OOS 64-1 - The trailing edge is everywhere: the transient risk of extinction hidden inside climate envelope forecasts

Society urgently needs ecologists to improve forecasts of the future state of ecological systems and the benefits they provide to humanity. However, ecological systems are complex; e.g., the patterns we would like to use for prediction are scale-dependent. In contrast, the tools used for ecological forecasting often rely on the assumption of scale-invariance. For example, the most commonly used approach to quantify the threat of extinction posed by changing climate, species distribution modeling or climate envelope modeling, assumes climate responses are scale-invariant – that individual- and population-level climate responses match species-level climate responses inferred from occurrence data – and hence, that populations at the cool edge of a species’ distribution should benefit from warming (the “leading edge”), whereas populations at the warm edge should decline (the “trailing edge”). Alternatively, individual-scale climate responses might differ from species-scale climate responses; e.g., a plant that is soil moisture-limited throughout its distribution might respond negatively to warming everywhere. We investigated the scale-dependence of climate responses in an aridland pine (Pinus edulis), where scale refers to both species-level geographic distribution vs. individual performance and to space vs. time. We used presence-absence and tree-ring time series data from the US Forest Service’s Forest Inventory and Analysis plot network, including 1,558 growth time series collected in 977 plots. We compared P. edulis’ occurrence vs. growth with respect to spatially varying temperature and precipitation, and growth variation with spatial and temporal climate variation. We found that P. edulis grows less in warmer-than-average and drier-than-average years throughout its distribution. In contrast, its probability of occurrence is unimodal and symmetric with respect to spatial variation in temperature, and mostly a declining function of average precipitation. Its ubiquitous negative individual-level sensitivity to temperature variability indicates that instead of “leading edge” populations benefitting from warming, all populations will experience decreased performance – the entire distribution is effectively a “trailing edge”. This adds to a growing body of evidence for scale-dependent climate responses, in trees and a diversity of other organisms. We conclude climate envelope models underestimate the negative impact of climate change on biodiversity; specifically, they ignore the transient risk of extinction during the timeframe that range dynamics are governed by fast ecological processes (plasticity, demography) and instead predict equilibrium endpoints influenced by slow ecological processes (evolution, migration, community sorting). The lesson for ecological forecasting is clear: careful attention to the scale-dependence of organism-environment relationships is needed when making predictions in ecology.