Skip to main content

Coastal Biomass Observatory Services

Final Report Summary - COBIOS (Coastal Biomass Observatory Services)

Executive Summary:
The focus of the CoBiOS project was on the monitoring and prediction of eutrophication caused by nutrient pollution leading to high biomass algal blooms. High biomass algal bloom events occur each year in many places in the European waters, causing nuisance (smelly foam on beaches) and potentially dangerous situations when there are toxic species or when the biomass decays rapidly and sinks to the bottom to form pools of hypoxic matter. High biomass events often reach the public through newspapers and internet since they cause disruptions in the recreational use of beaches and coastal waters. This type of blooms influences the turn-over of fisheries and aqua-cultural operations in many ways, sometimes with large economic losses.

CoBiOS aimed to integrate satellite products and ecological models into a really operational and user-relevant information service on high biomass blooms in Europe’s coastal waters. Remote sensing can offer high quality harmonized Chl-a and Kd products including error statistics that can be used for the monitoring of bloom events. However, the observations lack complete coverage due to cloud interference. They also lack predictive value. During high biomass bloom events in coastal waters, the shape and location of blooms changes completely during a period of 2-3 weeks. Since blooms are brought to life by periods of sunshine, these are often but not always periods in which there is rather good coverage of satellite images. Therefore it is important to combine remote sensing products with hydro-ecological models.

Through theoretical analysis, a good understanding was achieved of the main parameter expressing biomass (Chlorophyll-a) and the various means of monitoring and modelling this parameter. The project has gained a deep understanding in how the parameter “transparency” is calculated in non-optical way in various ecological models and how this links to the optical descriptions of this parameter. Methods were proposed and implemented to drive ecological models with satellite observed (gap-filled) daily maps of transparency or its proxy “total suspended matter”. Using MERIS images, methods were designed to come to an ensemble mean Earth Observation product that significantly reduces uncertainty with respect to using a single map/method. Trials were successfully executed to test the existing and improved methods and models. In general providing TSM information to ecological models significantly improves the description of the underwater light climate and the predicted Chlorophyll-a. Modelled biomass development predictability was investigated under permutations of wind-fields, nutrients, model parameters, riverine inputs etc. Validation data was collected mainly from buoys and Ferrybox systems and used, together with consolidated validation methodologies, to validated improved outcomes of satellite products and ecological model outputs. After successful adaptation of all services to MODIS, four operational service lines were put into place. In each service line an Earth Observation service provides daily data to an ecological model service. Monitoring and modelling results are automatically placed on the CoBiOS webportal (http://cobios.waterinsight.nl) to give users an overview of past and current events and predictions of 3 days in the future. Based on the portal info, early warning bulletins can be generated together with time series plots and longer term animations. The system was successfully demonstrated to a number of users (from various stakeholder communities) who expressed the interest the use of the system after the lifetime of the project. CoBiOS partners participated and contributed to various coastal waters oriented European services discussions and contributed to a white paper expressing the need for a pan-European Coastal waters monitoring concept using satellite observations.

Project Context and Objectives:
The focus of CoBiOS is on the expression of eutrophication caused by nutrient pollution leading to high biomass algal blooms. High biomass algal bloom events occur each year in many places in the European waters, causing nuisance (smelly foam on beaches) and potentially dangerous situations when there are toxic species or when the biomass decays rapidly and sinks to the bottom to form pools of hypoxic matter. High biomass events often reach the public through newspapers and internet since they cause disruptions in the recreational use of beaches and coastal waters. This type of blooms influences the turn-over of fisheries and aqua-cultural operations in many ways, sometimes with large economic losses. CoBiOS aims to integrate satellite products and ecological models into a really operational and user-relevant information service on high biomass blooms in Europe’s coastal waters.

Remote sensing can offer high quality harmonized Chl-a and Kd products including error statistics that can be used for the monitoring of bloom events. However, the observations lack complete coverage due to cloud interference. They also lack predictive value. During high biomass bloom events in coastal waters, the shape and location of blooms changes completely during a period of 2-3 weeks. Since blooms are brought to life by periods of sunshine, these are often but not always periods in which there is rather good coverage of satellite images. Therefore it is important to combine remote sensing products with hydro-ecological models.

Main objectives:

1) To collect, review and consolidate existing knowledge and data related to optical measurements and modelling parameters of coastal waters biomass development.
2) To define and analyze methods to improve the underlying information production lines of CoBiOS. On the one hand we intend to improve the quality of Earth observation products by incorporating error statistics products and to expand the portfolio of EO-products with harmonized water transparency products which are suitable to drive ecological models. On the other hand we will expand the capability of ecological models with methods to predict relevant additional information such as: biomass transport vectors and growth/decay rates. We will study the best schematization to use EO-transparency data to force ecological models and we will study methods to compare model results in terms of Chl-a concentration with satellite observations of the same parameter.
3) To implement the improvements in models and EO-products and to test the enhanced output quality by running two series of hind cast trials: one with the original models and one with the improved models (forced by EO-transparency data).
4) To collect relevant in-situ validation data during the Near Real Time operational phase and to actively engage key-users of the CoBiOS services in the validation of the service performance and the quality of the EO-data products and model information products.
5) To demonstrate the novel CoBiOS information system by operationally running the models and EO products service chains during an extended Near Real Time demonstration phase.
6) To set-up and fill the CoBiOS web-portal with maps of high biomass algal bloom events.
7) To communicate the extent of the events by means of early warning bulletins to professional users.

The CoBiOS project is organized in the following Work Packages:

WP1: management
WP2: preparation and consolidation of data, protocols, validation formats etc.
WP3: theoretical development of EO-transparency products, model interfaces and assimilation methods
WP4: Implementation of tools and methods, testing and trial runs
WP5: Validation activities, including the collection of in-situ validation data during the demonstration trials
WP6: demonstration runs of the operational services
WP7: dissemination and service sustainability development

CoBiOS has been set up to achieve the objectives in 36 months. An incremental and iterative approach is implemented to ensure that, at the end, CoBiOS results perfectly fulfil the initial objectives. To this aim the project is split into 6 phases as follows:

Phase 1: Preparation (3 months)
Phase 2: R&D support to CoBiOS Services development processes (12 months)
Phase 3: Implementation of tools and methods (6 months)
Phase 4: First and second end-to-end trial and validation of results (6 months)
Phase 5: Third, operational demonstration trial and validation (6 months)
Phase 6: Conclusion of earlier started dissemination and service sustainability activities

Project Results:
During the first year of CoBiOS an inventory was made of available resources (satellite images, models, methods for Kd estimation from MERIS data, etc.). An overview was made of available satellite data (MERIS and MODIS in various levels of processing). At Brockmann Consult an ftp server and the MERCI system were opened for satellite data retrieval by project partners. Also at
Brockmann Consult the CoastColour archive was opened for the provision of (full resolution) satellite data. An overview of relevant MyOcean data has been made and set of initial procedures to enable data exchange and file access from MyOcean to CoBiOS partners has been established. This allows each partner to prepare conversion tools. Together with the other CoBiOS partners, the formats and initial interfaces to the CoBiOS end user portal have been defined and a start was made with the architecture design of the CoBiOS portal. For the purpose of harmonization of data and to understand the differences between e.g. Chl-a derived from satellite observations and Chl-a measured in-situ, a large number of national and international measurement protocols was collected, evaluated and archived. Many sources of in-situ data have been inventoried, including National Monitoring Programs, Ferrybox, SmartBuoys, and Open Repositories. The Catalogue provides information about the available in situ-measured parameters, the locations and time spans of measurement; it includes an estimate of the number of measurements per year, and an estimate of the data quality and quality check procedure. The source and availability of the data is always indicated, with contact persons and acknowledgement indications. A Validation Board has been introduced and early users’ recommendations have been taken into account. To provide user relevant services, CoBiOS has made an early start with the identification of user requirements. A major innovation in CoBiOS is to use EO products to introduce the vertical diffuse attenuation coefficient (Kd) into the ecosystem models. Therefore a study of current methods was conducted resulting in a report describing the theoretical best-practice to derive a Kd product directly from EO data, which may serve as a basis for further algorithm testing and development.

An important development during the second year of the project was the departure of Meris. This meant that we had to change the focus to operational use of MODIS. First the experiments were completed studying the harmonisation of satellite products. For MERIS images we have decided to use a standard for radiometric correction and output file formats. The results of several algorithms were bundled into one file for a number of days and a number of regions to allow us to come to an uncertainty product. Some of the algorithms were re-parameterized to adapt to local conditions. Using a number of different approaches, we have developed several KDPAR products. For MODIS we did the same but the number of proprietary algorithms is less and the quality of the results is somewhat affected by noise. Fortunately we have been able to develop algorithms for MODIS that provide approximately the same accuracy for total suspended matter and chlorophyll in the North Sea. In the Baltic Sea artifacts show up where high sediment areas are falsely recognized as phytoplankton blooms. The ecological models were tested for their ability to provide probability numbers in cases where we permutated wind speed and direction, various amounts of run-off, various amounts of nutrients and some intrinsic model parameters. An index was developed to assess the effect of the permutations on the dynamics of phytoplankton development. We have come to realize that the parameter Kd has different meanings in models and from satellite observations. In ecological models Kd is calculated from the components carbon, mineral particles and salinity. In remote sensing terms Kd is defined by optical properties. We have been able to bring the knowledge of optical properties to the coefficients that determine the Kd in the ecological models. Based on an extensive survey we have defined a set of methods to evaluate the difference between model results without remote sensing, and models results where the model was driven by remote sensing input. Software was developed to integrate the new methods in the existing processing chains of the service providers. From initial model experiments it was concluded that a number of parameters determine the prediction of the onset and duration of phytoplankton blooms. For example long-term variation in the wind fields and the initial settings of internal model parameters have large effects on the predictions.
In the third year all the service lines were developed towards operational status. Because MERIS-Kd was no longer an option, some of the models were driven alternatively by daily TSM fields derived from MODIS. Gapfilling was performed using the DINEOF method and software, implemented to the systems of the various partners. Especially for winter months the surface visibility frequency is very low, so the DINEOF method needed some adjustment to produce reasonable results for those months. To present the results from EO and models in a sensible way to users various tools were developed:
1) Early warning bulletins, showing daily results together with 2 previous days and predictions for 3 consecutive days in the future. Users found the bulletins very usefull, especially if they are annotated giving some expert judgment comments and interpretation guides.
2) The portal website was developed containing a side-by-side view of 2 user defined maps. The user can select Chlorophyll-a observations for one of the regions from one of the providers and/or model predictions for the region from one of the providers. The portal contains relevant background information and fact-sheets for all the services. A backbone system was set-up to receive and process the results from service providers.
3) Movies are being produced using an automatic movie generator tool to allow the visualization of processes during longer periods in the past.
4) Time series plots are generated and show to the user per OSPAR or HELCOM box.

After the production of systems and services, users were approached to test and use the systems for a certain period and provide feedback. To collect relevant feedback two questionnaires were developed (one of which on-line). In general the users appreciate the CoBiOS service, are interested in continuation after the project and some are interested to purchase services. From discussions with users from HELCOM and OSPAR it has become apparent that some of the CoBiOS products play a significant role in the monitoring and reporting of these gremia. Especially Chlorophyll-a as indicator for eutrophication and transparency measurements link to ongoing monitoring and decision making. The products of the service lines were validated against e.g. available in-situ data using consolidated validation measures and methods. Cross validation between satellite Chl and model Chl was performed with variable results. The performance of the service lines and systems was also validated. In general delivery within the day was achieved, with sometimes short interruptions because of hardware maintenance or software updates. A non-service period of 14 days was caused by the shutdown of the US administration which caused unavailability of MODIS data. We think that it should be noted that, although not asked in the questionnaire, we have the impression that National users are tempted to wait to see what form of services MyOcean can offer in this respect, because these inherently will be free of charge. The consortium has been very careful not to express negative opinions about MyOcean products to the National users in our area, but our own analysis shows that our models in general do not benefit from MyOcean met-ocean products while our EO products are suitable for the Northern European waters where the MyOcean products are not. Extreme care has to be taken that the anticipation on MyOcean operational products is not going to disturb the market and the user base created by projects such as CoBiOS. Although the project started with an orientation on MERIS, the projectteam has successfully demonstrated that biomass development in coastal areas can be monitored and modelled using MODIS-AQUA data as well. This provides an extra perspective on long-term service continuity since also succeeding sensors such as VIIRS and Sentinel-3 will be able to provide relevant data using the CoBiOS methods.

Potential Impact:
High biomass algal bloom events occur each year in many places in the European waters, causing nuisance (smelly foam on beaches) and potentially dangerous situations when there are toxic species or when the biomass decays rapidly and sinks to the bottom to form pools of hypoxic matter. These pools can resurface and form dead zones where massive marine life mortality occurs. On the other hand, it may be profitable in the future to monitor high biomass blooms because they could be harvested as biofuel or as fertilizer (since the world stock of phosphates is decreasing rapidly). High biomass events often reach the public through newspapers and internet since they cause disruptions in the recreational use of beaches and coastal waters. This type of blooms influences the turn-over of fisheries and aqua-cultural operations in many ways, sometimes with large economical losses.
The CoBiOS project aims to integrate satellite products and ecological models into a really operational and user-relevant information service on high biomass blooms in Europe’s coastal waters. The service aims to reduce economic losses by giving timely warnings for high algal biomass development and predictions on algal biomass movement and fate. CoBiOS will provide strategic information on which new biomass harvesting initiatives can be based.

The ecological state of surface water can be read from several simple measurements and indicators such as the concentration of Chlorophyll-a (Chl-a) as proxy for algal biomass and Secchi disk depth, suspended matter concentration or turbidity, all as measurement or proxy for water transparency. In-situ measurements are expensive, time consuming (and therefore unrepresentative for the natural variability in large coastal waters) and unsuitable to monitor changes in NRT to allow for timely warning for events of foam formation, oxygen depletion and dead zone formation.

Remote sensing can offer high quality harmonized Chl-a and Kd products including error statistics that can be used for the monitoring of bloom events. However, the observations lack complete coverage due to cloud interference. They also lack predictive value. Therefore it is important to combine remote sensing products with hydro-ecological models. Such an integration of information sources will provide novel and detailed information for many types of analyses and potential commercial spin-offs.

CoBiOS has established these information services which can be used to:

• Provide information on the state (EO) and evolution and fate (models) of near coastal high biomass blooms
• Evaluate and predict the probabilities of nuisance (foam, biofouling of off-shore installations and ships), harmfulness (red tide, fish kills etc), decayed biomass accumulation,
• Follow patterns of nutrients pollution (eutrophication)
• Predict (and ultimately prevent) potential hypoxia and anoxia events/locations that might lead to oceanic dead zones and massive fish kills.
• Predict when and where bloom harvesting would be economically feasible
• Indicate locations/periods where fishing would benefit from decreased transparency (resulting in netting invisibility)
• Indicate areas where blooms have lower than normal intensity due to pollution or variations in insolation due to climate change.

In a wider context, CoBiOS services will provide novel information to be used in the framework of EU directives, the most important of which is the Marine Strategy Directive – (EMS, 2008). The European Marine Strategy framework directive (EMS) aims to achieve good environmental status of the EU's marine waters by 2021 and to protect the resources upon which marine related economic and social activities depend. The Marine Strategy will constitute the environmental pillar of the future maritime policy from the European Commission, designed to achieve the full economic potential of oceans and seas in harmony with the marine environment. The other important EU directive that has links with the CoBiOS service products is the EC Bathing waters directive (2006).
The directive has requirements relevant to monitoring of bacteria, assessment (water quality evaluation), cyanobacterial risks and monitoring of other parameters such as the proliferation of macro-algae and/or marine phytoplankton. Regional Seas conventions such as OSPAR and HELCOM also have links with the CoBiOS services because they are concerned (amongst others) with pollution of the marine environment with nutrients and the consequent eutrophication.

CoBiOS results are being disseminated through the project webportal (cobios.waterinsight.nl) and have been presented on a number of user relevant meetings, such as meetings of HELCOM, OSPAR and EUROGOOS.

The consortium has decided, supported by the interest of key users, to keep the portal (and the services) alive for one year after the project to make the change from MODIS to Sentinel-3 based services. Together with related FP7 projects a white paper was written (currently send to National Delegates) to propose an operational coastal waters monitoring service using Copernicus satellites, data and infrastructures.

List of Websites:

public website: cobios.waterinsight.nl

Contact:

Dr. S.W.M. Peters

Water Insight bv

E: peters @ waterinsight.nl

M: +31 6 41903163

P: +31 317 210004