Separating sensitivity from exposure in assessing extinction risk from climate change

Abstract

Predictive frameworks of climate change extinction risk generally focus on the magnitude of climate change a species is expected to experience and the potential for that species to track suitable climate. A species' risk of extinction from climate change will depend, in part, on the magnitude of climate change the species experiences, its exposure. However, exposure is only one component of risk. A species' risk of extinction will also depend on its intrinsic ability to tolerate changing climate, its sensitivity. We examine exposure and sensitivity individually for two example taxa, terrestrial amphibians and mammals. We examine how these factors are related among species and across regions and how explicit consideration of each component of risk may affect predictions of climate change impacts. We find that species' sensitivities to climate change are not congruent with their exposures. Many highly sensitive species face low exposure to climate change and many highly exposed species are relatively insensitive. Separating sensitivity from exposure reveals patterns in the causes and drivers of species' extinction risk that may not be evident solely from predictions of climate change. Our findings emphasise the importance of explicitly including sensitivity and exposure to climate change in assessments of species' extinction risk.

Figure 1. Schematic illustration of climate breadth and mean climate change.

Two example climate variables are shown from six used. Climate variables within grid cells showing current values (circles) and predicted values after climate change (triangles). The cells occupied by two species are shown (blue, red points) for both current and predicted values. (a) The sensitivity of a species is defined by the volume of climate space corresponding to the species geographic range, which is derived from the width of the values experienced along each environmental axis (edge lengths of rectangles for each species' points). (b) Covariation may inflate the environmental width of those axes and so a rotation is used to produce orthogonal environmental axes for calculating sensitivity. (c) The exposure of a species is measured as the arithmetic mean of the Euclidean distances (blue, red lines) between the current and predicted environmental values of occupied cells. In this example, one species (blue) is sensitive and exposed, with a narrow environmental width and a high mean displacement across the cells in which it resides. In comparison, the second species (red) is neither sensitive nor exposed, having broad environmental widths and short environmental displacements.

Figure 2. Identifying sensitive, exposed and at risk species.

Mean climate breadth and exposure values across GCMs for both amphibians and mammals under RCP 4.5 Species (grey points) with high exposure may not be sensitive to climate change (blue region) and hence are better able to withstand climate change across their ranges. Species with narrow climate breadth may be sensitive to climate change but not exposed (red region) and their risk would need to be reassessed as estimates of climate change across their ranges are updated. We identify species as vulnerable when they are projected to experience a high magnitude of climate change which they are expected to be relatively unable to withstand (varying purple regions). Dashed horizontal and vertical lines define thresholds used to place species into broad categories of sensitivity and correspond to the most extreme 10, 20 and 50% of climate breadth (smallest values) and exposure values (highest values). The contour lines in black show the proportion of species that are both exposed and sensitive (falling in the region above and to the left) for different sensitivity and exposure thresholds. As examples, approximately 1% of species fall under the combination of the 10% exposure and sensitivity thresholds, approximately 3% of species fall under the combination of the 20% exposure and sensitivity thresholds and approximately 20% fall under the combination of 50% exposure and sensitivity thresholds.

Figure 3. Global patterns of climate breadth and mean climate change.

Relative magnitudes of average climate breadth and exposure values within cells for amphibians (a) and mammals (b). As exposures were not qualitatively different across RCPs, only the results for RCP 4.5 are shown. Species in yellow areas are predominantly exposed (E) to climate change and species in magenta areas are predominantly sensitive (S) to climate change, having narrow climate breadth. Red areas are characterised by communities in which both (B) high exposure and high sensitivity may be found. Geographical patterns in sensitivity and exposure are not congruent. Extinction risk may be low in regions where exposure is high where species in those regions are not sensitive to climate change (yellow areas). Similarly, extinction risk may be low in regions that are highly sensitive to climate change where species in those regions are not exposed (magenta areas). Extinction pressure is likely to be highest in those areas where high values of sensitivity and exposure occur together (B - red areas). The saturation of the colours indicates the overall severity of conditions in each cell. Mean cell values were calculated and maps were generated in R.