5.5.1 Modifications
to Leopold Matrix
Leopold matrix was one of the first methodologies which were not sector-specific. It evoked great interest in USA during the early seventies. Many EIA practitioners and scholars attempted revision of the methodology to suit their own needs. As mentioned before the Leopold matrix scheme does not allow arriving at a total impact score.
Some scholars ignored the Leopold premise and suggested that the magnitude scores as well as the importance scores for each row (or column) can offer insight into impact assessment and interpretation (Lohani and Halim 1982). Based on such an erroneous premise Lohani and Thanh (1977) suggested a modification of the Leopold matrix to arrive at a total impact score. They suggested a scheme in which relative weights are assigned to each developmental activity on a ranking scale (The weightage values are ordinal). The impact on the environment owing to the project activity 'i' may thus be given by the ith column total
Where Pi = relative weight of the ith activity;
Iij = importance of the interaction between ith activity and jth cell;
Mij = magnitude of impact on the jth environmental attribute caused by the ith activity.
Total impact score for the project may be given as
When 'n' number of activities act on 'm' number of environmental parameters.
Lohani and Thanh (1977) argued that if the relative priority of a development activity is determined, then the total value of a particular activity is the vertical sum of the column represented by that activity multiplied by the priority value. This seems to be logically weak. Use of a ranking scale (priority value) instead of a weighting scale has weakened the promise. Moreover the scheme of representing impact by the product of 'importance' and 'magnitude' does not have any physical significance. It has already been pointed out in this chapter that ordinal values cannot be summed or averaged. In the Lohani and Thanh's modification such ordinal values have been multiplied with corresponding 'importance values and magnitude values to arrive at the impact value (score). This representation is not backed by interpretation and hence is not acceptable.
Parker and Howard (1977) modified Leopold matrix by replacing the two scales of 1-10 with a single scale of 0-5. The modified scale was used to measure temporal and importance aspects of the impacts. One important drawback of the scheme is that it does not consider impact magnitude at all. The application of their scheme was however limited to site-selection for exploratory drilling in Antarctica.
EIA scholars, including Rau (1980) and Westman (1985), often felt that arriving at a grand index is advantageous as it allows a quantitative comparison of alternatives. Rau (1980) proposed the following summarisation scheme for Leopold matrix:
Where,
Mij = magnitude of the jth action on the ith environmental factor; and,
Wij = importance weighting of the jth action the ith environmental factor
It may be observed that Rau's scheme is somewhat similar to that of Lohani and Thanh (1977) excepting that Rau did not incorporate any scheme of prioritising the activities. Moreover Rau tried to justify the grand indexing by interpreting the total project impact as a quality-of-life indicator. His logic was that the term 'Mij' represented the magnitude of impact of the jth action on the ith quality-of-life factor. He strengthened his argument by stating that Wij is the importance of the interaction between the jth action and the ith quality-of-life factor. Hence the product MijxWij is a representative index.
Fischer and Davies (1973) proposed a matrix methodology comprising three steps as:
1. Environmental baseline evaluation;
2. Compatibility matrix; and,
3. Decision matrix.
The baseline evaluation matrix identifies the important environmental elements, their present condition and their susceptibility to management. The environmental parameters are assessed on a scale of 1 to 5 and only the parameters receiving an importance score of 4 or 5 are assessed in the compatibility matrix, which is an open cell matrix.
The compatibility matrix assesses impacts of the project activities on the chosen attributes on a scale of -5 to +5, with '0' denoting no-impact. Short term and long-term impacts are suffixed as S and L respectively. Once again parameters in the impact score range of -5 to -4 and +4 to +5 only are considered in the final decision matrix. The final step brings together in one format all the necessary information for decision-making. Although this is called a matrix method, the final decision matrix is actually a checklist and this method can therefore be termed a stepped matrix-checklist method. Like other matrices this scheme is heavily reliant on the subjective evaluation of the assessors. Moreover at each step parameters receiving a score of 3 or less (in a scale of 0-5) is screened out. This may lead to a situation where projects with significant overall impact but having low impact on individual parameters may be preferred.
Phillip and Defillipi (1976) adopted BEES into matrix format. The methodology was developed for use in wastewater management projects but can be easily adopted for use in other types of projects. Like BEES a total parameter importance value of 1000 is distributed to 15 parameters in 4 categories. Relative importance is assigned to each activity-parameter interaction in such a way that the sum of all relative importance for a particular parameter is 1.00. This methodology is highly replicable but it shows inflexibility. The magnitude values are assigned using a scale of 0-1 but no scheme of value function curve is provided.
In India a rudimentary form of a quantitative matrix, called the CMPDIL matrix, is in use in the coal mining sector. A joint committee of CMPDIL and MOEF proposed the matrix in the year 1986. This sector-specific matrix is an 11 by 12 matrix showing project activities on rows and environmental attributes in columns. Temporal aspects are incorporated by including 'site preparation' and 'mine construction' in the list of activities. The activities considered are as given below.
The environment is divided into the following parameters:-
A total weight of 1000 is distributed to these environmental parameters. This matrix, in the name of considering socio-economic factors, assigns an importance value of 250 to non-environmental parameters such as health, population, services and aesthetic value.
As if that was not enough, two more parameters, employment and literacy, have been added to the list of environmental parameters and they have been assigned importance weightage of 75 and 50 respectively making the total weight equal to 1125. This has resulted in a situation where environmental assessment is done with one third of the total weight being assigned to parameters not belonging to biophysical environment category. The revised matrix now comprises 14 entries under environmental parameters with a total weight of 1125.
Leopold matrix was one of the first methodologies which were not sector-specific. It evoked great interest in USA during the early seventies. Many EIA practitioners and scholars attempted revision of the methodology to suit their own needs. As mentioned before the Leopold matrix scheme does not allow arriving at a total impact score.
Some scholars ignored the Leopold premise and suggested that the magnitude scores as well as the importance scores for each row (or column) can offer insight into impact assessment and interpretation (Lohani and Halim 1982). Based on such an erroneous premise Lohani and Thanh (1977) suggested a modification of the Leopold matrix to arrive at a total impact score. They suggested a scheme in which relative weights are assigned to each developmental activity on a ranking scale (The weightage values are ordinal). The impact on the environment owing to the project activity 'i' may thus be given by the ith column total
Where Pi = relative weight of the ith activity;
Iij = importance of the interaction between ith activity and jth cell;
Mij = magnitude of impact on the jth environmental attribute caused by the ith activity.
Total impact score for the project may be given as
When 'n' number of activities act on 'm' number of environmental parameters.
Lohani and Thanh (1977) argued that if the relative priority of a development activity is determined, then the total value of a particular activity is the vertical sum of the column represented by that activity multiplied by the priority value. This seems to be logically weak. Use of a ranking scale (priority value) instead of a weighting scale has weakened the promise. Moreover the scheme of representing impact by the product of 'importance' and 'magnitude' does not have any physical significance. It has already been pointed out in this chapter that ordinal values cannot be summed or averaged. In the Lohani and Thanh's modification such ordinal values have been multiplied with corresponding 'importance values and magnitude values to arrive at the impact value (score). This representation is not backed by interpretation and hence is not acceptable.
Parker and Howard (1977) modified Leopold matrix by replacing the two scales of 1-10 with a single scale of 0-5. The modified scale was used to measure temporal and importance aspects of the impacts. One important drawback of the scheme is that it does not consider impact magnitude at all. The application of their scheme was however limited to site-selection for exploratory drilling in Antarctica.
EIA scholars, including Rau (1980) and Westman (1985), often felt that arriving at a grand index is advantageous as it allows a quantitative comparison of alternatives. Rau (1980) proposed the following summarisation scheme for Leopold matrix:
| Total impact on the ith environmental factor from all actions |  |
| Total impact of the jth action on all environmental factors |  |
| Total project impact |  |
Where,
Mij = magnitude of the jth action on the ith environmental factor; and,
Wij = importance weighting of the jth action the ith environmental factor
It may be observed that Rau's scheme is somewhat similar to that of Lohani and Thanh (1977) excepting that Rau did not incorporate any scheme of prioritising the activities. Moreover Rau tried to justify the grand indexing by interpreting the total project impact as a quality-of-life indicator. His logic was that the term 'Mij' represented the magnitude of impact of the jth action on the ith quality-of-life factor. He strengthened his argument by stating that Wij is the importance of the interaction between the jth action and the ith quality-of-life factor. Hence the product MijxWij is a representative index.
Fischer and Davies (1973) proposed a matrix methodology comprising three steps as:
1. Environmental baseline evaluation;
2. Compatibility matrix; and,
3. Decision matrix.
The baseline evaluation matrix identifies the important environmental elements, their present condition and their susceptibility to management. The environmental parameters are assessed on a scale of 1 to 5 and only the parameters receiving an importance score of 4 or 5 are assessed in the compatibility matrix, which is an open cell matrix.
The compatibility matrix assesses impacts of the project activities on the chosen attributes on a scale of -5 to +5, with '0' denoting no-impact. Short term and long-term impacts are suffixed as S and L respectively. Once again parameters in the impact score range of -5 to -4 and +4 to +5 only are considered in the final decision matrix. The final step brings together in one format all the necessary information for decision-making. Although this is called a matrix method, the final decision matrix is actually a checklist and this method can therefore be termed a stepped matrix-checklist method. Like other matrices this scheme is heavily reliant on the subjective evaluation of the assessors. Moreover at each step parameters receiving a score of 3 or less (in a scale of 0-5) is screened out. This may lead to a situation where projects with significant overall impact but having low impact on individual parameters may be preferred.
Phillip and Defillipi (1976) adopted BEES into matrix format. The methodology was developed for use in wastewater management projects but can be easily adopted for use in other types of projects. Like BEES a total parameter importance value of 1000 is distributed to 15 parameters in 4 categories. Relative importance is assigned to each activity-parameter interaction in such a way that the sum of all relative importance for a particular parameter is 1.00. This methodology is highly replicable but it shows inflexibility. The magnitude values are assigned using a scale of 0-1 but no scheme of value function curve is provided.
In India a rudimentary form of a quantitative matrix, called the CMPDIL matrix, is in use in the coal mining sector. A joint committee of CMPDIL and MOEF proposed the matrix in the year 1986. This sector-specific matrix is an 11 by 12 matrix showing project activities on rows and environmental attributes in columns. Temporal aspects are incorporated by including 'site preparation' and 'mine construction' in the list of activities. The activities considered are as given below.
| 1. Site preparation; | 2. Mine construction; |
| 3. Coal/O.B. extraction; | 4. Coal/O.B. transportation; |
| 5. Formation of O.B./solid waste dumps; | 6. CHP/Coal preparation plant; |
| 7. Workshop; | 8. Fan house and other structure; |
| 9. Land reclamation; | 10. Afforestation/plantation; and, |
| 11. Community development. |
The environment is divided into the following parameters:-
| 1. Land use; | 2. Ecology/forest; |
| 3. Air quality; | 4. Major surface water course; |
| 5. Ground water; | 6. Water quality; |
| 7. Noise; | 8. Ground vibration; |
| 9. Health; | 10. Services; |
| 11. Population/migration; and, | 12. Aesthetic value |
A total weight of 1000 is distributed to these environmental parameters. This matrix, in the name of considering socio-economic factors, assigns an importance value of 250 to non-environmental parameters such as health, population, services and aesthetic value.
As if that was not enough, two more parameters, employment and literacy, have been added to the list of environmental parameters and they have been assigned importance weightage of 75 and 50 respectively making the total weight equal to 1125. This has resulted in a situation where environmental assessment is done with one third of the total weight being assigned to parameters not belonging to biophysical environment category. The revised matrix now comprises 14 entries under environmental parameters with a total weight of 1125.