89
Over land, ESEMs have traditionally been viewed as tools to calculate surface momentum, energy and water fluxes for atmospheric models (AMs) given a set of land surface
characteristics (Avissar and Verstraete 1990). In the broader context of interactive climate
system modeling, however, land ESEMs become an interactive interface between atmosphere, biosphere and surface hydrology models. To produce the surface fluxes, ESEMs
require the provision of a number of surface variables, e.g. vegetation cover and characteristics (see section 3). In most present models, these are specified as input datasets.
In a coupled CSM such as that of Fig. 1, however, these can be interactively generated
by ecosystem models, which in turn make use of the surface climate variables and water
budget generated by the ESEM to simulate ecosystem dynamics.
A wide range of ecosystem models is today available, depending on the specific applications (e.g. Martin 1995). In the simplest models, such as that of Holdridge (1947), a
life zone, or life form, is uniquely assigned to given values of climatic parameters. In the
more complex gap-phase and layered-based succession models (e.g. Botkin et al. 1972,
West et al. 1981), forest dynamics is deterministically represented by processes of establishment, growth and mortality as well as stochastic events (e.g. fires). Ecosystem models
are also available which describe plant productivity, biogeochemical cycling of nutrients
and vegetation dynamics both for natural ecosystems and agricultural crops (Running
and Coughlan 1988, Parton et al. 1990, Rosenzweig 1990).
In present ecosystem models, climate variables are treated as external forcings, obtained either from observations or from output of atmospheric models. A basic problem
with interactive coupling of biosphere and atmosphere models is that each calculates
their own hydrologic cycle with different methods. This may result in an inconsistency
between the climate forcing and the vegetation response. This inconsistency may be
overcome if an ESEM is used as true interface between the models, whereby the ESEM
would calculate the surface hydrologic cycle for a given ecosystem in response to a given
climate forcing. This hydrologic cycle would be passed to the ecosystem model, which
would return the surface vegetation characteristics needed by the ESEM to calculate the
atmosphere/surface exchanges.
Coupling of atmosphere and surface hydrology essentially occurs through the partitioning of precipitation in evaporation, runoff and infiltration into deep soil. Present
climate models basically calculate runoff as a residual term accounting for precipitation,
evapotranspiration and the water balance of a conceptual surface soil layer of a few meters
depth. In reality, however, precipitation, evaporation, soil moisture and runoff affect each
other in very complex ways which depend on terrain morphology, soil characteristics and
interactions with deep water tables (Freeze 1978, Dooge 1986). Hydrology models are
thus needed to account for such effects in the calculation of runoff efficiencies and water
exchanges with deep soil (e.g Beven and Kirkby 1979, Abbott et al. 1986). ESEMs can
Over land, ESEMs have traditionally been viewed as tools to calculate surface momentum, energy and water fluxes for atmospheric models (AMs) given a set of land surface
characteristics (Avissar and Verstraete 1990). In the broader context of interactive climate
system modeling, however, land ESEMs become an interactive interface between atmosphere, biosphere and surface hydrology models. To produce the surface fluxes, ESEMs
require the provision of a number of surface variables, e.g. vegetation cover and characteristics (see section 3). In most present models, these are specified as input datasets.
In a coupled CSM such as that of Fig. 1, however, these can be interactively generated
by ecosystem models, which in turn make use of the surface climate variables and water
budget generated by the ESEM to simulate ecosystem dynamics.
A wide range of ecosystem models is today available, depending on the specific applications (e.g. Martin 1995). In the simplest models, such as that of Holdridge (1947), a
life zone, or life form, is uniquely assigned to given values of climatic parameters. In the
more complex gap-phase and layered-based succession models (e.g. Botkin et al. 1972,
West et al. 1981), forest dynamics is deterministically represented by processes of establishment, growth and mortality as well as stochastic events (e.g. fires). Ecosystem models
are also available which describe plant productivity, biogeochemical cycling of nutrients
and vegetation dynamics both for natural ecosystems and agricultural crops (Running
and Coughlan 1988, Parton et al. 1990, Rosenzweig 1990).
In present ecosystem models, climate variables are treated as external forcings, obtained either from observations or from output of atmospheric models. A basic problem
with interactive coupling of biosphere and atmosphere models is that each calculates
their own hydrologic cycle with different methods. This may result in an inconsistency
between the climate forcing and the vegetation response. This inconsistency may be
overcome if an ESEM is used as true interface between the models, whereby the ESEM
would calculate the surface hydrologic cycle for a given ecosystem in response to a given
climate forcing. This hydrologic cycle would be passed to the ecosystem model, which
would return the surface vegetation characteristics needed by the ESEM to calculate the
atmosphere/surface exchanges.
Coupling of atmosphere and surface hydrology essentially occurs through the partitioning of precipitation in evaporation, runoff and infiltration into deep soil. Present
climate models basically calculate runoff as a residual term accounting for precipitation,
evapotranspiration and the water balance of a conceptual surface soil layer of a few meters
depth. In reality, however, precipitation, evaporation, soil moisture and runoff affect each
other in very complex ways which depend on terrain morphology, soil characteristics and
interactions with deep water tables (Freeze 1978, Dooge 1986). Hydrology models are
thus needed to account for such effects in the calculation of runoff efficiencies and water
exchanges with deep soil (e.g Beven and Kirkby 1979, Abbott et al. 1986). ESEMs can