Selection and modeling of sustainable development indicators: a case studyof the Fraser River Basin, British Columbia

A large number of groups in British Columbia and Canada as a whole have exp ended considerable effort to define sustainable development goals and, to a lesser extent, select sustainable development indicators. Many hope that c ollection and monitoring of such indicators will provide important policy g uidance to decision-makers and provide a means for tracking sustainable dev elopment. This project was designed with the following goals: (i) to select indicators that could be linked to an operational definition for sustainab le development within the Fraser River Basin (British Columbia); (ii) to as sess the accessibility, quality and relevance of the best available data fo r developing such indicators; (iii) to apply modeling techniques for discer ning the linkages among indicators; and (iv) to provide recommendations for further indicator selection and modeling research. The major methodologica l conclusions reached were: (i) it is important to link sustainable develop ment goals to movements of a small slate of individual indicators as single indicators can rarely be linked to any specific sustainable development go al; (ii) the poor quality, inaccessibility and irrelevance of existing data are pervasive constraints to reliable indicator modeling; (iii) modeling i s most appropriate at aggregated spatial scales such as provinces or large watersheds, while modeling at smaller ecosystem-based spatial levels is fea sible but unreliable; and (iv) linking the use of deterministic and qualita tive modeling approaches is a useful means for projecting indicators and di scerning important policy linkages, while conventional statistical modeling approaches are frequently inappropriate because of unreliable or non-comme nsurable data. This has profound implications for indicator selection and m odeling. In contrast to much of the current indicator work, which relies on selecting a large number of detailed specific indicators, it would be more fruitful and less costly to focus attention on a small number of indicator s within selected indicator classes (such as economic, social, environmenta l or human health indicators). The precise specification of the indicator w ithin each of these classes is of less consequence. Indicator modeling work should focus on larger scale systems, as opposed to smaller ecosystem unit s. Such work is most suited to identifying qualitative policy trade-offs an d implications, rather than to forecasting specific indicators. Data gather ing efforts can be scaled down substantially and should be informed, but no t unduly constrained, by available model frameworks. Greater focus is requi red on modeling frameworks that can use incomplete data sets or qualitative information, and linking existing quantitative model structures to externa l qualitative models. (C) 1999 Elsevier Science B.V. All rights reserved.