Assessing the Response to Land Degradation Risk: The Case of the Loulouka Catchment Basin in Burkina Faso

Abstract

This work is concerned with land use assessment in a region of Burkina Faso, Western Africa. It can help to support the definition of politics promoting a sustainable development of the region. A spatial decision support model is built, based on a coupling of Multi-Criteria Decision Analysis (MCDA) on the one hand and a Geographic Information System (GIS) on the other hand. The Electre Tri method is used to sort the spatial units to ordered categories corresponding to various levels of response to the risk of landscape degradation. The model allows to aggregate physical, economical, socio-cultural and environmental aspects interpreted in terms of their impact on landscape preservation or degradation. Such a categorization leads to the determination of homogeneous zones in the region under study. It can possibly serve as a basis for allocating resources to the most promising sub-regions or the zones needing an urgent intervention.

Appendices

Appendix: Description of Criteria and Indicators

1.1 Criteria and Indicators for Assessing the ERO Principle (P ₁)

Four criteria were defined to account for Principle 1, i.e. erosion limitation

1.1.1 Adequate Pedo-Geomorphologic Choice (Cr ₁₁)

The soil texture or pedology (examples: sandy soil, clayey soil) and the geomorphology [example: slope, plain, plateau, shoal (“bas fond”)] must be in close relation with the type of cultivation. If cultivated in a poor soil, a cultivation requiring soils that are rich in nutritive elements will impoverish the poor soil even more until it may become inappropriate to any vegetation, while vegetation can protect soils against rains and winds. Such a situation triggers soil erosion. This criterion has only one indicator:

• i ₁₁₁ : comparison with agricultural aptitude map.

A pedo-geomorphologic map describing the agricultural aptitude of the area under study was obtained from BUNASOL.⁸ This map allowed us to assign a class of agricultural aptitude to each spatial unit. Table 12.22 shows the list of pedo-geomorphological categories on the BUNASOL map and the corresponding soils aptitudes. In each spatial unit represented on the map in Fig. 12.2, the type of pedo-geomorphologic zone which it belongs to is indicated in the upper part of each square. Hence, referring to Table 12.22, we know the degree with which each type of vegetation or cultivation is suitable to each spatial unit.

Table 12.22 Legend of agricultural aptitude of Loulouka watershed (source: BUNASOL)

Assessment w.r.t. indicator i ₁₁₁ is qualitative (ordinal). It consists, through field observations, of comparing actual cultivations (farmlands) made by the farmer on each spatial unit to those recommended by the agricultural aptitude map provided by BUNASOL. Such comparison results in a degree of appropriateness of the response of each SU w.r.t. the type of cultivation. This response is assessed on a three level scale of appropriateness (A, MA, NA) as announced above. Note that the pedo-geomorphologic map and Table 12.22 actually distinguish five levels of agricultural aptitude. We have merged “moderate aptitude”, “weak aptitude” and “moderate to weak aptitude” classes in Table 12.22 in one class that we label “moderate or weak aptitude”.

The rule used for assessing SU’s w.r.t indicator i ₁₁₁ reads as follows:

• Category A is for spatial units that belong to the “good aptitude” or “moderate or weak aptitude” class for a cultivation and on which this cultivation is actually present on more than $\frac{3} {4}$ of the surface of the spatial unit.

• A SU is assigned to category MA if less than $\frac{3} {4}$ and more than half of the spatial unit is occupied by a cultivation for which the SU has “good aptitude” or “moderate or weak aptitude” according with the BUNASOL map.

• A SU receives the NA mark in the following cases:

  • if the spatial unit belongs to the “good aptitude” or “moderate or weak aptitude” class for a cultivation practice and that less than half of the spatial unit is occupied by such a cultivation,

  • in case of agricultural unfitness of the spatial unit according with the BUNASOL map.

Table 12.23 summarizes the rule designed for assessing indicator i ₁₁₁.

Table 12.23 Evaluation rule for indicator i ₁₁₁

1.1.2 Adequate Application of Water and Soils Conservation Techniques (WSC) (Cr ₁₂)

Some WSC techniques (bund⁹—in French, diguette, stony cordon—in French, cordon pierreux) are used to brake the water run-off which carries away land, thus leading to erosion. Other WSC techniques (micro water harvesting or zaï,¹⁰ mulching¹¹—in French “paillage”) favor the fertilization of the soil on which a vegetation could then grow and ensure a form of protection against wind erosion and erosion due to water run-off. Mulching is no longer applied due to the scarcity of crop waste. Indeed, crop waste is used as animals fodder and also as energy source because of heating wood shortage. Hence, criterion Cr ₁₂ can be evaluated using the following three indicators:

• i ₁₂₁ : presence of bund,

• i ₁₂₂ : presence of stony cordon,

• i ₁₂₃ : presence of zaï.

The evaluation of spatial units w.r.t. each of these three indicators is qualitative (ordinal):

• For indicator i ₁₂₁ (resp. i ₁₂₂), the evaluation results from observing the presence or absence of bunds (resp. stony cordons) and, in case of their presence, from evaluating their state and, also, the distance between two bunds (resp. stony cordons) according to whether they lie on a slope, a plateau or a shoal. For these indicators, the assessment rules are the following. In case there are no bunds (resp. stony cordons) on a given SU, this is assigned to level NA on the scale of indicator i ₁₂₁ (resp. i ₁₂₂ ). In case bunds (resp. stony cordons) are present in a SU, we value as level A the situation in which bunds (resp. stony cordons) are in good state and they are regularly spaced; if one of these conditions is not fulfilled, the value assigned to the SU is MA.

• Indicator i ₁₂₃ is related to the presence of zaï in SU’s subject to zipella.¹² A SU subject to zipella receives the following marks on the scale of indicator i ₁₂₃:

  • mark A if zaï is present on a surface S that represents at least $\frac{3} {4}$ of the zipella surface;

  • mark MA when this surface S lies between $\frac{1} {2}$ and $\frac{3} {4}$ of the zipella surface;

  • mark NA if S is smaller than $\frac{1} {2}$ of the zipella surface.

  If there is no zipella on the SU, only indicators i ₁₂₁ and i ₁₂₂ are assessed for that SU.

These three indicators are aggregated using expert rules, yielding an assessment of each SU on criterion Cr ₁₂. The rules used are described in Table 12.24. The upper part of the table is used for SU’s in which no zipella shows up: only indicators i ₁₂₁ and i ₁₂₂ intervene (Aggregation 1). In case of presence of zipella in a SU, then the result of Aggregation 1 is aggregated with indicator i ₁₂₃ yielding the rule called “Aggregation 2”.

Table 12.24 Expert rules used for aggregating indicators in the case of criterion Cr ₁₂

1.1.3 Adequate Application of Soils Preparation Techniques (SP) (Cr ₁₃)

Ploughing, by loosening the soil, favors erosion. Hence it is important to plough perpendicularly to the direction of water flow. The fallow practice allows the soil to rest. This favors vegetation growing in the long-term, which helps reducing erosion. Criterion Cr ₁₃ has two indicators:

• i ₁₃₁ : fallow practice,

• i ₁₃₂ : ploughing technique.

The evaluations of SU’s w.r.t. indicators i ₁₃₁ and i ₁₃₂ are qualitative. For i ₁₃₁ (resp. i ₁₃₂), the evaluation results from the observation of the fallow practice (resp. the ploughing technique) used on each SU. The appropriateness of the response of each SU w.r.t. fallow practice and ploughing technique is assessed on the three levels scale (A, MA, NA) of each indicator. The expert rule used for aggregating indicators i ₁₃₁ and i ₁₃₂ into criterion Cr ₁₃ is displayed in Table 12.25.

Table 12.25 Expert rules used for aggregating indicators in the case of criterion Cr ₁₃

1.1.4 Soil Compaction Limitation (Cr ₁₄)

Compact nude soil forbids water infiltration. This induces run-off which takes away all solid particles from the ground surface. On the other hand, the soil overstamping by animals, either in habitual penning places (penning for the night), or due to seasonal move to summer pastures, or in watering places, has an impact on soil compaction. Cr ₁₄ has two indicators:

• i ₁₄₁ : presence of nude soil,

• i ₁₄₂ : animals overstamping.

The evaluation of SU’s w.r.t. i ₁₄₁ and i ₁₄₂ was performed as follows:

• For i ₁₄₁, mark A is assigned if less than a quarter of the SU is nude; mark MA is for SU’s in which more than a quarter and less than half the surface is nude, NA is given when more than half the spatial unit is nude;

• i ₁₄₂ is a binary indicator noting the presence (mark A) or absence (mark NA) of animals overstamping on the SU.

The results of theses evaluations obtained by means of indicators i ₁₄₁ and i ₁₄₂ were aggregated following three classes (A, MA, NA) of appropriateness response of the spatial unit w.r.t. the soil compaction limitation (see Table 12.26).

Table 12.26 Expert rules used for aggregating indicators in the case of criterion Cr ₁₄

1.2 Criteria and Evaluation Indicators of BIO Principle (P ₂)

Four criteria were identified to assess the limitation of biodiversity loss.

1.2.1 Limitation of Cultivated Surfaces Extension (Cr ₂₅)

Increasing the cultivated surface reduces biodiversity. In particular, it eliminates plants which are essential to soil reconstitution. Cultivated surface extension is most of the time made by using inapt soils left fallow. Historical data related to the extension of cultivated surface is not available. Therefore, we have opted for the quantification of cultivated marginal land. Criterion Cr ₂₅ has only one indicator:

• i ₂₅₁ : cultivation on marginal lands.

The evaluation w.r.t. i ₂₅₁ is qualitative. It rests on the observation of the percentage of marginal lands cultivation in each SU. Mark A is attributed if there is no marginal land cultivation, MA results when cultivation on marginal land remains below one third of the SU surface, while NA is attributed if at least one third of the SU surface consists of cultivated marginal lands.

1.2.2 Adequate Pesticide Usage (Cr ₂₆)

Each type of crop is associated a particular type of pesticide which protects it by taking on its enemies (vegetal or animal). Using pesticides that attack other organisms is not adequate since, sooner or later, it will have a negative influence on some plants which are sensitive to these pesticides. Moreover, the right matching between the type of pesticide and the type of cultivation is not sufficient. Also the quantity of pesticide used and the moment at which it is spilled may be more or less appropriate.

Remark 12.1

Pesticide usage is not easily observable unless the team in charge of the survey remains on the field during all the cultivation period. Obtaining precise and reliable information on pesticide usage from the population is not easy either. In our case we have not been able to gather the required information so that this criterion could not be taken into account.

Nonetheless this criterion Cr ₂₆ was analyzed; it has two indicators:

• i ₂₆₁ : appropriate matching pesticide-cultivation,

• i ₂₆₂ : frequency of pesticide usage.

These indicators are assessed on a qualitative scale. For indicator i ₂₆₁, mark A is attributed in the case where no pesticide is used, MA is for the case where the pesticide appropriate for the type of cultivation is used, while NA results if the pesticide used is not adequate to the type of cultivation. For indicator i ₂₆₂, mark A is assigned whenever no pesticide is used; the assessment of a SU is MA if the pesticide adequate for the crop is used once. For more than one use of pesticide per crop or in case of inadequate pesticide usage, we attribute mark NA. The expert rule designed for aggregating these two indicators is displayed in Table 12.27, using three classes (A, MA, NA) of appropriateness of response of SU’s w.r.t. pesticide usage.

Table 12.27 Expert rules used for aggregating indicators in the case of criterion Cr ₂₆

1.2.3 Preservation of Ecosystem Integrity (Cr ₂₇)

The presence of trees and forests has a positive impact on the preservation of the biodiversity. Criterion Cr ₂₇ has four indicators:

• i ₂₇₁ : presence of sacred grove (or copse),

• i ₂₇₂ : reforestation zone,

• i ₂₇₃ : protected forest,

• i ₂₇₄ : presence of trees stump.

For assessing a SU w.r.t each of the indicators i ₂₇₁, i ₂₇₂ and i ₂₇₃, we observe the surface percentage of the SU occupied either by a sacred grove, a reforestation zone, or a protected forest. The rule is the same for all three indicators. We attribute mark A (resp. MA, NA) if at least $\frac{2} {3}$ (resp. between $\frac{1} {3}$ and $\frac{2} {3}$, at least $\frac{2} {3}$) of the SU surface is occupied by a sacred grove, a reforestation zone, or a protected forest. The expert rule used for aggregating the indicators i ₂₇₁, i ₂₇₂ and i ₂₇₃ in order to assess the degree of appropriateness of the response of each SU w.r.t. the preservation of ecosystem integrity is given in Table 12.28.

Table 12.28 Expert rules used for aggregating indicators in the case of criterion Cr ₂₇

Remark 12.2

Indicator i ₂₇₄ could not be taken into account in the assessment of SU’s w.r.t. criterion Cr ₂₇. Assessing a SU w.r.t. this indicator, requires information about plants regeneration capacity as well as about the cutting technique applied to the plants. Some cutting techniques favor regeneration while others do not. In our case, these data on the vegetation of the area were not available so that we could not assess SU’s w.r.t. indicator i ₂₇₄.

1.2.4 Bush Fire Limitation (Cr ₂₈)

Bush fire is often practiced by farmers to clear land before cultivation by removing vegetation, or by animals farmers to eliminate straw and favor vegetation growth in view of feeding the animals in the beginning of the dry season. Bush fires have undesirable consequences such as the destruction of the vegetation, of animals, particularly the micro-fauna, and animals habitat. Bush fires also lead to soil erosion by the loss of vegetal coverage. Criterion Cr ₂₈ has only one indicator:

• i ₂₈₁ : presence of bush fire.

Indicator i ₂₈₁ has only two modalities; it encodes the absence of bush fire (mark A) or, on the contrary, the practice of bush fire (NA) on each SU.

1.3 Criteria and Evaluation Indicators of FER Principle (P ₃)

Criteria Cr ₃₉ and Cr ₄₀ represent the relevant aspects of the preservation of soil fertility:

1.3.1 Adequate Application of Cultivation Techniques (Cr ₃₉)

Crop rotation consists in varying the cultivations in a given field, alternating those which impoverish the soil and those which enrich it in some nutritive elements (for example, nitrogen). Crop rotation practice allows the soil to reconstitute. Likewise, choosing an adequate association of cultivations on the same SU (leguminous plants, bean, groundnut or others, on the one hand, and gramineae, maize, millet, sorghum or others, on the other hand) have a beneficial effect on the preservation of nutritive elements in the soil (nitrogen, organic matter, etc.). Two indicators account for criterion Cr ₃₉:

• i ₃₉₁ : practice of crop rotation,

• i ₃₉₂ : practice of crop association.

Indicators i ₃₉₁ and i ₃₉₂ are qualitative and binary. Regarding i ₃₉₁, mark A is attributed when crop rotation is applied, mark NA, otherwise. Regarding indicator i ₃₉₂, mark A is attributed when crop association is practised, mark NA, otherwise. Table 12.29 shows the aggregation rule used for assessing SU’s on criterion Cr ₃₉ by aggregating indicators i ₃₉₁ and i ₃₉₂. Criterion Cr ₃₉ is evaluated on a three classes (A, MA, NA) scale assessing the appropriateness of the response of the spatial unit w.r.t. the application of cultivation techniques.

Table 12.29 Expert rules used for aggregating indicators in the case of criterion Cr ₃₉

1.3.2 Adequate Practice of Soil Fertilization (Cr ₃₁₀)

The chemical fertilizer must be used in such a way to allow its absorption by the cultivated plants and to attenuate the risk of discharge in the environment. Organic manuring is recommended in association with NPK fertilizer [nitrogen (N), phosphorus (P), potassium (K)]. Two indicators are associated with criterion Cr ₃₁₀:

• i ₃₁₀₁ : presence of manure,

• i ₃₁₀₂ : use of chemical fertilizer.

Indicator i ₃₁₀₁ is assessed on a three degrees qualitative scale. The part of each SU surface enriched with organic manuring was observed. Mark A (resp. MA) is assigned if at least three quarters (resp. between half and three quarters) of the SU is enriched with manure; mark NA results otherwise. Regarding indicator i ₃₁₀₂, we have observed the quantity of chemical fertilizer used and compared it with the recommended norms. It appears that the norms for usage of chemical fertilizers are not respected in the region. Consequently, the evaluation w.r.t. i ₃₁₀₂ was brought back to a binary scale. Mark A is attributed if chemical fertilizer is used, mark NA results otherwise. The evaluation results w.r.t. indicators i ₃₁₀₁ and i ₃₁₀₂ are aggregated using a three classes (A, MA, NA) scale assessing the appropriateness of the response of each SU w.r.t. the practice of soil fertilization. The aggregation rule is displayed in Table 12.30.

Table 12.30 Expert rules used for aggregating indicators in the case of criterion Cr ₃₁₀

1.4 Criteria and Indicators of PRO Principle (P ₄)

The agricultural productivity is measured by the quantity of cultivation produced per hectare during a given period. Two criteria were identified to evaluate the presence of a good potential of agricultural productivity.

1.4.1 Technical Training of Farmers (Cr ₄₁₁)

It is necessary to train the farmers to good cultivation techniques. A single indicator is used to evaluate a SU w.r.t. criterion Cr ₄₁₁:

• i ₄₁₁₁ : presence of trained farmers.

Indicator i ₄₁₁₁ is binary. The evaluation of a SU results from enquiry about the proportion of farmers having received a training in cultivation techniques and in water and soils conservation techniques. Mark A is assigned if at least half the farmers on the spatial unit were trained, otherwise mark NA is attributed. The choice of an evaluation of binary type is justified by the “spread effect”: farmers generally imitate their neighbors’ behavior if their results are better than theirs.

1.4.2 Improvement of the Agricultural Production (Cr ₄₁₂)

If the farmer has at disposal sufficient financial means in order to buy products or rent equipment (manure, fertilizer, ameliorated seeds, tractor and plough rental, etc.) for exploiting his field, he can improve his production; his family can live on the harvest in a satisfactory way; a part of this harvest can be sold and the profit used to satisfy other needs. In this way, the farmer can dispense from practising extensive agriculture, which would push him to exploit marginal or lying fallow soil. On the other hand, if the farmer has at disposal other sources of financial income (market-gardening, gold washing, cattle sale, etc.), the need to practise extensive agriculture will disappear, helping to install an intensive sustainable agriculture which requires purchasing seeds and fertilizer. To evaluate a SU w.r.t. criterion Cr ₄₁₂, we use two indicators:

• i ₄₁₂₁ : excess production,

• i ₄₁₂₂ : practice of an activity constituting a source of income.

Indicators i ₄₁₂₁ and i ₄₁₂₂ are assessed on the three levels (A, MA, NA) scale. Regarding i ₄₁₂₁ (resp. i ₄₁₂₂), mark A is attributed if at least three quarters of the farmers in the SU have an excess production (resp. another profitable activity); mark MA is assigned when this fraction lies between half and three quarters of the farmers and mark NA results otherwise. The evaluation results w.r.t. indicators i ₄₁₂₁ and i ₄₁₂₂ were aggregated on a three classes (A, MA, NA) scale assessing the appropriateness of the response of the SU w.r.t. the improvement of the production. The aggregation rule used is displayed in Table 12.31.

Table 12.31 Expert rules used for aggregating indicators in the case of criterion Cr ₄₁₂

1.5 Criteria and Indicators of EN Principle (P ₅)

If the elementary needs of populations are satisfied, they will not need to overexploit and degrade their land. We have identified three criteria allowing to evaluate the intensity of satisfaction of the population’s elementary needs. In contrast with those described above, these criteria do not vary from a SU to another. They can be assessed at the level of the whole region, namely the Loulouka watershed. These criteria would be needed for comparing two different regions. Since we are concerned with a single region, we do not use and assess them. Nevertheless, for the sake of completeness, we give below a short description of the three criteria accounting for the satisfaction of population’s elementary needs and we try to propose some indicators that help to assess them.

1.5.1 Respect of Socio-Cultural Spaces (Cr ₅₁₃)

It is important to take into account populations’ cultural practice. Let us mention for example the preservation of sacred groves (or copses) which represent particular spaces for the farmer in our study area. Actions to undertake for fighting against land degradation need to preserve these cultural spaces. More generally, recommended actions should not go against populations’ cultural values, otherwise they will hardly be sustainable.

1.5.2 Improvement of Education Level (Cr ₅₁₄)

Education helps populations to become more aware of the problem of land degradation. It makes them able to understand what is at stake and helps to convince them to act according with the principles of degradation limitation. Moreover, education allows the populations to diversify their professional abilities. In this way, agricultural activities will cease to be the only professional perspective, and this will help to avoid over-exploiting the land. Two indicators account for criterion Cr ₅₁₄:

• i ₅₁₄₁ : percentage of children in full-time education,

• i ₅₁₄₂ : presence of school infrastructures

1.5.3 Improvement of Socio-Economic Conditions (Cr ₅₁₅)

The farmers population need to attain food self-sufficiency in order to be able to draw financial income from their work. They also need to feel materially able to start actions or projects which will go along the lines of improvement of their social condition (e.g. by selling excess food). Two indicators can be proposed to account for criterion Cr ₅₁₅:

• i ₅₁₅₁ : living conditions,

• i ₅₁₅₂ : populations’ health level.

Editors’ Comments on “Assessing the Response to Land Degradation Risk in the Loulouka Basin”

Metchebon, Pirlot, Yonkeu and Some present in their study case a fully worked out example of what they call a decisional map—the association of MCDA methods with a GIS—in order to structure and assess land degradation risks in Africa. This chapter presents thus a thorough discussion on how to, both, formally model and visually illustrate a decision problem in order to achieve an effective decision aiding for all stakeholders in a territorial management problem.

Main aspects relevant for the purpose of this handbook are, on the one hand, the detailed analysis of the modelling process of the decision alternatives, the evaluation criteria and the performance evaluations (see Sect. 2.2). On the other hand, the decision problem statement illustrated here consists in sorting into a predefined set of ordered categories following an Electre Tri approach (see Chap. 4, Sect. 4.2.3). As such, this study relates especially to the chapter by Mercat-Rommens, Chakhar, Chojnacki and Mousseau (Chap. 13) presenting a similar approach for building and evaluating a geographical nuclear risk map, and to the chapter by Lué and Colorni (Chap. 14).

Without a public mandate, this assessment study of the land degradation observed in the Loulouka catchment basin (Center North of Burkina Faso in Western Africa) does not directly imply any institutional and/or public decision makers. Instead, one of the authors of the study is a recognized environmentalist expert, who actively contributed to previous land degradation studies concerning the region under review (Yacouba et al., 2002; Yonkeu and Kiniffo, 2004). Here he acts as factual decision maker for providing all required preferential information such as the assessment criteria weights for instance. The target group for the decision recommendations are, in this case, local farmers and public authorities who will have to plan and undertake adequate land conservation actions.

The claimed specific objective of the case study, by the way appearing genuinely in many spatial decision problems (see Mercat-Rommens et al., Chap. 13, and Luè and Colorni, Chap. 14), is essentially to elaborate a structured and participative assessment of the land degradation state via a hierarchy of evaluation criteria based on relevant environmental indicators and taking into account all stakeholders—local population, authorities and experts—points of view.

The natural decision problem statement appearing in this kind of geographical decision aid problems consists, in a first problem structuring step, in describing geographical spatial units with respect to multiple evaluation criteria, and, in a second aggregation step, in sorting these spatial units, once assessed on all the relevant criteria, into four ordered categories assessing the degree of ability of each unit to respond to the risk of land degradation: inadequate, weakly adequate, moderately adequate, adequate.

As so often in a GIS integrated multiple criteria decision analysis, decision alternatives correspond to 229 identified contiguous spatial units—25 ha squares—that cover the region under review. These spatial units are, for the MCDA purpose, evaluated on a complex set of criteria in order to judge their response level to the land degradation risk they present. This part certainly represents the most specific and interesting aspect of this case study.

Five fundamental objectives for limiting land degradation within a framework of sustainable development and management of the region are here considered: limit soils erosion, preserve biodiversity, preserve soils fertility, favour a good level of agricultural productivity, satisfy elementary needs for social welfare. From climatic, as well as anthropogenic factors, undermining theses objectives, are derived 12 criteria like limited extension of cultivated soils, bush fire limitation, and technical training of farmers for instance. The response level to land degradation risk in each spatial unit is eventually assessed via the observation of 23 performance indicators like presence of nude soil, practice of crop rotation and appropriate matching of pesticide and type of cultivation for instance. Considering the large part of imprecision of these indicators, only three ordinal response levels:—not adequate, moderately adequate, adequate—are used as measurement scale on all the criteria.

The usage of the Electre Tri method for multiple criteria based sorting of the spatial units is a rather natural application in a GIS application. One of the most critical aspect in using this method concerns setting adequate criteria weights. The authors have, similar to the Mercat-Rommens et al. approach (see Chap. 13), used Simos’ method (Simos, 1997), not without practical difficulties. Why not simply start by default with considering the decision objectives as more or less equally important? This way, each criterion affecting a specific objective may be again considered as equi-significant contribution within the importance of the relevant objective. If the corresponding sorting result is not convincing the decision maker, differentiating the importance of the objectives, or the significance of some of the criteria may become useful. A similar comment may be addressed to the criteria weights used in the chapter on Choosing a cooling system for a nuclear power plant in Belgium (Chap. 8).

One may question at this point that, a sorting problem into k = 4 categories, when only d = 3 performance levels are discriminated on each criteria, is indeed well conditioned. Normally the marginal performance discrimination should allow to clearly map the global ordered outcome categories on each criterion’s measurement scale. Indeed, how will it be possible to construct consistent majority sorting situations—the ground statements of the Electre Tri method—if not all the outcome categories may be actually discriminated? With $k = 4 > d = 3$, this study shows here a rather special application context of the Electre Tri method; an application needing both the category limiting profiles as well as the majority cut level, to be handled in a non standard way. It appears by the way that the final discussion of the results actually concentrates more or less on solely two global sorting categories: not adequate or adequate, regarding the response to the risk of degradation, such that we anyhow come back to a sounder situation where $k = 2 < d = 3$.

Due to the unusual mapping of three marginal ordered categories onto four global categories, the validation process of the Electre Tri exhibits specific practical difficulties. This fact adds on to the impression that in this case study, the Electre Tri method is somehow used outside of its genuine usage. In the absence of an explicit decision, as is in this application the case, a multiple criteria descriptive decision aid approach, instead of the classic prescriptive decision aid approach, would perhaps provide even more convincing and tangible results.

Finally, this application illustrates again the great relevance and usefulness of integrating MCDA approaches into a GIS systems. Especially, in the context of a territorial management problem, such systems may well provide a very useful and effective decision aid for all private and/or public stakeholders.