COS 8-6 - Can genomic vulnerability predict climate change-induced population decline?

Abstract: Climate change is threatening global biodiversity. As extreme climate events become more common, populations risk decline due to increasing maladaptation. Evolution may be able to rescue populations undergoing demographic decline; however, this requires sufficient adaptive genetic variation. Establishing which populations have insufficient adaptive genetic variation is of value for conservation. Genetic offset, a metric of genomic vulnerability to climate change, has been used to predict which populations may be maladapted to future climate. The promise of this algorithm is considerable, but its accuracy at predicting population decline both due to extreme climatic events and predicted future climate change averages has not been well established. Genetic offset requires the additional expense of sequencing, which may prove to be a barrier in a conservation context were resources are limited. Thus, it is important to establish if genetic offset models predict demographic decline better than approaches that only use climatic information.



Here we use extreme drought across the range of Mimulus cardinalis to assess the ability for genetic offset to predict population decline. We conducted whole-genome sequencing on 55 populations sampled prior to the record-setting 2012-2015 drought in California and Southern Oregon. This yielded 2,156,697 SNPs across 347 genomes. We associated these SNPs with temperature, precipitation and moisture deficit to identify 294 SNPs that were most highly associated with climate. We then calculated genetic offset to both the 2012-2015 drought and to future average 2041-2070 climate projections. Genetic offset was more severe when using 2041-2070 projections, but parts of the range of M. cardinalis already showed conditions similar to the future projections. We found that central populations are particularly vulnerable to extreme multi-year drought.



To assess the predictive power of genetic offset, we leveraged seven years of demographic surveys from across the range of M. cardinalis. We estimated population growth rate from 2012 to 2016 across 12 sites and use genetic offset to predict demography. Populations with greater genomic vulnerability also had greater population declines, implying that these populations are likely at greater risk of extinction because of a deficit of adaptive genetic variation. Genomic data was critical to this inference, since using only climate data to predict population decline did not yield a significant association. Overall, this work demonstrates that genetic offset is a powerful tool to predict populations decline that may be worth the additional investment in a conservation context.