Global climate models (GCMs) are the best tools for simulating the Earth’s climate and forecasting climate change. However, in simulating climate across the entire globe, GCMs have less ability to accurately simulate the climate of very localised and specific regions of the world. This is because details like coastlines, topographic variation and land-use type all have an important influence on the regional climate, but the resolution of GCMs is too low to sufficiently represent these details and capture their influence.

It is important to simulate this more localised regional climate because this is the scale at which society needs to make decisions about adapting to climate change across different regions. This is why we use regional climate models (RCMs) to infer high-resolution climate information derived from low-resolution GCM climate information – to make the results ‘locally relevant’.

However, evidence about where and when RCMs provide new and more plausible information about both the current climate and future changes due to climate change projections has been lacking. This matters because RCMs are expensive in terms of time and computing resources, so it is important to better understand the ‘added value’ they provide compared to low-resolution GCM projections.

This research focused first on the seasons and locations in Australia where a suite of cutting-edge RCMs improved the simulation of the present-day climate compared to GCMs. The researchers found that, overall, RCMs simulated the Australian climate more accurately than GCMs, particularly for some regions like the heavily populated and economically important east coast.

The innovative part of this research is that it was the first to develop a method that shows where and when RCMs simultaneously add value to modelled representations of the present-day climate while at the same time making different projections about future climate change compared to GCMs. When RCMs show both of these attributes, it suggests that they confer plausible improvements in future climate projections, relative to GCMs. The researchers called this new quantity the ‘realised added value’ shown by RCMs.

For example, RCMs project a drier future for the Australian Alps in winter than the GCMs. These alpine regions were also an area where the RCMs produced better regional simulations of the present-day climate than the GCMs. Together, these two qualities suggest that RCMs provide strong “realised added value” in this region, which provides valuable information for activities such as water resource management. Overall, realised added value from RCMs averaged across climate variables and seasons was high across the majority of Australia.

This demonstrates that overall the RCMs consistently provide added value across Australia compared to GCMs. As a result, researchers and policymakers can obtain plausible improvements in future climate projections from the current generation of available RCMs.

Furthermore, this suite of RCMs are a key part of the coordinated regional downscaling experiment (CORDEX) program, which aims to improve the generation and evaluation of RCM simulations for Australasia and globally. Having a multi-model ensemble of RCMs for Australasia through CORDEX yields significant value compared to uncoordinated or single-model studies for different regions using different methods.