Box 18A Data availability

Industrial accidents

Data on accidents provide valuable lessons not only on how an accident happened but also about the clean-up or remediation process. They are important as inputs to predictive analyses for risk management and emergency planning. A better knowledge of the behaviour of various contaminants in the environment and more comprehensive data on small spills and accidental discharges are necessary to establish a reliable picture of the impacts of accidents on the environment. Until this information is available, it is difficult to be sure of the appropriateness of the priorities adopted for risk management (see Chapter 30).

Most data are currently collected on a national basis. Few international databases exist. There is a paucity of information for Eastern Europe, while Northern Europe is covered best of all. The main sources of data on major accidents worldwide are Lloyd's Weekly, the Loss Prevention Bulletin, the Rhine Commission Reports, the French Survey of Accidents, the MHIDAS database at UKAEA, the AHE database at USEPA, the ENVIDAS database, the FACTS database at TNO in The Netherlands and MARS of the European Commission's Joint Research Centre. The information compiled usually includes a description of the location and nature of the accident, as well as any known causes. Information on the type of release as well as its size is also usually given, although this information can be very approximate.

The impacts of individual accidents on human health are usually well reported, and especially so if the accident has caused fatalities. However, there are often surprising discrepancies between sources concerning consequences even in major accident reporting, and varying opinions on causes and effects, not only as part of legitimate scientific debate, can cloud a clear understanding of impacts. Numbers of human fatalities are widely used to assess the significance or scale of an accident (see Chapter 30). However, information on environmental damage is usually restricted to observations of the more obvious short-term impacts, such as immediate fish kills. More detailed information on long-term impacts is often not given or not known. (This has been well researched recently by Lindgaard-Jorgensen and Bender (1992) for the Community Documentation Centre on Industrial Risk.) There is a particular lack of information on the impacts of accidents affecting the terrestrial environment. For this purpose the technical journals and periodicals are a more useful source of information, and they tend to report the results of long-term post-accident monitoring of the affected environment in more detail.

Often accidents are reported incompletely and much important information is missing. A standard system of reporting would help to alleviate this problem and enable accidents to be compared. Overall, there are significant gaps in the data on accidents that should be filled, the most important of which are ecological effects, long-term recovery of the environment and clean-up actions.

Within the context of ecological effects and environment recovery there is a significant paucity of information on the ecotoxicology of chemicals (see Chapters 17 and 38). While the chemical and physical characteristics of most chemicals are well known, their complex environmental behaviour is poorly understood. Furthermore, baseline information is usually insufficient to be able to assess the environmental effects of accidents. Data from long-term environmental monitoring is very scarce. Clean-up actions are often reported incompletely, if at all. In some cases no clean-up is undertaken following an accident, while in other cases details are simply not recorded.

Natural hazards

In general, data on natural hazards are widely available from a number of sources. However, the type of information varies considerably according to the reporting purpose, and information on environmental impacts is seriously lacking.

Data are collected at the regional, national and international levels. Most reliable data come from developing countries, as well as certain parts of Eastern Europe and the USA. This is because of the propensity for natural disasters to occur in these regions. Some of the main sources include: the major re-insurance companies (such as Swiss Re), the International Red Cross and Red Crescent Societies, the UN Department of Humanitarian Affairs (DHA), the UN International Decade for Natural Disaster Reduction (IDNDR), the Centre for Research on the Epidemiology of Disasters (CRED), the European Programme on Climatology and Natural Hazards (EPOCH), the Bradford University Disaster Prevention and Limitation Unit, the Disaster Research Centre at the University of Delaware and the Natural Hazards Centre at the University of Colorado.

The classification of the impact of a natural hazard depends on the source of data. International sources such as the UN have traditionally classified 'impact' as the number of deaths and/or economic loss, whereas re-insurance companies focus on the number and value of insurance claims. The information records usually list the affected country and region together with the type of disaster. The level of 'damage' given is based on the chosen definition of impact.

It has not been routine to report upon the environmental impacts of natural hazards. Although reports on the link between natural hazards and environmental impact are occasionally released, no systematic reporting structure exists. There is a particular lack of information on the potential interaction between natural hazards and human activities as a cause of environmental impact. A significant issue which monitoring should take into account is the transfer of impact geographically and often transnationally (eg: drought, irrigation and water resource depletion elsewhere; deforestation, soil erosion, floods; coastal erosion, coastal defences, loss of sediment supply elsewhere, erosion elsewhere).