Environmental impact assessment using the evidential reasoning approach

Abstract

Environmental impact assessment (EIA) problems are often characterised by a large number of identified environmental factors that are qualitative in nature and can only be assessed on the basis of human judgments, which inevitably involve various types of uncertainties such as ignorance and fuzziness. So, EIA problems need to be modelled and analysed using methods that can handle uncertainties. The evidential reasoning (ER) approach provides such a modelling framework and analysis method. In this paper the ER approach will be applied to conduct EIA analysis for the first time. The environmental impact consequences are characterized by a set of assessment grades that are assumed to be collectively exhaustive and mutually exclusive. All assessment information, quantitative or qualitative, complete or incomplete, and precise or imprecise, is modelled using a unified framework of a belief structure. The original ER approach with a recursive ER algorithm will be introduced and a new analytical ER algorithm will be investigated which provides a means for using the ER approach in decision situations where an explicit ER aggregation function is needed such as in optimisation problems. The ER approach will be used to aggregate multiple environmental factors, resulting in an aggregated distributed assessment for each alternative policy. A numerical example and its modified version are studied to illustrate the detailed implementation process of the ER approach and demonstrate its potential applications in EIA.

Introduction

Environmental impact assessment (EIA) is concerned with the systematic identification and evaluation of the potential impacts (effects), both beneficial and harmful, of proposed projects, plans, programmes or legislative actions related to the physical–chemical, biological, cultural, and socio-economic components of the total environment. The primary purpose of the EIA process is to encourage the consideration of the environment in planning and decision making and to ultimately arrive at actions which are more environmentally compatible (Canter, 1996).

Since its introduction in the United States in the late 1960s, EIA has been adopted and implemented by many developed and developing countries (Sowman et al., 1995, Leu et al., 1996, Bojórquez-Tapia and Garcı´a, 1998, Barker and Wood, 1999, Chen et al., 1999, Hopkinson et al., 2000, Weston, 2000, Jay and Handley, 2001, Steinemann, 2001, Appiah-Opoku, 2001, Tran et al., 2002, Henne et al., 2002). Numerous EIA methodologies have been developed such as interaction matrices, networks, weighting-scaling (or -ranking or -rating) checklists (Canter, 1996), multicriteria/multiattribute decision analysis (MCDA/MADA) (Parkin, 1992, Marttunen and Hämäläinen, 1995, McDaniels, 1996, Hokkanen and Salminen, 1997a, Hokkanen and Salminen, 1997b, Hokkanen and Salminen, 1997c, Hokkanen et al., 1998, Hokkanen et al., 1999, Kim et al., 1998, Rogers and Bruen, 1998, Salminen et al., 1998, Wang and Yang, 1998, Lahdelma et al., 2000, Lahdelma et al., 2002, Janssen, 2001, Kwak et al., 2002, Pun et al., 2003), input–output analysis (Lenzen et al., 2003), life cycle assessment (LCA) (Tukker, 2000, Brentrup et al., 2004a, Brentrup et al., 2004b), AHP or fuzzy AHP (Ramanathan, 2001, Goyal and Deshpande, 2001, Tran et al., 2002, Sólnes, 2003), fuzzy sets approaches (Munda et al., 1994, Parashar et al., 1997, Enea and Salemi, 2001, Hui et al., 2002), Rapid Impact Assessment Matrix (RIAM) (Pastakia, 1998, Hagebro, 1998, Pastakia and Jensen, 1998, Pastakia and Bay, 1998), and data envelopment analysis (DEA) (Wei et al., 2004).

Since EIA problems are often characterized by a large number of identified environmental factors, most of which are qualitative in nature and can only be assessed on the basis of human judgments, EIA methods must be able to deal with various uncertainties that are inevitably involved in subjective judgments due to human being’s inability to provide accurate judgments, or the lack of information, or the vagueness of meanings about environmental factors and their assessments. There is a clear need to develop methods which can be used to handle various uncertainties such as ignorance and fuzziness simultaneously. This paper is dedicated to exploring an EIA method based on the evidential reasoning (ER) approach which is developed on the basis of decision theory and the Dempster–Shafer (D–S) theory of evidence (Dempster, 1967, Shafer, 1976) and can be used to model various uncertainties in EIA process.

The rest of the paper is organized as follows. Section 2 gives a brief description of the Dempter–Shafer theory of evidence, which is the basis of the ER approach. In Section 3, the ER approach for EIA will be fully investigated, including the identification of environmental factors, the ER distributed modelling framework, the description of the recursive ER algorithm, the development of a new analytical ER algorithm, and the utility interval based ER ranking method. Section 4 presents the examination of an example and its modified version using both complete and incomplete assessment information to show the detailed implementation process of the ER approach and its potential applications in EIA. The paper is concluded in Section 5 with a discussion about the features of the ER approach. The derivation of the analytical ER algorithm is provided in Appendix A.