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

moderate effect of average annual temperature on population growth rate (Fig 8). Supporting these findings, the direction and degree of bias of LTRE approximations differed across popu- lations’ climate conditions (S4 Fig in S1 Appendix), depending on whether the relationship between climate and population growth rate was linear, concave, or convex (Fig 7). For exam- ple, in Fig 7B, the S and C populations’ response to annual precipitation is highly nonlinear; these populations’ LTRE approximation predicted higher future population growth rate than is likely to occur. By contrast, the E1 and E2 populations’ response is quite linear; these popula- tions’ LTRE approximation predicted lower future population growth rate than is likely to occur. Average annual temperature was the most important driver of changes in population growth rate, with weaker effects of other climate variables. Average annual temperature strongly increased population growth rate at all populations, though the effect size, Δλc, varied (Fig 8). In fact, increases in average annual temperature are predicted to increase population growth rate so substantially that they are sufficient to counter the negative effects of changes in all other climate variables at all populations besides the N population. Coldest month tempera- ture was the most important climate variable that reduced population growth rate, though again Δλc varied across populations (Fig 8). Among-population variation in climate responsiveness to average annual temperature is

predicted to drive differences in climate change effects across populations. The multiple regression indicated that for average annual temperatures, climate responsiveness, rather than change in climate conditions, will be the key determinant of change in population growth rate (Table 2). For example, at the N population, climate responsiveness to average annual temper- atures is low (Fig 7A). Low climate responsiveness leads to a small Δλc at the N population (Fig 8), resulting in weaker overall effects of climate change at this population (Δλ is low, Fig 5). Fig 6. Sensitivity to climate variables at current (2008–2022) climate conditions. Sensitivities were calculated using a perturbation approach using kernels from the current time period. We show the mean sensitivity across 500 bootstrap replicates; bars indicate the 95% confidence intervals on sensitivities. As in other figures, populations are arranged by increasing current average annual temperature. https://doi.org/10.1371/journal.pone.0247290.g006 PLOS ONE Nonlinear responses of population growth rate to climate change PLOS ONE | https://doi.org/10.1371/journal.pone.0247290 March 3, 2021 12 / 20