Geoscience Reference
In-Depth Information
Modeling continues to be one of the most rapidly evolving fields
in climate science. As stated in the first edition of this textbook and
reinforced here, any attempt to define the “state of the art” is destined to
become becoming quickly dated. In recognition, we limit ourselves to an
overview of the major directions of modeling the Arctic climate system
and some of the key challenges. Basic features of model architecture are
outlined, but discussion is minimized to help maintain flow.
9.1
General Model Types
Single Column Models (SCMs)
: As the name suggests, these have been developed
to examine processes in a single column, such as that extending from the surface
through some depth in the atmosphere, or from the surface to some depth of the soil.
Examples are radiative transfer models and models of active layer thickness in per-
mafrost regions. An attraction of SCMs is that detailed physics can be incorporated
with relatively modest computer resources. Frequently, SCMs are applied to a two-
dimensional grid of horizontally distributed points or cells (with no communication
between grid cells). They are also commonly used to test parameterizations (semi-
empirical representations of complex processes such as convection) for subsequent
inclusion in global climate and numerical weather prediction models. SCMs can be
forced from observations, synthetic data (artificial data with statistics that mimic
reality) or output from another model.
Land Surface Models (LSMs)
: LSMs address interactions between the land
surface, atmosphere, and underlying surface. In a typical “stand alone” application
(i.e., in which the LSM is not directly coupled to another model), time series of
basic variables (generally downward shortwave and longwave radiation, precipi-
tation, near surface winds, humidity, and near surface air temperature) represent
model forcings. The model ingests these forcings and generates output state vari-
ables and fluxes, including soil moisture, soil temperature, snow water equivalent,
runoff, latent and sensible heat fluxes, and upward shortwave radiation. LSMs are
run variously for a single column or over a two-dimensional grid (as with SCMs).
An attraction of LSMs for Arctic studies is that the output variables, such as evapo-
transpiration and soil moisture, are often sparsely observed. LSMs are often cou-
pled with river routing schemes so as to provide simulated discharge. LSMs of vary-
ing complexity also represent a component of coupled global climate models and
numerical weather prediction models (and hence atmospheric reanalyses). LSMs
are run at varying resolution. For “stand-alone” applications to a two-dimensional
grid, a horizontal resolution of 50-100 km could be considered typical. LSM have
become increasingly complex; some have evolved to simulate biogeochemical
cycles (e.g., carbon and nitrogen cycles), land use, and vegetation dynamics.
Sea Ice and Coupled Ice/Ocean Models
: Sea ice models have been widely
used to examine the dynamic and thermodynamic interactions inherent to the sea
ice system. The earliest sea ice models were thermodynamic only (ice motion was
