| Home > Research > Ifsdocs > PHYSICS > |
Chapter 7. Land surface parametrization
IFS documentation Front Page
Chapter 1. Overview
Chapter 2. Radiation
Chapter 3. Turbulent diffusion and interactions with the surface
Chapter 4. Subgrid-scale orographic drag
Chapter 5. Convection
Chapter 6. Clouds and large-scale precipitation
Chapter 7. Land suface parametrization
Chapter 8. Methane oxidation
Chapter 9. Climatological data
REFERENCES
Previous Section
7.2 Tiles and surface fluxes
7.2.1 Tile and vegetation characteristics
Grid-box surface fluxes are calculated separately for the different subgrid surface fractions (or "tiles"), leading to a separate solution of the surface energy balance equation and skin temperature for each of these tiles. This is an analogue of the "mosaic" approach of Koster and Suarez (1992). Note that the tiles at the interface soil-atmosphere are in energy and hydrological contact with one single atmospheric profile above and one single soil profile below. Each grid box is divided into 8 fractions: two vegetated fractions (high and low vegetation without snow), one bare soil fraction, three snow/ice fractions (snow on bare ground/low vegetation, high vegetation with snow beneath, and sea-ice, respectively), and two water fractions (interception reservoir, ocean/lakes). The tile for "high vegetation with snow beneath" is a combined tile with a separate energy balance and evaporation model for the high vegetaton and the underlying snow. A mixture of land and water (ocean/inland water) tiles is not allowed, i.e. a grid box is either 100% land or 100% sea.
In each grid box two vegetation types are present: a high and a low vegetation type. An external climate database, based on the Global Land Cover Characteristics (GLCC) data that has been derived using one year of Advanced Very High Resolution Radiometer (AVHRR) data and ancillary information (Loveland et al. 2000; http://edcdaac.usgs.gov/glcc/glcc.html; see also Chapter 9). The nominal resolution is 1 km. The data used provides for each pixel a biome classification based on the Biosphere-Atmosphere Transfer Scheme (BATS) model (Dickinson et al. 1993), and four parameters have been derived for each grid box: dominant vegetation type, TH and TL, and the area fraction, AH and AL, for each of the high- and low-vegetation components, respectively.
The coverage Ci for the tile i depends on the type and relative area of low and high vegetation, and the presence of snow and intercepted water. In the absence of snow and interception, the vegetation coverage of high (cH) and low (cL) vegetation are calculated as AHcveg(TH) and ALcveg(TL), respectively, with cveg a vegetation type dependent coverage (see Table 7.1). The bare ground fraction cB is the residual.
|
(7.1) |
Each vegetation type is characterized by a series of (fixed) parameters as detailed in Table 7.1:
| • A minimum canopy resistance, rs,min; |
| • A leaf area index, LAI; |
| • A vegetation coverage, cveg; |
| • A coefficient, gD, for the dependence of the canopy resistance, rc, on water vapour pressure deficit; |
| • The root distribution over the soil layers, specified by an exponential profile involving attenuation coefficients, ar,and br; |
The numerical values for the parameters of Table 1 are based both on experiments conducted as described in van den Hurk et al. (2000) and on literature review, in particular Mahfouf et al. (1995), Manzi and Planton (1994), Giard and Bazile (1999), Dorman and Sellers (1989), Bonan (1994), Pitman et al. (1991), and Zeng et al. (1998).
The presence of snow and intercepted water dynamically modifies the coverage fractions. The coverage of snow, csn, is linearly related to the snow mass per unit area (abreviated to snow mass in the following), S (units
or m). The interception reservoir fraction, cl,
is given by Wl/Wlm , with Wlm,
the maximum value for the intercepted water in the grid box, defined from
the leaf area index contributions from the high and low vegetation tiles.
The water contents of the interception reservoir, Wl (units
m), and S are prognostic quantities in the model. Snow cover
is assumed to be overlying vegetation and bare ground with the same fraction.
The interception reservoir occupies an identical fraction of all snow-free
tiles.
|
(7.2) |
In the expressions above the minimum snow mass that ensures complete coverage of the grid box is
and the maximum water over a single layer of leaves or over
bare ground is 
. The leaf area index LAI, is specified in Table
7.1 as a function of surface type. The full set of fractional tile coverages
is given by Eqs. (7.3) and (7.4), where the indexing
of the tiles is detailed in Table 7.2. Since a mixture of
land and ocean tiles is not allowed, a grid box is either 100% water (open
water and ice, with ice fraction ci):
|
(7.3) |
or 100% land (tiles 3 to NT, where NT=8 is the number of tiles):
|
(7.4) |
Apart from the fractional gridbox coverage, each tile has a couple of additional parameters (see Table 7.2):
• The skin conductivity,  , provides the thermal connection between the skin level and the
soil or snow deck. For high vegetation,  , is different for a stable and unstable stratification of the temperature
gradient between the skin level and the upper soil or snow layer.
This difference is considered to represent the asymmetric coupling
between the ground surface and the tree canopy layer: an effective
convective transport within the tree trunk space for unstable conditions,
and a limited turbulent exchange for stable stratification (Bosveld
et al. 1999). |
| • A small fraction fRs of net short-wave radiation that is transmitted directly to the top soil or snow layer. The remaining fraction of the short-wave radiation (1 - fRs) is absorbed by the skin layer. |
| Index |
Tile |
 unstable(W m-2K-1) |
 stable(W m-2K-1) |
fRs |
Resistance scheme |
|---|---|---|---|---|---|
| 1 |
Open water |
 |
 |
0 |
Potential |
| 2 |
Ice water |
58 |
58 |
0 |
Potential |
| 3 |
Interception reservoir |
10 |
10 |
0.05 |
Potential |
| 4 |
Low vegetation |
10 |
10 |
0.05 |
Resistance |
| 5 |
Snow on low vegetation/bare ground |
7 |
7 |
0 |
Potential |
| 6 |
High vegetation |
 |
 |
0.03 |
Resistance |
| 7 |
High vegetation with snow beneath |
 |
 |
0.03 |
Canopy and snow resistance |
| 8 |
Bare ground |
15 |
15 |
0 |
Resistance |
The resistance scheme describes the way of coupling with the atmosphere: Potential denotes atmospheric resistance only; Resistance denotes aerodynamic resistance in series with a canopy or soil resistance; Canopy and snow resistance denotes a canopy resistance for the vegetation and an extra aerodynamic coupling to the snow surface (see Figs. 7.1 - 7.2 and Subsection 7.2.2 ). For tiles 6 and 7,  and  represent the aerodynamic coupling between the canopy and the
soil in the unstable and stable cases, respectively, and the factor
5 represents the longwave radiative exchanges. Unstable/stable
refers to the temperature gradient between the skin layer and the
top soil or snow layer. |
|||||
Finally, the surface albedo,
, is similar for all land tiles within a grid box except for
those covered with snow (see the snow scheme description below). The climate
database provides the snow-free background albedo on a monthly basis. Long-wave
emissivity, 
, outside the window region is equal to 0.99 for all tiles;
emissivity in the window region is tile dependent and varies between 0.93
and 0.98 (see Table 2.1 in Section 2.5.5 for more details). The remaining
surface characteristics (roughness length for momentum, z0m,
and heat, z0h) are similar for all land tiles within a
grid box and specified in the climate database (Chapter 9).7.2.2 Surface heat and evaporation fluxes
A resistance parameterization is used to calculate the turbulent fluxes. Momentum exchange is parameterized with the same roughness length for all tiles, but with a different stability correction for each tile. The resistance scheme for water vapour and heat exchanges is different for different tiles (see Fig. 7.2 ). For ocean, sea ice and snow on low vegetation, the turbulent fluxes of heat and water vapour are given by
|
(7.5) |
|
(7.6) |
with
the air density, cp the heat capacity
of moist air, g the acceleration fo gravity, 
, TL, qL,
zL the wind speed, temperature, humidity and height of the
lowest atmospheric model level, and CH,i the turbulent
exchange coefficient, that varies from tile to tile because of different
atmospheric stabilities. See Chapter 3 for a description of the
exchange coefficients where different roughness lengths for heat and momentum
are assumed and a Monin-Obukhov formulation is adopted for the stability
dependence.
For high and low vegetation, an additional canopy resistance rc is added:
|
(7.7) |
with
and i indicating the high or low vegetation
tiles. rc is a function of downward shortwave radiation
Rs, leaf area index LAI, average unfrozen root
soil water 
, atmospheric water vapour deficit Da and
a minimum stomatal resistance rs,min, following Jarvis (1976):
|
(7.8) |
f1 is a hyperbolic function of downward short-wave radiation only:
|
(7.9) |
where
, 
and 
.Function f2 is defined as
|
(7.10) |
where the soil moisture at permanent wilting point and at field capacity,
and 
, respectively, are defined in Table 7.5. 
is a weighted average of the unfrozen soil water
|
(7.11) |
where Rk is the the fraction of roots in layer k and the fraction of unfrozen soil water,
, is a parameterized function of the soil temperature of layer
k , Tk, as specified in Section
7.5.2. Table 7.1
lists the coefficients ar and br which
are used to calculate the root fraction Rk according
to Zeng et al. (1998):
|
(7.12) |
where zk+1/2 is the depth of the bottom of layer k (in m; z1/2 = 0 m). Contributions from levels exceeding the column depth are added to the deepest soil layer in order to ensure that
. Table 7.3 lists the distribution
of the roots over the four soil layers..
| Vegetation index |
1 |
2 |
3 |
4 |
5 |
6 |
7 |
8 |
9 |
10 |
11 |
13 |
16 |
17 |
18 |
19 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Layer 1 |
24 |
35 |
26 |
26 |
24 |
25 |
27 |
100 |
47 |
24 |
17 |
25 |
23 |
23 |
19 |
19 |
| Layer 2 |
41 |
38 |
39 |
38 |
38 |
34 |
27 |
0 |
45 |
41 |
31 |
34 |
36 |
36 |
35 |
35 |
| Layer 3 |
31 |
23 |
29 |
29 |
31 |
27 |
27 |
0 |
8 |
31 |
33 |
27 |
30 |
30 |
36 |
36 |
| Layer 4 |
4 |
4 |
6 |
7 |
7 |
14 |
9 |
0 |
0 |
4 |
19 |
11 |
11 |
11 |
10 |
10 |
A dependence on atmospheric humidity deficit (Da=esat(TL )-eL, with e the vapour pressure) is included according to
|
(7.13) |
where gD depends on the vegetation type (Table 7.1), and is non-zero for high vegetation only.
Evaporation from the interception reservoir is given by Eq. (7.6) only when the amount of water in the interception reservoir, Wl, is sufficient to sustain potential evaporation during the entire time step
. If Wl is limited, an additional resistance
rl, analogue to rc in Eq. (7.7), is introduced. rl
is calculated from the potential evaporation of the previous time step.
Note that this type of flux-limiter is a time-step dependent feature of
the model numerics.Bare-soil evaporation uses a resistance approach, an analogue to the canopy transpiration formulation (Eq. (7.7)). The soil evaporation resistance, rsoil, is
|
(7.14) |
with f2 given by Eq. (7.10), and rsoil,min = 50 s m-1. By this parameterization, evaporation from bare ground is treated similar to a single leaved canopy with a minimum resistance rsoil,min, extracting water from the upper soil layer only, and not experiencing any additional stress due to limited radiation or dry air. Eq. (7.14) shuts off evaporation when the top soil moisture reaches permanent wilting point. When compared to observations over semi-arid areas, an alternative relative humidity formulation (Mahfouf and Noilhan 1991; Viterbo and Beljaars 1995), that does not have a similar limitation, gave excessive evaporation (van den Hurk et al. 2000).
A special treatment is included in the calculation of evaporation over high vegetation with snow underneath (see Fig. 7.2 ). Evaporation takes place from both the canopy component in the tile (Eveg,7) and from the snow lying under the vegetation. The canopy evaporation uses a canopy resistance and saturation specific humidity at the canopy skin temperature, while the snow evaporation Esn,7 is parameterized with an additional constant aerodynamic resistance ra,sn and saturation specific humidity at snow temperature Tsn. The evpaoration from tile 7 is the combination of the canopy transpiration and the snow evaporation:
|
(7.15) |
where
is the humidity at the connection point of the three
resistances (Fig. 7.2 ). After elimination
of 
, E7 can be rewritten
as:
|
(7.16) |
The first term in the equation above is interpreted as Eveg,7 and is treated in the standard way (i.e., implicit in the tile skin temperature). The second term is interpreted as evaporation from snow (Esn,7) and is handled explicitly. The values of ra,sn depend on the stability of the subcanopy layer and are functions of
and 
(see Table 7.2); ra,sn
= 67 s m-1 and ra,sn = 220 s m-1
for an unstable and stable subcanopy layer, respectively. In spring, the
latent heat flux of that tile, LvEveg,7+LsE
sn,7 will be dominated by snow evaporation since the frozen
soil under the snow deck will lead to very large values of rc.The grid box total sensible and latent heat fluxes are expressed as an area weighted average:
|
(7.17) |
|
(7.18) |
with Hi given by Eq. (7.5), and Ei by Eq. (7.6) for ocean, sea-ice and snow on low vegetation, Eq. (7.7) for dry high and low vegetation, the interception reservoir (with rc replaced by rl) and for bare soil (with rc replaced by rsoil) and Eq. (7.16) for high vegetation with underlying snow.
Next Section
Previous Section
 |
05.04.2002 |
 |
© ECMWF |
