Difference between revisions of "Monitoring the water quality of coastal waters with automatic equipment"

From MarineBiotech Infopages
Jump to: navigation, search
(European policies and operational services)
(Data analysis, numerical modelling and data assimilation)
Line 33: Line 33:
 
==Data analysis, numerical modelling and data assimilation ==
 
==Data analysis, numerical modelling and data assimilation ==
  
Developing forecasting tools ... (CA p 8)
+
Developing forecasting tools for coastal zones necessitates increasing synergy between traditional and newly available data, modelling with very fine resolution and advanced data assimilation techniques. Extensive data analyses and model validations are a central task. GKSS will take the major responsibility of the information technology and model systems used in COSYNA. It will create hard- and software for collecting, storing and efficiently managing the incoming data flows including first quality checks. However, individual components of the model system or the assimilation module will be developed in strong cooperation with KDM partners including the AWI. Data streams based on observations and model simulations will also be established going from the centre at GKSS to sub-nodes at selected partners within the consortium.
 +
 
 +
Data analyses as have to rely on linear as well as non-linear methods for detecting common patterns or statistical characteristics (e.g. dimensionality) in multi-variate data sets. We will quantify the spatial as well as temporal cross-correlations of the continuous time-series from sensors at multiple positions. In addition, novel techniques based on Lagrangian transport modelling, which link different reference point can be applied (Callies et al., in prep; Brandt et al, in prep). Statistical results will not only lead to a better scientific understanding of specific mechanisms but also to guideline data assimilation and model validation experiments. One of the major problems here is the short predictability limit of ocean weather and sea state due to short physical temporal scales and atmospheric predictability limits, thus enhancing negative consequences from forcing inaccuracies, uncertainties in initial and lateral boundary conditions, and inappropriate model representation of physical and ecological processes. Some of these limitations also affect the reliability of long-term simulations.
 +
 
 +
Specific objectives are to: (1) improve the quality of existing physical and biogeochemical models for North Sea, German Bight and Wadden Sea by resolving relevant processes at adequate spatial scales in the entire computational domain, (2) enhance exploitation of existing data streams, (3) develop up-to-date analysis of as much as possible observations from different platforms and providing guidance for data assimilation, (4) provide improved forecast based on increasing the synergy between advanced numerical models, data assimilation methods and traditional and new observational platforms, (5) provide new knowledge and estimates of the state of coastal zone (e.g. ocean turbulence, stratification, transport of particulate organic material and total suspended matter, estuary-coast-offshore exchange, water quality status, source and sinks for major carbon and nitrogen species).
 +
 
 +
The complex dynamics of coastal regions and exchanges with the atmosphere, near-shore and offshore regions present a major scientific challenge. A drastic shift in ocean physics and in forcing functions is observed in the coastal zone due to (1) rapid change of coastal shape and bathymetry, (2) dominance of another set of physical processes, and (3) multiple forcing. Reliable offshore boundary conditions of coastal models provided by large-scale ocean models have the potential to extend predictability on shelves and enhance the representativeness of forecasting systems. However, small errors in initial and boundary conditions could trigger unphysical processes in the models, therefore substantial care must be given to improving their quality and correcting their outcomes through data assimilation.
 +
 
 +
Dynamical balances in the coastal zone seem prohibitive to use some relatively simple assimilation schemes (e. g. Cooper and Haines 1996), which have proved useful in open ocean modelling. Therefore a wide range of methods are presently used to assimilate data into coastal models including sequential methods (e.g. simple nudging, optimal interpolation, reduced order optimal interpolation (De Mey and Benkiran 2002), ensemble-based optimal interpolation (Lamouroux et al. 2006), and various forms of Kalman filter (Evensen 2003) including the extended, singular evolutive, and ensemble Kalman filters (Mourre et al. 2004, 2006) as well as adjoint-based methods (based on the minimization of cost functions, Taillandier et al. 2007). This variety of approaches reveals, not in the last place, the diversity of data available for the coastal zone. Their complementarities have not yet been well established but promising results exist (Mourre et al 2006). Recent studies demonstrate that some ocean fields can be sufficiently represented by a limited number of multivariate Empirical Orthogonal Functions (EOFs) facilitating the selection of the “optimal set” (Sparnoccia et al. 2003). The existing experience with data assimilation in circulation models of North Sea, German Bight and Wadden Sea is still limited, but the observational (discussed before) and modelling (Dick et al. 2001, Stanev et al. 2003, Gayer et al. 2006, Staneva et al. 2007) practices give an encouraging motivation to propose the below activities. Data assimilation will be used to estimate past (hindcast), present (nowcast) and future (forecast) conditions in the North Sea and in its coastal regions and will provide measures of uncertainty. The systems of interest will include the shelf area, estuaries, tidal flats, river plumes, straits and sills (cmp. Figure 2).
  
 
==Description of a facility==
 
==Description of a facility==

Revision as of 10:50, 20 November 2008

Category:Revision


Background

Within Europe most maritime countries have monitoring programmes in place to fulfil their regulatory, EU or otherwise conventional commitments. Up till now no European structure has been set up or is in place to harmonize these observations in such a way that a comparison between different regions or a Pan European view is possible. A very limited overview can be obtained from the Annual reports of the EEA (European Environmental Agency), e.g. regarding nutrient concentrations in coastal waters. With the upcoming WFD, directives for transitional waters and the marine strategy requirements for pan European approaches are increasing.

Attempts to set up integrated networks of observations have been made on regional scales e.g. in the Baltic (BOOS), and in the North Sea and adjacent Atlantic (NOOS). However, still these networks do hardly cope with the ecosystem approach as promoted by ICES (International Council for the Exploration of the Sea). Generally most networks intend to observe the primary parameters needed to improve our meteorological models or physical water parameters to improve our transport and current models. What is urgently needed is to extend these lists of parameters to important ecological parameters such as turbidity, phytoplankton and zooplankton biomass, occasional biomass of fishes, macro-benthos, and nutrients in surface and deeper waters. In order to obtain reliable data which can as well be used to drive ecological models intensive quality assurance procedures have to be implemented and integrated into the overall data management schemes.

The present article aims at presenting possible approaches and at showing the possibilities which are currently available. This includes ships-of-opportunity observations, in combination with the CPR (Continuous Plankton Recorder), satellite observations to extend the spatial scale of high frequency ground truth observations, and finally the incorporation through data assimilation of these observations into combined transport and ecological models.

European policies and operational services

Over the last decade, increasing national and international effort has been established worldwide in monitoring, forecasting and assessing the state of coastal systems. Around the coasts of the United States, for example, eleven regional Coastal Ocean observing systems have been established as part of the US-Integrated Ocean Observing System (IOOS) (Ocean.US, 2002). In Europe, EuroGOOS (www.eurogoos.org) has fostered the planning and implementation of prototype systems for various European shelf seas, e.g. the Mediterranean Operational Oceanography Network MOON (Pinardi and Flemming 1998), the Baltic Operational Oceanographic System BOOS (Buch and Dahlin 2000), or the North West Shelf Operational Oceanographic System NOOS (Droppert et al. 2000). A prominent example of a national effort is the Previmer Observing System (url: http://www.previmer.org) in France, which is set up by a consortium of scientific, public and industrial partners.

The identified needs for these systems comprise aspects of climate change, marine operations, national security, sustainable management, ecosystem conservation and restoration, public health and mitigation of natural and man-made hazards (see e.g. Frosch et al., 1999). Generally, they are based on existing mooring stations of governmental observational systems. Although these stations deliver impressive long-term time series, they capture the time domain at only a single point. Furthermore, reliably measured parameters are often limited to physical state variables. Chemical parameters or even biological are, if any, available in many cases only for limited time periods. Integration of further chemical and biological state variables as well as information from remote sensing images is always put forward as an essential target. As an example of a European prototype system, the GMES Services network MarCoast (Marine & Coastal Environmental Information Services) is developed to deliver satellite-based services in the field of marine and coastal applications. Services integrate detection and monitoring technologies involved in water quality, oil spill and meteorological information into a durable network.

In Europe, present activity focuses mainly on the harmonisation of existing technology and methods between different countries. Only few countries invest significantly into further development of integrated coastal observation systems (e.g. Previmer in France), although new technologies have been brought to pre-operational maturity, e.g. the development of FerryBox-systems onboard of ships-of-opportunity (Petersen et al., 2007).

In the US and Canada significant effort is raised to increase nowcast/forecast capabilities for coastal waters by a multiplatform sampling approach (Schofield et al. 2002, von Alt el al. 1994, Simonetti, 1998) combined with a suite of data assimilative models (see below). Temporal and spatial density of in-situ data is enhanced by using autonomously operating underwater vehicles (AUV, Glider) (von Alt et al. 1994, Simonetti 1998) with remote control of actual tracks by automated systems (Fiorelli et al. 2006). The installation of undersea laboratories, connected to the landsite by fiberglass broadband width and electric power cables (Schofield et al. 2002, Purdy et al. 2003) offers a complementary, albeit costly approach to permanently observe the environment from the seabed. As power supply and data transfer rates pose no real limitations, the submerged unit can be used as node, which supply basic infrastructure for a vast number of scientific partners to study both episodic events and long-term trends. Despite the individual research flavor of each observatory, these high-resolution regional sites will be unified through a coherent suite of broadband measurements (Schofield et al., 2002). Spectral images from remote sensing are becoming increasingly part of routine coastal observation to derive spatial distributions of dissolved and particulate (abiotic and biotic) material. Their usage for operational purposes, on the other hand, is limited a relatively low temporal resolution and by the presence of clouds. Also validity of reflectance values as proxies for different physical and biological variables is still a matter of ongoing research, so that in-situ measurements are still indispensable. Therefore, we still face the challenge to develop and integrate reliable and, in part, new technologies for long-term observation of water chemical and biological state variables.

In all cases effective data management is prospected or implemented to store, access, distribute and present data in nearly real-time and with documented quality. For operational purposes effective data management must provide tools for data integration from different sources and interfaces to data assimilation schemes and numerical model schemes. Data accessibility and presentation plays also an important role to inform, participate and educate the general public (Schofield et al. 2002).

Global and regional activities such as the Global Ocean Data Assimilation Experiment (GODAE) and Marine Environment and Security for the European Area (MERSEA) give a practical demonstration of gathering up-to-date near real-time observation systems and ocean data assimilation providing regular information about the state of oceans and regional seas, including suitable initial and boundary conditions needed for near coastal models. The activities carried out in the frame of the Global Ocean Observing System (GOOS) have recently been extended towards the coast (Coastal Ocean Observing Panel, COOP/JPICO) with a focus on coastal management, environmental protection, ports and shipping, and monitoring. Documenting and forecasting change in coastal environments will require integration of physical, chemical, biological and geological observations and modelling and could provide scientific products and guidance to a wide range of users.

The European Cluster in Operational Oceanography is developing marine prototype operational systems (in the Arctic Sea and North Atlantic- TOPAZ and in the Mediterranean Sea-MFS) together with capacity building activities (in the Baltic Sea-PAPA, in the Mediterranean Sea- MAMA, in the Black Sea- ARENA). In line with this strategy the European Commission recently launched the European COstal sea Operational observing and forecasting system Project (ECOOP) aiming to build up a sustainable pan-European capacity in providing timely, quality assured marine services (including data, information products, knowledge and scientific advices) in European coastal and shelf seas. It is thus a pivotal part of Europe's contribution to issues affecting the global environment and the safety of our planet now also addressed within the GEOSS initiative. GEOSS and GMES are ambitious concepts, which reconcile the political needs associated with environment and safety issues with the scientific and technological capacities offered by information technologies and Earth observation technologies.

Numerous examples already exist around Europe demonstrating advancements in existing operational observing and forecasting systems. Advanced observing technologies have been developed and demonstrated in projects like EuroROSE, Ferrybox, ADRICOSM and Poseidon. In the North Sea, and in particular in the area of German Bight a huge potential of observing platforms, data transmission and processing, as well as numerical modelling expertise exists. The most advanced level of systems operability has been demonstrated in the activities of BSH and in their product’s delivery. Research institutes and academia usually carry out in parallel, sometimes coordinated, sometimes complementary activities. However, any future development of observational systems and forecasting requires strong cooperation, setting standards, and mutual involvement. Sharing data, technologies and costs would shorten the research time, improve the products and reduce costs. A logical next step is to consolidate and integrate the current observing and model systems with a common information platform and quality standards.

Current challenges in operational coastal oceanography are that the existing systems are not harmonised, quality assessment is still a problem, real-time data transfer, data exchange between organisations and near-real-time availability to modellers are not solved. Assimilation of various sources of near coastal data is still a complex scientific problem.

Data analysis, numerical modelling and data assimilation

Developing forecasting tools for coastal zones necessitates increasing synergy between traditional and newly available data, modelling with very fine resolution and advanced data assimilation techniques. Extensive data analyses and model validations are a central task. GKSS will take the major responsibility of the information technology and model systems used in COSYNA. It will create hard- and software for collecting, storing and efficiently managing the incoming data flows including first quality checks. However, individual components of the model system or the assimilation module will be developed in strong cooperation with KDM partners including the AWI. Data streams based on observations and model simulations will also be established going from the centre at GKSS to sub-nodes at selected partners within the consortium.

Data analyses as have to rely on linear as well as non-linear methods for detecting common patterns or statistical characteristics (e.g. dimensionality) in multi-variate data sets. We will quantify the spatial as well as temporal cross-correlations of the continuous time-series from sensors at multiple positions. In addition, novel techniques based on Lagrangian transport modelling, which link different reference point can be applied (Callies et al., in prep; Brandt et al, in prep). Statistical results will not only lead to a better scientific understanding of specific mechanisms but also to guideline data assimilation and model validation experiments. One of the major problems here is the short predictability limit of ocean weather and sea state due to short physical temporal scales and atmospheric predictability limits, thus enhancing negative consequences from forcing inaccuracies, uncertainties in initial and lateral boundary conditions, and inappropriate model representation of physical and ecological processes. Some of these limitations also affect the reliability of long-term simulations.

Specific objectives are to: (1) improve the quality of existing physical and biogeochemical models for North Sea, German Bight and Wadden Sea by resolving relevant processes at adequate spatial scales in the entire computational domain, (2) enhance exploitation of existing data streams, (3) develop up-to-date analysis of as much as possible observations from different platforms and providing guidance for data assimilation, (4) provide improved forecast based on increasing the synergy between advanced numerical models, data assimilation methods and traditional and new observational platforms, (5) provide new knowledge and estimates of the state of coastal zone (e.g. ocean turbulence, stratification, transport of particulate organic material and total suspended matter, estuary-coast-offshore exchange, water quality status, source and sinks for major carbon and nitrogen species).

The complex dynamics of coastal regions and exchanges with the atmosphere, near-shore and offshore regions present a major scientific challenge. A drastic shift in ocean physics and in forcing functions is observed in the coastal zone due to (1) rapid change of coastal shape and bathymetry, (2) dominance of another set of physical processes, and (3) multiple forcing. Reliable offshore boundary conditions of coastal models provided by large-scale ocean models have the potential to extend predictability on shelves and enhance the representativeness of forecasting systems. However, small errors in initial and boundary conditions could trigger unphysical processes in the models, therefore substantial care must be given to improving their quality and correcting their outcomes through data assimilation.

Dynamical balances in the coastal zone seem prohibitive to use some relatively simple assimilation schemes (e. g. Cooper and Haines 1996), which have proved useful in open ocean modelling. Therefore a wide range of methods are presently used to assimilate data into coastal models including sequential methods (e.g. simple nudging, optimal interpolation, reduced order optimal interpolation (De Mey and Benkiran 2002), ensemble-based optimal interpolation (Lamouroux et al. 2006), and various forms of Kalman filter (Evensen 2003) including the extended, singular evolutive, and ensemble Kalman filters (Mourre et al. 2004, 2006) as well as adjoint-based methods (based on the minimization of cost functions, Taillandier et al. 2007). This variety of approaches reveals, not in the last place, the diversity of data available for the coastal zone. Their complementarities have not yet been well established but promising results exist (Mourre et al 2006). Recent studies demonstrate that some ocean fields can be sufficiently represented by a limited number of multivariate Empirical Orthogonal Functions (EOFs) facilitating the selection of the “optimal set” (Sparnoccia et al. 2003). The existing experience with data assimilation in circulation models of North Sea, German Bight and Wadden Sea is still limited, but the observational (discussed before) and modelling (Dick et al. 2001, Stanev et al. 2003, Gayer et al. 2006, Staneva et al. 2007) practices give an encouraging motivation to propose the below activities. Data assimilation will be used to estimate past (hindcast), present (nowcast) and future (forecast) conditions in the North Sea and in its coastal regions and will provide measures of uncertainty. The systems of interest will include the shelf area, estuaries, tidal flats, river plumes, straits and sills (cmp. Figure 2).

Description of a facility

New features (CA p 12)

(p 12, Fig. 1.)

Conclusions

References

See also

Internal Links

External Links

The main author of this article is Colijn, Franciscus
Please note that others may also have edited the contents of this article.

Citation: Colijn, Franciscus (2008): Monitoring the water quality of coastal waters with automatic equipment. Available from http://www.coastalwiki.org/wiki/Monitoring_the_water_quality_of_coastal_waters_with_automatic_equipment [accessed on 7-07-2020]