Integrated Models in Hydrology and Meteorology - Transport in the Atmosphere-Vegetation-Soil Continuum - page 328

q a T a

r r

r r

E soil H soil

q sat ( T S ) T S

θ g

Figure 9.14 LSM without vegetation: the Manabe ( 1969 ) model. T s is the tempera-
ture of the soil layer and θ g is the volumetric moisture content of the soil (which has
maximum moisture content θ max. ). The aerodynamic resistance for water vapour is
modiied based on soil moisture availability, expressed as ηθθ

. Note that q sat ( T s ) = g /

ax

is not an independent model variable as it directly depends on T s .

This partitioning is a strong function of the canopy resistance. Because water availability
in the root zone is an important determining factor in the canopy resistance, not only the
energy balance of the surface needs to be tracked, but also the water balance.

LSMs are both used in weather forecast models and in climate models. But given
the longer integration times of a climate model, in those models more emphasis is
placed on the correct long-term behaviour (e.g., the soil should not dry out too far;
Hagemann et al., 2004 ).

9.2.3 Modelling of Vegetation
Roughly four levels of complexity can be distinguished in the treatment of vegetation
in land-surface models (see also Sellers et al., 1997 and Pitman, 2003 ).

LSMs Without Vegetation (First Generation)
The model of Manabe ( 1969 ), the irst LSM to be implemented in a climate model,
is the archetype of this group (see Figure 9.14 ). In this model the luxes of both heat
and water vapour solely originate from the soil surface. The evaporation is regulated
by modifying the aerodynamic resistance with a factor that depends on the relative
saturation of the (only) soil layer. If the soil dries out, the resistance for water vapour
transport increases, thus reducing evaporation. In the original Manabe ( 1969 ) model
the soil heat lux is ignored, as it did not simulate the diurnal cycle. In later applica-
tions this simpliication has been replaced by including a force-restore method for the
soil temperature or with a multilayer soil model for temperature (as in, e.g., the MM5/
WRF model; Chen and Dudhia, 2001 ). Although on time scales of years to months
this simple scheme provides appropriate luxes, on diurnal time scales larger discrep-
ancies will occur (Desborough, 1999 ; Trier et al., 2006).