Geoscience Reference
In-Depth Information
topography, vegetation, and precipitation using statistical-dynamical approaches.
Models in this group also sought to improve the basic parameterizations of
hydrologic processes, such as infiltration, surface runoff, subsurface runoff, and
snow processes. The representation of these processes was considered to be overly
simplified in the earlier generation of micrometeorological models.
Although models in this group sought greater realism with the aim of improv-
ing the area-average calculations of the values 1-7 in Table 24.1, they did not give
priority to including description of CO
2
exchange. Improvements in the represen-
tation of hydrological processes in land surface models continues (e.g., Koster
et al.
, 2000; Liang and Xie, 2001; Milly and Shmakin, 2002; Cherkauer and
Lettenmaier, 2003; Huang
et al.
, 2008), and there is now interest in introducing
improved hydrological representation into the group of models described in the
next section, including the impacts of subgrid variability in precipitation based on
the schemes of Shuttleworth (1988) and Liang
et al.
(1996b), for instance.
Improving representation of carbon dioxide exchange
Motivated by the need for a more comprehensive representation of the carbon
cycle in GCMs to address climate change issues, the development of a further
generation of land surface sub-models was fostered by the work of Bonan (1995),
Sellers
et al.
(1996), Dickinson
et al.
(1998), Cox
et al.
(1998), and Dai
et al.
(2003).
The characteristic feature of this group of models, which are illustrated
schematically in Fig. 24.7, is that they seek to include representation of the plant
physiological processes and vegetation dynamics, in an attempt to account for
carbon uptake by plants and the feedbacks between climate and vegetation. Pitman
(2003) provides a comprehensive discussion of land surface models designed for
coupling to climate models. Two main approaches are used in predicting seasonal
variations in vegetation dynamics, i.e., a plant physiological process-based
approach (e.g., Lu
et al.
, 2001), and a rule-based approach (e.g., Foly
et al.
, 1996;
Levis and Bonan, 2004; Kim and Wang, 2005).
Studies of plant biochemistry had suggested a different approach to modeling
stomatal control that is somewhat less empirical than the Jarvis-Stewart model,
and therefore hopefully more transferable from one plant species to the next and
(perhaps) only dependent on whether species are C3 or C4 plants in terms of their
photosynthetic function. In such models, the assimilation of carbon is viewed as
the controlling factor, and stomatal conductance is described by (sometimes a
derivative of ) the so-called Ball-Berry equation (Ball
et al.
, 1987), i.e.:
(24.7)
gmACPFg
=
(
)
+
s
n
s
l
e
min
where
g
min
is a prescribed minimum stomatal conductance;
m
is a slope parameter
(~9 for C3 plants);
A
n
is the net carbon assimilation;
C
s
, is the partial pressure of
carbon dioxide;
P
l
is atmospheric pressure adjacent to the leaf; and
F
e
is a humidity
dependent stress factor, which in the original Ball-Berry equation was set as
