Delving into climate prediction, we unravel the intricate tapestry of atmospheric models. The foundation lies in discretisation, which chunks down Earth's multifaceted systems into manageable simulations. Central to this is the dynamical core—it's the heart that pumps life into the models by churning through equations that depict atmospheric movements. Physical parameterisation, on the other hand, hones in on the finer details of micro-scale processes. Where climate models shine is in their long game—employing broader strokes for global patterns over extended periods, unlike the finely tuned, short-term focus of Numerical Weather Prediction (NWP). This document shines a spotlight on these contrasting approaches and traces the evolution toward innovative hybrid models. These models are the alchemists, merging the vastness of global perspectives with the granularity of regional insights, all to sharpen the foresight of our weather forecasts. #### Model Types ![[Pasted image 20240107190309.png]] ![[Pasted image 20240107190639.png]] ![[Pasted image 20240317184811.png]] Global Earth System / Atmospheric Models are run using a process called discretisation. ![[Pasted image 20240107170932.png]] ### Anatomy of an Atmospheric Models ![[Pasted image 20240107171213.png]] - **Dynamical Core** - Responsible for solving the PDEs that govern the atmospheric flows (handles most transport and exchange between grid cells) - **Physical Parameterization** - Handles all the processes not captured by the equations of motion, only applied through individual columns i.e. grid cells that extend through the verticals ### Climate Models vs Numerical Weather Prediction Models - Climate models require coarse global grid spacing (100km) and a different set of physical parameterisations that were relevant on long time scales - NWP require maximal performance and high spatial resolution (10km +) so that forecasts can be computed quickly and accurately. Physical parameterizations for NWP were generally minimal. - The line has been blurred in the past few decades. ### Global vs Regional Models ![[Pasted image 20240107172305.png]] ### Variable Resolution Models ![[Pasted image 20240107172450.png]]