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

larger snowpack could lead to a shorter growing season, resulting in negative effects of precipi- tation. In D. alaskana, we also see a strong lagged positive effect of coldest month precipitation on survival, which likely arises because precipitation in the coldest month reduces the proba- bility of fruiting, which then increases survival the subsequent year. Consistent with our find- ings, other work also shows a mix of positive and negative effects of precipitation on different vital rates [30]. Surprisingly, we see no evidence for precipitation x temperature interactions; other studies have suggested that such interactions might be critical in alpine species. For example, increasing temperature could accelerate melt-out and negate any negative impact of precipitation [31]. In spite of the mix of positive and negative effects of temperature and precipitation, popula- tion growth rate across the species’ range is projected to be higher under future climate condi- tions (higher temperature and precipitation). These positive effects appear largely mediated by increases in average annual temperature (Fig 8), likely resulting from the positive effect of average annual temperature on mean growth rate (Table 1). Consistent with our findings, a recent review showed that many high- and mid-latitude populations across many taxa have responded positively to increased temperature under current climate conditions. In particular, plants responded more positively to temperature increases than any other taxa studied (inver-

tebrates, fishes, amphibians, birds, or mammals; [32]). Our results further suggest that future climate change will reduce D. alaskana’s extinction risk, an unexpected result considering that uncommon species such as D. alaskana are more likely to go extinct than common species [33]. However, both current and future population growth rates were below unity for most populations (other than future population growth rate at the S and C populations, S2 Fig in S1 Appendix), suggesting that even with increases in population growth rate with climate change, most D. alaskana populations, besides those in the most favorable conditions, will decline to extinction. While current population growth rates less than one are somewhat counterintui- tive, it is possible that suboptimal historical conditions (climate or otherwise) have resulted in current low population growth rates. This prediction is consistent with the likely extirpation of at least one population from previously occupied habitat (A. Louthan, pers. obs.). Alterna- tively, lower-than usual growth rates during the years of our study, seed bank dynamics, or unaccounted for immigration could result in low estimates of current population growth rates. Table 2. Role of climate responsiveness v. rate of climate change. Climate variable Climate responsiveness Change in climate conditions Change in climate conditions x climate responsiveness Residual Average annual temp. (˚C) 0.84933 0.08188 0.06631 0.00248 Annual precip. (mm) 0.20552 0.39518 0.0799 0.3194 Coldest month temp. (˚C) 0.60387 0.03245 0.28828 0.0754

Coldest month precip. (mm) 0.12438 0.50025 0.22179 0.15358 Proportion of variance in the contribution of climate variables to change in population growth rate (Δλc) attributable to climate responsiveness v. the change in climate conditions (Δc, mean future, 2086–2100, minus mean current, 2008–2022, conditions), and the interaction between the two. Climate responsiveness is obtained by regressing median annual growth rates (median is across 500 bootstrapped sets of parameter estimates) on annual climate conditions using a GAM (Fig 7), and dividing Δλ by Δc, as in Fig 1. For all multiple regressions, the response variable, Δλc (contributions of climate variables to Δλ) is the median across 500 bootstrapped sets of parameter estimates. Bold indicates which effect has the highest proportion of variance for each climate variable. https://doi.org/10.1371/journal.pone.0247290.t002 PLOS ONE Nonlinear responses of population growth rate to climate change PLOS ONE | https://doi.org/10.1371/journal.pone.0247290 March 3, 2021 15 / 20