Geoscience Reference
In-Depth Information
Table 3 continued
Name
Country
Stations
Website
USCRN
USA
114
http://www.ncdc.noaa.gov/crn/
USDA-ARS
USA
4
VAS
Spain
3
http://nimbus.uv.es/
Table adapted from http://www.ipf.tuwien.ac.at/insitu/index.php/insitu-networks.html
in situ data for validation, it is also important to select appropriate validation measures
(Entekhabi et al. 2010b ).
The assimilation of satellite data for land surface applications has only gained signifi-
cance in the last decade; it started later than atmospheric and oceanographic data assim-
ilation (see various chapters in Lahoz et al. 2010a ). This can be attributed to: (1) a lack of
dedicated land surface state (water and energy) remote sensing instruments; (2) inadequate
retrieval algorithms for deriving global land surface information from remote sensing
observations; and (3) a lack of mature techniques to objectively improve and constrain land
surface model predictions using remote sensing data.
3 Models of the Hydrological Cycle
As discussed above, observational information has gaps in space and time. It is desirable to
fill in these observational gaps using a model. Such models can range from simple linear
interpolation to full land surface models (LSMs). Land surface processes are part of the
global processes controlling the Earth, which are typically represented in global general
circulation models (GCMs). The land component in these models is represented in (largely
physically based) LSMs, which simulate the water and energy balance over land using
simple algebraic equations or more complex systems of partial differential equations. The
main state variables of these models include the water content and temperature of soil
moisture, snow and vegetation. These variables are referred to as prognostic state vari-
ables. Changes in these state variables account for fluxes, for example, evapotranspiration
and run-off, which are referred to as diagnostic state variables.
Most land surface models used in GCMs view the soil column as the fundamental hydro-
logical unit, ignoring the role of, for example, topography on spatially variable processes
(Stieglitz et al. 1997 ) to limit the complexity and computations for these coupled models.
During the last decades, LSMs have become increasingly complex to accommodate for better
understood processes, like snow and vegetation. Along with a more complex structure often
comes a more complex parametrization, and several authors (Beven 1989 ;Duanetal. 1992 )
have stated that LSMs are over-parametrized given the data typically available for calibration.
At larger scales, these models often rely on satellite-observed parameters, such as greenness and
LAI (leaf area index). For field-scale studies, the LSMs are usually calibrated to specific
circumstances to limit systematic prediction errors. Model calibration or parameter estimation
relies on observed data and can be defined as a specific type of data assimilation (Nichols 2010 ).
Many LSMs have been developed and enhanced since the mid-1990s, with varying
features, such as sub-grid variability, community-wide input, advanced physical repre-
sentations and compatibility with atmospheric models (Houser et al. 2010 ). Some examples
of widely used LSMs are the NCAR Community Land Model (CLM) (Oleson et al. 2010 );
the Variable Infiltration Capacity (VIC) Model (Liang et al. 1994 ); the Noah Model (Ek
Search WWH ::




Custom Search