5.4.3
Multi-attribute Utility Theory
As has been mentioned before the ranking system of various alternatives owe much of its conceptual base to the multi-attribute utility theory. Trade-off analysis used in weighting-scaling checklists involves comparison of alternatives on the basis of their relative impacts on various environmental parameters.
The first step involved in such systems is to define environment in terms of a finite number of environmental attributes that can be measured. The baseline situation for each attribute is measured and the likely situations 'with' and 'without' project are predicted over different time periods. Once attribute levels are estimated or predicted their desirability are assessed from environmental viewpoint. Such desirability is obtained from utility function curves for each attribute.
Utility function curves are developed by using systematic aggregation procedure to sum up subjective evaluation by the individuals in a group of experts. Utility function curves are developed for every attribute under consideration. A unique feature that distinguishes this methodology from all other EIA methodologies, is the capability of the multi-attribute utility theory to deal with chances, that is, the probability of attaining a specific level by an attribute.
Individual attributes are assigned scaling values. Usually a relative importance is assigned in such a way that the sum total of all such importance values equal 1.
Environmental quality index (EQI) for the alternative under consideration may be calculated by using a mathematical expression of the nature
Where,
U(X) = total utility or the composite environmental quality index for alternative X
ki = scaling factor for attribute xi
Ui(xi) = is the utility function for attribute xi
xi = ith attribute (i=1,2,....N) N = number of attributes (or decision factors)
As may be noted from the above expression, the environmental quality index is somewhat analogous to the composite environmental impact score obtained with weighting-scaling checklists. However, in the form presented above the equation appears to have ignored the interdependence between the attributes. Usually the interrelationships between the attributes are dealt with by using more complex mathematical expressions.
Methodology developed by using multi-attribute utility theory puts high resource demands. The EIA team is required to understand the theory of the utility function concept. When compared to various weighting-scaling checklists the multi-attribute utility theory appears to be more complex.
The principal advantage of this methodology is that it facilitates decision making and incorporates probability and sensitivity analysis.
5.5 Matrices
Environmental impact matrices can be seen as two-dimensional checklists in the sense that they incorporate both a list of project activities and a checklist of environmental parameters likely to be impacted. According to Sheckells (1980) matrices represent a 'progression of methodology development' over checklists. While checklists attempt to estimate likely changes in the environmental attributes in relation to the entire project (or its alternatives) under study, matrices disaggregate such projects into a series of activities and then estimate the likely impacts due to each activity-attribute interaction. According to Smith (1993) 'interaction matrices were developed from a desire to link environmental factors with project alternatives.' Inclusion of two checklists, viz., the action checklist and the environmental item checklist, in the matrix vastly aids in impact identification as items in one list can be systematically linked to all items in the other list in order to identify the likely impacts (Bisset 1984a). Typically an interaction matrix comprises a grid diagram containing the project activities on columns and the environmental attributes on rows or vice-versa. The estimated impact on an attribute due to an activity is recorded in the cell common to both. Depending on the nature and extent of information (regarding impact to be recorded the cells may be filled either in a presentational manner - using symbols or numerical scores, or in a mathematical manner - using algebraic functions (Shopley and Fuggle 1984).
In some of the interaction matrix methodologies the entries in the cells are combined using an aggregation scheme to arrive at a total impact score.
In fact early matrices pre-date US-NEPA. Although Luna B. Leopold and his colleagues are often given the credit of developing the first comprehensive matrix methodology for EIA (Bisset 1984, Canter 1977, Lohani 1984, Wathern 1988, Westman 1985), G. White was one of the pioneers who advocated this approach way back in 1968 (White 1972, Gilpin 1994). White (1972) proposed an approach for 'organising scientific investigations to deal with environmental impacts'. The suggested approach was exemplified using a test case dealing with the construction phase impacts of a dam.
Undoubtedly, the most famous of all the matrix methodologies is the 'procedure for evaluating environmental impact' developed by Leopold et al. (1971). The open-cell interaction matrix developed by them is popularly called the Leopold matrix.
Leopold et al. (1971) envisaged project planning as an action programme comprising the following eight sequential steps.
1. Statement of objectives;
2. Statement of technologic possibilities for achieving objectives;
3. Identification of proposed action or alternatives;
4. 'Environmental Characterisation Report' describing ecological conditions prior to the proposed action;
5. Development of alternative engineering plans;
6. Identification of impacts related to each alternative, and analysis of magnitude and importance of impacts;
7. Assessment of impact (textual description and bases for conclusion); and,
8. Recommendations.
As may be observed Leopold et al. (1971) considered EIA as a phase of overall project planning. Although they put a much greater attention on step (6) it is important to note that they underscored the importance of baseline studies (step 4) and impact evaluation (step 7).
The horizontal axis of the Leopold matrix contains 100 types of actions that can 'cause environmental impacts' and the vertical axis contains a checklist of 88 environmental attributes that can be affected by the project activities. This results in a 8800 cell matrix linking the actions and the attributes.
As has been mentioned before the ranking system of various alternatives owe much of its conceptual base to the multi-attribute utility theory. Trade-off analysis used in weighting-scaling checklists involves comparison of alternatives on the basis of their relative impacts on various environmental parameters.
The first step involved in such systems is to define environment in terms of a finite number of environmental attributes that can be measured. The baseline situation for each attribute is measured and the likely situations 'with' and 'without' project are predicted over different time periods. Once attribute levels are estimated or predicted their desirability are assessed from environmental viewpoint. Such desirability is obtained from utility function curves for each attribute.
Utility function curves are developed by using systematic aggregation procedure to sum up subjective evaluation by the individuals in a group of experts. Utility function curves are developed for every attribute under consideration. A unique feature that distinguishes this methodology from all other EIA methodologies, is the capability of the multi-attribute utility theory to deal with chances, that is, the probability of attaining a specific level by an attribute.
Individual attributes are assigned scaling values. Usually a relative importance is assigned in such a way that the sum total of all such importance values equal 1.
Environmental quality index (EQI) for the alternative under consideration may be calculated by using a mathematical expression of the nature
Where,
U(X) = total utility or the composite environmental quality index for alternative X
ki = scaling factor for attribute xi
Ui(xi) = is the utility function for attribute xi
xi = ith attribute (i=1,2,....N) N = number of attributes (or decision factors)
As may be noted from the above expression, the environmental quality index is somewhat analogous to the composite environmental impact score obtained with weighting-scaling checklists. However, in the form presented above the equation appears to have ignored the interdependence between the attributes. Usually the interrelationships between the attributes are dealt with by using more complex mathematical expressions.
Methodology developed by using multi-attribute utility theory puts high resource demands. The EIA team is required to understand the theory of the utility function concept. When compared to various weighting-scaling checklists the multi-attribute utility theory appears to be more complex.
The principal advantage of this methodology is that it facilitates decision making and incorporates probability and sensitivity analysis.
5.5 Matrices
Environmental impact matrices can be seen as two-dimensional checklists in the sense that they incorporate both a list of project activities and a checklist of environmental parameters likely to be impacted. According to Sheckells (1980) matrices represent a 'progression of methodology development' over checklists. While checklists attempt to estimate likely changes in the environmental attributes in relation to the entire project (or its alternatives) under study, matrices disaggregate such projects into a series of activities and then estimate the likely impacts due to each activity-attribute interaction. According to Smith (1993) 'interaction matrices were developed from a desire to link environmental factors with project alternatives.' Inclusion of two checklists, viz., the action checklist and the environmental item checklist, in the matrix vastly aids in impact identification as items in one list can be systematically linked to all items in the other list in order to identify the likely impacts (Bisset 1984a). Typically an interaction matrix comprises a grid diagram containing the project activities on columns and the environmental attributes on rows or vice-versa. The estimated impact on an attribute due to an activity is recorded in the cell common to both. Depending on the nature and extent of information (regarding impact to be recorded the cells may be filled either in a presentational manner - using symbols or numerical scores, or in a mathematical manner - using algebraic functions (Shopley and Fuggle 1984).
In some of the interaction matrix methodologies the entries in the cells are combined using an aggregation scheme to arrive at a total impact score.
In fact early matrices pre-date US-NEPA. Although Luna B. Leopold and his colleagues are often given the credit of developing the first comprehensive matrix methodology for EIA (Bisset 1984, Canter 1977, Lohani 1984, Wathern 1988, Westman 1985), G. White was one of the pioneers who advocated this approach way back in 1968 (White 1972, Gilpin 1994). White (1972) proposed an approach for 'organising scientific investigations to deal with environmental impacts'. The suggested approach was exemplified using a test case dealing with the construction phase impacts of a dam.
Undoubtedly, the most famous of all the matrix methodologies is the 'procedure for evaluating environmental impact' developed by Leopold et al. (1971). The open-cell interaction matrix developed by them is popularly called the Leopold matrix.
Leopold et al. (1971) envisaged project planning as an action programme comprising the following eight sequential steps.
1. Statement of objectives;
2. Statement of technologic possibilities for achieving objectives;
3. Identification of proposed action or alternatives;
4. 'Environmental Characterisation Report' describing ecological conditions prior to the proposed action;
5. Development of alternative engineering plans;
6. Identification of impacts related to each alternative, and analysis of magnitude and importance of impacts;
7. Assessment of impact (textual description and bases for conclusion); and,
8. Recommendations.
As may be observed Leopold et al. (1971) considered EIA as a phase of overall project planning. Although they put a much greater attention on step (6) it is important to note that they underscored the importance of baseline studies (step 4) and impact evaluation (step 7).
The horizontal axis of the Leopold matrix contains 100 types of actions that can 'cause environmental impacts' and the vertical axis contains a checklist of 88 environmental attributes that can be affected by the project activities. This results in a 8800 cell matrix linking the actions and the attributes.