difficult because it is not certain what or how frequently to monitor. The regulation of air pollution provides a good example. Airsheds are significantly affected by variables beyond the control of the institutional design (for example, wind velocity and direction, air temperature). In addition, variations on short time scales are consequential. As a result, monitoring is needed not only of long-term (e.g., annual) averages, but also on short time scales (e.g., hourly maximum limits).
Imperfect understanding also raises significant management problems because of different interpretations of data from monitoring. Scientific experts and resource users are likely to disagree about interpretation, especially when there is a significant divergence of values and interests among the appropriators (Dietz and Stern, 1998; National Research Council, 1996, 1999). The allocation of rights to resource flows and the transferability of these rights also can be problematic when knowledge is lacking about what the appropriate limits should be for resource use (see Tietenberg, Chapter 6; Rose, Chapter 7) and when diverse appropriators disagree about how cautious to be under uncertainty.
Wilson (Chapter 10) suggests that the more uncertain and variable the resource stocks are, the more effort needs to be put into frequent monitoring of the stocks, notification of users and managers when stocks are at levels that should be a cause for concern, and procedures for redefining the limits of resource withdrawals. Many researchers have concluded that uncertainty based in ignorance requires flexible institutions that adjust to improved understanding, allow users to quickly redefine the limits of resource use when the resource stocks require it, and incorporate low-cost conflict resolution methods. It is not yet clear, however, how best to design decision processes that can create the needed flexibility and responsiveness to conflicting demands (National Research Council, 1996; Wilson, Chapter 10).
Environmental systems do not neatly match the boundaries of the social systems within which they are managed. It is thus unlikely that the rules of any one social system will be adequate for resource management. It is necessary to link institutions both horizontally (across space) and vertically (across levels of organization). The need for vertical linkage is especially critical for resources of large size or high complexity or whose use results in extensive negative externalities for other common-pool resources (Karlsson, 2000). Higher level institutions may support the authority for local enforcement and provide resources for local monitoring or enforcement. When favorable conditions do not exist for local monitoring and enforcement, external authorities can help by providing information, long-term contracts, and enforcement mechanisms, taking into consideration the views of local resource users (Morrow and Hull, 1996). The most extreme challenges of linkage probably arise for global resource management (e.g., the atmosphere, the oceans, global biodiversity). Here, a global interest exists in managing resources