Physical Climatology
Physical climatology explains climate through the mechanisms that produce it: radiative transfer, circulation, moisture transport, and the exchanges coupling atmosphere, ocean, land, and ice. It stands against a purely descriptive climatology that catalogues what the statistics are without asking what generates them.
The distinction has a practical consequence rather than a philosophical one. A descriptive account can tell you the climate of a place has changed. Only a process account can tell you whether the change will continue, and that is the question almost everyone actually has.
Why process understanding is the load-bearing part
Extrapolating a statistical relationship outside the range it was fitted on requires believing the relationship is structural rather than incidental, and only process understanding can supply that belief. This is exactly the weakness of statistical downscaling under a changing climate — the transfer function is fitted on the observed record and assumed to hold in a climate that record does not contain.
It is also what makes a model result interpretable. A model reproducing observed rainfall is weak evidence on its own, because a model can produce the right total through compensating errors — too much convective rain and too little stratiform, or the right amount at the wrong time of day. Checking that the mechanism is right is a much stronger test, and it is why I find diagnostics like the CAPE time series or the water balance closure more informative than a correlation against observed precipitation.
The Galapagos as a worked case
The archipelago is a good demonstration of why the process account is not optional. It sits on the equator and is arid at sea level, which is not what equatorial position alone would predict. The explanation requires several mechanisms acting together: the Humboldt Current and equatorial upwelling delivering cold water, that cold SST suppressing deep convection while supporting a stratocumulus deck, and a strong inversion capping the boundary layer.
The consequence is a climate where the highlands receive substantial moisture from fog interception during the garúa season while the lowlands stay dry — and where ENSO can invert the whole arrangement by warming the surface ocean. No amount of statistical description of the rainfall record recovers that structure, and without it the ENSO response looks like an anomaly rather than the same mechanism running the other way.
See also: climate for the concept, applied climatology for the decision-facing counterpart, hydroclimate, and regional climate modeling for the tooling.