This project takes future climate projections and uses them to produce estimates of future species distributions. This process involves several steps.
Generating species distributions
Species occurrences were vetted for accuracy and plausibility. These occurrences were compared statistically to current climate data at 0.01° geographic resolution (often referred to as “1 km” resolution) to produce a statistical model that predicts species survival given specified climate values. This model is known as a Species Distribution Model or SDM and one was created for each species.
Using the SDM and current climate conditions, a climate suitability map was created for each species. These suitability maps were then clipped by experts to remove areas with suitable climate that did not currently host the species. This produced an estimate of the current species distribution.
Projecting future climate
Future climate projections must account for a broad range of factors, such as population growth, land use, energy demand, and new technology. The IPCC has defined a set of scenarios to represent potential futures.
This project uses two of the four IPCC scenarios. RCP8.5, a scenario with continuing high emission of greenhouse gases, was selected to represent a "business as usual" future. RCP4.5, a scenario in which greenhouse gas emission stabilises before 2100, was selected to represent a lower emission future.
18 different global climate models or GCMs were used to generate climate projections for each selected RCP and each decadal time step from 2015 to 2085.
Projecting future species distributions
The SDM for each species was then applied to future climate projections to create climate suitability maps. These suitability maps were then clipped based on a realistic dispersal rate to remove areas that had newly suitable climate but could not be reached by the species. A diagram demonstrating this process is shown below.
Dispersal and suitability for freshwater fish and crayfish, and selected stream-dependent species of frogs and turtles, was calculated using stream flow data in addition to climate factors.
The dispersal-clipped projections of different GCMs were summarised using the median or 50th percentile as a "best estimate", and the 10th and 90th percentiles to give an indication of variability between models. This summarising across models produced the projected species distribution maps displayed in CliMAS' map interface.
Generating biodiversity values
Maps for the species in each given taxa were reduced to a binary value of present or not-present. Calculating the number of different species present at a given point (with 10th, 50th and 90th percentile summaries across GCMs) produced the taxa biodiversity maps available in the map interface.
Projected species distributions were overlaid with region borders for each of the reportable regions. This allowed the species potentially present in the region at a given projected future point to be listed.
Similarly, region borders were used to clip temperature and rainfall values for current and future projections. This provided numerical climate summary data with 10th 50th and 90th percentiles, and this information is available along with species presence summaries from CliMAS' reports generation interface.