Integrated Models in Hydrology and Meteorology - Transport in the Atmosphere-Vegetation-Soil Continuum - page 339

within such a land-use type (e.g., LAI, albedo) a single value is assigned that is sup-
posed to be representative for that land use type. Then for each of these land use types
('tiles' or 'clumps') the LSM produces luxes which are averaged to a grid-averaged
lux based on the relative abundance of the given land-use type (e.g., Bonan, 1995 ;
Viterbo and Beljaars, 1995 ).
In the so-called mosaic approach the grid cell is subdivided into regular smaller units.

•

Each of these units can have its own unique combination of parameters, as far as spa-
tial information with suficient detail is available, for example, from remote sensing
or soil maps. Each of the mosaic elements will produce its own lux, thus contributing
to the grid-averaged lux (e.g., Avissar and Pielke, 1989 ; Seth et al., 1994 ; Ament and
Simmer, 2008 ).

Here the tile method is discussed in some more detail. The range of tiles distin-
guished varies between models. As an example of a tiled LSM, the Tiled ECMWF
Surface Scheme for Exchange over Land (TESSEL), is discussed (Viterbo and Bel-
jaars, 1995 ; van den Hurk et al., 2000 ). The following are the eight tiles that are used,
including the characteristics that distinguish them:

1. Low vegetation (without snow): large aerodynamic resistance, small canopy resistance,
shallow rooting zone
2. High vegetation (without snow): small aerodynamic resistance, canopy resistance sensi-
tive to VPD, deep rooting zone
3. Liquid water in the interception reservoir: no stomatal control
4. Bare soil (without snow): extraction solely from upper soil layer, resistance depends on
soil moisture content
5. Snow on bare soil or low vegetation: change in albedo, phase change before evapora-
tion
6. High vegetation with snow beneath: higher roughness than snow on low vegetation,
albedo mostly determined by vegetation that protrudes through snow; phase change
before evaporation
7. Sea-ice: difference in albedo and roughness compared to open water;
8. Oceans and lakes: low albedo, conservative surface temperature.

The relative proportions of low vegetation, high vegetation, bare soil and open
water do not change in time. Those fractions are taken from a global database. But
the interception reservoir and snow cover are dynamic. If rain ills the interception
reservoir, the proportions of low and high vegetation are reduced in favour of the pro-
portion taken up by the interception reservoir. Furthermore, whether or not each of
the vegetated or bare soil tiles is covered with snow does change in time. This implies
that the model needs to keep track of the snow cover accurately.
The tiling affects only the surface. Above the surface, all tiles are coupled to the
same atmospheric values for wind, temperature and humidity, valid for that particu-
lar grid box. And below the surface all tiles share the same soil parameters and state
variables (soil moisture and temperature).