Download Modeling Biogeochemical Processes in Marine Ecosystems

OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
MODELING BIOGEOCHEMICAL PROCESSES IN MARINE
ECOSYSTEMS
Grégoire, M.
GHER, University of Liege, Sart-Tilman B5, B-4000, Liege, Belgium
Oguz, T.
METU, Institute of Marine Sciences, P.O. Box 28, Erdemli, 33731, Icel, Turkey
Keywords: Marine biogeochemical cycles, ecosystem, mathematical modeling, food
web, trophic levels, primary production, new and regenerated production, plankton,
phytoplankton blooms, microbial loop, organic matter generation and decomposition,
inverse modeling, data assimilation, general circulation, mesoscale eddies, upwelling,
vertical mixing, entrainment.
Contents
1. Introduction
2. General Classification of Biogeochemical Models
3. Models Complexity in terms of Food Web Structure
4. Physical Controls on Biogeochemical Models
5. Inverse Approaches in Biogeochemical Modeling
6. Regional Applications of Coupled Physical-biogeochemical Models
Glossary
Bibliography
Biographical Sketches
Summary
Ecosystems are dynamic systems and the study of their functioning and evolution
requires the development of dynamic biogeochemical models. A wide application and a
pronounced development of this type of model has taken place during the last two
decades due essentially, to the development of computer technology, which has
permitted the handling of very complex mathematical systems, and also to the
increasing need of understanding marine ecosystems and their dynamic response to
environmental stresses from localized pollution to global climate changes.
A review of the progresses realized in ecosystem modeling, starting from simple NPZtype models to more complex models, describing in parallel the biogeochemical cycles
of different biogenic nutrients, is made in this paper. The evolution of the complexity
introduced to the model's structure and in the parameterization of biochemical laws, is
analyzed. The increasing coupling of ecosystem models with hydrodynamical models in
order to assess the role of physical processes on the ecodynamics is discussed. The
efforts made to develop ecosystem models more and more consistent with the available
observations by, for instance, improving the calibration of the most sensitive model
parameters and/or by determining the "optimal" model structure using data assimilation
techniques are outlined.
©Encyclopedia of Life Support Systems (EOLSS)
261
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
Finally, some regional applications of the coupled physical-biogeochemical models for
various oceanic basins and coastal and shelf seas, as well as the recent attempts at
incorporating them into the global carbon cycle models are described.
1. Introduction
The oceans exhibit highly diverse and variable ecosystems governed by complicated
sets of physical-biogeochemical interactions between the atmosphere, the surface ocean,
and its interior, on a variety of spatial and temporal scales. The marine biogeochemical
cycling involves organic matter generation by photosynthesis by primary producers
constituting the first trophic level of the food web, its transfer to higher trophic levels by
the feeding activities of animals, and its decomposition back to inorganic forms (Figure
1).
Figure 1. Schematic simplified representation of the organic and
inorganic matter cycling in the ocean.
Particulate non-living organic material (the so-called detritus) is formed through the
natural mortality of phytoplankton and zooplankton, or through the production of fecal
pellets. Dissolved organic matter is formed by soluble organic materials released during
excretion and exudation. These organic materials are then decomposed by microbial
processes. The growth of primary producers is usually limited by the availability of one
or several biogenic elements such as nitrogen (NO3+NH4+NO2), bioavailable iron (Fe),
phosphate (PO4), and dissolved silicon (Si(OH)4). The silicon cycle is relatively simple,
affects primarily diatoms and involves only inorganic forms. The phosphorus cycle is
also relatively simple; organic phosphate is converted back to an inorganic form which
then becomes available again for uptake by phytoplankton. It is a rapid process, and
therefore phosphorus is generally not limiting in the marine environment. Recycling of
nitrogen is a more complex process. Organic nitrogen is regenerated in the water
column by bacterial activities and zooplankton excretion in the form of ammonium,
which is then oxidized to nitrite and then to nitrate in the nitrification process occurring
in the oxygenated part of the water column. In the anaerobic systems, mostly in
sediments, nitrate is consumed as an oxidizer instead of oxygen. This process is referred
to as the denitrification, and results in conversion of nitrate by bacteria, into first nitrite
©Encyclopedia of Life Support Systems (EOLSS)
262
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
and then to nitrogen gas, thus leading to the loss of nitrogen from the system. The
atmospheric nitrogen gas may also be converted to organic nitrogen compounds by
some phytoplankton species (e.g., the blue-green algae).
The ocean is the largest reservoir of carbon with rapid exchanges with the atmosphere,
and it is the largest net sink for anthropogenic atmospheric CO2. A fundamental process
regulating the air-sea balance of CO2 and the amount of CO2 fixed in the ocean is the
biological pump; namely the efficiency of photosynthesis, the foodweb structure and the
amount of new nitrate entered into the euphotic zone, coupled with remineralisation at
depth. In general, CO2 is converted from inorganic to organic carbon by the
photosynthesis of the phytoplankton. This is then consumed by the higher trophic levels,
and some CO2 is recycled as inorganic bicarbonate formed by interaction of free
dissolved CO2 with water. Some losses may occur from the ocean surface in gaseous
form. Respiration and remineralisation processes also contribute to CO2 production in
the water column. (see Marine biogeochemical cycles: effects on climate and response
to climate change).
All the biogeochemical processes and interactions between living and non-living
components of the ecosystems cannot possibly be explored through observations alone.
The satellite-based observations, which are the only means of synoptic information on
regional and global scales, are restricted to the upper few meters of the water column,
and their correct interpretation requires additional knowledge on the physicalbiogeochemical processes taking place in the ocean interior. Comprehensive
observational programs for basin-scale measurements are economically not yet feasible,
and have to make compromises between temporal and spatial coverages as well as on
the number of variables to be measured. The observations alone are therefore unable to
provide a complete description of ecosystems. Observations generally provide
distributions of concentrations and/or biomass, but hardly yield details on the spatial
and temporal properties of the rates and processes controlling these distributions. Some
mathematical tools are necessary to interpolate and extrapolate the available data to
other parts of the region, and to complement missing data in a dynamically consistent
way. The relative importance of different factors and their response to different
conditions can be most efficiently analyzed using models, by varying individual factors
independently.
The ecosystem models may therefore be defined as mathematical tools that help to
further understand, conceptualize, and predict marine environmental processes using a
simplified representation of the real world in the form of a series of differential
equations. Biogeochemical modeling has made considerable progress during the last
two decades thanks essentially to the development of computer technology, and
increasing public concern about environmental stresses spanning from localized
pollution to global climate changes.
In the following sections, highlights from some important modeling initiatives are
presented in order to describe the progress made in marine biogeochemical modeling
within the last ten years. The biogeochemical models are first classified according to
their objectives. The structural complexity introduced into the food webs and nutrient
cycles are then described. It is followed by a description of efforts on the coupling of
©Encyclopedia of Life Support Systems (EOLSS)
263
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
upper ocean physics and biogeochemistry, and of the role of physical controls on the
biological production. The next section deals with inverse approaches used in the
biogeochemical modeling for a more realistic estimation of the model parameters by
means of assimilating the available data. The rest of the paper describes the regional
applications of the coupled physical-biogeochemical models for oceanic basins and
coastal and shelf seas, as well as the recent attempts of incorporating them into global
carbon cycle models.
2. General Classification of Biogeochemical Models
A wide spectrum of biogeochemical models is available, ranging from simple budget
models, to more complicated process, heuristic, predictive and inverse type models. The
budget models deal with estimation of the fluxes of water, nutrients and other materials
in a region over annual or multi-annual time scales. Using the available data, these
models attempt to make a budget of the system involving all major external and internal
sources and sinks, as well as the dominant internal transformation processes of the
variable in question (Figure 2).
Figure 2. Cycling and transport processes that occur in the cycle of trace metals in the
Adriatic Sea (above), and a simple budget model for the Adriatic Sea (below).
©Encyclopedia of Life Support Systems (EOLSS)
264
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
(Reprinted from Journal of Marine Systems, 25, Tankere S. P. C., Price N. B. and
Statham P. J., pages 270 and 272, copyright 2000, with permission from Elsevier
Science).
They therefore require a substantial amount of data. Comparing fluxes through the
system allows the budget models to identify key processes and to make tentative
predictions about the consequences of environmental changes.
Some models, called process models, are devoted to process oriented studies (Figure 3).
They are designed to identify specific physical, chemical and biological processes, and
to determine the relative influence of major environmental forcing functions. They are
generally constructed on the basis of field or laboratory studies to test a hypothesis, to
explore general principles on functioning of an ecosystem, and to predict the outcome of
idealized scenarios. In most cases, they are developed in isolation from the total system
with a limited number of variables and parameters, and by making use of a specific
empirical dataset collected at a specific location. They constitute important building
blocks of more sophisticated models.
Figure 3. Diagrammatic representation of the internal metabolism of phytoplankton
cells. Cellular constituents are divided according to their cellular function, into small
metabolites S, reserve products R, and structural products F. (redrawn from Lancelot
and Billen, 1989).
Heuristic models are concerned with the behavior of whole ecosystems of a region by
incorporating their physical and biogeochemical aspects. They are used to elucidate the
dynamics underlying a particular set of observations. They integrate information on
what are considered to be the most important variables and processes for the entire
ecosystem. They have well defined spatial boundaries and temporal scales, and their
behavior depends very much upon the scales selected. They contain a large number of
physical, chemical and biological processes, variables and parameters, many of which
are poorly understood. They are constrained, to some degree of accuracy, to reproduce
time series of one or more measured state variables. These models are used to draw
inferences about the dynamics that led to the observations. They can be either
descriptive (e.g., flow or network analysis) or dynamic (e.g., simulation models). Most
©Encyclopedia of Life Support Systems (EOLSS)
265
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
of the available modeling studies fall into this group, for which some examples will be
given in the following sections as we classify them below according to their structures.
Heuristic models may also be designed to predict responses of the system to both
natural variability and human activity (e.g., reduced freshwater runoff, increased
nutrient input), and to understand how critical biogeochemical processes are influenced
by environmental variables (Figure 4). This approach is particularly useful in
extrapolating the present state of the ecosystem to the future, and allows prediction of
the system's functioning under new conditions. The predictive capability of ecosystem
models can be used as management tools for investigating alternative management
plans and response scenarios to climatic and man-made environmental changes.
Figure 4. Diagrammatic representation of a complex ecosystem model which has been
used to hindcast the eutrophication of the North Sea during the last 40 years from 1955
to 1993.
©Encyclopedia of Life Support Systems (EOLSS)
266
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
This is a generic model that describes both the pelagic and benthic ecosystems and the
coupling between them in terms of the significant bio-geo-chemical processes affecting
the flow of carbon, nitrogen, phosphorus and silicon. It uses a 'functional group'
approach to describe the ecosystem whereby biota are grouped together according to
their trophic level, and sub-divided according to size and feeding method. The pelagic
food web describes phytoplankton succession (in terms of diatoms, flagellates,
picoplankton and inedible phytoplankton), the microbial loop (bacteria, heterotrophic
flagellates and microzooplankton), mesozooplankton (omnivores and carnivores) and
fish predation at a fixed grazing rate. The benthic sub-model contains a food web
capable of describing nutrient and carbon cycling. The model contains about hundred
parameters.The North Sea has been divided in 130 interior boxes interacting at their
common boundaries via matter and energy exchanges. The hydrodynamical transports
of matter by advection/diffusion are derived from a corresponding hydrodynamical
simulation using a general circulation model. (Reprinted from Neth. Journal of Sea
Research, 33, Baretta J.W., Ebenhoh W. and Ruardij P., copyright 1995, with
permission from Elsevier Science).
Early ecosystem models have generally been zero-dimensional, box models, integrated
over some appropriately defined region of the sea. However, one now realizes that
biological populations have complex spatial distributions, significantly affected by the
hydrodynamic processes on similar space and time scales. Therefore, marine
biogeochemical systems demand the time-dependent, three-dimensional models
describing multiple interactions between the state variables. The success of ecosystem
simulations are usually limited by our theoretical and experimental knowledge of the
performance of all components of the ecosystem. Despite some limitations and
problems of initialization, calibration and validation, when applied correctly, the 3D
models provide useful insights into the internal processes of the system. The description
of these processes can be used to predict the behavior of the ecosystem in relation to
environmental change (Figure 5). The theoretical and experimental limitations of these
models can be avoided to some extent by carrying out parallel studies on process
models and more systematic measurements.
a global entrainment by the mean flow, i.e., horizontal and vertical advection:
∂wy
∂z
∇ H .(uy ) +
a slipping through the flow due to the diffusion of y, i.e., horizontal and vertical
diffusion:
λH ∇ 2H y +
∂
∂y
(ν )
∂z ∂z
or to its migration,
i.e.
∂ ( wsy y )
∂z
.
©Encyclopedia of Life Support Systems (EOLSS)
267
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
Figure 5. Schematic representation of a coupled 3D physical-biogeochemical/ecosystem
model including its interactions with the atmosphere, the land and the sediments. The
evolution in space and time of any 3D biogeochemical state variable y is described by
equation 1. This equation illustrates the different ways according to which the
hydrodynamics influences the ecodynamics. In particular, the transport of any
biogeochemical variable results from the superposition of:
Qy is the local production-destruction term resulting from biogeochemical interactions.
A sound understanding of the oceans requires a combination of observations and
models. Models are initialized by observations, solved by integrating forward in time,
and then compared with a set of observations. If the simulations are successful, the
models predictions and data have only minor discrepancies. The model results can then
©Encyclopedia of Life Support Systems (EOLSS)
268
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
be used to diagnose the processes governing the observed ecosystem. If the model
simulations are inconsistent with the data, the food web structure and/or parameters of
the model are then revised until the simulations are able to provide a desirable accuracy.
Alternatively, as a recently growing discipline in marine ecosystem modeling, inverse
methods offer a novel approach to determine the model inputs (e.g., parameters, forcing
functions) directly from the observations. More precisely, ocean data assimilation refers
to the estimation of marine fields of interest by melding data and dynamics in accord
with their specific uncertainties (Figure 6). Data assimilation provides a powerful
methodology for state and parameter estimation.
Figure 6. General process of data assimilation.
©Encyclopedia of Life Support Systems (EOLSS)
269
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
3. Models Complexity in Terms of Food Web Structure
Models, in essence, reflect idealized, mathematical forms of natural ecosystems. A
model is therefore never expected to contain all features of a real system, otherwise it
would be the real system itself. The important point is to formulate the system as
realistically as possible under given simplifications and approximations. Usually no
practical, objective method exits to prefer a particular model; the choice is somewhat
arbitrary, but it is dictated by the characteristics of the ecosystem under consideration,
the degree of complexity one wishes to introduce, and the available data structure and
observational knowledge. Therefore, several alternative models can be developed for
the same environment, and it is always questionable to set up the right complexity. The
simplicity versus complexity in representation of an ecosystem structure is also a matter
of choice depending on the problem at hand, and the personal preferences of modelers.
Some modelers are inclined to represent the ecosystem as simply and as idealized as
possible, while some others tend to keep as much biological details as possible. It is a
general assumption that a model may be made more realistic by adding more and more
connections, up to a point. Addition of new state variables and new parameters beyond
that point does not contribute further to improve simulations. Indeed, more parameters
imply more uncertainty, because of the possible lack of information about all the
processes. Given a certain amount of data, introducing additional new state variables or
parameters beyond a certain model complexity does not give a further advantage, but
only adds unaccounted uncertainty.
Biologically more complex models may be thought of as providing potentially better
representation of the system due to the great amount of state variables and process
equations. They, however, have a major drawback: i.e., the need tospecify a larger
number of unconstrained parameters. They are also more difficult to solve and
understand, and often it is quite impractical to carry out detailed sensitivity analyses.
Simpler models may be an oversimplified representation of the ecosystems, but they are
easier to understand, contain a smaller number of parameters and are more amenable to
detailed sensitivity studies. Simpler models may be preferred in the process-oriented
studies designed to thoroughly elucidate the dynamics. On the contrary, those aimed at
simulating observations, or to predicting a future state of an ecosystem as realistically as
possible unavoidably require some degree of sophistication.
Ideally, marine ecosystem models should comprise not only food web representations,
but also biogeochemical cycles of major nutrients (i.e., nitrogen, phosphorus, carbon
and silicon). However, most of the earlier models specified the entire foodweb structure
in a very simple way by single phytoplankton and zooplankton groups, without
providing details either at the species level or size fractionation. They were also
supported by a grossly aggregated organic matter (detritus) compartment including both
particulate and dissolved substances and a single inorganic nutrient (usually nitrogen)
(i.e., NPZ-type models). In spite of the apparent simplicity in the mathematical
description of the ecosystems, such grossly-aggregated models had some problems and
difficulties in the representation of biochemical processes and in the specification of
metabolism rates (e.g., growth, grazing, mortality). However, such simplified nitrogenlimited ecosystem models have been applied to numerous marine ecosystems around the
world.
©Encyclopedia of Life Support Systems (EOLSS)
270
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
One of the simplest refinements of the NPZ-type grossly aggregated models is to divide
the phytoplankton and/or the zooplankton compartments in several species groups
or/and to consider several limiting nutrients (e.g., nitrogen, phosphorus, silicon). The
phytoplankton compartment was often found necessary to be represented in the form of
at least two species groups, as diatoms and flagellates, or two size groups as small and
large phytoplankton. Such distinctions allow the simulation of the succession of blooms
under different light, nutrient and zooplankton grazing conditions, and therefore gives
rise to a more realistic representation of seasonal variations of the primary production
and phytoplankton biomass. In some cases, this distinction was also accompanied by the
introduction of silicon as another limiting nutrient, since the growth of diatoms depends
on the efficiency of the silicate pump. The zooplankton compartment was customarily
differentiated into two groups, as microzooplankton and mesozooplankton. In some
cases, it was necessary to introduce more detailed structure in the form of several
classes of microzooplankton, of mesozooplankton and higher predators. For instance,
the ecosystem models applied to eutrophic regions have been refined to take into
account opportunistic species, such as mixotrophic dinoflagellates and gelatinous
carnivores. This is the case for the north-western shelf of the Black Sea and the
European continental shelf of the North Sea.
By the early 1980s, the importance of the microbial loop in controlling the organic
matter recycling and plankton productivity became evident by measurements. Following
these observational studies, a description of the microbial loop was introduced in
ecosystem models and the NPZ-type models were extended to include bacteria,
dissolved organic matter and microzooplankton. Some studies incorporated further
sophistication by representing the refractory and labile fractions of DOM separately,
and dividing the POM into small and large groups. Such fractionations led to a more
realistic organic matter recycling. For instance, in the western Mediterranean, the small
POM and the labile DOM are mostly regenerated, while the others are exported below
the mixed layer.
In environments with a limited ventilation, such as the Black Sea and the Arabian Sea
deep waters, the degradation of the biological production leads to the formation of
hydrogen sulfide, a toxic substance for all marine living resources. In the transition zone
between the oxic and anoxic waters, the oxidization-reduction potential (redox
potential) of the water decreases sharply due to oxygen deficiency and the appearance
of hydrogen sulfide at its lower boundary. As a result, this layer contains a rich set of
redox reactions involving oxygen, nitrogen, sulfur, carbon, manganese and iron, and so
its lower and upper interfaces are characterized by important gradients of biological and
chemical variables. Also, models describing the nitrogen and sulfur cycles (e.g.,
hydrogen sulfide, total elemental sulfur, thiosulfates, sulfates, total organic nitrogen,
ammonium, nitrite, nitrate) coupled with oxygen dynamics, and sometimes to the
manganese cycle as a catalyzer for the redox reactions have been developed for such
oxygen-deficient waters.
Figures 7 to 10 give the schematic representation of ecosystem models of different
complexity applied in the eutrophic Black Sea basin with different aims.
©Encyclopedia of Life Support Systems (EOLSS)
271
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
The comparison of these four pictures shows clearly that the structure of the model is
adapted according to the objectives of the studies. Also, for instance, the ecosystem
model given in Figure 7 has been coupled with a 1D vertically integrated physical
model in order to simulate the complex modifications of the Black Sea ecosystem
structure and functioning caused by eutrophication. The model has been applied in the
Danube's discharge area and is described by 33 state variables and 26 processes linking
them. The biogeochemical model described in Figure 8 has also been coupled with a 1D
physical model to provide a unified representation of the dynamically coupled oxicsuboxic-anoxic system of the interior Black sea.
Figure 7. Structure of a complex ecosystem model applied on the north-western Black
Sea. Solid lines show an idealized description of the foodweb representative of the
situation before eutrophication. Dashed lines correspond to the incorporation of
autotrophic opportunistic species. The gelatinous food chain is represented by dotted
lines. Primary producers are described as three different functional groups according to
their different physiological properties and trophic fates: diatoms (DIA), nanoflagellates
(NAN) and autotrophic opportunistic species (OPP). The secondary trophic level is
composed of bacteria (BAC), micro-zooplankton (MUZ), copepods (COP) and the giant
omnivorous dinoflagellates Noctiluca scintillans (NOC). Among carnivorous
organisms, only the jellyfish Aurelia aurita (AUR) and the exotic ctenophore
Mnemiopsis ledyi (MNE) are considered explicitely. DOM and POM respectively are
dissolved and particulate detritus. Most of the parameters are determined experimentally
based on processes-oriented studies. (Reprinted from Kluwer Academic Publishers,
Sensitivity to change: Black Sea, Baltic Sea and North Sea, E. Ozsoy and A. Mikaelyan
(Eds), NATO ASI Series, 27, 1997, Van Eeckhout D. and Lancelot C., Modeling the
functioning of the north-western Black Sea ecosystem from the 1960 to present, figure
4, page 460, copyright 1997, with kind permission from Kluwer Academic Publishers).
©Encyclopedia of Life Support Systems (EOLSS)
272
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
Figure 8. Prey-predator interactions between plankton groups of the Black Sea pelagic
foodweb (the biological model, 54 parameters, left), and (right) pathways of
biogeochemical transformations constituting elements of the nitrogen and redox cycles
in the model (about 21 parameters). The biological model represents the pelagic
foodweb with two groups of phytoplankton (diatoms Pd, and phytoflagellates Pf), two
classes of herbivorous/omnivorous zooplankton (microzooplankton Zs, and
mesozooplankton Zl), nonphotosynthetic free-living bacterioplankton B, the gelatinous
carnivore group Zm of medusae Aurelia aurita, and an opportunistic species group Zn
representing the omnivorous dinoflagellate Noctiluca scintillans. DON represents the
dissolved inorganic nitrogen, and D are the detritus. The numbers refer to
remineralisation (1), ammonification (2), first step of nitrification (3a), second step of
nitrification (3b), first step of denitrification (4a), second step of denitrification (4b),
oxidation of dissolved manganese by nitrate (5), oxidation of dissolved manganese by
dissolved oxygen (6), oxidation of hydrogen sulfide by particulate manganese (7),
oxidation of ammonium by particulate manganese (8). The dotted lines indicate oxygen
dependency of nitrification, and the broken lines show couplings of the biological
model with the nitrogen cycle model (Reprinted from Global Biogeochemical Cycles,
14 (4), Oguz, T., Ducklow H.W., Malanotte-Rizzoli P., pages 1336 and 1337, copyright
2000, with permission from American Geophysical Union).
The model relates the annual cycle of plankton production in the form of a series of
successive phytoplankton, mesozooplankton, and higher consumer blooms to organic
matter generation and to the remineralisation-ammonification-nitrificationdenitrification chain of the nitrogen cycle, as well as to anaerobic sulfide oxidation in
the suboxic-anoxic interface zone.
©Encyclopedia of Life Support Systems (EOLSS)
273
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
Figure 9. Schematic representation of a classical 5 state variable ecosystem model
showing the state variables and the interactions between them. The version presented
here has been applied in the Black Sea. The mathematical expression of the interaction
terms is written on the arrows.
©Encyclopedia of Life Support Systems (EOLSS)
274
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
Figure 10. Schematic representation of a refined version of the 5 state variables
ecosystem model presented in Figure 9. This model is still defined by a nitrogen cycle
described by 13 state variables. The food web is composed of two branches: the linear
food-chain leading to fish (i.e., microphytoplankton, mesozooplankton) and the
microbial food chain (nanophytoplankton, microzooplankton, bacteria, dissolved
detritus, ammonium). (Reprinted from Journal of Marine Systems, 16, Gregoire, M.,
Beckers J.M., Nihoul J. C. J. and Stanev E., page 90, copyright 1998, with permission
from Elsevier Science).
Models described in Figures 9 and 10 are not as complex, because they have been
coupled with a 3D physical model in order to assess the influence of the hydrodynamics
on the ecodynamics. Figure 9 gives a classic NPZ-type model, whereas Figure 10 shows
©Encyclopedia of Life Support Systems (EOLSS)
275
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
a refinement of this model obtained by dividing the phytoplankton and zooplankton
compartments respectively in to three and two groups, by including an explicit
representation of bacteria, and by taking into account the possibility of a phytoplankton
growth limitation by the availability of dissolved silicate and inorganic phosphorus. The
sea surface phytoplankton concentrations simulated by these two models are shown on
Figure 11 and are compared with a satellite picture showing the sea surface chlorophyll
concentration. This comparison shows that the refined ecosystem model simulates more
intense phytoplankton blooms, which are in better agreement with the observations.
Indeed, some sensitivities studies of the model solution to the ecological parameter
values have shown that the grazing control on the phytoplankton growth is
overestimated in the 5 state variables model, due to the simplicity of the representation
of the phytoplankton and zooplankton compartment. Besides, the influence of
ecological interactions on the phytoplankton horizontal distribution is more pronounced
in the refined model, and is outlined by the presence of some localized bloom
corresponding to the pick of the development of one particular group (in the case of the
simple model, the phytoplankton distribution is essentially driven by the
hydrodynamics). However, one should note that due to feasibility and reliability
constraints, simple models, such as the NPZ-type model, are often used in 3D high
resolution coupled ecosystem-hydrodynamical models.
Figure 11. Sea surface distribution of the phytoplankton field (in mgchlam−3) simulated
by the NPZ-type ecosystem model (top-left) and by the refined ecosystem model shown
in Figure 10 (top-right). Below, sea surface distribution of chlorophyll concentration (in
mgchlam−3) reconstructed from satellite data (CZCS data).
The HNLC (high nutrient-low chlorophyll) regions covering more than 20% of the
world's oceans possess a fairly constant phytoplankton biomass all year long. In these
regions, spring bloom activity is rarely observed and nitrate stocks are never
©Encyclopedia of Life Support Systems (EOLSS)
276
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
significantly depleted. The observations suggest that the combination of iron deficiency
for the growth of net phytoplankton, and strong microzooplankton grazing control on
nanophytoplankton communities are two reasons limiting the primary production in
HNLC areas. Iron deficiency prevents phytoplankton growth rates from ever being
sufficiently large for a bloom to be triggered. In the equatorial Pacific waters, iron
supply was shown to determine the most prominent phytoplankton size fraction for
blooming, while grazing pressure was shown to keep the population below the system's
carrying capacity for smaller phytoplankton. The model simulations also suggested how
the 1992 El Niño event modified the ecosystem composition from larger phytoplanktondominated populations and relatively high levels of new production, to a recyclednutrients based system dominated by small phytoplankton and zooplankton.
An alternative approach for the development of complex food web structure is to
introduce a generic, organism size-dependent food web structure and associated rate.
This approach obviates, to a large extent, the difficulty of parameter estimations by
using the body size as an independent criterion to estimate most of the parameters. It has
been used in some highly sophisticated ecosystem models coupled with sediments
dynamics and describing the cycling of several biogenic elements in parallel with
several groups of autotrophs and heterotrophs.
There are also a limited number of model studies which incorporate specific details
about the population dynamics of particular zooplankton species to examine the role of
different copepod stages on the functioning of the ecosystems. Such individual-based
population dynamics models are designed to investigate the effect of different
behavioral, bioenergetic, and physiological characteristics on growth and development
of the copepods. For example, this approach has been used to understand how the
copepod Metridia lucens respondes to environmental conditions in the North Atlantic
Ocean (see Models and Functioning of Marine Ecosystems).
4. Physical Controls On Biogeochemical Models
Marine ecosystems with their endogenous time scales are confronted with the multiscale physical environment. Recent modeling efforts recognized the substantial role of
the upper ocean physics on the surface layer biogeochemical processes. A wide range of
space-time variations of the biochemical processes in the water column requires a
thorough representation of the upper layer physics; in particular, currents, temperature,
salinity and vertical mixing. The biological production is modified by physical
processes such as mixing and entrainment of nutrients from below, light penetration,
confinement of phytoplankton in the euphotic zone by stratification, the input of
nutrients by rivers discharges, lateral diffusion of nutrients or plankton by
barotropic/baroclinic instabilities giving birth to filaments, meanders, or eddies (Figure
12). The fact that biological production in the euphotic zone is closely related to the
upper layer physical processes demands coupling of the ecosystem model with a
hydrodynamic model capable of simulating all these features of the surface layer. (see
Nested interdisciplinary three-dimensional models of the marine system).
©Encyclopedia of Life Support Systems (EOLSS)
277
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
Figure 12. Typical vertical structure of the water column (a) in winter, (b) in summer. In
winter, the mixing layer depth is usually deeper than the euphotic layer and thus, the
development of phytoplankton is limited by the light availability. In summer, however,
the mixing layer is reduced to the first upper meters of the water column by the presence
of a seasonal thermocline which inhibits the vertical exchanges. As a result, the surface
layers are depleted in nutrients and the phytoplankton develops below the thermocline at
the base of the euphotic layer using inorganic nutrients mainly produced by the in-situ
regeneration of detritus, or brought by lateral advection/diffusion.
The vertical resolution in the coupled physical-ecosystem models is achieved either by
introducing a few interactive, homogenous layers or a series of vertical levels with
typically 5-10 m spacing. In the case of the multi-layer representation of the water
column, the layers are assumed thoroughly homogeneous and mixing takes place
through the entrainment and diffusion processes across the interfaces between the
layers. The entrainment rate is computed in terms of the upper layer flow characteristics
using a Kraus-Turner type, bulk mixed layer turbulence parameterization. The
performance of the layer-averaged models were found to depend critically on the mixed
layer depth variations. This is because specification from observations is a highly
ambiguous matter unless continuous time series measurements exist. Moreover, the
mixed layer ecosystem models can not resolve the summer biological production once
©Encyclopedia of Life Support Systems (EOLSS)
278
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
the mixed layer generally becomes shallower than the euphotic zone and the production
is confined to subsurface levels below the seasonal thermocline. The layer-type models
thus underestimate the biological production during the periods of shallow mixed layer
structures. On the other hand, in the case of the deeper mixed layer, primary production
tends to be underestimated, because of the averaging of the light limitation function
over the deep mixed layer.
Multi-level representation of the water column, on the other hand, removes these
deficiencies of the layer averaged representation and gives rise to a more realistic
representation of the upper ocean biogeochemical structure. The vertical mixing is
represented by the vertical diffusion process whose depth and time dependent
coefficient is either specified empirically in some functional form in terms of the water
column stratification characteristics, or computed using a turbulence closure model.
They simulate stronger and more active summer production below the seasonal
thermocline (known as the mid-water chlorophyll maximum) in summer months.
Higher vertical resolution in the multilevel models not only provides some nitrate inputs
from the nitracline layer, but also allows a more efficient recycling within the lower part
of the euphotic zone, below the seasonal thermocline. Under these conditions, the
subsurface chlorophyll maximum layer develops at a rate depending on the local
zooplankton grazing and the light intensity. During such periods of stratification, the
turbulence may be strongly inhibited in the upper layer. However, internal waves may
provide some upwards and downwards displacements of plankton and nutrients. The
high frequency variability of the wind stress induces short-term strong vertical mixing
which generates nutrient pulses into the euphotic zone and provides additional
contribution to the biological production process.
More detailed and accurate representation of the temperature stratification and the PAR
within the water column are other advantages of the vertically-resolved models. The
penetration of PAR into seawater depends on the intensity of the surface light field and
on the absorbing properties of water and the substances present in it. The studies
indicated that the traditional, non-spectral models of light attenuation and
photosynthesis can overestimate daily primary production in the water column as much
as 50% and spectral properties of PAR must be considered if higher accuracy is
necessary. Comprehensive models are available that divide PAR into a large number of
wavebands and explicitly include the wavelength-dependent properties of various
parameters which affect light attenuation and photosynthesis. However, the majority of
biological models deal with the non-spectral form of the irradiance.
The seasonal development of a thermocline, in combination with increasing solar
radiation in spring, is accepted as a general mechanistic requisite for the development of
blooms. There are, however, some observations suggesting that the spring bloom can
precede the onset of vertical water column stability, i.e., prior to the formation of the
seasonal thermocline. Indeed, simulations with a vertically-resolved, coupled physicalbiogeochemical model using turbulence closure parameterizations have shown that the
intense vertical mixing delays the occurrence of the bloom until the mixing layer is
shallower or comparable to the euphotic layer (Figure 13). Once this level was
maintained, the bloom then developed rapidly within the upper part of the deep mixed
layer receiving sufficient radiation to promote the photosynthesis, until it became
©Encyclopedia of Life Support Systems (EOLSS)
279
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
limited by the reduction in the surface nutrients. Therefore, vertically resolved models
have an advantage over the bulk mixed layer models. They need not to delay the bloom
until the mixed layer shallows up to a certain depth to provide a sufficiently strong solar
radiation, which is the case in the mixed layer averaged models.
Figure 13. Seasonal evolution of the vertical distribution of the turbulent kinetic energy
(top), the temperature (middle) and the phytoplankton (bottom) in the central basin of
©Encyclopedia of Life Support Systems (EOLSS)
280
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
the Black Sea, simulated by a NPZ-type ecosystem model coupled with a 3D
hydrodynamical field. This picture shows that the phytoplankton bloom occurs before
the establishment of the seasonal thermocline.
In coastal waters, modeling studies focused on the response of phytoplankton growth to
the stratification controlled by local factors (e.g., river input, wind and tidal stirring) and
on the possibility of predicting the timing of the bloom as in the deep waters using the
Sverdrup critical ratio model. The model results have shown that, contrary to the
classical concepts of stratification and blooms in deep pelagic systems, phytoplankton
population growth does not necessarily increase as the surface layer shallows. The
results reveal a more complicated relation than in the deep waters depending on the
surface layer depth and various competing processes, such as (i) the interaction between
turbulent mixing and sinking in the surface layer, (ii) maximization of the average net
growth rate in the surface layer, (iii) minimization of turbulent leakage from the surface
layer, and (iv) minimization of the production occurring below the surface layer.
Whereas persistent stratification greatly increases the probability of a bloom,
semidiurnal periodic stratification does not increase the likelihood of a phytoplankton
bloom over that of a constantly unstratified water column. Furthermore, the surface
layer depth, thickness of the pycnocline, vertical density difference, and tidal current
speed all contribute in equal proportions to producing conditions which promote the
onset of phytoplankton blooms.
The formation of subsurface phytoplankton patches as a result of wind-forced motions
in many fronts is another example of the strong link between the biological and physical
dynamics. Simulations with coupled physical-biological models showed that the
formation and maintenance of subsurface frontal patches was strongly dependent on
water column properties such as frontal characteristics, nutrient gradient, euphotic layer
depth, and historical wind stress. The model results indicated that the phytoplankton
respond to cross frontal nutrient fluxes and wind-driven vertical mixing and advection
on a time scale of the order of about a week. To the first order, the phytoplankton
dynamics at the wind-perturbed front are controlled by vertical fluxes of nutrients.
However, details of the structure and biomass of the subsurface patch are strongly
influenced by the wind direction. Winds aligned with the frontal jet weaken the crossfrontal density gradient, causing a flux of nutrients into the warmer waters of the front,
and eroding the chlorophyll patch through deep vertical mixing. Winds aligned against
the frontal jet, on the other hand, enhance the cross-frontal density gradient and isolate
the subsurface patch from vertical mixing.
5. Inverse Approaches In Biogeochemical Modeling
The capacity of dynamical coupled physical-ecosystem models to simulate
interdisciplinary processes over specific space-time windows, and thus forecast their
evolution over predictable time scales depends upon the availability of relevant
observations. The models require various types of datasets to initialize and continually
update the physical and biogeochemical state of the ocean, to provide relevant
atmospheric and boundary forcing, to parameterize and to properly represent the subgrid scale processes, growth rates and reactions rates. A main reason for the inability of
the ecosystem models to simulate the observed biogeochemical fields is the need to
©Encyclopedia of Life Support Systems (EOLSS)
281
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
specify values for a large number of external parameters (e.g., half saturation constants,
zooplankton grazing, specific growth and mortality rates). Most of these biological
parameters also depend on the complexity of the model structure. They are generally
poorly measured and known, differ greatly for different plankton species groups, and
from region to region as well. In practice, these parameters are adjusted for each
specific modeled ecosystem in order to obtain the "best" consistency between computed
and observed state variables. The calibration may be carried out by trial and error
procedure (i.e., subjectively) or by means of mathematical methods which yield the
"best" fit (i.e., objectively) for selected parameters of the model. The classical approach
is to tune each of these parameters subjectively until the model reproduces the observed
data reasonably well for each ecosystem. A moderately complex model usually requires
a hundred individual experiments to generate a parameter set that provides a successful
and realistic simulation. The task of determining the parameter set can thus quickly
become tedious even for a simple model configuration. In this approach, the calibration
of 3D ecosystem models is even more difficult because the calibration of the parameters
are based on only several observed vertical profiles, and may not be representative of
the entire basin.
The accelerating rate of development in the coupled physical-ecosystem modeling
studies, together with the recent advances in measurement systems and data assimilation
research efforts, gave a major thrust to the data assimilation studies. The data
assimilation strategy in ecosystem models helps to calibrate the models, to determine
the simplest possible model structure that is consistent with the data, to identify
limitations of existing model formulations and to give insight in to how to remedy this
problem, and to determine the most critical data types constraining the models. The
essence of the data assimilation procedure is to minimize the misfit between the data
and the corresponding control variables (e.g., model parameters, initial/boundary
conditions, forcing functions) under certain constraints satisfied by the model. The
implementation of data assimilation procedures requires a dynamical description of the
behavior of the ecosystem (forward model), observations that provide information on
the ecosystem dynamics (data), and a technique used for incorporating the observations
into the model dynamics (inverse model) (Figure 6). The most common inverse
techniques adopted in ecosystem modeling studies are based on non-linear optimization
algorithms such as adjoint variational methods, simulated annealing, or regression
methods.
Due to the large number of parameters often needed to be specified in the models, data
assimilation procedures becomes particularly useful if some redundant parameters are
removed from the model, prior to data assimilation. For instance, the nondimensionalization of the model equations helps to estimate the first order processes and
their parameters, and allows a more optimal and efficient solution to the inverse
problem. On the other hand, assimilation of a limited number of data sets can only
constrain a few linear combinations of the parameters, and therefore is generally not
sufficient to provide a realistic representation of the entire ecosystem structure. Also, it
was shown that assimilating surface chlorophyll concentration alone from satellite data
was sufficient to constrain primary production within the photic layer, but inadequate to
constrain its regeneration in surface water and its export towards deep water.
Simultaneous assimilation of surface chlorophyll, temperature and nitrate concentration
©Encyclopedia of Life Support Systems (EOLSS)
282
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
data was found to give a significant improvement over model simulation in terms of a
more realistic trophic system within the euphotic layer.
The application of a data assimilation algorithm to a specific ecosystem model does not
always guarantee the realism of a simulation, unless the model is appropriately set up to
describe all the basic biological processes. The most important constraint is the number
of independent data types which are used in parameter estimations. When a specific data
set from the Ocean Weather Station Papa in the Gulf of Alaska was applied to three
different ecosystem models of varying complexity, the data were able to resolve up to
10 independent model parameters. This number was, however, less than the number of
unknown parameters even for the simplest NPZ-type model. This implies that a more
complicated ecosystem structure, such as the 7-compartment configuration (i.e., the
phytoplankton and the zooplankton biomass, nitrate, ammonium, dissolved and
particulate detritus and bacteria), cannot be justified, unless additional measurements
are made available. A similar study designed to reproduce the Bermuda Atlantic Time
Series (BATS) data using a 7 compartment model also arrived at similar conclusions.
They indicated that some of the model parameters couldn't be estimated independently
and that the seven-compartment model used was not appropriate for simulating the
BATS ecosystem. Testing various more simplified versions of this model, its four
compartment reduced form with only 11 parameters to be estimated was able to
describe more successfully the annual cycle in BATS data as compared with the original
model.
The assimilation of both physical and biological data in 3D coupled models is relatively
new. A challenging test of this approach has been applied to the Gulf Stream frontal
region characterized by highly complex mesoscale physical and biogeochemical fields.
This study has shown the importance of the submesoscale "event scale" processes in
physical-biological interactions. It further demonstrated that the high resolution 4-D
field estimation through observations is necessary to capture the essential physicalbiological interactions inferred by the data. It has been found that jet meandering and
winds cause only slight enhancement in phytoplankton biomass at the Gulf stream front
while ring-stream interaction events are an important process for generating
phytoplankton patches at the front (see Synoptic/Mesoscale Processes).
6. Regional Applications of Coupled Physical-Biogeochemical Models
The three dimensional, regional or global scale coupled physical-ecosystem models are
the only possible approach to interpolate observations from different times and locations
so that a dynamically-consistent picture emerges. The regional models are developed as
a compromise between the structural complexity required to accurately reproduce
detailed local measurements, and the conceptual simplicity required for large scale
applications and interpretations. There is also a practical constraint due to the
computational facilities that impose limits on the number of explicitly resolved
ecosystem variables and/or vertical and horizontal resolution of the model. These
limitations and restrictions put a major constraint on the widespread use of the high
resolution, state-of-the-art ocean circulation-biogeochemical models. The applicability
of eddy-resolving coupled models for basin wide and ocean scale applications studying
©Encyclopedia of Life Support Systems (EOLSS)
283
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
annual structure and interannual variability of the system is therefore quite limited and
began only within the last few years.
6.1. Models for Open Ocean Studies and Basin Scale Applications
A common feature of the current three dimensional interdisciplinary models simulating
the basin-scale variations of the physical and biogeochemical environment is the
relative simplicity in the treatment of both plankton and circulation dynamics. Due to
computational limitations, and feasibility and reliability constraints, 3D ecosystem
models generally consider the biogeochemical cycle of only one type of biogenic
element (mostly nitrogen) together with a limited number of aggregate variables (Figure
9). Simplifications are introduced to the circulation models either by invoking the
primitive equation dynamics with multi-layer configuration or the quasi-geostrophic
dynamics using a vertically better-resolved water column. The first approach considers
a more simplified and crude vertical structure in the form of several homogeneous,
interactive layers. However, since it uses fully primitive equation dynamics, it is valid
for mesoscale dominated circulation problems. On the other hand, the other approach is
only applicable for open ocean dynamics where non-linearity is not important. Its
relative computational efficiency, however, allows the introduction of a high number of
vertical levels to resolve the water column. Both of these approaches therefore offer
computational efficiency for the long term (decadal) integrations. The choice depends
on the nature of the problem under consideration. Some problems (generally those
occurring on coastal and shelf regions) are inherently non-linear and require the
introduction of a high vertical resolution; they therefore dictate implementation of
vertically-resolved, multi-level primitive equation circulation models coupled with a
biogeochemical model of similar vertical configuration. This type of model is relatively
more expensive, and time consuming.
An application of the layer-type coupled model was made to investigate the seasonal
variations of the surface circulation and biological production in the Arabian Sea. The
simple physical and ecosystem structures were defined by introducing phytoplankton,
zooplankton, nitrate and detritus compartments using a two layer system over an
infinitely deep and motionless water body. The mixed layer dynamics were embedded
inside the first layer to compute temporal changes of the mixed layer thickness
depending on the entrainment, detrainment and vertical mixing. The particular objective
of this study was to understand the relative contributions of upwelling, vertical mixing,
horizontal advection and biological remineralisation in generating nutrient sources for
the blooms, the role of light limitation in controlling the onset of blooms, and the
relative contributions of predation and nutrient depletion for their decays. For the
climatologically mean conditions of the Arabian Sea, three different types of bloom
formation have been identified in response to physical processes of upwelling,
detrainment and entrainment. Upwelling blooms were strong, long-lasting events that
continued as long as the upwelling persisted. Detrainment blooms were found to be
short-lived, intense events that developed when the mixed layer shoaled abruptly,
thereby increasing the depth-averaged light intensity available for phytoplankton
growth. They occurred mostly in spring during the mixed layer shallowing phase. In
contrast, entrainment blooms were weaker because entrainment steadily thickens the
mixed layer, which in turn decreases the depth-averaged light intensity.
©Encyclopedia of Life Support Systems (EOLSS)
284
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
Alternatively, several studies adopted the vertically-resolved, multi-level models with
primitive equation dynamics. In the case of basin-scale applications, like in the Northern
Pacific or Atlantic, a compromise for a rather economical and practical use of the
models was made by having a relatively coarser horizontal grid resolution. The
emphasis was then given to the larger scale features of the circulation by ignoring the
role of mesoscale dynamics on the biogeochemistry of the basin. In some of these
models, a further sophistication has been introduced by implementing a higher level
Mellor-Yamada type turbulence closure parameterization for vertical mixing. An
application of such a sophisticated hydrodynamical model coupled with a sixcompartment ecosystem model involving the phytoplankton, zooplankton, nitrate,
ammonium, particulate and dissolved detritus compartments has been given for the
North Pacific, extending from 122oE to 72oW and from 23oS to 63oN. It used a 2ox2o
grid interval with 45 vertical levels compressed toward the surface to properly resolve
the seasonal variation of the mixed layer. The model was found to successfully simulate
various observed features of the circulation and ecosystem in different domains of the
North Pacific. But it also had some deficiencies such as an underestimation of surface
chlorophyll concentrations in the subpolar regions and an overestimation in the
subtropics and equatorial region.
A limited number of three-dimensional, coupled models resolving mesoscale circulation
features and implementing multi-compartmental plankton groups and multi-nutrient
limitations are also available for smaller basins, such as the Black Sea (Figure 10),
Baltic Sea, the North Sea (Figure 4) and for various sub-domains of the oceanic basins.
Such modeling studies suggested that the primary production may be increased by at
least 30 percent through the mesoscale dynamics. One of the two main mechanisms
invoked to explain such enhancement of the biological productivity is the upward
vertical circulation associated with eddy dynamics. This process can transfer nutrients to
the enlightened surface layer in nutrient limited conditions. The second mechanism is
related to the eddy-induced mixed layer shoaling. In light limited conditions, it can
locally increase the mean exposure time of the photosynthetic organisms to solar
radiation. In both cases, mesoscale dynamics can play a role in enhancing primary
production by reducing the dominant stress.
The contribution of mesoscale eddies to open ocean primary productivity has been little
explored simply because the open ocean observational programs are limited in number,
and usually provide biweekly to monthly measurements, and therefore are not able to
capture sporadic processes on a weekly time scale. For example, the subtropical gyre of
the North Atlantic has long been thought to be a mid-latitude "ocean desert" due to its
downwelling vertical motion and lack of nutrient supply into its surface waters. But it
becomes evident that it sustains some degree of primary productivity due to eddyinduced nutrient transport on horizontal scales of several tens to hundreds of kilometers,
and over a time scale of several weeks. Indeed, the studies in the Sargasso Sea of this
subtropical gyre have described the importance of eddy-induced upwelling mechanisms
associated with a passage of strong cyclonic eddies to meet the nutrient demand for new
production. The nitrate flux due to eddy pumping has been estimated to be an order of
magnitude greater than the diapycnal diffusive flux and transport due to isopycnal
transport. The mechanism of eddy upwelling has been explained by the vertical
perturbation of density surfaces, as a result of formation, evolution and destruction of
©Encyclopedia of Life Support Systems (EOLSS)
285
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
mesoscale features. Shoaling density surfaces lift nutrients into the euphotic zone,
which are then rapidly utilized by the biota. Deepening density surfaces serves to push
nutrient-depleted water out of the illuminated surface layer. The asymmetric light field
thus rectifies vertical displacements in both directions into a net upward transport of
nutrients, which is balanced by a net commensurate flux of sinking particulate material.
This idea has been put into a larger context to investigate the contribution of eddyinduced nutrient supply to the entire North Atlantic Ocean, by means of an eddyresolving coupled general circulation-ecosystem model including altimeter data
assimilation, to achieve a more realistic representation of the real ocean eddy field. The
comparison of the simulations performed with and without mesoscale eddies in the
circulation field has indicated that eddy activity was able to increase the total nutrient
input into the euphotic zone by about thirty percent in the subtropic and mid-latitudes.
The same conclusion has been found in the North Equatorial Current (NEC) system
which is a region of strong mesoscale activity (e.g., strong currents, eddies, fronts).
Modeling studies have shown that meanders and eddies formed by instabilities cause
upward motion of water transporting nitrogen into the euphotic zone. This upwelled
nitrogen induces an enhancement of about 30% of the primary production and of the
phytoplankton biomass primarily during the period of eddy-eddy interactions at the
edge of eddies. In non-upwelling regions or when upwelling events are seldom present,
the horizontal advection of particulate organic matter is shown to support the
production. Moreover, the model studies indicated doubling of the primary production
in the jet zone when the eddies are resolved.
In addition to the transfer of nutrients to the upper enlightened layer, mesoscale
dynamics may also induce the shoaling of the mixed layer and thus enhance the
phytoplankton development by locally increasing the mean exposure time of the
photosynthetic organisms to solar radiation. For instance, modeling studies applied to
the north-western Mediterranean Sea have suggested that eddies, which are developed at
the periphery of the dense water patch, participate very actively in the restratification of
the surface waters following deep convection. Indeed, mesoscale upward motions cause
shoaling of the mixed layer in the trough of the meanders, which then acquires higher
rates of solar radiation and thus increase primary productivity. Phytoplankton biomass is
then subducted from these sources towards the crest of the meanders. The center of the
patch, on the other hand, is the least productive region where strong and deep vertical
mixing is shown to inhibit photosynthesis (see Synoptic/Mesoscale Processes).
6.2. Models for Coastal Regions and Shelf Seas
Recent decades have seen a degradation of the environmental quality in various basins
of the world's oceans caused by eutrophication and pollution problems resulting from
increased anthropogenic inputs of terrestrial origin (e.g., mineralized nutrients, organic
and inorganic pollutants). Such problems particularly affected the coastal zone located
at the interface between the continents and the oceans and thus exposed to increasing
socio-economic pressures, led to dramatic alterations of the structure and functioning of
the ecosystem. In particular, the semi-enclosed seas and enclosed inland water bodies
were the regions most sensitive to natural and anthropogenic perturbations.
©Encyclopedia of Life Support Systems (EOLSS)
286
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
Since the early 1980s, satellite-derived observations of the ocean surface temperature
(Advanced Very High Resolution Radiometer, AVHRR), pigment distributions from
CZCS (Coastal Zone Color Scanner) and SeaWiFS (Sea-viewing Wide Field-of-view
Sensor) provide evidence of enhanced biological production in coastal regions.
Increased production near-coastal boundaries are often related to eutrophication
phenomena due to high anthropogenic load of nutrients and organic materials from
rivers. Coastal waters thus tend to present very different patterns of phytoplankton
biomass variability from the open ocean waters. The connections between physical
variability and phytoplankton bloom dynamics are less well established for these
shallow systems. Predictions of biological responses to physical variability in these
environments are inherently difficult because the recurrent seasonal patterns of mixing
are complicated by aperiodic fluctuations in river discharges and the high frequency
components of tidal variability. In addition, shallow coastal waters have high
concentrations of suspended particulate matter that originate from resuspension of bed
sediments as well as riverine inputs of terrigeneous material. Light attenuation by
suspended particulate materials acts as a major control on photosynthesis. On the other
hand, nutrient availability is relatively high in coastal waters because of terrigeneous
inputs as well as rapid recycling rates, both in the water column and the benthic
boundary layer. Moreover, benthic-pelagic coupling is also a controlling factor for
changes in the phytoplankton population in coastal waters. Offshore flowing jets,
mesoscale eddy fields embedded within the slowly moving current systems, meandering
currents as well as other transport mechanisms due to wind-induced upwellings,
episodic mixing of deep water by winter convection, and diapycnal diffusion can all
lead to highly complex biological structures in coastal waters. Therefore, shallow shelf
waters are less predictable and have more complex bloom dynamics than the open
ocean.
Because the eutrophication-induced biological production has severe consequences for
local tourism, fishery and economy, ecosystem modeling studies devoted to coastal
regions and shelf seas have received a particularly great interest. Indeed, ecosystem
models are necessary to assess the ecosystem's vulnerability to contaminants of
anthropogenic origin, and to calculate the transfer and accumulation of toxic substances
from one level of the food web to the next. Besides, the study of the eutrophication
problems and their impacts on the basin scale ecosystem cannot be made without
considering the physical processes leading to the mixing and transport of pollutants and
biogeochemical constituents discharged by the rivers. The studies of the coastal threedimensional coupled physical-biogeochemical models have been approached either by
exploring the dynamics of certain physical mechanisms governing the regional
ecosystem, or by understanding the details of the ecosystem's functioning under given
physical conditions.
Several ecosystem models have been developed to study the eutrophication problems in
various heavily polluted regions such as the western part of the Seto Inland Sea, the
Italian coast of the northern Adriatic receiving the inputs from the River Pô, the Baltic
Sea, the North Sea, and the Black Sea. For instance, the North Sea has been the subject
of a series of modeling attempts during the last two decades to hindcast the long-term
transformation of the North Sea ecosystem with the gradual build up of eutrophication
since 1950s. They reveal that the increase in primary production associated with
©Encyclopedia of Life Support Systems (EOLSS)
287
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
eutrophication was not simply proportional to the increase in the anthropogenic nutrient
load. It was shown that, once the ecosystem is perturbed, it shows a complicated
response to many factors such as anthropogenic nutrient load, zooplankton grazing,
temporally varying nutrient limitations, and domination of different phytoplankton
groups at different seasons.
A study of the sensitivity of the Black Sea's ecosystem dynamics to the macroscale and
mesoscale dynamics has been made by means of a three-dimensional, mesoscale
resolution hydrodynamical model coupled with different ecosystem models describing a
simple nitrogen based structure defined by the phytoplankton, zooplankton, nitrate,
ammonium, pelagic and benthic detritus (Figures 9 and 10). A more sophisticated
version of the biological model has also been developed to test the roles of phosphate
and silicon in the growth of three different phytoplankton groups. This model has
further involved two zooplankton groups and the microbial loop. The results illustrated
a highly complex spatial variability of the phytoplankton annual cycle imparted by the
horizontal and vertical variations of the physical and chemical properties of the water
column. In particular, the seasonal variability of the north-western shelf circulation
induced by the seasonal variations in the Danube discharges and the wind stress
intensity have been found to have a major impact on the primary production repartition
of the area. In agreement with observations, the results indicated a rapid and efficient
recycling of particulate organic matter in the oxygenated layer of the water column.
More than 90% of the particulate organic nitrogen produced in the euphotic zone was
shown to be recycled in the upper 100 m of the water column and about 79% in the
euphotic layer. Comparison of the seasonally-averaged phytoplankton biomass
computed by the models with the satellite-derived estimates of the seasonal surface
chlorophyll field suggested the best agreement under the simulations with the mesoscale
resolution of the hydrodynamical models, and with the refined ecosystem model having
a better representation of the biological processes.
The continental margin zones are well known as regions of enhanced biological
production, partly fuelled by nutrients brought by river discharges and boosted in many
regions by upwelled nutrients from the deep sea. The shelf-edge processes influence the
cross-shelf exchanges of biogenic nutrients, and plankton, as well as larvae transport.
The continental shelf-deep sea interactions are therefore of crucial importance, and the
coupled hydrodynamics-ecosystem models are needed to study in detail the flow
structure over the steep continental slope, the shelf edge processes and the transport of
biogeochemical constituents along and across the continental slope. A study in the Gulf
of Lion region of the Western Mediterranean indicated the complexity and temporal
variability of the mechanisms governing the exchanges across the continental margin
zone. In early winter, new production is very low, and sediment release represents a
major source of nitrate to the margin, which is not locally consumed biologically, but
carried away to the open sea by deep offshore flow along the bottom. Later in the
winter, as the deep mixing intensifies in the region, the surface waters of the open sea
are enriched and advected onto the margin. During the spring bloom event, new
production is more active on the shelf than offshore due to an additional contribution of
the Rhone River, leading to stronger depletion of nitrate in the margin as compared with
the open sea. For this period, the new production is sustained in the coastal zone by both
the Rhone River discharge and oceanic input.
©Encyclopedia of Life Support Systems (EOLSS)
288
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
Modeling studies have examined the influence of the freshwater and nutrient discharges
on the formation of inner shelf frontal structures and associated alongshore and crossshelf current systems. Alongshore currents developed in response to cross-shelf density
gradients were strong, and typically about 100 ms-1 at the surface. The cross-shelf
circulation is characterized by a surface-intensified offshore current near the coast and
multiple cells inside the frontal zone. In addition to being recycled through biological
processes, the nutrients in the frontal zone were shown to be supported physically by
two sources: (i) horizontal advection from the river discharge, and (ii) advectioninduced upward nutrient flux outside the frontal zone. Multiple cells of the frontal zone
acted like retention zones that recirculated the nutrients in the vertical and also restricted
marine organisms from cross-frontal exchanges. The physical processes thus caused the
formation of a large concentration of bottom-rich nutrients in the weak current region
within the frontal zone. Cross-shelf distributions of the biological fields were further
significantly modified by upwelling and downwelling favorable winds depending on
their amplitude and duration. The downwelling-favorable winds tended to enhance the
vertical mixing and caused a more vertically uniform pattern of nutrients, phytoplankton
and zooplankton. The upwelling-favorable winds, on the other hand, tended to speed up
the offshore migration of phytoplankton and nutrients in the upper layer and the onshore
advection in the lower layer. This process increased the vertical gradients at the expense
of reduced horizontal variations within the shelf.
6.3. Models Embedded in Global Ocean Carbon Cycle Studies
A novel understanding of marine productivity is necessary to appraise the role of marine
ecosystems in global changes. The present ocean carbon cycle models consider an
extremely simplified ecosystem configuration for the oceanic uptake of atmospheric
CO2. In the past, the majority of these models assumed no variations in the biological
pump over the whole ocean, and its role was parameterized indirectly by forcing the
model output towards observed surface nutrient and other biogeochemical observations.
However, this technique predicted unrealistically high nutrient accumulation in the
equatorial region, and suggested that a considerable amount of the new production flux
to the ocean interior must be carried by dissolved organic matter. Indeed, it was thought
that if the dissolved organic matter has a lifetime of years and decades, its advection
would carry nutrients away from the regions of strong upwelling and thereby avoid
nutrient trapping. Furthermore, specification of one globally uniform phosphorus uptake
rate for the annual net effect of biological processes generally provided a better
agreement between the simulated and observed seasonal amplitude of CO2 partial
pressure (pCO2). However, the model yielded an almost complete depletion of
phosphate in the surface waters of the subtropical gyres, due to deficiency of the
circulation field in bringing sufficient phosphate to the surface layer. The enhanced
incorporation of phosphate into particulate organic matter and its export from the
surface layer in upwelling areas resulted in a strong PO4 depletion that was only
supported by lateral PO4 advection, as in the subtropical gyres. In a general sense, the
coupling of the global scale carbon cycle models with sophisticated biological models
allows a better representation of the pCO2 repartition. These studies have revealed that,
even under the apparent simplicity of the plankton models, they reproduced regional
differences in seasonal oceanic pCO2 as expected, and generally model results were in
qualitatively good agreement with the observations. Moreover, zooplankton dynamics
©Encyclopedia of Life Support Systems (EOLSS)
289
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
was shown to cause an important part of the regional variations. The interplay between
the mixed layer depth (which influenced the local phytoplankton growth rate) and
reduced zooplankton abundance allowed the development of strong phytoplankton
blooms, while, in other regions, the North Pacific for instance, relatively shallow mixed
layer depths supported a higher zooplankton standing stock. The seasonal variations of
the phytoplankton populations were followed by immediate responses in the
zooplankton stocks. This study confirmed that the seasonal effect of marine biology
might be captured by simple plankton models. However, simple ecosystem models with
globally uniform parameter settings seemed not to be able to successfully simulate the
observed different seasonal cycles of primary production and nutrients between
different regions. More complex ecosystem models must be developed if one wishes to
parameterize the difference between ecosystem dynamics at high latitudes of the
Southern and Northern Hemispheres. In this case, simple hydrodynamic models, usually
global box models, have to be used.
A more complex global carbon and nitrogen cycling model represented major oceanic
basins in the form of interactive boxes (Figure 14). Each of these boxes involved three
homogeneous layers representing surface, intermediate and deep water masses. This
type of model gives a detailed description of the seasonal cycles of the surface mixed
layer ecosystem and specifies different parameters for different regions. The biological
model included the phytoplankton and the zooplankton biomass, the organic and
inorganic parts of detritus, dissolved inorganic nitrogen, oceanic total inorganic carbon,
total alkalinity and atmospheric carbon dioxide. The model also included formation of
particulate organic matter and CaCO3 in the upper ocean, their export to depth,
degradation of particulate organic matter and dissolution of CaCO3 in the deep ocean,
and exchange of CO2 between the ocean and atmosphere. The model, primarily
considered the processes of gas exchange between ocean and atmosphere, advective
transport, entrainment and detrainment at the lower boundary of the mixed layer,
primary production of phytoplankton, phytoplankton and zooplankton mortality, detritus
remineralisation, gravitational sinking of particles, degradation of detritus, and
inorganic carbon chemistry processes.
Figure 14. Global box model simulating nitrogen and carbon cycling in the ocean
including the biological pump.
©Encyclopedia of Life Support Systems (EOLSS)
290
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
The model includes a detailed description of the seasonal cycles of the oceanic upper
mixed layer (UML) ecosystem and a refined parameterization of the role of seasonal
variations of marine biota in atmospheric pCO2 regulation. The difference between
ecosystem dynamics at high latitudes of the Southern and Northern Hemispheres is also
parameterized. The ocean is represented as a stack of three homogeneous outcropped
boxes (deep, intermediate and surface water masses) in each hemisphere. The model
prognostic variables include phytoplankton P, zooplankton Z, the organic part of
detritus D, dissolved inorganic nitrogen N, inorganic carbon C, the inorganic part of
detritus H, total alkalinity A, and atmospheric carbon dioxide Ca. The principal
processes considered by the model are: gas exchange between ocean and atmosphere,
advective transport, entrainment and detrainment at the lower boundary of the UM,
primary production, phytoplankton and zooplankton mortality, grazing of zooplankton,
detritus remineralisation, gravitational sinking of particles, degradation of the detritus
hard parts, and inorganic carbon chemistry processes. (Reprinted from Global
Biogeochemical Cycles, 14 (4), Popova, E.E., V. A. Ryabchenko, M.J.R. Fasham, page
479, copyright 2000, with permission from American Geophysical Union).
Also, there are some attempts to simulate the ecosystem characteristics of various
contrasting regions of the global oceans such as the Southern Ocean, the North Pacific,
and the North Atlantic by implementing a globally standard biogeochemical model
structure. One such a model described the global ecosystem by ammonium, nitrate,
multiple-size classes of phytoplankton, zooplankton, bacteria, detritus and dissolved
organic nitrogen with a variable chlorophyll-to-biomass (nitrogen) ratio. It was shown
that the inclusion of multiple size classes of phytoplankton and detritus, as well as the
use of a variable chlorophyll-to-nitrogen ratio enabled the model to predict important
biological differences not only temporally within a site, but also spatially between the
different regions.
Glossary
Aggregation:
Assimilation of
data:
Autotrophs:
Biota:
Calibration:
Ciliates:
Compensation
depth:
Unification of system components that are homogeneous in some
properties into blocks, each being a new component with
properties defined by the aggregation laws.
Incorporation of information from observations into the model, in
order to improve model predictions.
Organisms able to fix carbon from inorganic to organic form using
light or chemical energy.
Plant and animal life of a region.
Adjustment of a number of model parameters in order to obtain the
best accordance between computed and observed state variables.
(microns to several hundreds of microns). Present in all marine
regions, they can be extremely important. They use cilia for
locomotion, and eventually food capture. They feed on small
phytoflagellates and zooflagellates, small diatoms and bacteria.
The depth at which the photosynthetic system is at the
compensation point. In field work, the compensation depth can be
approximated from the depth of 1% of the surface radiation.
Compensation depths may vary from a few meters in turbid coastal
©Encyclopedia of Life Support Systems (EOLSS)
291
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
Compensation
point
Copepod:
Continental shelf:
Critical depth:
Ctenophores:
Denitrification:
Detrainment:
Diapycnal:
Diatoms:
Dinoflagellates:
Dissolved Organic
Matter:
Downwelling:
DOM:
Eddies:
Entrainment:
Euphotic zone:
waters to over 150 m in some tropical seas.
The photosynthetic system is at the compensation point when the
daily average net production (i.e., the gross photosynthesis
production minus respiration) is zero.
One of the predominant forms of zooplankton.
Gradually sloping submerged extension of a continent, composed
of granitic rock overlain by sediments. Has features similar to the
edge of the nearby continent.
The depth at which the integral net production (the gross
photosynthetic production - respiration) is zero.
Carnivores species (with a size ranging from centimeters to
meters) which feed on fish eggs and fish larvae and compete with
young fish for smaller zooplankton prey such as copepods.
Process in which nitrate is reduced into nitrite and nitrogen gas.
It is the opposite process of the entrainment. It represents mixing
of upper layer (less dense) waters into the lower layer (more dense)
across the interface between two water parcels with different
densities. Detrainment results in changes in the properties of the
lower layer, whereas no changes take place on the characteristics
of the upper layer.
Processes occurring across isopycnal surfaces.
Unicellular planktonic algae with cell size ranging from about 2µm
to over 1000 µm (some species form larger chains or other forms
of aggregates). All species have an external skeleton made of silica
and therefore diatoms need silicon for their growth.
autotrophic dinoflagellates unicellular algae with a size ranging
from a few tens of microns to fractions of millimeters and existing
mostly singly (only a few species form chains), with flagella which
give them motility not only vertically (which can also be assured
by
buoyancy
control
as
in
diatoms)
but
also
horizontally.heterotrophic dinoflagellates (up to 1mm or more for
Noctiluca, often observed in dense swarms in coastal areas) feed
on bacteria, diatoms, other flagellates, ciliates and for Noctiluca,
fish eggs.
The division of the organic matter to its dissolved part and
particulate part is arbitrary. Usually, this separation can be made
by filtering over 0.5 µm filters. The dissolved substances are not
retained by the filter.
Downward movement of surface water.
Dissolved Organic Matter.
Patches of turbulent flow.
One-directional movement of water through a boundary between
two water bodies with different densities. The denser water mixes
into the less dense water. This process results in changes in the
properties of the less dense fluid, while the denser fluid keeps its
original properties.
Region where the net rate of photosynthesis is positive. The lower
limit of the euphotic zone is usually assimilated to the
©Encyclopedia of Life Support Systems (EOLSS)
292
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
Eutrophication:
Flagellates:
Heterotrophs:
Initialization:
Isopycnal:
Kraus-Turner bulk
mixed layer
dynamics:
Labile dissolved
organic matter:
Medusae:
Mesoplankton:
Microplankton:
Mixing layer:
Mixotroph:
Nanoplankton:
New production:
Nitracline:
Nitrification:
Nutricline:
Nutrient:
NPZ-type model:
compensation depth.
Anthropogenic enrichment of natural waters with inorganic
nutrients (ammonium, nitrate, phosphate) by which phytoplankton
growth is stimulated. Eutrophication leads to increased biomass,
decomposition, and in the worst cases, results in oxygen depletion
and mass mortality.
Planktonic species. Flagellates may be autotrophs, heterotrophs, or
mixotrophs, with some other species being parasitic or symbiotic.
Organisms unable to fix carbon from the inorganic to organic
form, dependenton already existing biomass to develop.
Introduction of initial conditions for each state variable of the
model.
Surfaces characterized by a constant density
It is a method of calculating entrainment and detrainment
velocities across the base of the surface mixed layer and the layer
immediately below. Depending on the rate of entrainment
(detrainment), the mixed layer undergoes deepening (shallowing).
Part of the dissolved organic matter which is rapidly consumed by
bacteria and so is characterized by a fast turnover.
are common in open sea and coastal waters. They are carnivorous
and have a size ranging from a few millimeters to a few meters,
possibly with tentacles extending down to several tens of meters.
Planktonic organisms of intermediate body size between 0.2 and
20 mm.
Small-sized (between 20 and 200 µm) planktonic organisms.
The zone being actively mixed from the surface at a given time and
corresponding generally to the zone in which there is strong
turbulence directly driven by surface forcing.
Organism which has the capability of both autotrophic and
heterotrophic functions, presumably switching from one to the
other according to food availability.
Planktonic bacteria and algae between 2 and 20 µm in size.
Part of the primary production based on newly incorporated
inorganic nutrients imported into the euphotic layer from the
continent (e.g., by the rivers) and/or from the deep waters (e.g., by
winter mixing/upwellings).
Layer where the nitrate concentrations present an important
vertical gradient.
Process in which ammonium is oxidized into nitrite and nitrate.
Layer where the nutrient concentrations presents an important
vertical gradient.
Any nourishing substance. As used in the Marine Sciences
literature, it usually refers to the important and commonly
measured elements needed for the growth of plants phosphorus,
nitrogen and silicon.
Ecosystem model defined by a nitrogen cycle described by 5 state
variables which are the phytoplankton and the zooplankton
biomass, nitrate, ammonium and detritus.
©Encyclopedia of Life Support Systems (EOLSS)
293
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
Photosyntetically Available Radiation
PAR:
Particulate Organic The fraction of the total organic matter that can be filtered or
centrifuged from seawater. Hence the difference between
Matter:
particulate and dissolved organic matter is arbitrary. Usually, a
substance is considered as particulate if it is retained by a 0.5 µm
filter.
CO2 partial pressure.
pCO2:
The pelagic environment extends from the sea surface to depths
Pelagic:
but has no relation with the bottom layer. The pelagic environment
supports two basic types of marine organisms the plankton,
which is, to a large extent, transported by currents in the sea and
the nekton consisting of free-swimming animals more or less
independent of water movements. Phytoplankton refers to vegetal
planktonic organisms and zooplankton to animal ones.
The process by which autotrophs bind light energy into the
Photosynthesis:
chemical bonds of food with the aid of chlorophyll and other
substances. The process uses carbon dioxide and water as raw
materials and yields glucose and oxygen.
Unicellular algae with a size ranging from a few microns to a few
Phytoflagellates:
tens of microns and existing mostly singly (only a few species
form chains) with flagella which give them motility not only
vertically (which can also be assured by buoyancy control as in
diatoms) but also horizontally.
Particulate Organic Matter.
POM:
Primary producers: Autotrophs.
Fixation of carbon from inorganic to organic form, whether by
Primary
photo- or chemoautotrophy.
production:
by a population is its total elaboration of new flesh (or plant
Production:
material), irrespective of whether the individuals containing that
production survive to the end of the time interval over which
production is calculated
Layer where the density presents an important vertical gradient.
Pycnocline:
The essential nutrients and carbon dioxide are taken up (and
Redfield ratios:
released, when living matter is metabolized) in approximately
constant ratios to each other and the oxygen produced (or
consumed). The ratios calculated by Redfield are 1P 16N 106C
138O2, where P,N,C, and O2 stand for phosphorus, nitrogen,
carbon, and oxygen concentrations.
They approximate the upper parts of the water column by a
Reduced-gravity,
number of homogeneous layers. The deeper part of the water
layered models:
column is assumed very deep and motionless. The advantage of
this assumption is to simplify the solution of the dynamical
equations governing the motion.
Part of the organic matter which is not remineralized in the water
Refractory
column due to its very slow turnover rate, and sinks to the bottom.
dissolved organic
matter:
Part of the primary production based on inorganic nutrients
Regenerated
recycled in the euphotic layer by bacteria remineralisation or by
production:
©Encyclopedia of Life Support Systems (EOLSS)
294
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
-
TO ACCESS ALL THE 37 PAGES OF THIS CHAPTER,
Visit: http://www.eolss.net/Eolss-sampleAllChapter.aspx
Bibliography
Baretta J.W., Ebenhoh W. and Ruardij P. (1995). The European Regional Seas Ecosystem Model, a
complex marine ecosystem model. Neth. J. Sea Res. 33, 233-246.
Carlotti F. and Radach G. (1996) Seasonal dynamics of phytoplankton and Calanus finmarcicus in the
North Sea as revealed by a coupled one-dimensional model. Limnology and Oceanography 41, 522-539.
Gregoire M., Beckers J.M., Nihoul J. C. J. and Stanev E. (1998). Reconnaissance of the main Black Sea's
ecohydrodynamics by means of a 3D interdisciplinary model. Journal of Marine Systems 16, 85-105.
[This paper describes the results of a 3D coupled ecosystem-hydrodynamical model applied to the Black
Sea. Focus is put on the influence of macroscale physical processes on the space-time distribution of the
primary and secondary production].
Jorgensen S. (1988). Fundamentals of ecological modeling, 391pp. Elsevier. [This presents all the
necessary steps in the development of a mathematical model: its conceptualization, parameterization,
calibration, initialization and validation].
Lancelot C. and Billen G. (1985). Carbon-nitrogen relationships in nutrient metabolism of coastal marine
ecosystems. Advances in Aquatic Microbiology 3, 263-321.
Leonard, C.L., McClain C., Murtugudde R., Hofmann E., Harding L.W.Jr. (1999) An iron-based
ecosystem model of the central equatorial Pacific. Journal of Geophysical Research 104, 1325-1341.
©Encyclopedia of Life Support Systems (EOLSS)
295
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
Levy M., Memery L., Madec G. (1999) The onset of the spring bloom in the MEDOC area: mesoscale
spatial variability. Deep-Sea Research I 46, 1137-1160.
McCreary, J.P. Jr., K. Kohler, R. R. Hood, D. B. Olson (1996) A four-compartment ecosystem model of
biological activity in the Arabian Sea. Progress in Oceanography 37, 193-240.
McGillicuddy D.J. Jr., Robinson A., Siegel D., Jannasch J., Johnson R., Dickey T., McNeil J., Michaels
A.F., Knap A.H. (1998) Influence of mesoscale eddies on new production in the Sargasso Sea. Nature
394, 263-266.
Moisan, J.R., Hofmann E., Haidvogel E. (1996) Modeling nutrient and plankton processes in the
California coastal transition zone 2. A three-dimensional physical-bio-optical model. Journal of
Geophysical Research 101, 22677-22691.
Nihoul J. and Djenidi S. (1991). Hierarchy and scales in marine ecohydrodynamics. Earth-Science
Reviews 31, 255-277. [This paper describes the ecohydrodynamic hierarchical organization in marine
ecosystems and describes the effects of major marine hydrodynamic processes on the ecosystem
dynamics]
Oguz T., Ducklow H.W., Malanotte-Rizzoli P. (2000) Modeling distinct vertical biogeochemical structure
of the Black Sea: Dynamical coupling of the oxic, suboxic and anoxic layers. Global Biogeochemical
Cycles 14, NO. 4,1331-1352. [This paper describes the application of a one-dimensional, vertically
resolved, physical-biogeochemical model to provide a unified representation of the dynamically coupled
oxic-suboxic-anoxic system for the interior Black Sea].
Oschlies A. and Garcon V. (1999) An eddy-permitting coupled physical-biological model of the North
Atlantic. 1. Sensitivity to advection numerics and mixed layer physics. Global Biogeochemical Cycles 13,
135-160.
Popova E.E., V. A. Ryabchenko, M.J.R. Fasham (2000). Biological pump and vertical mixing in the
Southern Ocean: Their impact on atmospheric CO2. Global Biogeochem. Cycles, 14, 477-498. [This
paper describes a global box model simulating nitrogen and carbon cycling in the ocean and involving a
detailed description of the seasonal cycles of the oceanic upper mixed layer ecosystem. The model is used
to estimated the global ocean new production].
Prunet P., Minster J.F., Pino D.R., Dadou I. (1996) Assimilation of surface data in a one-dimensional
physical-biogeochemical model of the surface ocean 1. Method and preliminary results. Global
Biogeochemical Cycles 10, 111-138.
Robinson A. (2000). Data Assimilation for modelling and predicting coupled physical-biological
interactions in the sea. In Robinson A.R. and Lermusiaux P. (eds.), Workshop on the assimilation of
biological data in coupled Physical/Ecosystem models, Globec Special Contribution N°3, pp 2-13. [This
paper describes the general process of data assimilation (e.g., methodologies, purposes) and gives several
examples of regions where data assimilation has been successfully applied].
Six K.D. and Maier-Reimer E. (1996) Effects of plankton dynamics on seasonal carbon fluxes in an ocean
general circulation model. Global Biogeochemical Cycles 10, 559-583.
Tankere S. P. C., Price N. B. and Statham P. J. (2000). Mass balance of trace metals in the Adriatic Sea.
Journal of Marine Systems, 25, 269-286. [The paper presents a first order mass balance of six different
trace metals (Mn, Fe, Pb, Zn, Cu, Ni) for a 1-year period for the different compartments of the Adriatic
Sea].
Van Eeckhout D. and Lancelot C. (1997). Modelling the functioning of the north-western Black Sea
ecosystem from the 1960 to present. In: Sensitivity to change: Black Sea, Baltic Sea and North Sea, E.
Ozsoy and A. Mikaelyan (Eds), NATO ASI Series, Kluwer Academic Publishers, 27, 455-468. [This
paper analyses the changes that have occurred in the northwestern Black Sea ecosystem resulting from
eutrophication, and describes the results of a highly complex ecosystem model applied in the area to
simulate the modifications of food-web structure and functioning].
Biographical Sketches
Temel Oguz was born on 13 November 1952 in Aydın, Turkey. He has B.Sc and M.Sc degrees in
Physics from the Middle East Technical University (METU), and Ph.D degree in Dynamical Meteorology
©Encyclopedia of Life Support Systems (EOLSS)
296
OCEANOGRAPHY – Vol.III - Modeling Biogeochemical Processes in Marine Ecosystems - Gregoire, M., Oguz, T.
and Oceanography form the University of Reading (United Kingdom). Dr. Oguz joined Institute of
Marine Sciences of the METU as an Assistant Professor in 1981. He was promoted to Associate Professor
in 1986 and full Professor in 1992 at the same institution. Dr. Oguz is recognized for his work on Black
Sea circulation and ecosystem dynamics. His major interests within the last 10 years were mainly
concentrated on studying the transports and ecological processes that control biogeochemical cycles in the
Black Sea using coupled physical-biogeochemical models with data assimilation capabilities. More
specifically he has been working on circulation dynamics, their effects on the biogeochemical transports,
and understanding biogeochemical processes in the Black Sea.
Recent publications:
Oguz, T., H. Ducklow, P. Malanotte-Rizzoli, J.W. Murray V.I. Vedernikov, U. Unluata (1999) "A
physical-biochemical model of plankton productivity and nitrogen cycling in the Black Sea". Deep Sea
Research I, 46, 597-636.
Oguz, T. and S. Besiktepe (1999) "Observations on the Rim Current structure, CIW formation and
transport in the Western Black Sea". Deep Sea Research, I, 46, 1733-1753.
Oguz, T. and B. Salihoglu (2000) "Simulation of eddy-driven phytoplankton production in the Black
Sea". Geophys. Res. Letters, 27(14), 2125-2128.
Oguz, T., H. W. Ducklow, P. Malanotte-Rizzoli (2000) "Modeling distinct vertical biogeochemical
structure of the Black Sea: Dynamical coupling of the oxic, suboxic and anoxic layers". Global
Biogeochemical Cycles, 14, 1331-1352.
Oguz, T. and J.W. Murray (2001) "Simulation of Suboxic-Anoxic interface zone structure in the Black
Sea". Deep Sea Research I, 48, 761-787.
Marilaure Grégoire was born on June 29, 1972 in Huy, Belgium. She graduated in engineering in 1995
(University of Liège). From 1995 until 1999, she worked as a research assistant, National Fund for
Scientific Research (FNRS) Belgium and obtained a PhD in applied sciences (University of Liège) in
1999. She is currently working as a researcher, FNRS Belgium, in the department of geological fluid
dynamics, on physical and ecological modelling of the marine environment. She develops coupled
ecosystem-hydrodynamical models applied to the Black Sea in a sustainable development perspective.
The study of the desertification in the Aral Sea Region is also one of her main concerns.
Recent publications:
Grégoire M., Beckers J.-M., Nihoul J. and Stanev E., 1998. Reconnaissance of the main Black Sea's
ecohydrodynamics by means of a 3D interdisciplinary model. Journal of Marine Systems, 16:85-105.
Delhez E., M. Gregoire, J.C.J. Nihoul and Beckers J.M., 1999. Dissection of the GHER Turbulence
Closure Scheme, Journal of Marine Systems, 21., pp 379-397.
Nezlin N., Kostianoy A. and Gregoire M, 1999. Patterns of seasonal and interannual changes of surface
chlorophyll concentration in the Black Sea revealed from the remote sensed data. Remote Sensing of
Environment, 69, pp 43-55.
Beckers J.M., Gregoire M., Nihoul J.C.J. Stanev E., Staneva J. and Lancelot C., 2000. Hydrodynamical
processes governing exchanges between the Danube, the North-Western continental shelf and the Black
Sea's basin simulated by 3D and box models. Estuaries and Coastal Shelf Science, in press.
Stanev E., Beckers J.M., Lancelot C., Letraon P.Y., Staneva J., Peneva E. and Gregoire M., 2000.
Coastal-open ocean exchange. Black Sea examples from survey, satellite data and modelling. Estuaries
and Coastal Shelf Science, in press.
©Encyclopedia of Life Support Systems (EOLSS)
297