UNITED NATIONS ENVIRONMENT PROGRAMME
ENVIRONMENTAL OBSERVING AND ASSESSMENT STRATEGY
REFERENCE PAPER ANNEXES

CONTENTS

ANNEX 1
REVIEW OF PAST ACTIVITIES, PRESENT GAPS AND LESSONS LEARNED paragraphs 1-2

I. CRITICAL REVIEW OF UNEP OBSERVING AND ASSESSMENT PROGRAMMES
Review of UNEP's work launched before UNCED paragraphs 3-5
Global Environment Monitoring System (GEMS) paragraphs 6-7
- GEMS/Air paragraphs 8-10
- GEMS/Water paragraphs 11-13
Global Resource Information Database (GRID) paragraphs 14-15
- GRID/Sioux Falls paragraphs 16-17
INFOTERRA paragraphs 18-19
State of the Environment Reporting paragraph 20
Other information activities paragraph 21
UNEP in the post-UNCED period paragraphs 22-33
Comparisons with Other Programmes paragraph 34
- Global Atmosphere Watch paragraph 35
- U.S. Environmental Monitoring and Assessment Program (EMAP) paragraph 36
- U.S. Water Quality Program paragraphs 37-39
- Weather paragraphs 40-42
- USEPA Air Quality Program paragraphs 43-44
Conclusions from Case Studies paragraph 45

II. OTHER ENTITIES ENGAGED IN OBSERVING, ANALYSIS AND REPORTING paragraphs 46-53

III. USEFULNESS OF DATA COLLECTED paragraphs 54-67

IV. CRITICAL DATA GAPS paragraphs 68-81

ANNEX 2
POLICY-ORIENTED ENVIRONMENTAL ASSESSMENT paragraphs 1-4

1 METHODOLOGICAL FRAMEWORK paragraphs 5-8
1.1 DPSIR framework paragraphs 9-10
1.2 Data for Integrated Environmental Assessment paragraph 11
1.3 Modelling tools and scenarios; dynamic analysis paragraphs 12-14
1.4 Presentation paragraphs 15-17
1.5 Key areas of expertise paragraph 18
1.6 Key aspects of Integrated Environmental Assessment paragraphs 19-21

2 CRITICAL GAPS paragraph 22
2.1 Critical gaps in expertise paragraphs 23-31
2.2 Critical data gaps paragraphs 32-35
- Driving forces paragraph 36
- Pressures paragraph 37
- State paragraph 38
- Impacts paragraphs 39-40
- Responses paragraph 41

3 COLLABORATION FOR IMPROVED OBSERVING AND ASSESSMENT paragraph 42
3.1 Organization of data monitoring and handling paragraphs 43-47
3.2 Disseminating data and expertise between collaborating centres paragraphs 48-50

1. The first section of this annex gives an historical overview of the development of the different UNEP observing and assessment programmes, together with an evaluation of their impacts and shortcomings. The first part covers UNEP's initiatives in monitoring and assessment launched prior to the UN Conference on Environment and Development (UNCED) in 1992, including the recent status of long-established programmes. The second part considers UNEP's new work in the past 6 years. A third part provides some comparisons with case studies of other programmes, and draws some conclusions.

2. The second section summarizes the other entities engaged in observing, analysis and reporting and UNEP's interactions with them. The third section looks at the issue of useful data and less useful data. The fourth section provides a brief review of critical data gaps which a UNEP observing and assessment strategy should aim to fill.

I. CRITICAL REVIEW OF UNEP OBSERVING AND ASSESSMENT PROGRAMMES

Review of UNEP's work launched before UNCED

3. UNEP was established as one of the outcomes of the UN Conference on the Human Environment, held in Stockholm in June 1972. During the 1970s and 1980s, the organization very effectively played a leading role in putting on the environmental agenda many urgent and emerging issues, such as ozone depletion and trade in hazardous wastes, and facilitated the creation of many important international environmental conventions on, for example, ozone depletion, climate change and biodiversity.

4. The Stockholm conference called for the establishment of the Earthwatch system to keep the global environment under review. At the time, environmental observing was not yet a major routine activity in most countries. Comprehensive and relevant data on the global environment were scarce. Consequently, during its first years of operation, UNEP focussed its observing and assessment programme on strengthening environmental monitoring through the Global Environment Monitoring System (GEMS) programme and on harmonizing existing efforts for global overviews.

5. In retrospect, UNEP's past monitoring and assessment programme has been a mix of successes and failures, as is probably true of any complex programme. However, the critical comments in this review need to be placed in the context of the developing understanding worldwide of how environmental observing and assessment should work. Many programmes of 25 years ago can be criticized in hindsight. The real issue is to learn the lessons and move forward. For instance, UNEP often failed in its catalytic role because it focussed on providing data rather than policy-relevant information, as will be evident in the discussion and selected case studies below, but this was quite common at the time. Another contributing factor was the thinly spread resources of UNEP. UNEP has repeatedly been instructed to implement ambitious mandates, but the support of governments to the Environment Fund has not followed. Even well designed and established programmes could not become self-sustaining, and when UNEP funding declined, they collapsed. Despite this, in terms of cost effectiveness, UNEP may in fact have accomplished much more than better funded national programmes.

Global Environment Monitoring System (GEMS)

6. Initially, GEMS focussed on developing monitoring methodologies, establishing quality control systems and global databases, capacity building and technical support, and the production of status and trends overviews in specific environmental compartments. Some of these assessments, such as on Ozone, contributed directly to the launching of international environmental conventions.

7. GEMS programmes were implemented through and with sister organizations in the UN system. This respected the intent of UNEP’s mandated role to catalyze, coordinate and promote routine environmental observing in a large number of countries and by UN organizations. They were instrumental in catalyzing world-wide observing and assessment and promoting their incorporation in UN programmes. However, they only covered a limited number of environmental sectors/themes such as freshwater quality, urban air pollution, and food contamination. Furthermore, UNEP was unable to secure adequate funding to ensure their long term operation and global coverage, as the following case studies of their recent status prepared by the World Resources Institute illustrate.

GEMS/Air

8. GEMS/Air, in its early years, helped governments in some parts of the world in designing the current monitoring networks for ambient air quality. More recently, the US Environmental Protection Agency has provided significant support. The GEMS/Air database has been housed for the last five years at the USEPA's Data Center in Triangle Park, North Carolina. The USEPA donated database, software, maintenance, and other labour to provide a home and distribution center for GEMS/Air data. USEPA rarely received data directly from the 48 participating GEMS countries. Instead this data passed through the GEMS/AIR office of WHO in Geneva. Because it is administered by WHO, most of GEMS/AIR sites were originally set up in cooperation with local health ministries and were sited to help characterize human health effects within a GEMS city. The sites thus often bore no relation to environmental observing networks within a country established by other agencies. Indeed, GEMS/Air has never received data from the vast bulk of air quality monitoring stations in the world. In the mid-1990s, an attempt was made to add a capacity building component to the programme that would have given something back to the participating cities through a city twinning programme, but then UNEP's general budget decline hit. As a result, the GEM/AIR database cannot provide answers to many pertinent questions about trends in global air quality or the extent of exposed populations. In addition, there does not appear to have been quality control efforts, such as inter-calibration of equipment, to insure comparability of data among countries.

9. Until recently, EPA distributed at no cost about 50-75 copies a month of the GEMS/Air database along with software to browse and extract the data. It has since combined this database with its domestic data and continues to distribute both for free from its Internet site. But USEPA has stopped receiving data from WHO and has been told that WHO has contracted to develop its own software and plans to sell the data USEPA has previously distributed freely.

10. As presently constituted, GEMS/Air seems to have no clear purpose. It does not respond to an international constituency, supports no comprehensive policy assessment, and does not serve those few cities that have contributed to its database. Finally, instead of working with international partners -- such as the USEPA -- it has abandoned them.

GEMS/Water

11. The GEMS/Water Collaborating Centre in Burlington, Ontario is hosted by the National Institute of Water Research of Environment Canada. Like USEPA for GEMS/Air, Environment Canada, as a contribution to the global community, operates GEMS/Water. The system was intended to characterize changes in water quality along a watershed and the pollutant loadings introduced into the coastal zone by rivers. Although under-funded, GEMS/Water has recently undergone a rebirth due to greater interest, especially within UNEP. For the first time in years, new water quality data has been posted on the Internet for down-loading (1990-1993). However, many of the 600+ original stations of GEMS/Water no longer participate, and the system is not in any sense a comprehensive water quality database that could be used to answer many water-related policy questions.

12. In addition to data collection, GEMS/Water -- like GEMS/Air -- had another purpose, that of building capacity for data collection. But the capacity was solely for the purpose of contributing to the global system -- and did not help meet local or national needs. As a result, it did not engender local and national support for data collection. Like GEMS/Air, GEMS/Water does not today seem to have a clearly defined local or global audience and purpose.

13. The GEMS programme was not designed to meet the information needs of local and national decision-makers. Because these systems do not meet those needs, they have failed to develop a constituency -- especially in North America -- for continued action and cooperation, regardless of any other successes. At the same time, the very limited coverage of the data collected in the GEMS system makes its relevance to international assessment activities very questionable. Even the two global assessments of freshwater quality did not use GEMS/Water data. Given that the locus for action in environmental policy is primarily at the local and national level, it is hard to say what successes can be ascribed to GEMS. The GEMS/Air database is irrelevant to the assessment of urban air quality in any city, much less at a national level, or to characterize human exposure globally.

Global Resource Information Database (GRID)

14. With the emergence of remote sensing (RS) and geographic information systems (GIS) in the early 1980’s, UNEP spearheaded the application of these systems in international environmental observing and assessment through the establishment of the Global Resource Information Database (GRID) system. GRID provides an environmental data management system for UNEP, the UN system, international organizations and governments. Through a network of cooperating GRID centres, it archives, collates and disseminates digital information which has been extracted from maps, satellite images, statistical tables and other sources within and outside the UN system. All these data are linked together through computerized Geographic Information Systems and Image Processing Systems (IPS). The basic uses of GRID, in addition to its data management function, are:
- helping to find environmental data;
- supporting environmental decision-making;
- technically assisting regional and national capacity-building;
- supporting the analysis of national environmental conditions and trends and assessing risks and vulnerabilities.
At a more advanced stage GRID has supported environmental impact assessments.

15. The GRID centres are very diverse in their functions, structures, institutional relationships, and the extent of UNEP support. Some have attracted major outside support, are highly cost-effective in terms of UNEP input, and are considered very successful, even leaders in their field. Others are fledglings requiring further capacity-building. It is not possible to review all of them here, but a case study of one centre illustrates some of the issues for this strategy.

GRID/Sioux Falls

16. The Global Resource Information Database (GRID) node at the US Geological Survey's EROS Data Center in Sioux Falls, South Dakota, has been successful as a catalyst, coordinator, and liaison to operational divisions of the U.S. government. The GRID office convenes specialists to catalyze new analytical activities within other agencies, helping to legitimize and sometimes partially fund such efforts. For example, a small amount of money from UNEP/GRID was used by Dr. Tom Loveland to leverage much greater U.S. agency funding for his ground-breaking land cover characterization maps; WRI and Dr. Uwe Deichman were able to use partial UNEP/GRID funding to create geo-referenced population databases.

17. GRID/Sioux Falls has helped to promote the use of geo-referenced databases and it does make such data available to those analysts or agencies that request them. But the office works essentially independently from other GRID nodes; the GRID system does not function as a global network or data system in the sense that was originally envisioned. Even if GRID had become a repository of geo-referenced environmental data--the role originally envisioned for it--the changing information culture brought about by the Internet would nonetheless have rendered the need to collect data in central repositories obsolete. And the North American GRID node, while successful in catalyzing the creation of global level databases, has had less success in promoting the creation of local or national level geo-referenced databases. GRID, like GEMS, was not designed originally to meet the information needs of local and national decision-makers, and thus failed to develop a constituency, at least in North America. This error has been corrected in some regions through the recent ENRIN activities described below.

INFOTERRA

18. INFOTERRA has similarly been a casualty of the changing global information system. When UNEP was established in the 1970s, there was a great need to facilitate access to the then scarce and scattered environmental information and experience. An environmental information referral service, called INFOTERRA, was created to direct questions to the best sources of environmental information, with information guides and thesauri prepared to facilitate the process. National INFOTERRA focal points, usually within environmental ministries, were established and trained in information retrieval. Today there are over 170 such focal points officially linked to UNEP through INFOTERRA, and a steady flow of requests pass through the system. However new information technologies such as the Internet are radically transforming information access, making multiple sources of environmental information readily available, and part of the original rationale for INFOTERRA has vanished. There is still a need for referral systems with a human interface able to find answers to vague or tentative questions, queries for non-written information or in non-Western languages, etc., but a separate network for this is no longer so necessary. The referral function is important, but is more relevant to a library than to the environmental observing and assessment system as such.

19. The network of national focal points, while of variable quality since so few resources have been available to maintain it, is still a valuable resource in which UNEP has made a major investment in capacity-building over the years. It represents the only relatively complete UNEP network at the national level. Its functions can be revised and adapted to the needs of the new strategy.

State of the Environment Reporting

20. Given the need for scientifically valid information for sound environmental management, most UNEP programme areas also undertook topic-specific diagnostic studies and assessments relating to watersheds, land degradation, regional seas, biodiversity, etc. However, a coherent UNEP-wide strategy to integrate these assessments and to ensure that they contributed to a global environmental overview was never implemented. UNEP's major state of the world environment reports up to 1992 were prepared by a small group of senior advisors and consultants under the direct supervision of the Executive Director, with little relation to other UNEP assessment activities.

Other information activities

21. While the original UNEP Earthwatch programme concentrated on the observing and assessment of natural resources and chemicals, other UNEP programmes focused on the industrial side of the equation, implementing assessments regarding industrial pollution, health risks, etc. A number of databases were developed for topics such as cleaner production, environmental law, etc. Information networks and clearing house systems were developed and maintained. With the recent creation of UNEPnet, a start has been make to link these disparate systems, but it is clear that further integration and streamlining are necessary to make these information sources more relevant and accessible to the users.

UNEP in the post-UNCED period

22. In the 1990s, UNEP was confronted with the entry of many new players in the field, both at policy and assessment levels. On the policy level, UNCED created the Commission on Sustainable Development and installed a new coordinating mechanism within the UN, the Inter-Agency Committee on Sustainable Development. A special UN International Negotiating Committee was given the mandate to negotiate the Climate Convention. Although UNEP still plays an important role in the scientific assessment of climate change (as co-host of the IPCC with the WMO), the Convention is a policy framework that is rather detached from UNEP. More generally, the proliferation of issue-oriented multilateral conventions has resulted in an increase in the number of institutional players and policy fora and has exacerbated the problem of policy coordination between the various conventions.

23. In early 1992 in the run-up to UNCED, UNEP instituted a UN system-wide Earthwatch Coordination function to coordinate, harmonize and integrate observing, assessment and reporting activities across the UN system. After UNCED, UNEP was designated by the Inter-Agency Committee on Sustainable Development as Task Manager for Earthwatch in reporting to the Commission on Sustainable Development (CSD). Earthwatch has stimulated increased strategic planning and collaboration with a wide range of partners in the UN system and beyond. It succeeded in obtaining UN system-wide agreement to better coordination and collaboration, in building strong cooperation and policy development on information for decision-making in support of the CSD, in advancing work on indicators of sustainable development, in building an extensive system-wide Earthwatch web site, in providing UN system inputs to the GEO reports, in establishing close collaboration between the Global Observing Systems and their sponsors, and in supporting the Integrated Global Observing Strategy. However the minimal support that UNEP has been able to provide to Earthwatch Coordination has prevented it from reaching its full potential.

24. Just after UNCED, and in response to the expansion in the number of institutes and organizations dealing with environmental observing and assessment, and the identified need to improve access and disseminate the resulting data for assessment purposes, UNEP established the Environmental and Natural Resources Information Networking (ENRIN) component in GEMS in 1993. ENRIN’s stated objective was to build institutional linkages and facilitate the flow of information for environmental assessment between governmental, inter-governmental and non-governmental institutions in developing countries and countries with economies in transition. The ENRIN projects were established to catalyze the formation of cooperative assessment networks as a conduit for information flows and a mechanism for regional consultations on priorities, policies and implementation of assessments of transboundary environmental issues. ENRIN also provided a channel for technical assistance to participating institutions to build their capacities in data management and analysis for environmental reporting. A number of these assisted institutions have subsequently indicated a desire to become members of the GRID network.

25. Over the years, as environmental problems became better known, governments and the public began to ask how the issues could be resolved or managed. The definition of the task of international environmental assessment was therefore expanded from the scientific assessments of status and trends, towards the provision of integrated information on critical linkages between the environmental/ecological and human/economic subsystems. Agenda 21 confirmed the need for this type of environmental data and information for decision making purposes.

26. Consequently, in response to the adoption of Agenda 21, UNEP developed a new series of biennial assessment reports, the Global Environment Outlook (GEO) series, to evaluate the environmental implementation of Agenda 21. The GEO process represents a radical change, providing integrated, forward looking assessments through global participation and consultation. It revealed the need for better integration of UNEP’s observing and assessment activities, which historically had followed rather separate tracks. It also identified a need to redefine observing activities that reach beyond the state of the environment proper. The long term goal was to provide a set of integrated policy relevant assessments and reports from the subregional and regional to the global level, drawing on the expanding networks developed by the GEO and the ENRIN projects, input from UN specialized agencies through Earthwatch, sectoral assessments conducted by the Programme Division of UNEP and supported by GIS/RS data sets derived through the GRID system. The first GEO report was released in 1997 to support the 5-year review of Agenda 21 and had considerable impact. The second will be issued in early 1999.

27. However, especially after UNCED, many intergovernmental (World Bank, UNDP, regional organizations, development banks, etc.) and non-governmental organizations have engaged in global and regional environmental data collection and assessment activities. At the same time, UNEP was faced with a steady decline in funds to execute its work programme, including its observing and assessment activities. As a result, UNEP lost the leading role it had played during the 1970s and 1980s. This weakened UNEP’s position in the UN system and precipitated a review of UNEP’s role in international environmental policy making.

28. In response to UNEP’s political and financial difficulties, the UNEP Governing Council formulated, in the 1997 Nairobi Declaration, a new policy to revitalize the Programme. It restated that UNEP should continue to be the principal United Nations body in the field of the environment and that the role of the United Nations Environment Program is to be the leading global environmental authority that sets the global environmental agenda, that promotes the coherent implementation of the environmental dimension of sustainable development within the United Nations system and that serves as an authoritative advocate for the global environment .

29. At the same time, the Governing Council decided to add to the governance structure of UNEP by establishing a High-Level Committee of Ministers and Officials as a subsidiary organ of the Governing Council. In this way it sought both to strengthen UNEP’s executive powers so as to enable it to act more swiftly and effectively, and to enhance its policy support through increased high-level policy involvement and participation. The decision also reflects the wish for a strengthening of regionalization and decentralization within UNEP through the increased involvement and participation of regional ministerial and other relevant forums within UNEP, complementary to the central coordination role of the Programme’s headquarters in Nairobi . Even more important for UNEP’s effectiveness was that the Governing Council succeeded in agreeing on policy priorities and a corresponding allocation of budgets, while the previous decline in the fulfilment of financial commitments by donors was halted.

30. So after a difficult period, UNEP may again regain its former leading role, but only if it is able to adjust to new circumstances and policy needs. Its role in raising awareness and initiating the development of new international environmental policies and law has become relatively less important. For most issues, some global or regional frameworks have been developed. Although the identification of newly emerging issues remains important, many issues have matured to the policy formulation or policy implementation stages of the policy life-cycle.

31. As a consequence, assessment of policy options and monitoring of policy implementation and environmental progress have become more important. Moreover, priority setting, policy coordination (between issue-oriented policies) and the integration of environmental issues into sector policies have become prominent matters to be addressed in integrated environmental assessments. This was clearly illustrated in GEO-1.

32. More than before, UNEP’s observing and assessment activities should serve the attainment of all its stated goals and not just those related to continual review of the state of the global environment. However, at the same time it becomes more important to demarcate clearly its field of work, including in observing and assessment. This is essential if it is to keep activities focussed and effective and to define its relations with other actors in the field.

33. The noted maturation of international environmental policy development not only has implications for the contents of environmental observing and assessment, but also for their methodologies. The assessment of policy options and the development of effective strategies for their implementation require the increasing involvement of stakeholders in the policy development process and a gearing of environmental observing and assessment activities to their needs. As the stakeholders are the ones on which the effectiveness of environmental policies depend, their involvement in policy development is crucial for securing their support for the implementation of policies. Moreover, by involving the envisaged users of the outcomes of the assessment in the process, there is a better guarantee that the information produced is properly geared to their information needs and that it will be used in policy development. A third reason to include stakeholders is because they have expert knowledge about their own domain and/or region that is very relevant to the assessment process (e.g. to assess the vulnerability of sectors or regions to environmental pressures or the feasibility and acceptability of particular policy options).

Comparisons with Other Programmes

34. Some of the external reviewers were requested to compare UNEP's experience with that in their own regions, such as in North America. While the lessons from these case studies are enlightening, any comparison of UNEP programmes with those in the United States needs to consider the great difference in resources available to UNEP and to U.S. Federal programmes. In terms of cost effectiveness, UNEP may come out reasonably well.

Global Atmosphere Watch

35. The Global Atmosphere Watch (GAW) has stations around the world to assemble information on the background atmospheric composition so as to be able to characterize the atmosphere by location and over time. GAW is a special purpose network of existing monitoring stations to achieve global coverage. Because it monitors the atmospheric baseline, its purpose can be served by a small number of observing stations distributed around the globe. Unlike the GEMS system, the GAW network does have a clear policy audience--those concerned with atmospheric chemistry and the relevant global conventions on global warming and ozone depletion. It depends for its functioning on the strong cooperation of many national efforts--each of which has independent reporting responsibilities, but each of which is also used by the World Meteorological Organization for a global synthesis.

U.S. Environmental Monitoring and Assessment Program (EMAP)

36. It is not only international monitoring schemes like GEMS that have fatal flaws. The United States Environmental Protection Agency, for example, promoted the creation of the Environmental Monitoring and Assessment Program (EMAP). EMAP was intended to provide policy makers with information on the state of U.S. ecological domains through a complex scientific sampling scheme. The effort was finally abandoned, in part because, even in principal, it would only provide information at a very high cost. Equally, however, it became clear that EMAP would probably never provide useful information to decision-makers--local or national. EMAP was a classic case of a monitoring system designed by scientists for scientific purposes and lacking a clear connection to policy-relevant issues.

U.S. Water Quality program

37. The US Environmental Protection Agency has created an exemplary observing system for national water quality by building up from local and state water quality monitoring systems. The US EPA helps localities to define issues and set up systems to meet their own needs; at the same time, it also sets standards and harmonizes systems. It assembles local information--indeed, many states use the EPA Office of Water Internet site to disseminate their own findings--makes it publicly available, and uses it to analyze national water quality issues including a national indicator map of water quality and vulnerability.

38. The Internet site also makes available access to other databases that are relevant to understanding water issues in each watershed and for the nation as a whole. These databases include air quality, water use and availability, population, and agency jurisdictions.

39. The power of this observing system is that it builds from, and includes, all data that are collected and considered relevant by local policy makers and managers. Even though the provision of the data is mandated by law, the system rewards cooperating institutions by publishing the data in a common form, providing comparisons across space and time, and giving local information a national context and a national purpose alongside local and state purposes.

Weather

40. One of the most powerful examples of local information aggregated to regional, national, and global levels is that of the U.S. National Weather Service and the World Weather Watch of which it is a part. While a centralized organization, the Weather Service has 52 Weather Service Forecast Offices, 204 Weather Service Offices, 13 River Forecast Centres, and 3 National Centres distributed around the country. Reporting stations are manned by its own employees as well as a variety of volunteers, private sector employees, and employees of other public sector institutions. These stations serve to document climate, river flows, or weather. There are about 12,000 observing stations run on a cooperative basis.

41. Local weather observations are valuable to the public, farmers, utility planners, businesses, and local government. Forecasts received back from the National Weather Service provide all of these with necessary information to plan both normal activities as well as to provide advance response to coming hazardous weather conditions (e.g., hurricanes, floods, etc.) at the local, regional and national levels (as well as contributing U.S. data to the global weather picture). The National Weather Service not only coordinates local observations and supports local and regional non-official forecasting, it also mobilizes and manages strictly observational stations that are more reasonable under national management (buoys, regional radar, satellite observations, aircraft) as well as providing forecasts at all geographic levels, including global. It is the official source of severe weather information and disseminates its observations and forecasts through many means, including traditional media, NOAA Weather Radio, and private weather services.

42. The National Weather Service has a clear purpose and audience. It serves decision-makers at all levels from private and local to global and public. And it repatriates information to local institutions from which it was derived.

USEPA Air Quality program

43. Observations of urban air quality are useful to local and regional public sector managers, policy makers and the public. In the United States, over 4800 local air quality monitoring stations provide real time information to local policy makers on what can become a health emergency requiring a quick policy response. Analogous to weather stations, this real time information is also accumulated and provided to the USEPA so that they can assess the national experience, compare cities, and formulate policies, as well as to harmonize and calibrate information. Essentially U.S. national assessment of air quality has the same relationship to local air quality observations as climate assessments do to weather data in meteorology--the long-term builds directly on the short term.

44. The USEPA sets health guidelines and operating parameters for cities, and they in turn are obligated to supply information to the national system. These data and resulting analyses are made available to all and support both local and national policy responses. The programme illustrates how national goals and guidance can inter-relate with and support local observations and assessments for mutual benefit.

Conclusions from Case Studies

45. Past experience in observing systems as illustrated in the case studies seems to support several conclusions:

(a) Observing systems designed to support global purposes seem to work best when they also serve local and/or national purposes; that is, they are built on the back of local/national systems and add value by aggregating and summarizing data in a global context, and they return something to the local/national entities that are the source of the data, usually information that provides geographical context or is based on more detailed analysis, but also sometimes standards, training or other forms of capacity building, or direct financial support. Purely global systems that have dense enough networks to be useful are rare, and the available evidence suggests that adequate financial resources to support global networks are unlikely to be sustained unless the networks also serve local/national needs.

(b) Observing systems that provide specific information products in a form useful to decision-makers seem to be far more successful than those that simply collect and make available data; indeed, the greater the variety of information products provided and audiences served, the more robust the support for the observing system and the greater its success. The World Weather Watch and the televised daily weather maps we take for granted are the classic example. Thus analysis and simplification or visualization--the process of turning data into information, into products useful to decision-makers--is a critical component of successful observing efforts.

(c) Most older global monitoring systems were designed around a centralized database and management; in the Internet age, distributed databases are equally feasible and may fit better with the concept of broad participation and ownership in a monitoring system, eg, a more bottom-up approach that helps to ensure the system also serves local/national needs. Some combination of the two types may be needed. Centrally-coordinated systems can provide sampling frameworks, classifications and global models, and feed back results in a wider context. Local networks spread ownership and involvement, and can provide unique types of information to be cross-checked with more standard sources.

II. OTHER ENTITIES ENGAGED IN OBSERVING, ANALYSIS AND REPORTING

46. A summary of the other entities engaged in observing, analysis and reporting and UNEP’s interaction with them can help to put UNEP's strategy in perspective. Space does not permit describing all UNEP's interactions, so only highlights are given here.

47. In the 26 years since Stockholm, the world has seen a largely uncoordinated explosion of environmental data, information and related research activities, driven partly by new technological developments and possibilities, and partly by increasing political and public attention to environmental issues. A plethora of governments, UN organizations, research organizations, NGO’s, the private sector, etc. are involved in environmental observing and assessment. Some major UN organizations, international research organizations and development banks now have observing and assessment programmes in place that are far greater in scope than UNEP programmes. The major global assessments and reports are listed in Table 1.

48. There have recently been initiatives to bring some order to international efforts, including UNEP's own work through Earthwatch. One such major initiative is the Integrated Global Observing Strategy (IGOS) that addresses the common interests of major space-based and in-situ measuring systems for global observation of the earth. This includes observations relating to climate and atmosphere, oceans and coasts, and the land surface of the earth in an integrated framework. It strives to build upon strategies of existing international programmes and upon current achievements. An IGOS Partnership has recently been established between the Space Agencies represented in the Committee on Earth Observation Satellites (CEOS), UNEP, FAO, UNESCO and its IOC, WMO, the International Council for Science (ICSU), the Global Observing Systems (GCOS, GOOS, GTOS), and major international research programmes (WCRP, IGBP). IGOS is implementing some pilot projects including the Global Ocean Data Assimilation Experiment (comprising GOOS/GCOS/WCRP, CNES, ESA, NASA, EUMETSAT, NOAA and NASDA), Global Observation of Forest Cover (the above partners plus GTOS, IGBP LUCC, FAO, CAS/CCRS, INPE, EC, ASI) and Disaster Management Support (NOAA, EC, BNSC, ASI, STA/NASDA, NASA, CSA, CNES, WGISS, CCRS, Council of Europe, DARA, ESCAP, EUMETSAT, GTOS, IDNDR, INPE, RPA, PLANETA, WCRP, WMO).

49. UNEP has responded to this increasing complexity with a number of other structures as well, adding the network of GEO collaborating centres and the ENRIN activities, amongst others, to the existing GEMS, GRID and Earthwatch coordination. One of the selection criteria for the collaborating centres was a linkage to one or more of the above networks or processes.

50. A specific response to this explosion was the production by the UNEP HEM (Harmonization of Environmental Measurements) office in 1994 of a directory of environmental information and assessment programmes. HEM’s goal was to regularly update and expand the directory. Unfortunately, again due to funding constraints, the programme was aborted soon after publication of the first directory.

51. Other organizations have conducted similar exercises, but mainly in sectoral areas (air, water, oceans, forests). The UN Department of Economic and Social Affairs (DESA) and Earthwatch, in their 1995 review of the implementation of the Agenda 21 chapter on Information for Decision-making, brought some of this information together in an overview of information and assessment activities, which is available on the Earthwatch web site and is regularly updated. However, at present there is no comprehensive summary of all these activities, including those outside the UN system, that brings out their complementarities.

52. Basically there are four (maybe more) systems working in parallel:
- the statistical systems, with the UN Statistics Division and regional statistical offices, such as the Regional Economic Commissions and Eurostat, etc. developing statistical frameworks for environmental data, including indicators;
- the whole remote sensing and geographic information systems community working with geo-referenced data;
- the scientific community collecting data for research purposes, such as IGBP, ICSU, etc. and the three Global Observing Systems for operational data collection; and,
- a plethora of sectoral data exercises conducted by everybody from grassroots groups to UN organizations and convention secretariats.

53. Despite this rather complex picture, UNEP has, over the years, established more or less strong links, be it sometimes ad hoc, with the major entities engaged in observing, analysis and reporting. UNEP should identify the key programmes with which effective collaboration has to be established as well as the modalities of this collaboration.

III. USEFULNESS OF COLLECTED DATA

54. An analysis of which data collection efforts have resulted in useful data and which in less useful data can provide useful background to the new observing and assessment strategy. The following summary description of both positive and negative experiences highlights some of the lessons that can be learned from such an exercise.

55. Within the framework of the GEO programme, UNEP and DPCSD (DESA) convened an expert consultation in 1996 to identify the core global data sets needed for global environmental assessment and assess their availability. The initiative built on an International Symposium on Core Data Needs for Environmental Assessment and Sustainable Development organized by UNEP and UNDP in Bangkok, 15-18 November, 1994. The UNEP/DPCSD meeting report lists in detail the availability of core global data sets and identifies the gaps. The meeting identified the next steps needed to fill the gaps and to complete the descriptive information. It proposed combining the available core data sets on a CD-ROM for widespread distribution and use. Funding constraints prevented UNEP and DESA from carrying this initiative forward.

56. Instead, again within the framework of GEO, a smaller initiative was launched to assess the availability of a core of data sets at the global level. The global information was also compared with available national and regional data sets. The work is ongoing with a workshop held in Budapest in September 1998. The results will be reported in GEO 2. GRID centres and the GEO collaborating centres are actively involved in this exercise.

57. The preliminary findings of the above exercise include the observations that:
- The FAO Agrostat data base is one of the major important/useful environmental databases;
- WMO climate data sets are extensively used in climate change related research;
- World Bank data, as recently put together in the excellent indicators report, are very useful but cover mainly the socio-economic field and are at a country level (which is not always the most appropriate for international environmental assessments);
- Global databases are often derived from other data and analyses of data collected for other purposes (extrapolations, survey data, results from modelling efforts, etc), with such data recycled endlessly and not kept up to date;
- The correlation between global data sets, as for example reported in the World Resources Report, and the national and regional data sets is very low (<30 %) because of conversion problems and poor coordination between different data sources within countries;
- For many global data sets, matching sets at the regional and national level are not available;
- A major gap is the lack of correlation/integration between geo-referenced data sets and numeric data;
- Global coverage of point data (such as water quality, urban air pollution, forest status, soil degradation) is scanty and often only available at a level of aggregation that is not very useful for management purposes;
- very few new global data sets have been developed since Rio;
- GIS (spatial) data sets are mainly used by the research community and then mainly in developed industrialized countries.

58. UNEP played a pioneering role in data collection related to air and water quality through GEMS. However, it was not able to maintain the momentum, nor did it build local justification or motivation for continued support. While some of the GEMS methodologies may still be highly relevant, the data assembled in the GEMS global databases have little present value (see case studies above).

59. In addition some national databases related to land cover have been created for many countries of the world. The experience shows that UNEP can successfully catalyze environmental data collection and dissemination, but in the absence of ongoing funding for institutionalizing such efforts, their long term sustainability is in doubt.

60. With respect to spatial data, the experience of the UNEP Global Resource Information Database (GRID) system, particularly Sioux Falls, Geneva, Bangkok and Arendal, shows that global data sets available on the web are in great demand, albeit mainly by the scientific community. GRID has played a unique role for over a decade in providing access to spatial data. Numerous users have accessed and benefitted from the GRID system. Recently, GRID has played a catalytic role in the development of globally consistent and comprehensive data sets related to:
- Global land cover using AVHRR 1km.
- Global topographic database
- Global drainage basins boundaries database
- Global population distribution database (Africa and Asia)

61. Remote sensing is the most readily available source of true global coverage. However, these data must be translated into the thematic information and validated using ground truthing information from countries. The capacity to make effective operational use of remotely-sensed imagery has fallen far behind the technologies for producing images. The imbalance in funding for in situ data collection necessary to interpret satellite imagery, and for the analysis and assessment of the results, relative to the investment in new satellite hardware, needs to be corrected.

62. Development of globally consistent and comprehensive spatial data sets is time-consuming and expensive. It is possible only through international partnerships. However, within the cluster of partners, at least one organization is needed which has sufficient resources -- hardware, software and peopleware -- and long term commitment, to pull together the rich sources of data and expertise of partners.

63. The majority of data development efforts are fragmented. Only a handful of efforts are under way to generate truly globally consistent data sets for major environmental issues. Such global data sets are needed to address global, regional, transboundary, and many cases national concerns. However, global data are at a scale that limit their uses at national and local levels. More efforts need to be directed towards consistent nested sets of data at different spatial scales to improve understanding of the relationships between local, national and global problems.

64. User demand for environmental information is normally underestimated, as users do not know ‘what to ask for’ unless they ‘see’ it and realize its potential. Knowledge of availability of data sets stimulates demand and new applications.

65. The international data management system, as it develops, can provide an example for national strategic databases containing georeferenced and statistical data for improved decision making. The national level database could provide thematic information for planning and policy analysis and for performing comparative analyses with other countries.

66. The Framework for Environmental Statistics, developed by the UN Statistics Division has the potential to generate one of the most important data collections, if only because there is a well established institutional structure for statistical data collection in most countries. It is essential to note that any data collection effort that is not institutionally underpinned will not be sustainable and might not prove useful for long term status and trend analysis. The lack of an appropriate institutional "home" for many kinds of environmental data collection is a serious handicap to building the long-term time series of data needed to assess many kinds of environmental change.

67. Recent developments in information and sensing technologies offer the prospect for developing new and more cost-effective approaches to data collection worldwide, including participatory approaches with grass root groups and data assembly and feedback over the Internet. The World Resources Institute is experimenting with this approach in the Global Forest Watch and other initiatives.

IV. CRITICAL DATA GAPS

68. There are significant gaps between the information provided by present observing systems and what is needed by decision-makers. There is far less information available for marine environments than for terrestrial environments, for example, and far less information for natural or unmanaged environments than for managed environments such as agricultural systems. Information about human activities that affect the environment is generally more abundant than information about the status or health of the environment itself. And the industrial countries are, not surprisingly, information-rich compared to most developing countries. Even so, the most serious problem in many developing countries is not the lack of information, but the failure to share existing information among agencies and to use information in decision-making processes.

69. Existing observing systems often place greater emphasis on short-term phenomena (for regulatory compliance, for example) rather than long-term trends that may reveal serious underlying problems; and there is a large imbalance in the amount of data collected compared to the amount of analysis of such data for policy-relevant purposes, reflecting the dominantly science-driven (as opposed to policy-driven) nature of many data collection efforts. Some examples may help to illustrate the nature of the problems.

70. There have been several attempts to develop consensus lists of key data gaps. The 1994 UNEP/UNDP symposium on Core Data Needs for Environmental Assessment and Sustainable Development Strategies identified data needs for Land Use; Demographics; Hydrology; Infrastructure; Climatology; Topography; Economy; Soils; Air Quality; and Water Quality. A WRI-sponsored international workshop developed a similar list of key data needs: Economic Activities; Land Cover; Land Use; Population; Emissions; Pollutant Sources; Stocks and Harvests: Soils; Actual Productivity; Other Resource Offtake; Hydrology; Biodiversity Quality: Biodiversity; Pollution Loads; Exogenous Stresses. In both lists, for example, information on land use, water supply, and water quality were identified as a key data gap, and in both efforts there was a strong finding that all of these important data be geo-referenced so that spatial analysis could be performed. More broadly, the global environmental conventions -- climate, biodiversity, desertification -- imply the need for far more detailed information on energy use patterns and greenhouse gas emissions, on the health and functioning of ecosystems, and on the pattern of human activities that contributes to land degradation.

71. Land use, for example, is a measure of human intervention into numerous earth processes. Land use patterns affect hydrology, soil, soil erosion, habitat, biodiversity, water quality, and ecosystem health (degradation), as well as the capacity of forests or other ecosystems to absorb greenhouse gas emissions. While attempts have been made to map land cover at a 1 km resolution for the globe, these maps do not accurately reflect variations in human management of those mapped wild and managed ecosystems--degraded and intact forests are mapped the same way, and subsistence agriculture is not distinguished from high intensity farming. Managers at all levels from local to global need to understand how land is used if they are to promote sustainable practices and understand the other environmental impacts that can be expected from specific land use practices. Thus the lack of a digital land-use map significantly hinders more environmentally sustainable development and even characterization of existing environmental stresses.

72. Water quality information provides another example. There is very little useful information on the quality of the water in the world's rivers and underground sources. Even within nations, an accurate characterization of water quality is often at best spotty. Yet water quality information is important for reasons of human health, for industrial and agricultural purposes, for pollution control, and for maintaining the viability of freshwater and estuarine ecosystems. We need to understand what chemicals, nutrients, and sediments are being transported and in what quantities, if pollution is to be documented and contained; excessive nutrients and large amounts of sediments, for example, can pose a serious threat to coastal ecosystems including coral reefs and hence to the viability of many fisheries. Such information is important locally and regionally (since many watersheds cross national boundaries), as well as for global assessments.

73. In some instances, the lack of information reflects not just missing data collection but more fundamental scientific uncertainty. So observing systems do need to serve scientific goals as well as managerial, decision-making needs. But the balance has been to too great an extent in the other direction--a major emphasis on data collection for research purposes, and too little emphasis on data collection and analysis to meet current managerial and policy needs. Given scarce financial resources, the world will not support parallel observing systems -- for scientific and managerial purposes -- so that designing these systems to serve both purposes is critical. This is where UNEP can play a leading role.

74. Over the past five years, UNEP’s assessment programme has made important progress by establishing the GEO series and its related network of collaborating centres. The GEO process has revealed critical gaps in expertise and data. Without claiming to be comprehensive, the following paragraphs give a comparatively detailed overview of these gaps. At a more strategic level, these gaps relate to:
- lack of guidance and coordination of environmentally related observing (fragmented, incomplete and overlapping);
- too little emphasis on observing of causes and impacts;
- too little attention to problem solving, i.e. existing and alternative policies and actions.

75. Critical gaps in environmental data collection globally include a lack of current and consistent data sets on almost all significant aspects of the global environment. The majority of present global data sets suffer from the following limitations:
- consistently small scale and coarse spatial resolution;
- data derived from numerous source materials that vary in quality and time;
- based on variety of analysis methods;
- frequently out-of-date.

76. The majority of global assessments are based on an "expert opinion approach" rather than a "data-based analytical approach". The majority of global reports are "recycling" the same old data. In this connection, some of the conclusions drawn and lessons learned from the FAO Global Forest Assessments are worth mentioning, as FAO has a long history of conducting such an assessments, in some form or other, since its inception in 1945. "The quality of international statistics is, generally speaking, bad because the statistics are not available at the national level or the collected statistics are of poor quality and the comparability of the statistics is often poor."

77. The situation has not changed significantly in the last twenty years as concluded in the most recent FAO Global Forest Assessment 1990 report (which itself was only published in 1997):
- No developing country had carried out a national forest inventory containing information that could be used to generate reliable estimates of the total woody biomass volume and change;
- The data were about ten years old, on average;
- Where countries have carried out more than one forest cover assessment, they did not use appropriate techniques for change assessment;
- Only a few countries have reliable estimates of actual plantations, harvest and utilization.

78. In order to overcome the above limitations, a number of agencies have collaborated in developing globally consistent and comprehensive spatial data sets such as the global land cover characteristics database using NOAA-AVHRR satellite data, global Digital Elevation Model (DEM), global drainage basins boundaries database and global population database. These data sets are in huge demand by a variety of users.

79. Recent experience by UNEP, but also by many other organizations involved in regional or global data collection, particularly Eurostat, UNECE, EEA and OECD, shows that the problem of critical data gaps will not be solved for a long time to come with the current approach to data collection. The complexity and rapid evolution of environmental problems and knowledge mean that new data sets are required constantly. See for example the Eurostat experience with consumption and production indicators. Furthermore the link between scientifically valid data sets and information useful for policy and management purposes is certainly not automatic and often requires complex manipulations, such as modelling, expert systems, etc.

80. The global data issue can be divided into four streams:
- Lack of basic data in many developing countries because the essential observing systems are not in place;
- Lack of methodology and institutional structures to use global information derived from remote sensing sources (see also Mission to Planet Earth, The Japanese led Global Mapping initiative, etc.);
- Excessive or irrelevant data: data sets collected for scientific purposes (which may be valid in themselves) but with no specific policy application;
- Lack of analytical capability to derive useful information from observing data.

POLICY-ORIENTED ENVIRONMENTAL ASSESSMENT
(based on material submitted by RIVM)

1. Policy-oriented environmental assessment should form a basis for the formulation of the main components of UNEP’s strategy for observing and assessment. Section 1 describes the most relevant concepts and methodologies. Section 2, in contrast, identifies a few major gaps in both current assessment (expertise) and observing practices. Finally, Section 3 addresses some topics related to collaboration for improved assessment and observing.

2. In terms of methodology, three levels of environmental assessment can be identified, each requiring a higher level of integration: (i) describing the current physical state of the environment; (ii) explaining the changes - which requires changes in the physical state to be related to their main causes; and (iii) advising actual policy making processes - which also requires information on possible impacts of environmental change and potential responses. Finally, environmental assessment should be set in the context of sustainable development assessment, taking into account economic and social objectives.

3. The distinction between these different levels also has a historic context. In the late 1980s, environmental assessment tended, world-wide, to focus mainly on the quality of the environmental media (largely descriptive, concentrating on past trends and current conditions). In the same period, international activities were oriented towards building up monitoring networks and harmonizing monitoring activities. Although it is still necessary to produce such information, alone it is of limited value for decision makers. In response, environmental assessment gradually expanded to include the explaining and advising levels discussed above. Today, further development should focus on these levels. In addition, for future success, information on environmental issues should also be clearly positioned within a sustainable development context.

4. These changes in observing and assessment needs are in fact also related to changes in the general position of environmental problems in the decision making cycle. During the past few years, several global and regional environmental problems have moved from the stage of problem identification to the stages of policy setting and implementation. These stages require a more integrated type of information, and environmental assessment in many countries has started to respond to this need.

1 METHODOLOGICAL FRAMEWORK

5. Assessment of complex global and regional issues with a large environmental component requires an Integrated Environment Assessment framework that goes beyond traditional disciplinary boundaries of the natural and social sciences to incorporate the tools and insights of different sciences and different detailed issues. It is directly aimed at distilling practical messages for policy from available scientific understanding and data. Integration within the context of UNEP’s assessment activities should cover many dimensions, among them (i) capturing as many cause-effect relationships of a phenomenon as possible, (ii) addressing the cross-linkages between different phenomena, and (iii) integration across the different relevant geographical scales.

6. For Integrated Environment Assessment, communication is an essential element. On the one hand, integrated insights from the scientific community are communicated to the decision makers (or other target groups) in particular, and to society in general. On the other hand, specific questions, experiences and lessons learned by decision makers are an input for scientific assessment. This means that the ideal Integrated Environment Assessment is a cyclical and participatory process.

7. To support international decision making processes, it is important to explore how the insights of scientific assessments and their meaning for policies can be communicated adequately to decision makers. The development of appropriate indicators plays an important role in this. Assessment should be made with the specific target groups in mind and - as far as possible - in a dialogue with those groups. The need for a participatory approach is already a key element in the current Global Environment Outlook process, for example through the use of regional expertise in collaborating centres, and the involvement of regional policy makers and target groups in regional and global consultation processes.

8. There are several analytical tools and methods available for combining information from the various disciplines and sources for Integrated Environment Assessment. This annex reviews: the conceptual framework of Driving Force-Pressure-State-Impact-Response; data for Integrated Environment Assessment; the role of modelling tools and scenario construction; presentation and indicators; and key areas of expertise.

1.1 DPSIR framework

9. During the past few years, the Driving Force - Pressure - State - Impact - Response (DPSIR) framework has become accepted as an organizational framework for structuring environmental assessment and forecasting activities. However, with its implied linear causality, it is of course an oversimplification of reality.

10. The main advantage of this framework is its simplicity and its focus on causal relationships. Looking at the activities identified earlier in this annex, describing requires only knowledge on the state of the environment; explaining requires information on state and the main pressures and, finally, advising requires full knowledge of the total DPSIR framework.

1.2 Data for Integrated Environment Assessment

11. To perform assessments, data for a large number of variables are needed for the present as well as for the past situation. Section 2 discusses the observing activities that generate these data. Integrated assessment often involves combining data on different things, for example, acid deposition and vulnerability (is the critical load being exceeded?) or pollution loading and population health (are the trends in the same direction?) or banning a group of substances and ambient concentrations (does banning help?). In order to carry out this sort of analysis, even roughly, the available data should permit comparisons over time, between things at different levels of scale (e.g. urban - regional - global), between different steps in the causality chain (e.g. driving force - impact) and between different issues (e.g. climate change - fresh water availability). Therefore, consistent measurement schemes are absolutely essential to go beyond first crude estimates and to proceed to evaluation of progress.

1.3 Modelling tools and scenarios; dynamic analysis

12. Some changes in natural or societal systems can only be properly assessed if sufficiently long time scales are taken into account. This is both true for assessing the impacts of human disturbance of natural systems (e.g. the carbon and nitrogen cycles or the stock of natural resources) as well as for the impact of the penetration of new technologies, governed by capital turnover rates. Here, models and scenario analysis are useful tools in assessing these structural changes in a coherent and consistent way.

13. Models are tools for organizing knowledge at the proper time/space levels. There are several ways in which models can add to environmental reporting and assessment. These include:
- analysis of the dynamic linkages and interactions between driving forces, responses, state of the environment, impacts and responses, as well as linking different scales of analysis.
- development of forecasts and early warning by extending an analysis in time and space and by implicitly showing time delays in environmental and social processes, facilitating risk assessment.

14. Scenarios, in combination with models, can also help in providing a consistent frame for analysing potential developments. Scenarios are mostly used to identify different potential future pathways and their implications and conditions. Most scenarios do not actually predict, but instead try to paint a set of consistent images of possible futures. A typical approach to scenario development is to prepare a business-as-usual or reference scenario, complemented by different options for policy intervention. In its assessment activities, UNEP should strive to encourage the development of internally consistent scenarios for use across different regions of the globe.

1.4 Presentation

15. In view of the objective of Integrated Environmental Assessment to integrate and communicate the insights from (often complex) scientific research to decision makers, it is crucial to use appropriate visualization and presentation tools.

16. Indicators are useful tools for this. Indicators are especially meant to represent a bridge between the wealth of detail and the need for interpreted information that focusses on the main changes and interactions. Much has been done by various organizations in developing and selecting environmental indicators (in the international context OECD and UN CSD among others). Above all, one important lesson from indicator development work is that indicators should be developed in the context of the particular assessment at hand. In fact, a list of indicators could be agreed upon with each specific target group as part of their guidance to the assessment process.

17. More dynamic presentation methods also exist, among which the interactive "meta-models". These simple models contain the results of much larger models, and present them in an interactive way to target groups. Their easy-to-use interfaces allow changing parameters and variables in order to visualize modelling results under various assumptions. This has been used successfully, for example, in support of international environment negotiations. In this way, the underlying behaviour of a system can be communicated to target groups as clearly as possible.

1.5 Key areas of expertise

1.6 Key aspects of Integrated Environment Assessment

19. In the past few years, considerable experience has been built up in Integrated Environmental Assessment. On the basis of this experience, a few principles for successful Integrated Environment Assessment can be derived.
- For assessment, adequate data and observing are an absolute prerequisite. This is even more true for evaluating and projecting trends.
- There is a need for coordination of observing and assessment activities in order to fulfil the objectives of Integrated Environmental Assessment. The compilers of assessments should play an active role in identifying needs for observing data.
- Partnerships are essential for success. Cooperation with sectoral/regional centres of excellence not only taps addition sources of information (both data and expertise), but should be used to maximize consensus on the scientific basis of the assessment, and on aspects such as consistency and trustworthiness of the information..
- Combining retrospective reports and outlooks offers the possibility of generating information that is relevant for several different stages in the policy-making cycle.
- Assessments should try to aim for as much consistency in time as possible with respect to indicators, data and methodology.

20. Integrated Environmental Assessment should be a demand-driven process, although demand can seldom be specified in advance with enough detail. Programmes should therefore be in close contact with clients.

21. The use of Integrated Environmental Assessment methodology has enabled several positive developments. One is that it can provide more concrete policy recommendations. It also enables illustration of the role of specific societal sectors, and the formulation of specific sectoral policies, e.g. for transport, agriculture and industry.

2 CRITICAL GAPS

22. There are several factors which may influence the significance of the assessments, including gaps in expertise and knowledge, related deficiencies in the assessments themselves, and limited availability and reliability of data. This section points out several critical gaps, most of which can only be filled simply by continuing assessing systematically both current and future risks to the environment and to development.

2.1 Critical gaps in expertise

23. For most global environmental issues, expertise seems to be best developed for state, while uncertainties and gaps in expertise increase towards the outward ends of the DPSIR chain. This is in particular true for driving forces (in which economic and social expertise should play a key role), impacts (both in terms of impacts on humans and ecosystems) and responses. The poor understanding of the dynamics of relevant processes, the weakness in empirical data to describe trends, and insufficient knowledge about interactions and feedback currently all limit the application of integrated assessment methodologies.

24. Most of the models currently used for Integrated Environmental Assessment are either based on a natural science approach or on an economic approach. Physical processes have long been subjected to formal modelling and consequently the results are often listed as "strong knowledge". This is also true to a certain extent for the economy, but great controversies exist between various modelling traditions. The complexity of an assessment increases significantly if an attempt is made to include the human and social domains. Because of the importance of human behaviour in environmental issues, it is tempting to include these domains in assessment methodology, but at the same time understanding the dynamics of social systems is difficult (e.g. due to the variety of cultural values, beliefs and attitudes) and developing predictive capabilities almost impossible. The science is therefore often labelled as "weak". In the past few years, the search for methods to bridge the gap between the different layers of knowledge has intensified and interesting approaches have been developed. Nevertheless, further research in this field is still needed. Often different tools will have to be used to deal with assessment of human behaviour. Such tools include scenario development (e.g. using the POLESTAR environment), expert panels, and perspective-based scenario analysis.

25. An important factor for assessment is the selected geographical scale (e.g. country; region, world). Obviously, global environmental assessment needs to focus on a high level of aggregation. However, for most processes in the environmental, economic and human domains, the underlying analysis needs to be more detailed to respect important local and regional interactions and responses. For reporting, the approximately 20 large sub-regions as distinguished by UNEP for GEO can be chosen as the basic unit, but for most assessments these are too large to be considered homogeneous for the inclusion of processes acting at a sub-region level. An appropriate level of analysis has therefore to be determined. The problems related to spatial aggregation are particularly important for climate change, land and water. Many models should improve on geographic detail. The coarse resolutions of the current integrated assessment models of climate change and acidification do not allow coverage of all the necessary administrative, spatial and temporal heterogeneity and diversity that determine the driving forces and responses.

26. An important challenge in integrated assessment is finding a way to address catastrophic events of low probability of occurrence. Such events are often extremely important in their impact on environmental change, but our understanding of the relevant processes is still limited. Moreover, most models used in environmental assessments assume that real-world processes can be described in terms of deterministic mathematical qualities, while in reality many processes are stochastic by nature. This is, for instance, very relevant for climate change and water assessments (perhaps related to risks of forest fires, extreme water scarcity due to changes in rainfall patterns, harvest failure, etc.).

27. Obviously, there are also specific gaps in expertise for the various environmental issues. For acidification, for instance, attention has only recently been paid to the sensitivity of ecosystems to acid deposition outside Europe and North America. In the case of fresh water assessment, studies of nutrients are still under development.

28. Limitations in model calibration due to lack of historic data present yet another problem. Even if one can find relevant data, these data sets are often not comprehensive enough or are too unreliable to apply in a thorough calibration and validation process. The next section reviews the issue of data availability.

29. As integrated assessments evolve over time and include policies in their analysis, it becomes obvious that the links between policies and the environment are often difficult to quantify. The first and basic problem is that the degree of implementation of policies is hardly measured. This problem is encountered in attempts to characterize current policy initiatives in GEO-1 and GEO-2, in assessing the effect of multilateral agreements, and in defining a scenario without additional policies. In order to integrate the fragmentary measurements of policy implementation, and to repeat the assessment over time, a widely applicable and accepted conceptual framework is needed.

30. Finally, our knowledge of the integration of different issues is still limited. It is extremely important to continue building expertise here, since in reality ecosystems do not experience stress from single environmental problems in isolation but rather from, simultaneously, stress caused by nitrogen eutrophication, acidification (by sulphur and nitrogen oxides and ammonia), climate change, forest clearing, burning, etc.

31. It is not possible to identify and describe in detail here all critical gaps in knowledge and data for all environmental issues.

2.2 Critical data gaps

32. The foundations for the empirical base for integrated environment assessment - the basic layer of core data sets - has by and large already been laid out by the IEA/GEO Core Data Working Group in 1995/96 on the basis of early work by RIVM. The challenge before UNEP now is to make explicit the need for specific data for the Integrated Environmental Assessment programme activities and subsequently to match demand and supply. In practice this requires the collection and conversion of huge and highly diverse data sets into usable formats, verification of the data at the national or regional level, filling gaps in data and, finally, making the data available in a form convenient for those performing the assessments.

33. Integrated Assessment tools require vast amounts of empirical data for calibration and validation, not only state-of-the-environment data as such, but also data on underlying driving forces. A broad, consistent and spatially detailed historical database for integrated assessment over a long time period deserves, therefore, high priority and should include reliable gridded data sets.

34. To illustrate the major data shortcomings for current observing and to define priorities for action, the core data list is discussed below, following the DPSIR framework. A critical factor is the need for data consistency and harmonization. Typical inconsistencies relate to differences in definition of content, spatial resolution and references years. Often different geographical scales are used in different DPSIR elements. GIS tools may help translate these from one level to another, but sometimes the raw collected data simply lack enough spatial resolution.

35. Fortunately, in the past few years some progress has also been made in data collection and observing, including the development of:
- Grid-based population maps, although spatial detail, updating and extension with demographic data still need attention;
- Improved, multipurpose, comprehensive databases, such as the World Development Indicators (World Bank) and FAOSTAT (FAO); in addition to their further improvement, they have been extended and are readily available to the public;
- The new 1 km x 1 km global land-cover map (DISCover), currently being validated;
- Comprehensive environmental data reporting systems in an increasing number of countries; these have benefited from development of UN Statistics Division methodology and they open up a potential wealth of data at the country level.

Driving forces

36. Most of the information on the forces driving environmental change comes from statistical data that countries, regional bodies and global agencies (UN and others) collect, i.e. size and composition of human population (the 'consumers') and production and consumption patterns within the economy. In addition, population size and density are widely used as a tool to scale environmental pressures from the country or district level to grid cells. Such 'gridding' is needed to address the local variation in environmental pressures and in the sensitivity of ecosystems. Existing global population data need to be made consistent, improved, updated and extended with attributes including rural/urban split, age distribution and health indicators. Furthermore, the method for distributing economic activities (data from UNDP, World Bank, UNSD) over grid cells needs to be checked against available information on physical parameters within the economy (e.g. production) and compared with geographic layers (e.g. land cover)

Pressures

37. The most important information source for environmental pressures are derived data on emissions and resource abstraction, resulting from the combination of economic data with emission factors, plus data reported through national communications. The gridded emission data layers, still quite coarse and usually only available for a single year in the past, urgently need verification with historical information and updating to recent years and levels of higher detail. Deposition of emitted substances on ground and water surfaces also remains a rather weak area, partly because of the poor gridded emission data available, but also because of uncertainties in the atmospheric transport models used.

State

38. The most important data for the ‘state’ element are measurements of environmental quality (and quantity) of water, air and soils from station and/or remotely sensed data. This includes data on soils and land use, land cover, climate, water and air. Within the broad realm of observing, the empirical base for real environmental data is still weak. Soils and land use/cover are relatively well off, but global water and air monitoring are still very weak For global observing of biophysical issues, coordinating and steering bodies have been set up, i.e. GTOS (land), GCOS (climate) and GOOS (oceans), together labelled as G3OS and guided by the Integrated Global Observing Strategy (IGOS). Such coordination, already an active component of UNEP through Earthwatch, needs further strengthening and enlargement into all environmental compartments, feeding the assessment and early warning activities.

Impacts

39. Data needed on environmental impacts includes the results from calculations that combine exposure and sensitivity. For health impacts, this can be done by combining demographic data with dose-effect relations for specific pollutants, resulting in impacts on human health in terms of mortality, morbidity and disabilities. Poor demographic and health data at the country level are the main bottleneck. Better country coverage and higher spatial detail is needed to cater for the geographic differences in population characteristics. The generation of a comprehensive, harmonized mortality database would require a major effort, as shown by the compilation of the global Mortality Database of WHO, and also recently confirmed by the development of a WHO European Mortality Atlas.

40. Impact indicators for critical loads of acidification, climate change and soil erosion are under development now, although much more work still needs to be done to develop consistent and reviewed risk maps. Increasing emission levels coupled with highly variable vulnerability of ecosystems require much more analysis for completion of the global picture with enough detail. In addition, new methodologies are required to express the dimensions of biodiversity (habitat area and ecological quality), stimulating at the same time new data collection activities (habitat areas, occurrence of species).

Responses

41. Efforts to promote and organize measurements of the effects of conventions and other policies have only just begun, and there are few data available on societal response to environmental problems. This type of assessment requires ample information on environmental policies and conventions in place as a starting point. Then the effect on pressures and state of the environment needs to be estimated. Overviews of current environmental policies and directives are available for some countries, regions and the world as a whole, but these are far from complete and certainly not available in a form that allows policy efforts to be related to changes in environmental impacts. There is a need to improve, extend and link current efforts at regional and global levels and to start monitoring the effects of the policies in place. GEO-2 now shows how difficult, but also how useful, this is.

3 COLLABORATION FOR IMPROVED OBSERVING AND ASSESSMENT

42. The identification of the critical gaps described points to several important matters:
- There is a need for further development of networks of Integrated Environment Assessment practitioners (involving both generators of core data and assessment institutes) and for links to actual users. Such networks may promote a dialogue and the exchange of ideas, and provide a network to place individual activities in their broader context. The network of GEO Collaborating Centres could contribute to this.
- Centres of expertise could play an important role in both actual assessment and in filling critical gaps in expertise on real issues and in regional knowledge and data.
- Observing activities should be targeted towards actual assessment in order to reduce critical gaps here. Most international observing efforts were originally set up in the tradition of a technically-oriented observing of environmental compartments (describing). As assessment is moving towards the levels of explaining and advising, it is now time to reconsider observing activities.

3.1 Organization of data monitoring and handling

43. As suggested by the pyramidal "iceberg model of information", environmental assessment is based on a large body of underlying data. By means of assessment methodologies, these data are condensed and interpreted into the final information as reported in the assessments themselves.

44. It is important to realize that assessments by other bodies (such as FAO and UNDP) use to a large extent the same input data as UNEP. This means that overlapping activities can be avoided if observing is regarded as generic activity; the particular core data generated by one body can be used in the context of several assessments. Such collaboration also helps in ensuring consistency and building up consensus. As discussed in Section 2, the data currently used for UNEP’s assessment, and probably also for related international assessments, still contain several critical gaps.

45. The collective input data for global assessments are generated by an "underlying layer" of observing activities. Their results feed into global environment assessments as well into related assessments by other agencies.

46. For the environmental assessment process itself, a consolidated collection of input data is needed. The UNEP GRID system should be used to generate such a collection. In order to do so, the GRID centres need to collaborate with the providers of core data and the organizations involved in the assessment process itself.

47. Finally, a key element in the assessment processes is authorization. In the GEO processes, regional policy consultations have been used to allow national governments to comment on assessment activities and the data used. In future, the Collaborating Centres could be used more systematically to channel comments regarding data back to the original compilers.

3.2 Disseminating data and expertise between collaborating centres

48. The network of Collaborating Centres makes an important contribution to UNEP's assessments which could be expanded. At the same time, there are large differences in experience and type of expertise between these different centres. Consequently there is need for specific tools to disseminate knowledge and expertise within this network, and preferably to harmonize assessment methodologies as much as possible. These tools must accommodate a wide range of experience, focal areas and scientific insights.

49. Encouraging the dissemination of knowledge and expertise within the network is primarily the responsibility of UNEP. This could be supported by the development and application of a comprehensive, generic assessment tool, conceivably composed of three layers:
- a basic layer of data from observing systems consisting of exogenous inputs, scenarios and pre-run results of detailed models;
- a second layer consisting of independent expert models for priority issues, to be run separately but fed by the layer of basic data;
- a third layer of linked meta-models in which the expert models are represented by simplified images.