Toward Marine Environmental Predictions in the Mediterranean Sea Coastal Areas
367
physical environment is given by the light radiative propagation submodel which
normally follows a simple exponential law which itself is dependent upon inorganic matter and phytoplankton pigments concentration.
The pelagic submodel communicates with the benthic system through pore water
diffusive processes. The benthic system is described as a multiple layer system
with pore water nutrients, dissolved gases, benthic bacteria, zoobenthic organisms,
detritus, dissolved organic matter and particulate organic matter. Important developments in the future should consider benthic-pelagic coupling through resuspension processes of cohesive sediments. Coastal ecosystems almost certainly require
the complexity of interactions illustrated above while open ocean ecosystems may
be more simply schematized, as done in the past (Fasham et al., 1990). Another
level of complexity in coastal ecosystem modeling is given by the imposition of
land derived lateral fluxes of nutrients, particulate and dissolved organic matter and
inorganic sediments, as outlined in Fig. 17.9b.
In order to solve the initial and boundary condition problem given by equation
(29) we need to give initial estimates of each state variable, the lateral fluxes from
land sources and the vertical fluxes determined by atmospheric transport. The latter
can be relevant for open ocean areas and/or shelf areas in absence of importance
river sources. The second dominates over atmospheric transport in river runoff
dominated shelf areas. The land derived lateral flux is only considered for dissolved nutrients, dissolved and particulate organic matter and inorganic material.
The flux is parameterized in the form ofprescribed values, e.g.:
(31)
where Xo ' Yo are the nutrients discharge land points and C* is the measured concentration. For atmospheric input, the problem should be solved with the imposition of
observed nutrient fluxes at the surf ace,
K aCI = Qc
vaz z = o
(32)
where Qc is the flux ofthat particular nutrient at the surface.
The initial condition problem can be again posed by specifying climatological
estimates of state variables or by assimilation techniques. If the assimilation methods are being developed for the shelf areas hydrodynamic modeling, data assimilation for field estimation in ecosystem models is at its infancy. Initially the
observations have been used to solve an inverse problem where best estimates of
parameters for the biochemical rates were sought. Only recently, multidisciplinary
observations have been used to estimate three-dimensional initial fields for some of
the state variables. The large dimensionality and the multivariate character of the
field estimation problem does not allow simple approaches such as nudging or
univariate optimal interpolation schemes. However, experiments are starting which
consider the assimilation of satellite color data to estimate several initial ecosystem
dynamical model variables.
367
physical environment is given by the light radiative propagation submodel which
normally follows a simple exponential law which itself is dependent upon inorganic matter and phytoplankton pigments concentration.
The pelagic submodel communicates with the benthic system through pore water
diffusive processes. The benthic system is described as a multiple layer system
with pore water nutrients, dissolved gases, benthic bacteria, zoobenthic organisms,
detritus, dissolved organic matter and particulate organic matter. Important developments in the future should consider benthic-pelagic coupling through resuspension processes of cohesive sediments. Coastal ecosystems almost certainly require
the complexity of interactions illustrated above while open ocean ecosystems may
be more simply schematized, as done in the past (Fasham et al., 1990). Another
level of complexity in coastal ecosystem modeling is given by the imposition of
land derived lateral fluxes of nutrients, particulate and dissolved organic matter and
inorganic sediments, as outlined in Fig. 17.9b.
In order to solve the initial and boundary condition problem given by equation
(29) we need to give initial estimates of each state variable, the lateral fluxes from
land sources and the vertical fluxes determined by atmospheric transport. The latter
can be relevant for open ocean areas and/or shelf areas in absence of importance
river sources. The second dominates over atmospheric transport in river runoff
dominated shelf areas. The land derived lateral flux is only considered for dissolved nutrients, dissolved and particulate organic matter and inorganic material.
The flux is parameterized in the form ofprescribed values, e.g.:
(31)
where Xo ' Yo are the nutrients discharge land points and C* is the measured concentration. For atmospheric input, the problem should be solved with the imposition of
observed nutrient fluxes at the surf ace,
K aCI = Qc
vaz z = o
(32)
where Qc is the flux ofthat particular nutrient at the surface.
The initial condition problem can be again posed by specifying climatological
estimates of state variables or by assimilation techniques. If the assimilation methods are being developed for the shelf areas hydrodynamic modeling, data assimilation for field estimation in ecosystem models is at its infancy. Initially the
observations have been used to solve an inverse problem where best estimates of
parameters for the biochemical rates were sought. Only recently, multidisciplinary
observations have been used to estimate three-dimensional initial fields for some of
the state variables. The large dimensionality and the multivariate character of the
field estimation problem does not allow simple approaches such as nudging or
univariate optimal interpolation schemes. However, experiments are starting which
consider the assimilation of satellite color data to estimate several initial ecosystem
dynamical model variables.