Climate modelling is one of the key research methods in climate and climate change research. In addition to use in research on the climate system as such, climate modelling facilitates research on attribution and provides means for climate change projections, thus supporting the pursuant of climate change mitigation and adaptation.
Inherently, climate models need to encompass the overall global system as the atmosphere, ocean, sea ice and land surface system are all interconnected via a multitude of processes and across a wide range of temporal and spatial scales. As climate modelling studies typically look into how the climate system behaves and develops over time – from a few decades to a century or longer – the need of computing resources is often quite considerable. In practice, there are constraints on how spatially explicit global models can be made, as the amount of computations needed increases in proportion to the third power of model resolution increase. For example, doubling model resolution gives an 8-fold increase in the amount of computations needed for a simulation. Consequently, global climate modelling effectively targets relatively large scale aspects of the climate system, instead of being especially suited for addressing research questions related to extreme events, regional and local climate detail etc.
Regional climate modelling (a.k.a. dynamical downscaling) is one of the methods to complement global climate modelling in exploring such aspects. It involves running climate models over some specific (geographical) region, using some global climate model simulation as “boundary conditions”. A limited model domain allows an increase in model resolution and thus more detailed (regional) representation of the climate system and its processes, still keeping the computational cost down.
As global climate models, regional climate models are based on the fundamental understanding and physical laws (fluid mechanics, thermodynamics and radiation) that the real atmosphere-ocean-sea ice-land surface system abides by, and thus share the same challenges related to the representation of climate processes and their numerical formulation. However, as regional climate models need a set of boundary conditions, they are less “free-running” than global climate models. Thus, there are also differences in the “mathematics” of regional climate modelling, compared to global modelling. This has consequences for model development, simulation analyses and interpretation of results.
This presentation is about regional climate models as a technique, outlining their research potential fundamental limitations and applications. Having a background on these aspects is useful when pursuing qualitative and quantitative analysis of regional climate model simulation results.