MOC Regional and Urban Climate Modeling
Global models run at resolutions too coarse for almost any decision. This map covers the two families of methods that bridge the gap, and the applications I have used them for.
The split is between dynamical downscaling — running a physical model at higher resolution inside global boundary conditions — and statistical downscaling, which fits a transfer function from the observed record. The trade is clear and worth being explicit about: the statistical route is cheap and assumes its fitted relationship survives into a climate the fitting record does not contain, while the dynamical route makes no such assumption and cannot correct errors it inherits at its boundaries. Neither escapes the fact that a regional model is only as good as what drives it, usually ERA5.
At fine enough resolution the parameterisation question changes character rather than disappearing — convection-permitting modeling removes the convection scheme and puts the weight on microphysics and turbulence instead. At the urban end the same logic runs down to micro-scale resolution with large-eddy simulation.
Core workflow notes
- Regional climate modeling
- Dynamical downscaling
- Statistical downscaling
- Convection-permitting modeling
- ERA5 reanalysis
- WRF
- ARW core
- tropical WRF setup
- CER v2
Galapagos and DARWIN branch
- DARWIN project
- Galapagos refined analysis
- DARWIN WRF experiment suite
- Heavy rainfall in the Galapagos
Urban climate branch
- urban climate
- urban meteorology
- urban climatology
- urban atmosphere
- urban surface
- Urban heat island
- Anthropogenic heat
- microscale process
- PALM-4U
- large-eddy simulation models