Riparian restoration and genetic diversity of a foundation tree along the principal river of the Southwest

Background/Question/Methods

Cottonwood trees of the southwestern USA occupy a vital niche in riparian ecosystems, and their genetic variation has been shown to be important for a myriad of dependent community members and ecosystem processes. Despite widespread historical extirpation and habitat destruction in southwestern riparian systems, cottonwood trees (Populus fremontii) are currently a target organism for restoration efforts throughout the Lower Colorado River – in a region famous for environmental and hydrologic change.  It is currently unknown how restoration will depend on genetic variation and diversity in cottonwood trees. Because the original genetic stock for many of the cottonwood forests which once dominated the lower Colorado River has been lost, it is unknown how well restoration efforts might reconstruct genetic diversity in restored forests in the region. In this study we use a common garden planted to explicitly test the impact of genetic diversity on restoration success. Trees were planted in 2007 along the lower Colorado River (Cibola National Wildlife Refuge, Arizona) in 16 tree blocks which varied in composition from one to 16 genotypes. The sixteen tree genotypes were collected from across the entire natural range of P. fremontii in the Southwest in order to maximize tree phenotypic differences.

Results/Conclusions

After five years of growth, we find that tree productivity is very sensitive to tree genotype, but less sensitive to stand genotype diversity. Similarly, tree form and phenology are highly predictable by tree genotype. We find nearly identical patterns in restoration success by genotype when evaluating individual tree growth, biomass production at the stand level, or percent canopy cover measured from areal imagery. Interestingly, the most successful genotypes were not those genotypes collected from native trees at the planting location. The most successful genotypes were instead trees collected from locations at higher (but not the highest) elevations. Abiotic variables related to home-site temperature precipitation were also predictive of tree growth. These results may reflect widely-understood patterns in wind-pollinated species where genetic variation can be higher within local populations than among populations. Nevertheless, because genetic diversity of riparian trees has been so dramatically reduced in many arid regions, genetically diverse populations from more distant locations may perform much better than remaining small populations proximal to sites targeted for restoration.