A new method for assessing the sustainability of land-use systems (II): Evaluating impact indicators

Abstract

In the past decade, numerous indicators and indicator sets for sustainable agriculture and sustainable land management have been proposed. In addition to their interest in comparing different management systems on an indicator by indicator basis, land managers are often interested in comparing individual indicators against a threshold, or, in order to study trade-offs, against each other. To this end it is necessary to (1) transform the original indicators into a comparable format, and (2) score these transformed indicators against a sustainability function. This paper introduces an evaluation method for land-use-related impact indicators, which was designed to accomplish these tasks. It is the second of a series of two papers, and as such it links into a larger framework for sustainability assessment of land use systems. The evaluation scheme introduced here comprises (1) a standardisation procedure, which aims at making different indicators comparable. In this procedure indicators are first normalised, by referencing them to the total impact they contribute towards, and then they are corrected by a factor describing the severity of this total impact in terms of exceeding a threshold. The procedure borrows conceptually from Life Cycle Assessment (LCA) Impact Analysis methodology; (2) a valuation procedure, which judges the individual standardised indicators with regard to sustainability. This methodology is then tested on an indicator set for the environmental impact of a spinach production system in Northwest Germany. The method highlights mineral resource consumption, greenhouse gas emission, eutrophication and impacts on soil quality as the most important environmental effects of the studied system. We then explore the effect of introducing weighting factors, reflecting the differing societal perception of diverse environmental issues. Two different sets of weighting factors are used. The influence of weighting is, however, small compared to that of the standardisation procedure introduced earlier. Finally, we explore the propagation of uncertainty (defined as a variable's 95% confidence limits) throughout the standardisation procedure using a stochastic simulation approach. The uncertainty of the analysed standardised indicator was higher than that of the non-standardised indicators by a factor of 2.0 to 2.5.