Rapport
Résumé : AMORE (Advanced Modeling and Research on Eutrophication) is an interdisciplinary consortium composed of biologists and physical and ecological modellers focusing their research activities on coastal eutrophication in the Belgian coastal zone (BCZ) with special interest in harmful Phaeocystis colony blooms. The long-term objective of AMORE is to develop a mathematical model able to predict the magnitude and geographical extent of Phaeocystis colony blooms in the Eastern Channel and Southern Bight of the North Sea with focus on the BCZ in response to varying shortterm climate conditions and riverine nutrient (N, P, Si) loads. To achieve this objective AMORE has developed an integrated research methodology that involves and combines in an interactive way the collection of historical and new field data, process-level studies, mathematical modelling and data assimilation. Between 1997 and 2001 AMORE research focused on mechanisms through which a change in riverine nutrient loads induces modification in the phytoplankton community structure and how this change affects in turn the structure and functioning of the planktonic food-web and the related biogeochemical cycles. The knowledge gained was synthesized for integration in the existing ecological MIRO which in turn was coupled with the 3D-COHSNS hydrodynamical model developed for describing water transport in the studied domain. Progress achieved in our understanding of eutrophication mechanisms in the BCZ as well as our present ability to predict Phaeocystis spreading and magnitude in response to rive rine nutrient delivery are discussed in the present report. Among bottom up factors (light, temperature, nutrients) that currently control phytoplankton blooms we demonstrated that a light threshold of 12 μmole m-2 s-1 in the water column is critical for the onset of the spring phytoplankton bloom in BCZ. This threshold is reached between mid-February and mid-March and relies on physical processes determining the load of suspended matter. This light level corresponds to the light required by early spring diatoms for an optimised cell division rate. These early spring diatoms are also better competitive compared to Phaeocystis colonies at the low temperature of late -February early March (5-6°C). On the contrary, Phaeocystis colonies optimise their growth at higher temperature and light but are better flexible to light change due to a fast xantophyll cycling rate. Nutrients availability on the other hand is driving the succession of diatom species and Phaeocytis. Diatom blooms are characterized by the succession of three different communities characterized by a specific Si stoichiometry. The observed relationship between the silicification level of diatom species and ambient silicic acid strongly suggests that this nutrient is shaping the observed spring-to-summer diatom 3 succession in BCZ. Further observational evidence showed that the magnitude of the early spring diatoms is controlled by the availability of Si(OH)4 and PO4 to a less extent. This observation, also supported by 0D-MIRO model runs, suggests that “excess new nitrates” (i.e. left over after early spring diatom growth) but regenerated PO4 and Si (for diatoms only) sustain the growth of Phaeocystis colonies and the cooccurent diatoms (Guinardia sp). For the first time, PO4 limitation was demonstrated in the BCZ via the detection of alkaline phosphatase activity in spring. One major result is that this enzymatic activity is associated to mainly large particles including phytoplankton cells and their attached bacteria. The highly significative correla tion between alkaline phosphatase activity and Phaeocystis suggests that the colonies might play a major role in PO4 regeneration. Specific grazing experiments conducted during AMORE clearly demonstrated that Phaeocystis colonies are not grazed by indigenous zooplankton (Temora longicornis) but the reason was not identified. The combination of all biological activities and biomass measured in 1998 allowed calculating the budget of carbon transfer through the planktonic network. This calculation indicated that most of ungrazed Phaeocystis flows through the microbial network where Phaeocystis cells are grazed by microzooplankton and Phaeocystis-derived organic carbon is rapidly recycled. This conclusion was supported by independent microbiological assays which suggested that most of organic matter synthesised during the spring bloom was biodegradable per se. The very low bacterial growth yield of 0.1 indicates that most organic carbon taken up is mineralised rather than building biomass. This could be due to PO4 limitation although not experimentally demonstrated. Also in agreement with conclusions of carbon budget calculation, are the negative potential sinking rates obtained for Phaeocystis colonies larger than 250 μm. Based upon this carbon budget, it was further hypothesized that adult copepods would be in food shortage during Phaeocystis blooms which could impact negatively not only on the next generation of copepods but also on fish recruitment by starvation of fish larvae. This hypothesis was however challenged by additional field data on egg production by copepods suggesting alternate sources of good quality food for copepods. Yet the trophic pathways are even more complex due to the presence between April and June of an impressive mass of gelatinous zooplankton which trophic role was not identified. Numerical experiments included 0D and 3D modelling. Prior to its 3D implementation, the published version of 0D-MIRO was upgraded based on a synthesis of processlevel studies and making use of data assimilation technics. For this purpose the adjoint model of 0D-MIRO was set up and twin experiments were conducted to improve MIRO parameters and estimate possible error due to some unresolved physical processes in 0D. From this numerical work the independence of all MIRO 4 parameters was demonstrated and the most important parameters of the ecosystem functioning were highlighted. 3D-MIRO&CO was implemented by coupling the upgraded version of MIRO to the 3D-COHSNS hydrodynamical model. The modelled domain was bordered by latitudes 51°N and 52.5°N and included input from the main rivers within this domain (Rhine, Scheldt, Thames). For this first application the grid resolution was 4.5km. Simulations were run for the years 1995-1998 and sensitivity studies conducted to assess the relative importance of wind, tide, and river discharge on physicsical features and nutrient discharge on ecosystem dynamics. Main results obtained show that thee geographical extent of the plume of continental coastal freshwater in BCZ depends mainly on wind speed and direction, and is only weakly dependent on tide and river flow-rate (which affects absolute salinity but not dispersion of freshwater). In contrast to currently admitted opinion, freshwater found in the central BCZ at station 330 originate primarily from the Rhine discharge and salinity at station 330 is only slightly affected by discharge from the river Scheldt. However and contrarily to the salinity fields, plumes of nutrients from the rivers Scheldt and Rhine remain relatively distinct (not merged) because of winter depletion. Impact of Scheldt nutrient discharge on BCZ remains important, thought the Rhine discharge needs also to be considered. Simulations of biological variables with 3D-MIRO&CO demonstrated a number of observed processes such as the diatom-Phaeocystis succession and the related depletion of silicates and nitrates and gave for the first time a view of spatial variability within the domain. However, a number of model weaknesses were also identified which require in part a better understanding and parameterisation of biologial processes. These are for instances the parameterization of mesozooplankton feeding and of processes involved in phosphorus benthic diagenesis. Finally model results obtained with NO3 and/or PO4 reduction scenarios suggest that NO3 discharged by the Scheldt river should be the target nutrient to be reduced for obtaining a significant decrease of Phaeocystis colony blooms in the BCZ. NO3 discharged by the Scheldt river should be the target nutrient to be reduced for obtaining a significant decrease of Phaeocystis colony blooms in the BCZ.

Documents en relation

DI-fusion