Skip to main content

Ocean Strategic Services beyond 2015

Final Report Summary - OSS2015 (Ocean Strategic Services beyond 2015)

Executive Summary:
The OSS2015 project was set up to help specify and design the “Marine Biology” component of the COPERNICUS Marine Core Service (MCS), further to the MyOcean progress within the FP7 framework and in advance of a new era of global multispectral observations of the ocean (through COPERNICUS) in support of the scientific community dealing with ecology and climatology, and in support of policy makers & enforcers in charge of socio-economic challenges including the European blue growth strategy.

Due to its vastness, to its dynamics variability, to its interactions with land and atmosphere at turbulent boundary layers, its ecology complexity, its geomorphology specificities and its strains under anthropogenic stresses, the Ocean deserves a complete, efficient and comprehensive observation & monitoring system built upon in situ measurements’ networks and remote sensing from satellites, using equations and laws to interpolate between measurements that are always too sparse to resolve all processes, and to extrapolate all parameters. OSS2015 research was dedicated to a better understanding of the upper ocean biology, relevant and reliable information products and the organisation of data dissemination.

The success of the project, in support of environmental services, is assessed by the delivery of accurate values of bio-geochemical parameters which are relevant for the study of the marine ecosystems and of the carbon cycle, and the prototyping of a Collaborative Platform for scientific exploitation of Earth Observation (EO) data (i.e. to facilitate the data handling by scientists and to spur algorithm development). Some of the information has been evaluated and provided for the first time; reason why, specific dissemination methods and exploitation methods have been implemented to promote its use.

The key results of OSS2015 are:
1. A methodology was designed to optimally deploy observation floats at sea to complement EO;
2. Key factors driving marine eco-systems have been identified allowing to produce biochemical parameters that will help understanding & accurate prediction of marine ecosystems response to anthropogenic changes, useful for the implementation of European policies with regard to climate change (monitoring and adaptation to), marine environment protection & exploitation (marine resources management) like the Marine Strategy Framework Directive;
3. OSS2015 informs on the state of marine ecosystems for the global ocean, including climate monitoring, by delivering a unique time series of innovative biogeochemical products. This OSS2015 production supports the increasing marine biology knowledge, thanks to an easy access open to any interested user.
4. The OSS2015 Collaborative Research Platform is the start of a new virtual research centre which allows scientists working together

Results led to recommendations on a COPERNICUS “green” offshore marine service to be taken into account for the next building step of the Copernicus Marine service. After this prototyping phase, the OSS2015 service might be implemented in the framework of a Public-Private Partnership (in support not only of climatology, adaptation and mitigation of climate change, but of sustainable and responsible development of marine resources, by optimal deployment of observation means).

In the course of the project, OSS2015 consortium members have opened several additional perspectives:
• gathering of SMEs interested by further developments of the OSS2015 demonstrator and by its commercialisation to feed EO downstream services like algal blooms’ forecast or optimisation of aquaculture;
• expansion of OSS2015 to coastal area applications, using satellite-based information aside from the ocean- dedicated COPERNICUS satellites (High Resolution and Very High Resolution sensors such as Sentinel 2);
• Big Data analytics for the “green” ocean monitoring
Project Context and Objectives:
Objective 1: Develop and validate new tools for integration and/or assimilation of both EO and in situ data in biogeochemical models –in support of the modelling effort of the MCS

An accurate prediction of the bio-chemical composition of the Ocean upper layers, either a hindcast, a nowcast or a forecast, needs good theoretical models of the ocean properties and data (remote sensing and in-situ) to be assimilated for regular re-initialisation of the aforesaid models or estimation of the ocean state.

OSS2015 contributions to this objective are the following:

i. Models’ adequacy, and new geo-chemical models are to be developed (D42.4 deliverable), their accuracy to be assessed through sensitivity analysis (D22.4 deliverable) and matchups between models outputs and measurements (D43.1 and D43.2 deliverables);

ii. data relevance & accuracy, in particular measurements by bio-Argo floats, with specific quality controls –bio-Argo floats to be optimally deployed to maximise the information retrieval capacity of the network of floats (D22.2 and D42.2 deliverables), and EO - recommendations for future EO missions to be drawn from the results of OSS2015 (D53.2 deliverable).

A full in-situ and EO data assimilation exercise has been performed for the pilot site in the Ligurian Sea (D21.5 deliverable), demonstrating the feasibility of the approach.

Objective 2: Develop and validate new EO products relevant to the biogeochemistry of the ocean –in support of the continuous enhancement of MCS users’ satisfaction

New bio-chemical Earth Observation derived information, i.e. EO products, have been developed using advanced algorithms, implemented and validated (D23.1 D41.2 deliverables) and 15 year time-series of these parameters have been produced (D52.2 deliverable). Free and open access to these EO products is granted on-line via a web service.

The new products are relevant to ecosystem monitoring and research on the role of phytoplankton in global carbon cycle of the Earth:
• Ecosystem monitoring with applications to science, fishing, fish and shell fish farming and to
• Evaluation of the contribution of the phytoplankton to the global carbon cycle.

Objective 3: Develop an “on-demand” information, elaboration and distribution system –in support of the improvement of Information Communication Technology solutions for the MCS

The OSS2015 system has been designed as a collaborative platform as per a new paradigm in the process of developing new Earth Observation derived information (so-called “products”). It provides remote users with a complete set of tools to implement, validate and process new data processing algorithms by accessing directly remote sensing archives and processing facilities (D25.2 deliverables). The Collaborative Platform has already found practical uses outside OSS2015, such as supporting the activities of the Mission Performance Centre of Sentinel 3.


Project Results:
1 Main results overview

The aim of OSS2015 is to deliver marine "green" services including:
* Marine bio-resources evaluation and changes' monitoring
* Eco-region characterizations
* Environmental and hazard monitoring

These services are performed using primary data coming from Earth Observation remote sensing data, in-situ measurements and bio-geochemical models, using various data processing techniques: data assimilation in models, digital discrete-time signal processing of deterministic and/or random series (temporal and statistical analysis).

OSS2015 improved each component of this information production line:
* collection, indexing and storage of raw data (either EO or in-situ measurements) and pre-processed data
* data/information production : transformation of EO measurements (optics) in environmental variables with
o new, and validated, bio-optical models, i.e. inverse of optical states' equation
o data assimilation in bio-geochemical forecast models to "fill the gaps" in sampling (reconstruction of variables' fields) and correct errors of the estimators of bio-optical models
(data production also involves algorithm's calibrations and quality control)
* data exploitation techniques (clustering, time-series analysis, etc.)

New or improved users' services have been implemented for the order, generation and distribution of the new EO products, focusing on
o reliability of the distributed information (delivery of statements of validation)
o benchmarking of information accuracy by users, allowing users' involvement in assessing the data quality (Quality Assurance),
o users' participation to the algorithms' update (the Open Software culture),
o on-demand production, with a dramatic improvement in data ordering and delivery.

Together, these advances contribute to the design of a performing Copernicus Ocean "green" service, a component or an adjunct to the MCS


2 Modelling, assimilation and validation: from apparent optical properties to the constituents of seawater

2.1 Optical modelling

Remote sensing provides a highly valuable source of information on the Ocean bio-chemical state vide its optical characteristics. This information is available in the form of apparent optical properties (AOP) at the surface of the Ocean, whereas it is the results of attenuation and backscattering in the water column as deep as where external light penetrates. The vertical variability of the inherent optical properties (IOP) which is due to the variability in the concentration of ocean waters' constituents (bio-chemical, minerals and man-made components) needs to be resolved. Moreover, the identification of the various Optically Active Components (OAC) of sea water from their spectral signature remains a challenge, especially in turbid waters.

OSS2015 progress in these fields are the following:

a) A new model based on local empirical evidence, enabling to extrapolate the light attenuation depth profile from remote sensing data. The linear model is built on in-situ measurements. An experiment was carried out during the BP09 cruise in the Ligurian basin.

b) Improvement of the Gordon and Boynton (1997) method for getting sea water Inherent Optical Properties (back-scattering and absorption coefficients) from Apparent Optical Properties (ratio of upwelling to downwelling radiances R, attenuation coefficient Kv, and mean cosine µ). The inversion relied on an approximate solution to the equation instead of iterated applications of a forward model with adjusted trial phase functions, which is dramatically faster than Gordon's and has similar accuracy as evidenced by validation performed with the BP09 cruise measurements.

c) Calculation of Particle Back-scattering (bbp) coefficients by a new approach based on Loisel et al. 2006, which retrieves independently the coefficients at several wavelengths.

2.2 Bio-optical models design and validation

Recently developed algorithms were implemented and validated for the generation and delivery of EO-derived information:
* Chlorophyll concentration: the CIA algorithm of Hu et al. 2012 provides improvements over the standard OC4 algorithm for low values of the chlorophyll concentration.
* Particle size distribution is in turn retrieved from bbp (Loisel et al., 2006 and Kostadinov et al. 2009). Particle size distribution provides new information on the marine phytoplankton ecosystem.
* Phytoplankton groups (Spectral based approach (PHYSAT), Abundance based approach (Uitz et al.)
* Surface and column-integrated Particulate Organic Carbon (POC, Stramski et al. and Duforêt-Gaurier et al. 2010). POC is a relevant quantity to evaluate the contribution of the Ocean to the global CO2 budget.
* Net Primary Production (Morel-based algorithm, and Uitz et al. algorithm), useful to evaluate the productivity of the ecosystem, with application for climate change monitoring, and for fish and shellfish stock evaluation.


2.3 Optical Data Assimilation in theoretical bio-chemical models to map biochemical variables in the upper layers of the ocean

2.3.1 A demonstration in the Ligurian Sea

Statistical (quasi-)equilibrium of the ocean life, i.e. seasonal, might be represented by bio-optical models, yet the dynamics cannot be represented without equations of motions in bio-chemical models: optical data are linked to bio-chemical properties of the ocean, which obey dynamics equations instead of static equations.

Data assimilation in bio-chemical models allows:
o to fill satellite observation gaps (due to cloud cover) and
o to extend surface-based remote sensing data to the ocean upper layers and to derive bio-chemical variables, in space and time.
In addition, such a model is able to predict the future evolution of the Ocean.

The OSS2015 team has conducted a data assimilation experiment and in due course learned how to run the models, to optimise the data sampling (whether EO or in-situ measurements), to define initialisation procedures and boundary conditions. A "full-scale" demonstration was done using both remote sensing and in-situ data in the Ligurian pilot test.

HOPS, the Harvard Ocean Prediction System, a state-of-the-art, four-dimensional, bio-geochemical model was set-up and calibrated for the test site. The model bio-chemical variables include phytoplankton, zooplankton, detritus, nitrate, ammonium and chlorophyll-a concentrations. It is coupled to a physical model and statistical models.

The model was run in near real-time throughout the cruise and other data (satellite and in-situ data from CTD, glider, scanfish) when available were prepared and assimilated. This made it possible to carry out real-time forecast as realistic as possible. Before any new data assimilation, model predictions were compared with observations, thus model skill was estimated. The model predictive skills in the Liguria Sea were within 3 to 6 days.

OSS2015 demonstrated that real-time forecast of biogeochemical variables, using satellite and in-situ data, was feasible as a COPERNICUS service. The results are reported in Deliverable D25 and have been presented at the 7th EuroGOOS conference.

2.3.2 Assimilation of Nitrate profiles in bio-chemical models

A preparatory research has been conducted for data assimilation of Nitrate profiles acquired from Bio-Argo floats (i.e. the equivalent, for biogeochemical parameters, of the global Argo network, one of the most relevant sources of data for the oceanic operational systems). Nitrates and Light are the main factors driving phytoplankton growth. Whereas the available light can be determined with enough accuracy from remote-sensing data, reliable nitrate concentration data in the open ocean were until recently very scarce. The Bio-Argo floats provide therefore an interesting perspective for an improvement of bio-chemical model prediction skills.

As a first step toward nitrate data assimilation, the Chlorophyll-a (CHL) and Nitrate (NO3) concentration profiles measured by the Bio-Argo floats were compared to the outputs of the operational bio-geochemical model of MyOcean. The comparison exercise highlighted significant discrepancies between the model and the observations, especially on Nitrate profiles. Although the qualitative behaviour is generally in agreement, the model does not predict accurately the nutricline depth and the steepness.

This research led to the following conclusions and outlooks:
* Data assimilation of nitrate profiles is expected to have a significant impact on the model
* Calibration and quality control of Bio-Argo nitrate measurement should at the same time be consolidated in order to provide systematically reliable data.

3 Applications in geography, ecology and climatology

3.1 Bio-regions and application to optimization of in-situ measurements

The segmentation of the global ocean in bio-regions (ecosystems) allows to better model the climate modes' variability, and, practically, to optimise the in-situ and EO sampling strategy.

a- The geographic approach
Intra-annual variability is the first index to be used the definition of bio-regions (d'Ortenzio and d'Alcala, 2009, for the Mediterranean Sea, DR09), by clustering of Ocean Colour (OC) time series. The second index is the phytoplankton evolution patterns which occur most but not all years (inter-annual variability), allowing a more complete classification of the Mediterranean regions.

b- Sampling optimisation : definition of "Lagrangian" bio-regions
The method is based on the approach of DR09 to identify clusters in Ocean Colour time series. However, instead of applying it on a pixel-by-pixel basis, we derive time series by sampling OC images along trajectories obtained from model simulations. These trajectories mimic the behaviour of a Bio-Argo profiling float and they are dispersed following the flow velocities calculated by a model. Every 5 days, we sampled OC corresponding satellite images over the actual position of the numerical particles, to extract CHL (i.e. match up analysis). To each particle is associated a CHL time series over one year, and the ensemble of time series are finally grouped with the DR09 cluster analysis, giving a bio-regionalisation of the area. This method has been successfully implemented in the North Atlantic.

One application of the new bio-regions is to support the optimization of in-situ sampling strategies. Indeed, it has been shown that the trajectories of a float remain generally inside the same bio-region after deployment. This is especially true for Lagrangian bio-regions in the North-Atlantic. Floats deployed inside the same bio-regions are likely to provide similar measurements time series. The deployment strategy should therefore try to sample evenly each bio-region and avoid redundant sampling of the same bio-region. In addition, measurements inside intermittent bio-regions in the Mediterranean (occurring most but not all years) are particularly valuable. An optimal sampling strategy should put a specific focus on these regions.

3.2 Global nutrient concentration mapping

A new algorithm has been developed for estimating surface nitrate concentration from satellite and operational-model derived data. The algorithm employs local multiple linear regressions of observed nitrate concentration vs. satellite-derived surface temperature and chlorophyll concentration and mixed-layer depth obtained from an operational ocean-circulation model. The main findings of this research are:
* The new algorithm has been calibrated and validated with independent datasets for different time periods.
* While seasonal variability is somewhat underestimated, the algorithm successfully reproduces observed inter-annual variability in surface nitrate concentration at time-series sites, even for values well outside the range of the calibration data.

3.3 Satellite derived global time series

A major achievement of the OSS2015 project is the delivery of 15-year time series of new Earth Observation bio-optical information. These new EO products are available on the webserver of the OSS2015 service line.

OS2015 made significant progress regarding the methodology for the detection of trends in time series of Ocean products such as Chlorophyll and temperature. In particular, the impact of the choice of the binning method and temporal period has been analysed with an appropriate statistical tool.

The developed approach analysed the sensitivity of the trend detection over decades (1997 to 2010) to the choice of the binning period and the Chlorophyll algorithm (OC4, CIA or GSM). Although the Chl-a monthly mean-quantities tested are different, similar global patterns are observed for all of the algorithms. Results in long term analysis are relatively stable over time. Similarly, the choice of the temporal bin product does not affect the relative long significant trend estimation. This shows the robustness of the detected trend in Chl-a.

3.4 Bio-chemical modelling

The optimality-based variable-stoichiometry phytoplankton model of GEOMAR has been the focus of further developments in the context of OSS2015. These developments contribute to better understand and predict the behaviour of the "living Ocean".

The phytoplankton model has been calibrated to represent an 'average' phytoplankton community and solved for ambient conditions derived from satellite and World Ocean Atlas data in order to identify the main factors limiting phytoplankton growth. More precisely, the purpose is to determine quantitatively for each season and each region of the global Ocean, how far phytoplankton growth is limited by the availability of light, phosphorus (P) and nitrogen (N). The main output of this research is to produce monthly maps identifying limiting factors, as aggregated into seasonal averages. The main conclusions are:
* Northern high latitudes display seasonal variability of N-light co-limitation, with predominant N limitation during July-September.
* P limitation is negligible in most of the ocean. The Southern Ocean appears as a mainly light limited area.

The model was calibrated with in-situ measurements in the North Atlantic:
* Chlorophyll and Nitrate profiles from bio-Argo floats
* Zooplankton measurements

Model calibration led to a significant improvement of the agreement with observations. The bio-Argo data provide strong constraints on some of the phytoplankton parameters driving the chlorophyll distribution. Owing to the strong influence of zooplankton parameters on overall model behaviour, additional constraints in the form of older zooplankton measurements proved useful, particularly because of the long time period covered by these data. Taking advantage of the float measurement profiles was more difficult due to a mismatch with the model temporal and spatial range and resolution. A further comparison/assimilation exercise with a 3D model and/or multi-annual time series should be valuable.

4 Prototyping of services

4.1 A consumer/user/customer-oriented effort

4.1.1 User needs and the Socio-economic analysis

The OSS2015 effort was not only triggered by the EC space & environment policies (COPERNICUS, WFD, MSFD...) but by a mismatch between environmental/climate and fisheries scientists' requirements and the perceived requirements identified by the producers of oceanographic data products. There is a clear need for
o integrated coupled analysis from chemical and biological oceanography,
o high-quality time series (historic data) more than real-time and short-term forecasts
The project has been fed by interactions with users, through surveys, questionnaires, workshops and statistical analysis. User's needs and requirements have been compiled and compared to the service offered by the developed prototypes. In the course of the OSS2015 project, two user workshops were organised, as well as contacts with users and stakeholders, and a gathering of the leaders of other FP7 projects. Finally, the perspective for future developments to better fulfil the user's needs has been derived.

In parallel to the scientific and technical activities to fulfil known 'requirements' or requests, there was a real questioning and a dedicated analysis of the socio-economic impact that the projects outcomes (assimilation techniques, modelling improvement, news products availability, service-on demand ...) might generate after its completion. Two lines of socio-economic impact have been assessed by Frontier Economics:
* the "academic" one, which implies benefit in terms of knowledge improvement and better research for public interest and
* a second line which is oriented towards economical incomes and that should ensure sustainability of services.

For the latter, the potential benefit of the new service lines on specific concrete cases (e.g. salmon's disease, algal bloom impact on fisheries) was assessed. This study highlighted the important potential cost savings related to an improved forecasting accuracy.

4.1.2 Service-Oriented Architecture

In line with this customer-orientation of the R&D efforts, the information's production and delivery system has been designed according to the "service orientation" paradigm: computation and data storage on distributed systems implies the integration of different pieces of software which exchange data and control parameters. Most of the users/consumers of the academic world are also partners or suppliers of EO data, of bio-optical & biochemical models, of in-situ data and expertise.

4.2 In situ and satellite data management

Remote sensing requires the management of a large number of data (storage, share, processing and further dissemination). To fulfil this requirement of internal needs (data share between partners and validation), and in preparation of a service deployment, several actions have been undertaken:
a. A geodatabase with an easy-to-use search interface has been made available to the team partners
b. The INSPIRE Directive principles in regard to in-situ data management have been implemented.
c. A database of all EO products required by the team has been made available to the team partners, including an up-to-date reprocessed data set of MERIS, MODIS, SeaWiFS
d. A link has been established with the MERMAID database.

An easy-to-use satellite data access service line provides users with:
a. An access to reprocessed data set of Ocean Colour products (the GlobColour data set), providing in particular VIIRS products and Europe 1 km resolution products
b. The extension of the data set to the new OSS2015 marine products
c. A new interface with improved data access functions

4.3 Platform of "on-demand" services

A collaborative working environment providing algorithm development and processing and validation functions has been developed. This "Platform as a Service" approach brings innovation in the field of Earth Observation data processing and seems well adapted to support future research and development activities for research institutions or businesses.

5 Outreach and scientific networking

5.1 Outreach to COPERNICUS

Links with the Marine Core Service (MCS) MyOcean have been maintained throughout the project. OSS2015 contributed to the evaluation of MyOcean-2 scientific activities and to the organisation of the Science Days (October 2014). Recommendations towards MCS have been discussed at the occasion of the EuroGOOS meeting in Lisbon (November 2014):

[R1] assimilation methods and parameterisations to be "adapted" on the basis of the homogenous behaviours of the bioregions.
[R2] BioArgo data and Ocean Colour observations to be used in strong complementarity. Nota: use of bioregions (or of similar methods to group in situ observations) should provide a suitable tool to increase this complementarity, in particular for the chlorophyll concentration parameter;
[R3] NO3 Bio-Argo data availability should be encouraged and reinforced, as it is likely the most crucial parameter to constrain CPBMs;
[R4] the Quality Control of this data should be considered as high priority;
[R5] Assimilation techniques to be driven by temporal and spatial characteristics of the chlorophyll and NO3 fields, as highlighted in the OSS2015 analysis;
[R6] Maintain access to time-series of Chlorophyll-concentration, Particle back-scattering coefficient and size distribution, Phytoplankton Functional Types, Net Primary Production and Particulate Organic Carbon concentration, along with the connection with users to improve access and get feedback (evaluation and validation) on the datasets,
[R7] Update (say every one year) the datasets for all variables and make the connection with upcoming pair of Sentinel-3,
[R8] Real-time automated calibration system needs to be developed and very careful inter-calibration of all sensors should be done prior to use.
[R9] Realistic estimates of water clarity by the ocean colour satellite sensors to be used in both the physical model to improve estimates of MLD, but also in the ecosystem model in addition to NO3 information for estimating phytoplankton growth.
[R10] Real need for nitrate assimilation from bio-profilers
[R11] Use the 3 following types of metrics for the near future data analysis in MCS: classification, time series decomposition and trend analysis, oceanic provinces,
[R12] the MCS to issue statements of validity for each delivered product or data set,
[R13] A collaborative platform to be maintained and the MCS to ensure input data availability for the benefit of the whole chain of expertise (from data centre to users) -this also covers links with other data centres like SeaDataNet, Emodnet and EuroGOOS'.

5.2 Cross-breeding with other EU R&D projects

A white paper on the application of Copernicus to coastal zone services was written together with other FP7 projects. This white paper has been presented at the Copernicus user forum in March 2014 and made available through the OSS2015 website.

5.3 Scientific networking and outreach

Outreach toward the Ocean science community has been developed thanks to publications (4 peer-reviewed papers accepted during the project's duration, and three more in preparation), communications, and the organization of dedicated workshops and discussion sessions.

The OSS2015 project issued recommendations aimed at other scientific communities: recommendations for future spatial Earth Observation missions and Climate Change studies.

An effort to reach a larger public has been followed by the development of the project's website and through communications targeted toward policy makers and businesses, in particular in fish and shellfish farming industry.

OSS2015 also contributed to educational and cultural initiatives such as:
* "Mon Ocean et moi" educational website
* "Fête de la science" at the LOV
* "Adopt-a-float" program targeted toward French high-school children

Potential Impact:
1 Promotion of ocean "health" monitoring

1.1 Further to MyOcean -for the success of the first batch of COPERNICUS services

"Operational oceanography", core to the COPERNICUS Marine Service, is very close to meteorology -retrieval and forecasting of the physical properties of the ocean in Near Real Time (NRT) for ocean-going decision making, and to climatology -assessment and understanding of the trends in thermodynamic changes of the environment. Like meteorology and its side branch of aeronomy which deals with chemistry of the atmosphere as it impacts the atmosphere dynamics, there is a chemistry branch though chemical phenomena in the ocean are simple (evaporation & precipitation) and operational oceanography mainly deals with advection & mixing. What is lacking is the mapping of the marine biology and the forecast of its dynamics ; oceanic primary production accounts for half of the carbon fixation carried out on Earth (synthesis of organic and inorganic carbon compounds by phytoplankton, -in particular cyanobacteria and eukaryotes, via photosynthesis in the uppermost sunlit layers of the ocean, which is used in the food chain). For the mitigation of climate change, the ocean with its marine life has the same importance as land vegetation.

The marine biochemical information currently delivered by the precursor MCS, i.e. theoretical models outputs (statics/equation of optical states, dynamics or thermodynamics), is quite limited. OSS2015 allows going a step further by delivering information on the higher trophic levels of biomass, starting forecasts at various scales (from a few days to decades).

1.2 Fulfilment of environmental policies' requirements -the decision makers' needs

In recent years the protection of the marine environment has moved away from traditional concerns about pollution (introduction of external elements of anthropogenic origin, transport and evolution) towards protection of ecosystems, habitats and species occurring in the offshore area, which implies the collection of information on the aforesaid ecosystems more than their physical environment (temperature, salinity, currents).

The valuation of ecosystems comes from trade-offs by governments in allocating natural resources or stopping their use or preventing pollutions, stresses, etc. according to contradictory needs of environmental protection (conservation) and environment's exploitation to accommodate the population's increase and the growing expectation of mankind for well-being. Accounting for nature's capital in the decisions of individuals, communities, corporations, and governments is often too late, because undervalued, as the case of green tides in North Britany, or recognized only after they have been lost as was the case following the Amoco Cadiz, Erika or Prestige tankers' wreckage or following hurricanes, when huge restoration effort is required, as was the case of cod overfishing in Newfoundland, tuna depletion in the Med, or when biodiversity loss can't be overcome. Assets embodied in ecosystems are poorly understood, rarely monitored and can undergo rapid degradation without recourse.

Information about ecosystems' change in values is required by institutions in charge of policy- and decision-making; ecosystem parameters are fed into models to assess the impact on the value of change of the ecosystems' services to society and economy.

In this context, the Marine Strategy Framework Directive (MSFD) is the main pillar of the European marine service's need, targeting the conservation and sustainable management of marine ecosystems. As a framework directive, it seeks to do this by coordinating other policies to deliver a common objective of reaching or maintaining 'Good Environmental Status' by 2020 across Europe's marine environment:
- water eutrophication is minimised, alteration of hydrographical conditions and marine litter do not adversely affect the ecosystem, the sea floor integrity ensures functioning of the ecosystem, and concentrations of contaminants give no effects
- introduction of energy (including underwater noise) does not adversely affect the ecosystem
- the population of commercial (shell & fin)fish species is healthy, elements of food webs ensure long-term abundance and reproduction, and contaminants in seafood are below safe
- biodiversity is maintained, and non-indigenous species do not adversely alter the ecosystem

The OSS2015 contribution targets the set-up of a COPERNICUS service component dedicated to water quality and benthic + pelagic ecosystem's health. The service would rely on fluid dynamics and ecology (habitats and ecoregions) to deliver relevant bio-geochemical information to ocean users, policy makers and policy enforcers.

2 Scientific communities cooperation

Because of the thick corpus of theoretical knowledge and related research methodologies which has grown over the years with the expansion of public research agencies & university labs and the strengthening of research departments of governmental agencies for expertise maintenance, further to the split with governments' executive services, the academic science community is split in thematic areas, not to mention the dichotomies between engineering and knowledge-building, between pure and applied science, or between data-driven discovery and hypothesis tests (both based on evidence) and inductive-driven mathematical methods (based on inference). Each community of experts is self-contained with own rules, labs, events, journals & proceedings to facilitate knowledge/knowhow building, dissemination and preservation, and staff training. As such it might lead to inconsistencies between research outputs from different communities, further hampering progress.

Because targeting a specific goal of COPERNICUS service design, OSS2105 brought together Ocean imaging spectroscopy (remote sensing specialists), ocean surveyors (oceanographers and their engineers), specialists in marine organisms and ecosystem dynamics, data scientists, climatologists, numerical analysts and modellers, and ITC experts.

Progress could not be done without pollination from farther disciplines. E.g. remote sensing of the marine ecosystems relates to land cover surveys, except that it introduces another depth dimension (z) as light penetrates in the water and brings life below the sea surface.

3 Virtualisation of the R&D in the innovation process

Because of its multidisciplinary character vs. the physical partition of specialised labs, OSS2015 was driven to the set-up of a collaborative data retrieval/ analysis/ and computing platform to facilitate exchanges between stakeholders. This platform sides with an information production and distribution line.

The platform, i.e. a Collaborative Working Environment (CWE) as per the best ICTs, acts as a virtual R&D centre in support of the COPERNICUS service (as if an adjacent research centre was built in the COPERNICUS Marine service facility, relying on a network of partner scientists working on a collaborative platform, sharing results for permanent peer-validation, devising standards and updating the data processors.

The collaborative platform is an on-demand facility which allows partners:
1. to access algorithms, documentation, data & data bases and software,
2. to run
- proprietary software with proprietary data on the server (by cloning tasks onto multiple virtual machines to accommodate as many users as possible),
- proprietary software with their own data,
- users' software or versions of the proprietary software that has been updated by the users
3. to validate results
4. to monitor the progress-of-work, to share data, to build standards, etc.

4 Towards maturity - focus on information reliability

Within OSS2015, remote sensing of marine biochemical properties leaves the labs to enter the socio-economic policies' arena (from understanding the Earth behaviour to responsibility): policy makers & enforcers need reliable information and checking means. Uncertainties of estimates shall be reliable to relevantly assess marine ecosystems health and to compute credible fluxes (differentials between states) between biologically mediated carbon boxes.

Validation of information is an ethical, commercial and regulatory obligation: ethical because it shall be good science and fit-for-purpose on customer's/user's behalf; commercial because of due care in products' liability; regulatory because it can be used in cases with legal implications of a bad product's use (consistent application of methods and comparability when applied to different data). Difficulty stems from the introduction of uncertainties at each step of information production, when assimilating optical indexes in models: from the in-situ measurements or the measurements from sensors on-board satellites, the sensors' calibration, the corrections, the data fusion when possible, the data transformation in information through equations of states or differential equations, etc.

Yet, a "statement of validity" shall be attached to methods and information that is delivered to users/ customers, as the case in pharmacology. This statement of validity is a strong recommendations of OSS2015 project to the upcoming version of the MCS.

List of Websites:
Public website: www.oss2015.eu

Contact:
Dr. Odile Fanton d'Andon
Tel: +33 (0) 492 96 75 03
Email: oha@acri-st.fr