DRAFT October 26, 1998
Introduction
Monitoring has become a cause célèbre among scientists, managers, and policy makers and has occasioned a flurry of activity, discussion, and publications. A glance at the literature indicates that while the subject is not new (see Spellerberg 1991) there is considerable attention on what, how, and when to monitor (see Ecological Applications, May 1998). Despite it being a subject which has been around for many years, monitoring continues to remain a conundrum.
Much of the difficulty with monitoring stems from the multiple purposes that it is intended to serve (Ringold et al. 1996; USDA Forest Service and DOI/BLM 1994). For some, monitoring is viewed as a way to resolve some of the uneasiness among constituents seeking resolution to environmental conflict. The public often views monitoring as a watchdog to detect adverse conditions in a sufficiently timely manner to allow ameliorative steps. For example, monitoring is expected to provide data to inform and evaluate the impact of implemented forest management on a given population of spotted owls or endangered salmon. This information is expected to be of sufficient quality to provide statistically significant verification and sufficient accessibility to allow any management plan re-formulation necessary to stop further negative impacts. In contrast, the scientific community may use monitoring to provide insight into ecosystem functions and data to inform their recommendations to the public and policymakers. Finally, policy makers benefit indirectly from the assurance monitoring provides for their constituents and scientists. More directly, monitoring demonstrates active attention to a given environmental issue.
In this paper, we take the position that monitoring is first and foremost a tool for managers. That is, the principal role of monitoring is to illuminate decision making. Monitoring that works does so in three ways: (1) by providing an accurate assessment of the status of the resources being managed, (2) by validating that management decisions are correctly interpreted and implemented, and that such decisions achieve desired consequences, and (3) by providing improved insight into how systems operate. Any monitoring system that achieves these three objectives will almost certainly meet the additional expectations placed upon it by interested publics.
Large-scale regional plans such as the Northwest Forest Plan (NWFP) or Interior Columbia Basin Ecosystem Management Project (ICBEMP) create special monitoring problems. These ecosystem management projects cover relatively large, diverse areas and involve multiple social and ecological objectives. Their very nature suggests that monitoring designed to track the effectiveness of these plans inevitably will be complex and ambitious. There is a tendency to identify a large number of potential ecological and socioeconomic indicators and try to monitor them all. A large number of indicators is not inherently bad; for a complex system, it may be unavoidable. Difficulties can arise, however, in two ways: (1) without a integrated strategy for processing monitoring information, the multiple indicators deliver a cacophonous signal with no clear message, or (2) monitoring budgets are overstretched, leading to haphazard prioritization that drops important indicators in favor of those more easily measured, or worse, results in insufficient effort to monitor any single element effectively.
Some of these problems can be avoided by having a sound conceptual strategy guide the development of the monitoring plan. Herein, we offer our views on how to develop such a strategy using a decision analysis approach. Our report is organized along two major themes.
First, we examine the conceptual model that drives many conventional monitoring efforts and discuss some complicating issues that lead to confusion and debate for scientists and land managers alike. For example, poor articulation of scaling concepts and the use of ill-defined variables which are heavily value-laden (e.g., watershed integrity) can increase uncertainty. In the context of monitoring, such imprecision confounds appropriate quantification of ecological change. In response to some of these issues and complications, we propose an alternative approach for the design and context of monitoring plans in our second theme.
We propose the adoption of a formalized, quantitative process for the formulation of monitoring plans; in short, the design and use of a decision analysis approach explicitly linked to monitoring. We suggest that the lack of hypothesis formulation embedded in a decision-making framework creates the core source of the present monitoring quandary, and describe in general terms what such an approach might look like for the ICBEMP or NWFP. (We recognize that development of monitoring plans for the NWFP are well underway, while the ICBEMP effort is in the nascent stages of developing a monitoring plan.) This approach stresses the importance of clearly articulated of hypotheses germane to the land-management decisions. Extant statistical and decision theories combined with a rigorous interpretation of adaptive management provide the appropriate tools to effect accurate translation of scientific information into inference, management decisions, and policy. The proposed approach offers several advantages. The approach: (1) documents the decision making pathway (from concept to data, to management objective, to decision, to policy); (2) elucidates what and how information is used in decision making; (3) maps types and sources of uncertainty; (4) provides an objective method for prioritization of monitoring targets; (5) explicitly integrates research, management, and policy.
We begin with an overview of monitoring as it is commonly perceived.