Climate change impacts on population growth across a species’ range differ due to nonlinear responses of populations to climate and variation in rates of climate change

few studies quantify how sensitivity varies across populations across a species’ range (e.g., ref. [41]), and fewer still quantify whether this variation arises from local adaptation or simply var- iation in driver values across a species’ range. Past work suggests that populations should be relatively insensitive to climate drivers that were historically highly variable [42]. We see some support for this pattern, with the N population (which has likely experienced the highest his- torical climate variability, due to its inland, high-latitude location) less responsive to some cli- mate drivers than other populations (the S and C populations, which are coastal, were likely historically buffered from climate variability due to proximity to the ocean). Similarly, in Chinook salmon, juvenile survival in different populations depends on completely different climate drivers, which results in divergent responses to climate changes [43]. Whether differ- ential responses of salmon juvenile survival to climate arise from local adaptation or not is unclear. Regardless, the salmon study shows that variation in climate change effects on popula- tion growth rate can arise from differences among populations in the slope of the vital rate function v. climate relationships, rather than only by variation in climate response caused by among-population variation in mean climate conditions, as occurs in D. alaskana. In addition to local adaptation, other among-site differences in non-climate drivers, such as soil or eleva-

tion, are confounded with climate differences in our study, because we only have one year of demographic data per site. Spatial variation in some climate variables’ impacts was due to variation in climate change magnitude across space (Table 2). Other studies have found that rates of warming vary across space, with high-latitude, inland, or wind-protected locations experiencing higher rates of warming [4]. Additionally, rates of population declines in birds and mammals are greater in locations where temperature has increased more [5]. Our work also suggests that differential rates of warming across seasons could have implications for climate effects on population growth rate. We found that increases in average annual temperature increase future popula- tion growth rate, but increases in coldest month temperature actually decrease population growth rate (Fig 8). Consistent with this finding, other studies have shown that seasonality of climate changes, relative to the life history of the organism, matters; for example, egg viability of two butterfly species decreased with extreme high temperature, but high mean temperature had positive effects on larval growth rates [44]. Similar effects have been observed for pheno- logical responses [45] and for crop yield [46]. Our work has important implications for predicting effects of climate change on popula- tions across a species’ range. Effects of climate change on D. alaskana are universally positive and nonlinear, mediated primarily by spatial variation in climate responsiveness. Variation in

the rate of climate change also contributes to differential impacts of climate change across pop- ulations. Thus, our work suggests that current responses to climate may not continue indefi- nitely under continued changes in climate, but that populations may differ dramatically in their responses to future climate, even if their current responses to climate are very similar to one another. This is important because often conservationists only have data on current responses to climate, rather than future response to climate. All of these factors are likely to generate idiosyncratic responses to climate across populations. Thus, our work may provide a framework for understanding why many species do not show predictable shifts in their geo- graphic range with climate change [2, 47, 48]. Supporting information S1 Appendix. (DOCX) PLOS ONE Nonlinear responses of population growth rate to climate change PLOS ONE | https://doi.org/10.1371/journal.pone.0247290 March 3, 2021 17 / 20