Toward Marine Environmental Predictions in the Mediterranean Sea Coastal Areas
373
microzooplankton, dissolved and particulate matter, bacterial heteronanoflagellates
(HNF) biomass and phytoplankton species. This is a unique data set to calibrate
model parameters and understand which initial field are important to achieve predictability in different state variables, in particular the phytoplankton biomass and
its different groups.
The data set consisted of high temporal resolution (4 hours) profiles taken at the
station SI during a two days period. The model was initialized with temperature,
salinity, dissolved nutrient profiles from the observations, dissolved and particulate
organic matter, bacterial and etheroflagellates biomass and predicted the ecosystem
state variables for two days, under realistic extemal atmospheric forcing conditions, the same used for the hydrodynamics models. The results are shown in Fig.
17.12 and 17.13 for a two days periods during October 1995. It is evident that the
model is capable of capturing the bulk of the biomass values for both bacteria, diatoms and flagellates but it is missing some small temporal scale dynamics and the
large lateral advection of phytoplankton in the region. We see in fact that observations indicate large subsurface patches of diatoms entering the region between midnight and 6 a.m., thus forcing us to conc1ude that they were laterally advected into
the region represented by our water column model. Errors, defined as the root
mean square values of the difference between observations and model predictions,
range overall between 10% and 30%, normalized with the mean observed values.
This experiment shows that there is a potential predictability in the ecosystem
state variables at short time scales, in particular for the primary producers biomass.
The overall structure of the phytoplankton, bacterial and microzooplankton functional groups is captured and simulated correctly, hinting to the fact that this coastal
ecosystem is mainly driven by nutrient supply by the river sources. During summer
however, internal food web dynamics could prevail, such as microbial carbon
dynamics competing with phytoplankton uptake and benthic-pelagic coupling
which could become a relevant source of dissolved nutrients from benthic remineralization of particulate and dissolved organic matter.
The results of this modeling experiment are difficult to transpose to the future
environmental forecasting system for this area since the observing system requirements are high (many parameters need to be monitored) and measurement techniques not yet developed for automatic observations of such quantities. If the
model would have not been initialized with, let's say etheronanoflagellates biomass, the bacterial biomass would not have been maintained within the range of
observed values (see Fig. 17.13). The coastal observing system components for
model initialization should be carefully planned in the future and major technological development will be required to realize measurements in operational mode.
17.7 Conclusions
In this chapter we have described a potential scheme to develop and implement
environmental predictions in the coastal areas ofthe world ocean. The coastal envi-
373
microzooplankton, dissolved and particulate matter, bacterial heteronanoflagellates
(HNF) biomass and phytoplankton species. This is a unique data set to calibrate
model parameters and understand which initial field are important to achieve predictability in different state variables, in particular the phytoplankton biomass and
its different groups.
The data set consisted of high temporal resolution (4 hours) profiles taken at the
station SI during a two days period. The model was initialized with temperature,
salinity, dissolved nutrient profiles from the observations, dissolved and particulate
organic matter, bacterial and etheroflagellates biomass and predicted the ecosystem
state variables for two days, under realistic extemal atmospheric forcing conditions, the same used for the hydrodynamics models. The results are shown in Fig.
17.12 and 17.13 for a two days periods during October 1995. It is evident that the
model is capable of capturing the bulk of the biomass values for both bacteria, diatoms and flagellates but it is missing some small temporal scale dynamics and the
large lateral advection of phytoplankton in the region. We see in fact that observations indicate large subsurface patches of diatoms entering the region between midnight and 6 a.m., thus forcing us to conc1ude that they were laterally advected into
the region represented by our water column model. Errors, defined as the root
mean square values of the difference between observations and model predictions,
range overall between 10% and 30%, normalized with the mean observed values.
This experiment shows that there is a potential predictability in the ecosystem
state variables at short time scales, in particular for the primary producers biomass.
The overall structure of the phytoplankton, bacterial and microzooplankton functional groups is captured and simulated correctly, hinting to the fact that this coastal
ecosystem is mainly driven by nutrient supply by the river sources. During summer
however, internal food web dynamics could prevail, such as microbial carbon
dynamics competing with phytoplankton uptake and benthic-pelagic coupling
which could become a relevant source of dissolved nutrients from benthic remineralization of particulate and dissolved organic matter.
The results of this modeling experiment are difficult to transpose to the future
environmental forecasting system for this area since the observing system requirements are high (many parameters need to be monitored) and measurement techniques not yet developed for automatic observations of such quantities. If the
model would have not been initialized with, let's say etheronanoflagellates biomass, the bacterial biomass would not have been maintained within the range of
observed values (see Fig. 17.13). The coastal observing system components for
model initialization should be carefully planned in the future and major technological development will be required to realize measurements in operational mode.
17.7 Conclusions
In this chapter we have described a potential scheme to develop and implement
environmental predictions in the coastal areas ofthe world ocean. The coastal envi-