Climate change at the landscape scale: predicting fine-grained spatial heterogeneity in warming and potential refugia for vegetation
Contents
Abstract
Current predictions of how species will respond to climate change are based on coarse-grained climate surfaces or idealized scenarios of uniform warming. These predictions may erroneously estimate the risk of extinction because they neglect to consider spatially heterogenous warming at the landscape scale or identify refugia where species can persist despite unfavourable regional climate. To address this issue, we investigated the heterogeneity in warming that has occurred in a 10 km × 10 km area from 1972 to 2007. We developed estimates by combining long-term daily observations from a limited number of weather stations with a more spatially comprehensive dataset (40 sites) obtained during 2005-2006. We found that the spatial distribution of warming was greater inland, at lower elevations, away from streams, and at sites exposed to the northwest (NW). These differences corresponded with changes in weather patterns, such as an increasing frequency of hot, dry NW winds. As plant species were biased in the topographic and geographic locations they occupied, these differences meant that some species experienced more warming than others, and are at greater risk from climate change. This species bias could not be detected at coarser scales. The uneven seasonal nature of warming (e.g. more warming in winter, minimums increased more than maximums) means that climate change predictions will vary according to which predictors are selected in species distribution models. Models based on a limited set of predictors will produce erroneous predictions when the correct limiting factor is not selected, and this is difficult to avoid when temperature predictors are correlated because they are produced using elevation-sensitive interpolations. The results reinforce the importance of downscaling coarse-grained (∼50 km) temperature surfaces, and suggest that the accuracy of this process could be improved by considering regional weather patterns (wind speed, direction, humidity) and topographic exposure to key wind directions.
