Ensemble forecast of climate suitability for chaparral whitethorn (Ceanothus leucodermis) in the southern Sierra Nevada and Tehachapi Mountains (California, USA) based upon multiple (n=11) downscaled 2045-2065 A2 GCM projections

The Nature Conservancy (TNC) has derived climate suitability forecasts for most species of trees and shrubs considered to be ecological dominants of terrestrial Californian habitat types. Our plant projections are compiled as decision support tools to help Conservancy project staff, as well as our external partners, develop the necessary plans, priorities and strategies to successfully adapt to uncertain changes in future climate. In the recently completed Southern Sierra Partnership's 2010 Climate-Adapted Conservation Plan for the Southern Sierra Nevada and Tehachapi Mountains, species and habitat forecasts shown here informed the development of a regional conservation design that explicitly incorporates long-term conservation values based upon potential climate adaptation threats and opportunities.

Maps of our species forecasts rely on colors to distinguish those areas where climates appear to be suitable both today and in the future (i.e. blue = 'climate refugia'), in contrast to those areas where suitable climates are projected to be lost (i.e. red = 'climate stress'), or gained in the future (i.e. purple = 'expansion zones'). For ensemble forecasts, we treat all futures as equally likely, and use color saturation to denote relative levels of uncertainty between future projections, where dark colors imply > 80% model consensus (i.e. 'low uncertainty'), and light colors imply > 60% model consensus (i.e. 'moderate uncertainty'). For scenario forecasts (i.e. 'hot & wet', 'warm & dry, 'hot & dry'), projections are based upon individual climate models, and can be used collectively to bracket the potential extremes of future climate changes. We caution that species distributions are often limited by many factors other than climate, and ultimately, the climate vulnerability of each species will depend upon its exposure and sensitivity to climate impacts, as well as factors that influence adaptation potential, such as species attributes (e.g. dispersal ability, physiology, behavior, genetic diversity), landscape attributes (e.g. topography, hydrology) and management context (e.g. conservation status, habitat loss, fragmentation).

All species projections were derived in Maxent (http://www.cs.princeton.edu/~schapire/maxent/) using default parameters settings, plus removal of all duplicate records, and partitioning of observation data into training data (70%) and testing data (30%) for model validation. All species modeled had > 30 spatially unique observation records. Observation data came from statewide rapid and releve surveys, supplemented by specimens from the California Consortium of Herbaria (http://ucjeps.berkeley.edu/consortium/index.html). Modern climate data (1961-1990) used in forecasts represents PRISM 800m layers (http://www.prism.oregonstate.edu/). Future climate projections (2045-2065) are based upon downscaled model run averages from 11 general circulation models (GCMs) run under A2 emission scenarios by International Panel on Climate Change (IPCC) contributors (http://www.ipcc-data.org/). To compare projections across futures, and between species, continuous logistic model projections were converted into binary grids (0 = unsuitable, 1 = suitable) using species-specific thresholds (i.e. sensitivity = 0.95) derived from modern projections and then applied uniformly to all modern and future projections. Statewide projections were clipped to the spatial extent of the Southern Sierra Partnership's project boundary. All layers are presented in geographic (i.e. latitude, longitude) WGS 84.

The Nature Conservancy (TNC) uses the most current and complete data available. GIS data and product accuracy may vary. Using GIS products for purposes other than those for which they were intended may yield inaccurate or misleading results. TNC reserves the right to correct, update, modify, or replace GIS products without notification.