Gérard BEGNI, Deputy Manager, MEDIAS-France
Giovanni CANNIZZARO, EUFOREO project leader, TELESPAZIO
The so-called GMES (Global Monitoring for Environment and Security) initiative was born in mid-1998, when several space agencies and related organisations met together in Baveno and issued a Manifesto. This document highlighted the strategic importance of geographic information collected by Earth Observing systems to monitor our environment and contribute to our security in an independent and objective way. The European Commission, the European Space Agency and other organisations such as EUMETSAT and several national agencies further elaborated the idea. GMES aims now at helping Europe in realising its policy goals in various fields, such as monitoring the global and European environment, detecting natural catastrophes, development co-operation, civil protection including managing mass movements of refugees, the fight against fraud, etc.
GMES is highly embedded in the ERA (European Research Area). Together with GALILEO, GMES (Global Monitoring for Environment and Security) is one of the two pillars of the ESS (European Strategy for Space). GMES is now a key item of the Aeronautics and Space priority in the Community's Sixth Research Framework Programme (2002 - 2006). On October 29, 2002, Research Ministers of all Candidates Countries have signed a Memorandum of Understanding associating their countries with that Programme As the European Commissioner Philippe Busquin noted: "GMES is not only of strategic importance for space research, it will also help Europe to better project its values and policies in the world, for example to ensure sustainable development." European researchers, private companies and public authorities are expected to be major GMES partners..
At the global level, GMES will provide new verification tools to contribute to the precise monitoring of the implementation of international protocols, such as the Kyoto protocol on climate change, as well as security and international aid agreements. At the other end of the spectrum, GMES will help local authorities to pinpoint problems (e.g. shoreline erosion) and better react to catastrophic events (e.g. floods, mudslides, avalanches, forest fires). At EU level, beyond environment and security, GMES will provide new objective data to support of a broad range of Community policies - notably regional development, transport, agriculture, enlargement, development, and foreign policy etc. In all these areas, the 6th Framework Programme will bring players together, combine their needs and federate their efforts.
This ambitious programme aims at consolidating a European capability in a sector where Europe often depends on data provided from third sources In particular, the GMES initiative objective is to federate Europe's disparate activities in satellite observation and remote sensing, in support of Community policies. Currently, space data for information on environment and security is derived largely from experimental or operational satellite systems that are develop either in the framework of the ESA Convention, or from national or multilateral programmes among Member States. ENVISAT (the largest Earth observation satellite ever built and a key component of GMES) and SPOT may be quoted as key examples. Collection of in situ information sources, modelling techniques allowing to make a joint use of heterogeneous datasets, and the development of a dedicated ground-based infrastructure are also a key GMES objective. Combining space, land-based and airborne technologies, GMES will help develop innovative tools and applications in order to assist decision and policy makers in the above-presented domains.
Considering the above facts and objectives, GMES is now developed as a joint initiative of the European Space Agency and the European Commission. Other national or international structures (EUMETSAT, national space and environment organisations) are involved through a Steering Committee and other relevant structures. In November 2001 the ESA Ministerial Council approved a new 5-year ESA programme dedicated to GMES, called the EarthWatch GMES Service Element (GSE for short). GSE will deliver policy-relevant services to end-users, primarily (but not exclusively) from Earth Observation sources, bringing a key contribution to the above-presented operational objectives. So, the European Commission, ESA and other national or European structures are already working closely together to make sure that their respective programmes will allow the research, development and operational user communities across Europe to unite in a co-ordinated effort to establish an autonomous European Global Monitoring capability for Environment and Security purposes by 2008.
In order to lead GMES, a three-strand action plan was endorsed, and proper management structures were set up. Nine thematic priorities were defined.
Among the objectives of the first strand or "Initial Period", GMES should be able to deliver pilot information products and services on priority environment and security topics, to assess current capabilities, and to define future systems infrastructure.
Following some preliminary informal works, the European Commission issued a first formal GMES call for tenders at the end of the 5th Research Framework Programme. Under the co-ordination of EARSC, the European Association of Remote Sensing Companies, represented by Telespazio, several research institutes (among which MEDIAS-France representing the MEDIAS network dedicated to global change research), service providers and users proposed to set up EUFOREO (EU FORum on EO use for Environment and Security), a Thematic Network (TN) endorsed and funded by the EC.
The objective of EUFOREO is to set up and demonstrate a European Forum linking the major Space Agencies, Research Centres, Service providers, Manufacturers and Users at both national and European levels. The purpose of the EUFOREO network is to develop the case for the efficient implementation of EU and national policies for Environment and Security (including International Treaties), through the use of Earth Observation. In so doing, it will assist in building the GMES Action Plan, accompanying and supporting GMES in the present initial phase. In particular, the role of EUFOREO is to involve a wider European range of participants in the discussion process leading to the definition of GMES. EUFOREO participants include 18 members, among which: ESA, JRC, EUMETSAT, EARSC and EUROSPACE. Eight EU Member States are directly represented. They will provide the views of GMES significant stakeholders on GMES organisation and implementation.
EUFOREO being a Thematic Network will mainly contribute to GMES Strand 2 or "assessment period" on some specific issues (obstacles, data and service provision) and market segments.
So, EUFOREO role in GMES is based on its strong points:
EUFOREO started as a concerted action along with the four following steps:
The first step led to a study on "Matching Demand and Supply for Earth Observation Data and Information" carried out by three independent experts. The study has confirmed that EO technologies have enormous potential as sources of data and as analytical frameworks for both research and policy on the environment. As yet, however, EO data are not achieving their full potential in these areas. In part the problems relate to the technical limitations and complexity of EO data. More generally, however, they derive from the low levels of EO expertise and awareness of potential users, and the poor understanding of user needs by data providers and the EO service sector. As a consequence, an effective market for EO data has not yet emerged.
The second step is a contribution to the selection of the most promising services with respect to the GMES priority themes, while the third step allows to analyse these services with respects to hindering factors evidenced by the first step. The results of that step can be considered as available now.
EUFOREO also plans to organise specific workshops or working sessions to collect users view on proposed priority GMES services based on EO data and information, to receive feedback from the users on the analysis of obstacles on the use on EO data and information as well as on the existing gaps between demand and supply.
This presentation is a part of that plan applied to Central Europe, in order to present some EUFOREO results and strengthen the links between GMES and the countries 'in accession phase', which are fully associated to the 6th Research Framework Programme, as explained above.
All the services addressed by EUFOREO cannot be addressed in a single presentation. Among them, considering the major environmental concerns faced by Central Europe, the following four categories appear to present a specific regional interest and so deserve a specific attention in the framework of the GIS OSTRAVA 2003 colloquium:
The presentation will briefly describe the results and preliminary conclusions of the three first EUFOREO methodological steps about the four themes listed above. Emphasis will be made on the use of EO derived products in combination with products from other sources.
The dramatic development of anthropic activities in Europe (industry, buildings, communication roads, uncontrolled leisure activities...) are a major threaten to our environment. Such environmental concerns are particularly severe in some industrialised areas of Central Europe. National, regional and international legislation and conventions have been set up to mitigate environment degradation.. At European level, some examples can be quoted: Bern Convention, EC Habitats Directive, Convention on Biological Diversity, EC Biodiversity Strategy (1998) and its four sectoral Biodiversity Action Plans (2001), EU Sustainable Development Strategy (2001) and the Sixth Environmental Action Programme (2001).
The service is planned to provide information to enable monitoring of specific environmental indicators connected with Biodiversity and Landscape protection.
The main users are National Administrations who are responsible for the correct management of natural resources and need to acquire, conform, update and integrate different information about their territory. At the regional and local levels, the potential number of users is high. International Bodies concerned with biodiversity and Nature Protection are also key interested parties.
Within the European Union, it is mandatory to have a unified approach on implementing the 6th EAP and Natura 2000, taking into account natural ecosystem stratification in Europe. Hence, there is a direct link to a consistent environmental reporting performed by EUROSTAT. The relevant Directives and Conventions are relatively recent. They call for a co-ordinated effort in data collection, analysis and exchange between all Countries. Analysing harmonising needs (and what role satellite information can play) and standardising methods is currently underway.
As a consequence, the service should calculate selected indicators in accordance with methodologies to be standardised in all EU countries The service should also promote the use of a common approach to calculate environmental indicators.
Users require information ranging from species distribution to landscape analysis and modelling. It is extremely important to highlight real potentialities of EO data in this connection, in order to avoid false expectations.
The use of satellite data provides a synoptic and objective spatial database to complement existing national inventories that are not fully comparable in terms of mapping standards, accuracy and classifications system. The representation of products in cartographic format is complemented by statistical attributes that are scaleable to different administrative levels. Information derived from satellites provides statistics and maps of land cover/use change. This has to be cross-checked with in-situ information to determine accuracy limits. Most elements of this have been demonstrated to individual users in the framework of pilot projects. The role of EO data is largely in determining "pressure" indicators (agriculture, urbanisation, deforestation, etc ...) from which "state" indicators must be derived through modelling using in situ data.
There is a need to improve the understanding of which indicators for biodiversity have elements which are remotely detectable to acceptable standards and accuracy (e.g. fragmentation and habitat indicators). Methods then need to be tested and validated on how this can be done.
The products (satellite classification maps, vegetation change maps, land cover maps, habitat maps) are generated using a combination of historic and recent satellite imagery, fieldwork results and ancillary data. Shape, fragmentation, location, connectivity and heterogeneity of habitat objects are detected and monitored, which represent recognised parameters for the definition of ecological quality of protection sites. In terms of "Loss of Biodiversity" the following certified pressure indicators have to be computed:
Some service components from satellite data are operationally available, technically feasible, ready and working.
The product components are based on proven data processing techniques and existing satellites missions and sensors such as SPOT Vegetation, and HRVIR, Landsat 7 ETM, MODIS, ASTER, IRS and IKONOS. Most can be obtained/purchased from data providers, but for higher spatial detail the costs per unit area increase significantly - too high for some users (local, regional and some national). Nevertheless, some RTD work is still required to improve mapping accuracy and class differentiation for habitat maps. Furthermore, access to validation and ancillary data held by regional or national providers has to be improved and harmonised.
The proposed service should replace step by step existing ones based on conventional methods even if use of EO data is still a research issue, being neither automated nor operational. In particular, for habitat fragmentation indicators, the proposed service could replace the current assessment of conservation status for NATURA 2000 sites and identifies hot spots of pressure towards sensitive semi-natural areas in Europe.
Some impact is likely in highlighting vulnerability and hence identify damage reduction activities. For habitat fragmentation in particular, the service may have a positive impact on the protection of biodiversity in Europe and contribute to a harmonisation of environmental databases especially taking into account the forthcoming enlargement of the European Union
Floods have always been a major natural hazard in Europe. Causes are both natural (rainfall pattern and variability) and anthropic (land cover changes, including urbanisation). In addition, legal or illegal building or settlements in more or less exposed areas increase the threats on human lives. Recent events show that these hazards are still a topical question, causing severe human and economic losses in Europe.
The potential flood risk service would be valuable at both national level and in providing cross-border flood risk assessment services, combining all the data available in involved countries. The application has mainly an applicability to specific areas at local and regional level and some at cross-boarder level (e.g. France/Germany Moselle basin, or the recent flooding in Czech Republic and Germany). The needs for information are largely well established and there is agreement on the parameters necessary and the accuracy/frequency desired.
The users of the EO derived information are the modellers - the outputs of the models are important for civil protection services, city management services and insurance companies.. Model development and validation are often at the research stage, depending on catchment and availability of ancillary data. So, information from EO data is an input to a model - it will never be a service on its own.
As a consequence, the prevention products (vulnerability, hazard and risk maps, land use changes, ground occupation, soil moisture) can be derived in part from EO data. This is particularly of value over large catchment areas and areas of poor accessibility - here costs per unit area are relatively low. The accuracy of measurement often falls short of requirements but still offers a significant improvement to the modelling process.
The methodologies used to obtain the EO contribution to these products are mature. Current EO capabilities are well suited to the provision of land related information at the scale required, such as land cover, land use, soil state and elevation. Nevertheless, the accuracy of elevation data needs to be improved (up to a 5 cm vertical accuracy) for flood risk assessment. Advanced modelling tools and non-space data are needed, particularly historical series of hydrological and meteorological data. Expertise is required for advanced products and models, such as risk map generation, and here there is still a significant R&D activity. Associated terrestrial infrastructures may be expanded in terms of in-situ measurements systems and sensors. Some devices already exist in some regions but their number must be increased and extended to all risky areas, even allowing near-real time transmission of measured parameters.
In Europe, forest fires are considered as a major threat mainly in the Mediterranean area, due to both climatic conditions and anthropic pressure. The need of risk assessment is particularly felt here, since it allows to avoid loosing forested surfaces prevent and associated soil loss and degradation phenomena. Nevertheless, that hazard does exist in the whole Europe, mainly under increasing anthropic pressure (transports, leisure...), extending far East up to the Siberian forests.
The risk assessment service aims at supplying risk maps on a daily basis to indicate the probability of fire ignition and spread in Europe. Currently, there are several fire risk indices which provide this service based on assessments of vegetation status together with meteorological, geomorphologic and statistic parameters. The risk maps can be produced at national or regional levels with a typical 1 km spatial resolution. This service provides an awareness and alert for fire prevention and assists in the planning and mobilisation of fire fighting resources. In addition to the indices, some GIS functions can assist in managing fire-risk areas, overlaying thematic layers such as administrative boundaries, land use, water resources, topography with information on access (roads, tracks, fire breaks, etc).
The forest damage assessment service provides maps and area measurements of the burnt area following a fire. High resolution satellite imagery before/after the event enable the perimeter of the area to be determined. A multidate comparison of vegetation status can help in a more accurate identification of partially burnt areas but at a coarser level of detail. Combined with ancillary data such as of boundaries of land ownership, this service can create a data base to assess and map burnt area damage, to monitor vegetation re-growth and to plan recovery activities in the affected areas. The information provided can be used to help prevent illegal housing and land misuse during regeneration.
Many national programmes include measures to prevent and mitigate forest fires damage. There are also clear needs to monitor burnt areas. Some new or refined needs for related estimates (e.g. biomass burnt, economic loss, ecological loss) are still to be consolidated. For instance, forest fire burnt areas should be accounted for in carbon balance for the Kyoto Protocol (see below). At EU level the risk assessment service may support the implementation of a number of Community legislation acts, concerning the set up of a framework for co-ordinating the initiatives at national level, while the burnt areas monitoring service is relevant for the 6th EAP and the CFSP.
Concerning fire-risk, currently as many as 6 indices are applied in Europe. The use of one consolidated index is not favoured - nations like to have visibility of several indices including their own and those of their neighbours. Discussions amongst member state agencies responsible for forest fires ensure a consistent method of calculation, so at European level they may be coherently adopted. Concerning burn areas monitoring, methods are largely proven and tested. Often detection is clear but problems will always occur when areas are only partially affected by burning, as underlined above.
Risk is assessed through static (or structural) indexes - e.g. vegetation type, topography - and dynamic indices (meteorological conditions and vegetation status, road-net density) . EO data can contribute to the vegetation part of both, including DEM for morphological analysis. The advantage over non-space technology is felt particularly in countries where meteorological local networks are less developed. Satellite data offers accessibility in isolated areas where fires usually occur. Concerning burnt areas monitoring, measurement accuracy can be improved when dealing with small parcels of land in variable terrain, such as by using a DEM, but cost is a factor. The time taken to obtain clear post-fire optical imagery is not usually outside the requirements.
So, all the basic product components are operationally available, technically feasible, ready and working. A daily risk assessment service is provided in a number of countries and by the JRC. The application uses existing technologies. No key R&D needs were evidenced. However, near real time fire risk models (in order to guide and streamline operative fire fights) need to consolidate the different solutions already developed by Value Adding Companies. The main effort to be led is in the economic domain : cost per unit area needs to be competitive with aerial survey for similar services in terms of efficiency and accuracy.
Other forest related services studied within the EUFOREO framework are monitoring global carbon mass within forests, and monitoring and Measuring ARD (Afforestation, Reforestation, Deforestation). Such services may have either scientific (study of carbon cycle) or economic (public and private forest management) applications. Nevertheless, monitoring the clauses of the Kyoto Protocol are a key need that justify such services. Europe together with developing countries played a key role to have the Protocol endorsed during the UNFCCC/COP-7 in Marrakech. Europe as a whole as well as each country subscribed global and differentiated commitments. This requires a careful multidate monitoring - including defining the state of forests in 1990, the reference year.
The objective of the global carbon mass service is to produce maps showing estimated carbon mass sequestrated within forests. These maps are obtained by the combined use of biogeochemical carbon cycle models (which may necessitate some in situ information) and EO data that enable the following forest parameters to be evaluated and monitored : forest/non forest change detection (using SAR and optical data), biomass estimation (using mainly SAR data) ; in addition, estimates of leaf area coverage can be made (using mainly multispectral/hyperspectral data. The ARD measuring and monitoring service aims to provide maps and statistics over a 5 year period, of ARD activities on a global and local basis.
As explained above, at EU level, the two services are relevant to the following policies: 6th EAP, Kyoto Protocol implementation and monitoring, Regional Development, Regional estimation of carbon sources and sinks through mapping ARD activities and impacts modelling. An indirect link with other signed international agreements (desertification, biodiversity) can also be evidenced. At national and local levels, the service should support Kyoto protocol implementation at national level, including the fulfilment of the EU commitment of CO2 emission stabilisation and then reduction, and national forest inventory and survey.
It should be kept in mind that the flexibility mechanisms linked to the Kyoto Protocol will use forest as an "adjustment variable". As a consequence, non-institutional customers (e.g. oil industry, GHG big polluters, wood trade sector) are potential users. So, fast growing market is expected to develop, due to the COP-7 agreements. It may impact on vulnerability and damage reduction linked to deforestation process. National forest management regulation can also be addressed (including fires and storms damages mitigation, combat against desertification, biodiversity conservation, etc ...).In addition, beyond trading mechanisms, the Kyoto Protocol includes commitments to co-operation with developing countries. So, potentially all forest areas in the world (particularly tropical forests) have to be addressed.
In the Kyoto Protocol, the use of common and agreed definitions and methodologies for ARD within the framework of carbon sources/sink accounting, is essential for comparisons on a global basis. For example, an agreed simple definition of forest/non forest (e.g. tree spacing, canopy extent, etc) has to be set up.
The ARD mapping has to be made in two steps:
1. a first zoning identifying the areas characterised by a high potential of ARD activities (according to socio-economical, historical and geographical criteria)
2. systematic coverage of these identified areas (for example, 2 images per year acquired at strategic dates, depending on the area's forest types, geographic location and climatic influence, and temporal analysis of the NDVI evolution over a five year period.
In spite of success stories in the deforestation domain (such as for instance the JRC TREES project), an overall methodology is still in a research stage. Afforestation and Reforestation are more difficult to quantify than Deforestation. Likely cost and effort will be high for coverage on a global basis and for field input/validation. Sampling techniques will have used, bringing together low resolution EO systems for global coverage, medium/high resolution EO systems together with in situ data for local 'calibration'. Validation efforts are mandatory at the local scale, but the effort in terms of operation costs are certainly more affordable.
Keeping this in mind, the spatial coverage of satellite information is a key asset. When large areas are concerned, space technology is a valuable tool to monitor ARD, even though afforestation and reforestation are not so easy to determine. Proper correlation with relevant in situ data is mandatory. Nevertheless, EO techniques should pave the way to standardised methodologies. Such international programmes as GOFC should greatly help. At least EO data can provide: land cover and land cover change maps for ARD; burnt area mapping (see the forest fire damage in the previous chapter). At local scales, high resolution data may prove useful, for calibration/validation purposes. Archives availability and continuity of EO data allowing long term monitoring are particularly important.
As far as global carbon mass sequestrated within forests is concerned, the service is new and heavily relies on carbon exchange models that are currently in a research stage - particularly at regional scales. EuroCarboFlux is a key example of such a research programme. EO derived products are acknowledged to be a key input for such models, but many other parameters, including in situ measurements, are necessary. It is quite clear that this service and ARD monitoring service will use many EO derived products in common.
As in the ARD case, the use of commonly agreed methodologies and tools to assess the carbon sequestration by forests is of utmost importance to actually authorise the balance of carbon emission by the development of carbon sinks and to accordingly trigger the market for trading these authorisations.
EO technology does provide spatial, synoptic information that cannot be provided by any other technology . EO technologies have been proven for: forest change detection, biomass assessment with SAR technologies (e.g. SIBERIA project), LAI assessment with optical sensors such as Vegetation (soon MERIS). The evaluation of carbon sequestration based on first level information (forest change detection and simple carbon models) has already been applied in pilot projects and could be quickly operationalised for a systematic service on a large scale. Checking the final accuracy of operational procedures against the users' need is a mandatory step..
To reach a further step, a critical issue is synoptic coverage which matches well with key features of present EO capabilities. Spatial resolution is not at stake for the presentation of the information but is important for the capability of sensors to map actual forest changes and to correctly evaluate biomass. Improved accuracy is needed that requires some further developments. The application of EO retrieved information (biomass, LAI) within biogeochemical carbon cycle models is presently studied by several institutes in Europe and should be mature in the short to mid term (2 to 3 years). MODIS and MERIS could help.
There is a clear expectation from users for reliable methodologies and tools for the estimation of carbon sequestration. This implies a clear need to improve currently available services. Evaluation of carbon sequestration relies on mesoscale models that are still under development. Reliable input data to the models are missing.
The proposed service is new and should be sustainable as the commitments for carbon emission reduction engage on the long term. Affordability (cost acceptance with respect to the value of the service for the customer) is not yet resolved but considering the importance of the carbon issue, affordability should be reached. Most likely, low/medium resolution complemented by some high resolution images is the right way, both for carbon mass within forests and ARD monitoring services . The issue of in situ data may be the main bottleneck.
The following analysis show how the GMES implementation was prepared by the EUFOREO, considering four applications of major interest to Central Europe.
Data and products delivered by Earth Observation from space have a great potential to contribute to solve or mitigate environmental issues linked to these applications. Hindering factors do remain. These factors have to be overcome to make the best use of the unused or misused potential of these sources of information, together with information delivered by in situ sources. This would pave the way to more efficient GMES services.
Most of the countries in Central Europe will join the EU in 2004 to turn the EU 15+ region into a political reality. At that time, GMES will start its second 'strand'. GMES is planned to be fully operational in 2008. So, it has to be considered as a key integral component of the of the environment and security monitoring systems of these countries at that time horizon.
From a technological point of view, GMES services appear as a unique opportunity to consolidate the operational use of multi-layered geo-coded information in implementing tools for a common environmental policy in the so-called EU 15+ region. This is the very definition of ambitious operational GIS.
The skills of Central European scientists and engineers in applications of remote sensing and GIS technologies applications are widely known and acknowledged. So, GMES should be considered as a major issue in the GIS Ostrava 2003 Conference and follow-on events.
This presentation should be perceived as an invitation to scientists, engineers, service providers and end-users involved in geographic information management in Central Europe to bring a contribution of their own in the GMES initiative for the benefit of their own countries and the future Europe as a whole.