transplant trials (Ågren & Schemske 2012), inter-provenance crosses between the two populations tended to have higher fecundity than the local provenance when planted in Sweden (Ågren et al. 2013). Furthermore, several Italian alleles provided a fitness advantage over the Swedish alleles when plants were grown in Sweden (Ågren et al. 2013). Demographic history, in particular postglacial expansion in northern Europe reducing genetic variation most likely led to increased fixation of these maladaptive alleles.
Novel environments from land use change
In some situation, environmental change, such as mining, agriculture, and deforestation, has resulted in novel environments. This presents a risk that the local provenances may no longer be fit for these novel, local conditions. Evidence of this is shown in two metallophyte legumes,Mimosa acutistipula var. ferrea and Dioclea apurensis , which are promising candidates for the restoration of degraded areas of mineral rich soils in the Amazon, Brazil. Adaptive genomic variation showed that local provenances would perform best in moderately disturbed sites (e.g., sites altered by fire) (Carvalhoet al. 2020). However, no provenance in the sampled area matched predictions for highly disturbed sites with substantially altered conditions (i.e., ex-mining sites; (Carvalho et al. 2020). With no suitable local provenance, regional mixing across multiple adaptive units was recommended to capture genetic variation and enhance adaptability, thereby de-risking revegetation plantings. Together with previous findings of rapid, microgeographic adaptation post-mining (e.g. Antonovics and Bradshaw 1970; Antonovics 2006), these results highlight that provenance choice may vary dependant on site history, with environmental changes having the potential to significantly influence provenance fitness in novel environments.
Together, these examples highlight that current and future environmental changes may have reduced local provenance fitness, and thus increased the risk associated with using local provenances in particular revegetation contexts. Further, they highlight the need to consider current and future environmental change, climate and anthropogenic, as well as effects of evolutionary history (e.g. demographic history) when selecting provenances, to ensure both the short- and long-term fitness and productivity of revegetation plantings. One approach could be mixing local seed with seed from non-local provenances already occupying current or future climates, which may de-risk the long-term resilience of revegetation plantings.