# 216

Dear Amanuel Biset Belete, Asmare Belay Nigussie and Dawit Munye

I read your paper

Integrated AHP and GIS-Based Approach for Suitable Landfill Site Selection to Improve Solid Waste Management: A Case Study of Dessie Town, Ethiopia

My comments

A very important subject, and I was impressed by the care you took in gathering data, really you spent a lot of time in preparing the information to model and the problem solved by a MCDM model. Congratulations, a work well done on something that is not easy. Your explanations about determining the size of the sampling are excellent, the same as the comments on criteria and buffers, something that rarely is considered

Unfortunatelly, I cannot say the same regarding treatment of this problem by MCDM methods. There are many inconsistencies, that defy reasoning and common sense. Therefore, in my opinion, it was an excellent planning, but not accompanied by a reasonable thinking. I am not judging your work, only pointing out what I read, interpreted, and called my attention.

It took me some time to understand the problem because you do not define it at the beginning, that is, which are the selected sites, that apparently were decided by other means not related to MCDM, something naturally expected, and you define and beautifully explain the suitability of each one of the 12 criteria, but do not add the three criteria: wind, distance and size. I follow closely landfill constructions since long time ago, and I learnt that the stakeholders must make first a pre-selection of sites, based in size, and not only because economies of scale.

Therefore, preselected sites must have a minimum size that can assure landfilling for a certain number of years. However, surface is not enough, for you also need depths, that is, you must work with volumes or with capacity for compacted waste, something that you wisely consider. By the way, the minimum number of years of operation should also be a criterion. In this way, the initial matrix must reflect the advantages and disadvantages of each site, and the software must work to find the best solution that gives the best balancs, when all criteria and sites are considered simultaneously, something that certainly you cannot do with AHP. One of the reasons, is that this method works only with independent criteria, which certainly is not the case here.

Obviously, once this has been decided, the selection of the best site may start subject to many factors or criteria including: Minimum depth, soil quality, distances, treatment plant for leachate, lining, potential use of CH4 produced, etc. I think that three aspects, wind, size and distance must be considered as criteria and joint the set of the other 12. I do not understand why you consider them independent of them.

You have a very good list of criteria and very well documented, and so, the job is more than half done.

In my opinion, also shared by other colleagues, AHP procedure in selecting first criteria and prior to alternatives is incorrect. There could be alternatives or sites, that for some reason my not agree with those criteria or even demand new ones. Just as an example: Suppose that one of the preselected sites fills most conditions, good size and terrain to even increase it in the future, excellent clay terrain, but also demands very serious and time demanding hydrologic as well as stability studies, which are not required b y the other sites. If you do not know this circumstance, how are you going to decide a criterion that refers to it? AHP proceeds in the opposite sense, since it determines first the criteria and after the alternatives they must evaluate.

Another error, you probably know that AHP demands that all criteria be independent. In your case, prone and soil are related, since for instance, GWT is most probably related to residential, because they get their water from underground. Consequently, you cannot apply AHP. This is not my opinion or thinking, it was very clearly stated in writing by AHP creator, Thomas Saaty

You have 25 experts, a considerable number indeed, however, if you have 15 criteria (12 + 3), each criterion must be compared with the other 14. Say that one criterion is GWT and you are an expert in this field, and want to compare the importance of this GWT criterion against say a residential criterion, a social issue and needing a social expert. How can you reach an agreement with this expert if you are not proficient in social issues, and vice-versa, and this is repeated in all 15 criteria for all 25 experts? Every expert is discussing on something that he ignores. Where is the rational of this?

In page 7, in Assigning Criteria Weights, I believe that you do not use the right procedure, simply because weights measure the relative importance of criteria, buy they do not evaluate alternatives. What evaluates alternatives is the dispersion of values contained in each criterion, that is, its content, not the envelope, as per Shannon’s Theorem.

As you can see the method that you chose to solve this problem, is not able to use and profit the excellent data that you acquired.

These are my comments, and of course, I will be delighted in having your answer to them.

Nolberto Munier