detailed set of environmental accounts. The above comparison may give some sense of why accounting for environmental flows outside the marketplace is such a daunting task.
Some of the analytical questions involved in environmental accounting have been analyzed in the previous section. To construct actual accounts requires both obtaining accurate physical data (discussed in this section) and valuing the flows (discussed in the next section).
Accurate data on physical flows and stocks are a prerequisite for developing any accounting system and are the focus of national accounting systems under development in several European nations. In some areas, ample physical data are available as a by-product of regulatory monitoring and resource management systems. Appendix B lists a number of databases identified by the panel that may be of use in further work on supplemental accounts.
Three concerns are fundamental to understanding data and measurement requirements for the development of environmental accounts: (1) the dose-response relationship, (2) measurement of actual doses experienced, and (3) the fate and transport of residuals in the environment. The first, the dose-response relationship, is the physical relationship between the concentration of or exposure to an environmental change and the response of the subject experiencing the dose. The dose-response relationship is applied to many different situations, for example, the response of trees and crops to chemicals such as carbon dioxide, tropospheric ozone, or acid deposition and the response of humans to pollutants such as lead, particulate matter, or radiation.
Dose-response relationships are often difficult to determine because they may be affected by complex interactions and intervening factors. For example, there are extensive medical data on causes of death and, less universally, illness. To determine impacts of environmental changes on human or natural ecosystems requires separating out the different causes of premature death or illness. In some areas, such as the impact of tobacco or lead, the relationships are relatively well established; in other areas, such as the impact of particulate matter or ozone, much uncertainty persists. For many of these relationships, average exposure over the year is rarely the relevant measure. Damage may be related to extreme levels or to periods in which the subject is particularly sensitive to the agent; acute effects may differ from chronic effects related to long-term, low-level exposure.
Resolving these uncertainties about dose-response relationships is important for policy decisions, such as the level at which to set primary