- freely available
Sustainability 2014, 6(2), 935-951; doi:10.3390/su6020935
Abstract: A pressing question that arises during the planning of an ecological restoration process is: where to restore first? Answering this question is a complex task; it requires a multidimensional approach to consider economic constrains and the preferences of stakeholders. Being the problem of spatial nature, it may be explored effectively through Multicriteria Decision Analysis (MCDA) performed in a Geographical Information System (GIS) environment. The proposed approach is based on the definition and weighting of multiple criteria for evaluating land suitability. An MCDA-based methodology was used to identify priority areas for Forest Landscape Restoration in the Upper Mixtec region, Oaxaca (Mexico), one of the most degraded areas of Latin America. Socioeconomic and environmental criteria were selected and evaluated. The opinions of four different stakeholder groups were considered: general public, academic, Non-governmental organizations (NGOs) and governmental officers. The preferences of these groups were spatially modeled to identify their priorities. The final result was a map that identifies the most preferable sites for restoration, where resources and efforts should be concentrated. MCDA proved to be a very useful tool in collective planning, when alternative sites have to be identified and prioritized to guide the restoration work.
Deforestation is one of the most serious current environmental problems. During the period of 2000–2010, approximately 130,000 km2 of forest have been lost around the world each year . In response, the efforts to support and promote the recovery of forest ecosystems have increased . One of the most comprehensive approaches to forest restoration is the Forest Landscape Restoration (FLR) . FLR is defined as a process of restoration of goods, services and ecological processes that forests can provide at a broader landscape level. It fosters restoration, based on the better assets for both people and the environment, focusing on regaining the ecological integrity and enhancing human well-being .
FLR provides a framework to implement large-scale and long-term restoration attempts with the incorporation of diverse dimensions within them. It is a collaborative process that involves a wide range of actors since the planning stage . Important issues that arise in restoration planning are where to restore first and where we should focus the available resources and efforts. The decision should come from consensus among stakeholders to identify priorities and the most preferable sites for restoration. This is clearly related to concerns of an economic kind and the awareness that the available resources should be directed to areas and actions that are likely to provide the greatest benefits to both nature and people. However, actors’ opinions are often in conflict with each other, and economic and technical restrictions limit the land available for restoration planning . Hence, the problem is complex, involving different components (e.g., environmental, socioeconomic) and views (e.g., local people, experts), and has a clear spatial dimension. Dealing with similar problems requires the application of proper tools to provide robust and informed decisions.
Multicriteria Decision Analysis (MCDA) encompasses a set of techniques to support decision-making processes . It provides a framework to integrate multiple opinions and evaluation criteria, to weight them according to their importance, and select the most suitable courses of action. Malczewski  documented the increase in use of MCDA in synergy with the capabilities of the geographical information systems (GIS) in the last two decades. This occurred in a broad range of applications, such as water management [7,8], land use planning , transportation infrastructure [10,11], waste disposal [12,13,14], urban planning , environmental planning [16,17], and forest restoration [18,19,20,21].
This paper presents a spatial MCDA-based method for identifying priority areas for forest landscape restoration that considers environmental and socioeconomic criteria and evaluates them according to the opinion of different participant groups, as obtained through a survey.
The framework methodology was designed in the Upper Mixtec region, in Southern Mexico, one of the most degraded areas of Latin America, where the lack of environmental data and resources go together with the urgency of implementing a regional plan for restoration.
The Upper Mixtec region lies on the northwest side of Oaxaca State, in South-eastern Mexico. It covers 8,100 km², corresponding to 13% of the total area of the State. Most of the information is available at the municipal level. However, some municipalities belong to more than one region. For this reason, the boundary of the study area was arbitrarily set to include all the municipalities whose area belongs for at least 50% to the region, according to the municipal and regional division available for the State of Oaxaca . As a result, 124 municipalities are incorporated in the study area for a total of 11,631 km² (Figure 1).
The region is characterized by high topographic diversity with canyons, hills, intermountain plateaus, and valleys. The elevation ranges from 550 to 3,300 m. Native vegetation cover is composed of tropical dry forest, pine-oak forest, and xerophytes scrubs. However, crops and eroded areas are common throughout the entire area. Mean annual precipitation is 692 mm with rainfall being mostly concentrated in July and August with intense showers [23,24].
The Upper Mixtec is a region with severe socio-economic and environmental problems derived, in part, from wrong strategies of natural resources management. There is a critical problem of water supply, degraded areas, and bare soils (Figure 2). Population is highly impoverished and dispersed . Approximately 50% of the original forest cover was lost with a deforestation rate of 52.7 km2 per year, the current cover of forest falls around 6,592.47 km2 (56.68% of study area). Although some governmental and non-governmental efforts against deforestation have been successful, resources available to combat deforestation cannot include all areas requiring restoration [26,27]. In 2005, only 26 km2 were reforested in the whole Oaxaca State . Clearly, there is an urgent need to optimize the resources available for restoration. The main restriction to implement successful restoration projects in the Upper Mixtec region has been the lack of a good strategy of distribution for available resources, economic and technical, within a great area with high socio-ecological degradation . Selecting the best areas for restoration is a difficult task due to the many factors that should be taken into account and the particular points of view of different stakeholders that participate in the process. The use of GIS-based decision-making tools to generate a map of priority areas has not been tried in this region so far.
In the intelligence stage, we sought the opinion of stakeholders to identify and evaluate socioeconomic and environmental criteria . Following Eastman et al. , a thematic raster layer was developed for each of the criteria based on the available data for the study region. A constraint layer was also generated that represents the areas where restoration activities are not relevant: urban, existing forest and water zones. Finally, maps of land suitability were generated , reflecting the preference of the different stakeholders. In the design stage, we established a threshold based on the suitability levels to extract the most preferable sites (i.e., contiguous pixels with high suitability levels), which can constitute potential restoration sites, as shown by Orsi and Geneletti .
In the choice stage, we extracted the sites with highest suitability scores by each participant. Next, they were integrated into groups of stakeholders (academic, governmental, Non-governmental organization, and public) to create different maps of priority areas. Finally, an overall map of forest landscape restoration priorities was developed by combining the maps of the four groups.
|Stage||Steps and Activities||Tool|
|Intelligence||Identify and evaluate relevant criteria (Section 2.1)||Interview|
|Defining relevant criteria.|
|Assign interval values of desirability.|
|Assign values of importance|
|Create criteria layers set: factors and constraint (Section 2.2)|
|Modeling preferences (Section 2.3)|
|Deriving commensurate criteria layers.||GIS|
|Design||Design restoration options (Section 2.4)||MCE-GIS|
|Define threshold based on forest loss rate.|
|Identify and eliminate unsuitability area.|
|Create restoration options maps for each stakeholder.|
|Choice||Identification of priority areas (Section 2.5)||GIS|
|Priority areas by population sectors.|
|Map of priority areas for forest landscape restoration.|
2.1. Identify and Evaluate Relevant Criteria
Four groups of stakeholders were identified: general public (P), academic institutions (A), non-governmental organizations (NGO) and governmental officers (G). The general public included participants with knowledge about governmental programs for restoration and inhabitants from the region (including farmers, students, craftsmen). The selection of stakeholders was made following the method described by Geneletti . We elaborated a preliminary list of people with widely known work on environmental issues or professional experience in the study region, and sought advice from them to include additional suitable people. The number of final participants was determined according to Landeta , who pointed a minimal of seven participants per group and thirty as the maximum number, regardless possible desertions of participants or incomplete interviews.
A total of 54 interviews were conducted with stakeholders belonging to the academic (9), governmental (10), NGO (9), and general public (26) groups. The interviews were conducted individually to avoid bias in judgment influenced by dissimilar opinions . The process was divided in two stages. In the first stage, we conducted two-round interviews to stakeholders face-to-face or via internet or telephone to generate a set of environmental and socio-economic criteria. In the second stage, the interviews aimed at evaluating the ten most relevant criteria.
Following the method applied by Orsi et al. , stakeholders were asked the question: Which criteria should be considered to select a site for forest restoration? Answers from first round were clustered into individual criteria based on semantic similarity. A preliminary list of criteria with all stakeholders’ opinions was created and returned back to each participant. In the second round, we asked to reconsider preview opinion taking into account the whole list of criteria. The criteria were ordered by the number of citations. Finally, the ten most cited criteria that fulfilled the following requirements were selected: spatial character (i.e., suitability to be displayed on a map); availability of the data; and significant variability of this character over the study area.
Taking a decision based on one or more criteria implies that decision-makers can assign different degrees of desirability to the scores of the adopted criteria . For example, considering the criterion “distance to forest”, it could be decided that areas in close proximity to existing forest should be accorded priority or rather that areas far from existing forest should be given such priority. In this stage, stakeholders were asked: what levels of a criterion reflect a higher priority for restoration? Three possible answers were possible: (1) lower, (2) intermediate, and (3) higher ranges. The answers allowed us to classify each criterion under one of the following types: 1, cost (the lower the values of the criterion the higher the desirability); 2, intermediate (the more intermediate the values the higher the desirability); and 3 benefit (the higher the values the higher the desirability).
In order to account for the fact that criteria may have different relative importance in a decision process [36,37,38], stakeholders were asked the following question: How important is the criterion in the selection of a restoration site? The answers were ranked from low to high level of importance using a five-point scale from not relevant (key 1), to very important (key 5).
2.2. Criteria Layers: Factors and Constraints
A raster layer was generated for each one of the criteria selected. Two types of layer were constructed: factors and constraint. The factor layers represented each one of the ten criteria selected in the consultation stage, while the constraint layer represented restrictions to be considered during site selection. For this purpose, we acquired, compiled and processed a set of available spatial thematic data from the study area. Tabular, vector, and raster data were acquired from governmental agencies. In addition, we acquired a SPOT’05 satellite image (30 m resolution) that was analyzed through supervised classification by personnel of the GIS Laboratory of CIIDIR Oaxaca, National Polytechnic Institute. We generated raster-base layers using ArcGis 9.2 software . A resample of 30 m pixel resolution and WGS84-UTM-14N system re-projection were applied for each layer output. The minimum spatial unit of study, i, was defined as an area of 900 m2 (i.e., 30m x 30m), represented by a pixel and named “alternative”, which was assessed by criteria in terms of suitability.
2.3. Modeling Preferences
In this stage we integrated preferences and criteria layers and generated individual suitability maps, one for each stakeholder by following the basic steps of MCDA: (1) transformation of the original attribute values of criteria layers into comparable units (standardization) (Section 3.3.1), (2) prioritization of the criterion based on its relative importance (weight assignment), and (3) computation and assignment of suitability scores to each pixel (evaluation alternative) by using a decision rule (aggregation). We used the Spatial Multicriteria Evaluation Module (SMCE) of ILWIS 3.6® to build the criteria tree and to carry out multicriteria analysis .
For a multicriteria decision analysis process, in which we combine criteria and preferences, it is essential to combine criteria in the same scale of values. Therefore, original values of criteria must be transformed into comparable units . To this end, a standardization process was performed by considering value functions that convert the original criterion scores into degrees of desirability, according to the type of criterion suggested by participant (i.e., cost, intermediate, benefit) . The standardization process was conducted using the SMCE module and the interval method of linear scale standardization following formulas 1 to 3. The standardized values of each criterion ranged between 0 (least desirable condition) and 1 (most desirable condition) .
2.3.2. Weight Assignment
The relative importance value assigned to each criterion, known as weight, was used for criteria prioritization [36,38]. We applied the rank order method for translating the participants’ preferences expressed during the survey into quantitative values of importance that they were normalized to sum 1 [37,41].
The procedure was set up in the SMCE module using the information of the interviews. At first, the criteria were put in order of importance, from the most to the least important criterion . Then, the expected value algorithm (Formula 4) was applied to calculate the weight (Wk) for criterion (k), where n is the number of criteria and i is the position of the criterion within the ranking.
The aggregation brings together the results of previous steps, combining normalized criteria values and weights for each alternative . The aggregation was performed using the weighted linear summation method, which is a linear combination of standardized values by means of weights (formula 5).
2.4. Design of Restoration Options
The design of suitable sites (restoration options) was conducted by selecting the best pixels from the suitability maps generated in the previous stage of aggregation. As suggested by Orsi and Geneletti , this process was driven by the identification of total restoration target and the minimal area of the restored patches, and involved the following steps. (1) Identify the areas where there are no constraints; (2) Within these areas, select the pixels with the highest suitability values until reaching an area of 152,000 ha, which represents the target area for restoration interventions; (3) Eliminate all patches smaller than 0.54 ha (i.e., 6 contiguous pixels), in order to avoid selecting areas too small to make restoration feasible. These rules were agreed on the basis of the rules to support and fund restoration projects by the National Forestry Commission (quoted in Céspedes-Flores and Moreno-Sánchez ). One restoration map was generated for each stakeholder.
2.5. Identification of Priority Areas
Five maps of forest landscape restoration priorities were generated, one for each stakeholder group and one combining all the groups. Individual stakeholders’ maps of suitable areas for restoration were summed up within each participant groups. This was performed by using a “sum” operation in a GIS, which generates a map where each pixel contains a value tantamount to the sum of the values of the corresponding pixels in all the input maps. Subsequently, we extracted the most suitable pixels until the fulfillment of the required restoration area (152,000 ha), as described in the previous sub-section. Pixel values in the resulting maps are classified in a three-level scale: 1 (if all four stakeholder groups agree on the suitability of the site), 2 (only three stakeholder groups agree), and 3 (only two stakeholder groups agree).
3. Results and Discussion
34 different criteria were identified: 21 environmental criteria and 13 socioeconomic. Table 2 shows the most frequently cited criteria. The highest number of citations was given to the criterion “distance from forest”, which was suggested by 18 of the interviewees. 41% of the criteria listed were mentioned only once and accounted for 3.6% of total citations. The ten most frequently cited criteria that were considered for further analysis are shown in Table 3, which does not include criteria for which data were not available (i.e., rate of land use change, soil depth, land tenure, forest fragmentation).
Stakeholders’ answers on the relative importance of criteria (weight) for each group or sector are shown in Figure 3. Out of all criteria, the “marginalization index” (an index based on the availability of public services and infrastructures) was almost unanimously (i.e., 90%) considered as a benefit factor, while “distance from forest” received the second highest consensus with 82.5% of participants considering it as a cost factor. On the other extreme, the interpretation of criteria like “distance from rivers” and “risk of erosion” were poorly agreed upon by participants. Regarding weights (Figure 3), 66% of participants included environmental criteria in categories of “important” and “very important”, while only 46% of them included socioeconomic criteria in the same categories. The criterion “distance from forest” was assumed to be very important by 31 out of 54 participants. Within the public sector, the most important criterion was “index of marginalization” with 18 participants out of 26 rating it as very important.
|1||Distance from forest||64.3|
|2||Rate of land use change *||57.1|
|3||Risk of Erosion||53.6|
|4||Biodiversity index *, natural corridors *||46.4|
|5||Slope of terrain (%), rate of deforestation *.||42.9|
|6||Distance from crops||39.3|
|7||Human density, Insolation, distance from roads, distance from rivers.||35.7|
|8||Distance from urban settlements, marginalization index, social capital *||32.1|
|9||Soil Depth *||28.6|
|10||Land tenure *||17.9|
|11||Distance from ANPs, distance from reforested sites||10.7|
|12||Soil fertility *||7.1|
|13||Distance from greenhouses, migration rate, forest fragmentation *, humidity, biotype *, presence of pest *, social stability *, precipitation, forest fires *, seed dispersion *, presence of cattle *, biodiversity use *, temperature, elevation above sea level.||3.6|
* Criteria without available data for the entire region.
|Socio-economic||Distance from Crops||Euclidian distance (meters) from existing areas used for agriculture|
|Human density||Number of people per km2. Based on extent of municipal area.|
|Distance from roads||Euclidian distance (meters) from highways, roads and trails.|
|Distance from urban||Euclidian distance (meters) from the existing urban zones.|
|Marginalization index||Summary of nine socio-economic indicators of social exclusion .|
|Environmental||Distance from forest||Euclidian distance (meters) from the existing forest.|
|Risk of erosion||Index of risk for soil erosion per water action.|
|Slope of terrain||Percentage of maximum change in z-value of the surface from each pixel.|
|Insolation||Units of hour-watt per m2 of sun insolation. Calculated for the, summer and winter, solstice and equinox.|
|Distance from rivers||Euclidian distance (meters) from the not intermittent rivers.|
The extent of territory that was analyzed with multicriteria analysis, after removal of areas where forest restoration cannot take place was 11,347.83 km2 or 97.56% of the total study area. The non- restricted area comprises 283.17 km2 or 2.43% of the study area. The range of suitability values of the 54 individual maps generated was between 0.16 and 0.94. The thresholds applied to the suitability values to extract the most preferable areas until reaching the restoration target area ranged between 0.42 and 0.86, with an average of 0.70. The percentage of area selected for restoration option sites was an average of 1.35% of the total area, varying between participant maps in a range of ± 0.5%. Figure 4 shows an example of results.
The maps of the most preferred sites generated for each stakeholder groups (Figure 5) showed that most of these are found in the southern portion of the Upper Mixtec region. This part of the region includes 80% of the most suitable area identified by the academic group, 85% by governmental group, 87% by NGO participants and 97% by the public sector.
Figure 6 shows the final map of priority areas for Forest Landscape Restoration, obtained combining the opinion of all participants. The overall restoration priority area covers 151.22 km2 or 1.35% of the study area, in three levels of priority.
Most of this, that is 87.8% of the total, falls in the southeastern part of the Upper Mixtec region. The area with the highest priority (i.e., pixels selected as priority by all sectors or groups of participants) covers 6.804 km2 with 97.7% of these sites located in the south, the area with intermediate priority (i.e., pixels selected as priority by three sectors) covers 52.35 km2 and the area with low priority (i.e., pixels selected as priority by two groups of participants) covers 91.92 km2. Half of the area of highest priority is concentrated within the territory of two municipalities: San Mateo Sinduhi and San Pablo Tijaltepecthe.
FLR is a pragmatic approach [3,4]. It was developed in response to the unclear results of traditional efforts of forest restoration, which have been focused mainly on ecological theory, leaving aside the social domain [45,46]. However, positive effects of an FLR-based process can only come when stakeholders are actually involved and this may prove particularly complex because value judgments must be integrated in a spatial context [47,48,49]. This study proposes an approach to it that aims to account for the opinion of each stakeholder, while keeping statistical analyses and computational efforts at a minimum.
The proposed approach is based on the idea of generating multiple priority maps, one per participant, and then combining them through identification of areas of concordance (i.e., pixels identified as priority in several participants’ maps) to obtain a map for each participant group first and a final overall map later. This approach, by excluding the identification of unique sets of value functions and weights, allowed us to minimize the uncertainty intrinsically associated with those. In fact, the resulting maps assign priority for restoration only to those pixels that have high suitability for multiple actors. GIS-based MDCA has the advantage of letting the user deal with multiple criteria, and can include conflicting opinions that describe the decision problem from different angles . The method proposed here offers a simple yet rigorous way to incorporating different views and automatically generating highly agreed upon results (i.e., pixels selected are only those for which there is a broad consensus among stakeholders).
Nevertheless, the proposed approach has also various limitations. First of all, the need to keep the process of stakeholder involvement sufficiently straightforward imposed simplifications that have deeply affected the robustness of the analysis. In particular, the assessment of value functions has been oversimplified (i.e., only trends were considered, not specific values) with negative consequences on the final results in terms of absolute reliability. Second, our results reveal that there is a poor agreement among experts about which criteria should be used for a land assessment to select restoration sites, as already emphasized by Orsi et al. . Only three criteria (Distance from forest, rate of land use change, and risk of erosion) were mentioned by more than 50% of the people consulted (Table 2). Identifying relevant criteria by means of expert opinion should be taken with caution because the expert knowledge can be incomplete, imprecise and value-biased . Further, even when a set of criteria is eventually agreed upon by stakeholders, this may need to be modified considerably depending on the availability of data for practical application as shown by this study. For this reason, a great amount of effort should be invested in the acquisition, generation, and updating of the information needed for a study.
In fact, the number of criteria and participants to be included is another controversial question. We decided to use the ten more relevant criteria based on a balance between computational load and the need of considering the widest possible range of criteria. While the selection of any number of criteria can be judged as arbitrary, few criteria may lead to the exclusion of important parameters from the analysis and too many criteria may excessively reduce the role of the single criterion and make the evaluation of criteria’s relative importance biased (e.g., establishing the relative importance of 15 criteria may generate issues of consistency in the evaluation). On the other hand, the process to determine the number of participants in a survey suggested by Landeta is based on the error decrease, leaving aside the cost of effort and feasibility. Ianni and Geneletti , and Oliver  present, respectively, a participatory workshop and expert panel approach, as good options to reduce these costs and better manage bias. However, sometimes this is not possible; in our case, those options were not feasible to implement because it would have been impossible to bring stakeholders together from their place of residence because of logistic limitations.
The applied methodology allowed us to identify the priorities of stakeholders from different population sectors and to design some potential restoration sites. At the end, a tangible result, the map of priority areas, represents the most consensual sites to implement a sustainability restoration plan at the regional scale based on participants’ judgment. Multicriteria decision analysis within a GIS environment is a helpful tool in restoration planning for its capabilities to deal with multiple criteria, to model spatially conflicting preferences of stakeholders, and to generate tangible outcome maps. In addition, it fulfils the objectives of FLR involving local stakeholders’ value judgment that are related to ecological integrity and social needs. However, the presence of uncertainty and the limited number of criteria represent some significant limitations. While the proposed approach can account for the opinion of each participant, future applications could improve the survey structure to capture the quantitative opinion of participants on value functions, thus further reducing uncertainty. Finally, it is important to note that the final map is not meant to be an optimal solution, but only the most suitable according to the participants’ value judgments. This map is a useful tool to continuing the negotiations between actors and it is susceptible to changes that can enhance the feasibility of the practical restoration process and guarantee the greatest benefits to local communities. Further research and practical applications could point to the implementation of MCDA techniques for the management of adaptive policies for regional and local scale restoration.
This work has been carried out within the framework of the ReForLan project. The project has received research funding from the European Community’s Sixth Framework Program (FP6), contract number 032132. The views presented do not necessarily represent the opinion of the European Community, and the Community is not liable for any use that may be made of the information contained therein. We would like to thank Raúl Rivera (Laboratory of Geo-Spatial Analyst, National Polytechnic Institute, Mexico) for his advices on GIS. We are grateful to the anonymous reviewers who commented on an earlier version of the manuscript.
David Uribe conducted the fieldwork and performed most of the analyses. Davide Geneletti, Rafael F. del Castillo and Francesco Orsi contributed to the analyses and the interpretation of results. All authors designed the methodology and wrote the paper.
Conflicts of Interest
The authors declare no conflict of interest.
- FAO. Global Forest Resources Assessment 2010. Available online: http://www.fao.org/docrep/013/i1757e/i1757e.pdf (accessed on 15 November 2012).
- Hobbs, R.J.; Harris, J.A. Restoration ecology: Repairing the earth’s ecosystems in the new millennium. Restor. Ecol. 2001, 9, 239–246. [Google Scholar] [CrossRef]
- Lamb, D.; Stanturf, J.; Madsen, P. What is Forest Landscape Restoration? In Forest Landscape Restoration: Integrating Natural and Social Sciences; Springer: Dordrecht, The Netherlands, 2012; pp. 3–25. [Google Scholar]
- IUCN. What is forest landscape restoration? Available online: http://www.iucn.org/about/work/programmes/forest/fp_our_work/fp_our_work_thematic/fp_our_work_flr/ (accessed on 15 November 2012).
- Maginnis, S.; Rietbergen-McCracken, J.; Jackson, W. The Forest Landscape Restoration Handbook; Rietbergen-McCracken, J., Maginnis, S., Sarre, A., Eds.; Earthscan: London, UK, 2007; pp. 1–4. [Google Scholar]
- Malczewski, J. GIS-based multicriteria decision analysis: A survey of literature. Int. J. Geogr. Inform. Sci. 2006, 20, 703–726. [Google Scholar] [CrossRef]
- Yang, M.; Qian, X.; Zhang, Y.; Sheng, J.; Shen, D.; Ge, Y. Spatial multicriteria decision analysis of flood risks in aging-dam management in china: A framework and case study. Int. J. Environ. Res. Pub. Heal. 2011, 8, 1368–1387. [Google Scholar] [CrossRef]
- Sánchez, K.; Jiménez, F.; Velázquez, S.; Piedra, M.; Romero, E. Metodología de análisis multicriterio para la identificación de áreas prioritarias de manejo de recurso hídrico en la cuenca del río Sarapiquí, Costa Rica. Recursos Naturales y Ambiente. Available online: http://ciat-library.ciat.cgiar.org/articulos_ciat/Karla1.pdf (accessed on 14 February 2014). (In Spanish).
- Geneletti, D. Expert panel-based assessment of forest landscapes for land use planning. Mt. Res. Dev. 2007, 27, 220–223. [Google Scholar] [CrossRef]
- Jha, M.K.; McCall, C.; Schonfeld, P. Using GIS, genetic algorithms, and visualization in highway development. Comput.-Aided Civ. Inf. 2001, 16, 399–414. [Google Scholar]
- Caloni, N. Análisis Espacial de Evaluación Multicriterio en la Generación de Alternativas Viales Para el Trazado de la Autopista Luján-Mercedes, Buenos Aires, Argentina. B.S. Thesis, Universidad Nacional de Luján, Argentina, 2003. (In Spanish). [Google Scholar]
- MacDonald, M.L. A multi-attribute spatial decision support system for solid waste planning. Comput. Environ. Urban Syst. 1996, 20, 1–17. [Google Scholar] [CrossRef]
- Leão, S.; Bishop, I.; Evans, D. Spatial–temporal model for demand and allocation of waste landfills in growing urban regions. Comput. Environ. Urban Syst. 2004, 28, 353–385. [Google Scholar] [CrossRef]
- Geneletti, D. Combining stakeholder analysis and spatial multicriteria evaluation to select and rank inert landfill sites. Waste Manage. 2010, 30, 328–337. [Google Scholar] [CrossRef]
- Gomes, E.G.; Lins, M.P.E. Integrating geographical information systems and multi-criteria methods: A case study. Ann. Oper. Res. 2002, 116, 243–269. [Google Scholar] [CrossRef]
- Pereira, J.M.; Duckstein, L. A multiple criteria decision-making approach to GIS-based land suitability evaluations. Int. J. Geogr. Inform. Syst. 1993, 7, 407–424. [Google Scholar]
- Bojórquez-Tapia, L.; Díaz-Mondragón, S.; Ezcurra, E. GIS-based approach for participatory decision making and land suitability assessment. Int. J. Geogr. Inform. Sci. 2001, 15, 129–151. [Google Scholar] [CrossRef]
- Cipollini, K.; Maruyama, A.L.; Zimmerman, C.L. Planning for restoration: A decision analysis approach to prioritization. Restor. Ecol. 2005, 13, 460–470. [Google Scholar] [CrossRef]
- Kangas, J.; Kangas, A. Multiple criteria decision support in forest management—The approach, methods applied, and experiences gained. Forest Ecol. Manage. 2005, 207, 133–143. [Google Scholar] [CrossRef]
- Chazdon, R.L. Beyond deforestation: Restoring forests and ecosystem services on degraded lands. Science 2008, 320, 1458–1460. [Google Scholar] [CrossRef]
- Orsi, F.; Geneletti, D.; Newton, A. Towards a common set of criteria and indicators to identify forest restoration priorities: An expert panel-based approach. Ecol. Indic. 2010, 11, 337–347. [Google Scholar]
- INEGI. Marco Geoestadístico Nacional. Available online: http://www.inegi.org.mx/geo/contenidos/geoestadistica/m_geoestadistico.aspx (accessed on 22 December 2012).
- Ferrusquía-Villafranca, I. Geologic and paleontological studies in the Mixtec region: Geology of the Tamazulapan-Teposcolula-Yanhuitlán area, Upper Mixtec, Oaxaca, Mexio. Bol. Inst. Geol. 1976, 97, 1–160. [Google Scholar]
- García-Mendoza, A.; Tenorio, P.; Reyes, J. El endemismo en la flora fanerogámica de la Mixteca Alta, Oaxaca-Puebla, México. Acta Bot. Mex. 1994, 27, 53–73. (In Spanish). [Google Scholar]
- Toledo, V.M.; Solís, L. Science for the poor people. Program “Water Forever” from the Mixtec región. Ciencias 2001, 64, 32–39. [Google Scholar]
- Carabias, J.; Arriaga, V.; Cervantes, V. Public policies for environmental restoration in Mexico: Constraints, progress, setbacks and challenges. Bol. Soc. Bot. Mex. 2007, 80, 85–100. [Google Scholar]
- Altieri, M.; Anta, S.; Caballero, J.; Hernández, J. Altieri, M.; Anta, S.; Caballero, J.; Hernández, J. Manejo de Agua y Restauración Productiva en la Región Indígena Mixteca de Puebla y Oaxaca: Resultados de los Estudios y Recomendaciones Para Los Tomadores de Decisiones de las Comunidades y Organizaciones de la Sociedad Civil(In Spanish), 1st ed.; The World Bank: México, México, 2006; pp. 23–39. [Google Scholar]
- Céspedes-Flores, S.E.; Moreno-Sánchez, E. Estimación del valor de la pérdida de recurso forestal y su relación con la reforestación en las entidades federativas de México. Investigación Ambiental 2010, 2, 5–13. (in Spanish). [Google Scholar]
- Simon, H.A. The New Science of Management Decision; Harper and Row: New York, NY, USA, 1960. [Google Scholar]
- Bojórquez-Tapía, L.; Brower, L.; Castilleja, G.; Sánchez-Colón, S.; Hernández, M.; Calvert, W.; Díaz, S.; Gómez-Priego, P.; Alcantar, G.; Melgarejo, E.; et al. Mapping expert knowledge: Redesigning the Monarch Butterfly Biosphere Reserve. Conservat. Biol. 2003, 170, 367–379. [Google Scholar]
- Eastman, J.R.; Jin, W.; Kyem, P.A.K.; Toledano, J. Raster procedures for multicriteria/multiobjective decisions. Photogramm. Eng. Rem. Sens. 1995, 61, 539–547. [Google Scholar]
- Orsi, F.; Geneletti, D. Identifying priority areas for Forest Landscape Restoration in Chiapas (Mexico): An operational approach combining ecological and socioeconomic criteria. Landscape Urban Plan. 2010, 94, 20–30. [Google Scholar] [CrossRef]
- Geneletti, D. Incorporating biodiversity assets in spatial planning: methodological proposal and development of a planning support system. Landscape Urban Plan. 2008, 84, 252–265. [Google Scholar] [CrossRef]
- Landeta, J. The Delphi Method: A Technique to Forecast Uncertainty; Ariel: Barcelona, Spain, 1999; p. 224. [Google Scholar]
- Geneletti, D. Formalizing expert opinion through multi-attribute value functions: An application in landscape ecology. J. Environ. Manage. 2005, 76, 255–262. [Google Scholar] [CrossRef]
- Geneletti, D. Multicriteria analysis to compare the impact of alternative road corridors: A case of study in northern Italy. Impact. Assess. Proj. Apprais. 2005, 23, 135–146. [Google Scholar] [CrossRef]
- Gómez, M.; Barredo, J. Sistemas de Información Geográfica y Evaluación Multicriterio en la Ordenación del Territorio(in Spanish), 2nd ed.; Ra-Ma: Madrid, Spain, 2006. [Google Scholar]
- Belton, V.; Gear, T. On the meaning of relative importance. MCDA 1997, 6, 335–338. [Google Scholar]
- ArcGis, version 9.2; Environmental Systems Research Institute Inc. (ESRI): Redlands, CA, USA, 2006.
- ILWIS, version 3.6; International Institute for Aerospace Survey and Earth Sciences (ITC): Enschede, The Netherlands, 2009.
- Malczewski, J. GIS and Multicriteria Decision Analysis; Wiley: New York, NY, USA, 1999. [Google Scholar]
- Geneletti, D. Using spatial indicators and value functions to assess ecosystem fragmentation caused by linear infrastructures. Int. J.Appl. Earth Obs. 2004, 5, 1–15. [Google Scholar] [CrossRef]
- Zucca, A.; Sharifi, A.; Fabbri, A. Application of spatial multi-criteria analysis to site selection for a local park: A case of study in the Bergamo Province, Italy. J. Environ. Manage. 2008, 88, 752–769. [Google Scholar] [CrossRef]
- National Population Council (CONAPO). Marginalization Index at village level 2005. Available online: http://www.conapo.gob.mx/es/CONAPO/Indice_de_marginacion_a_nivel_localidad_2005 (accessed on 14 February 2014). (in Spanish).
- Dudley, N.; Aldrich, M. Five Years of Implementing Forest Landscape Restoration Lessons to Date; World Wildlife Fund International: Gland, Switzerland, 2007. [Google Scholar]
- Mansourian, S. Overview of Forest Restoration Strategies and Terms. In Forest Restoration in Landscapes: Beyond Planting Trees; Mansourian, S., Vallauri, D., Dudley, N., Eds.; Springer: New York, NY, USA, 2005; pp. 8–13. [Google Scholar]
- Keeney, R.; Raiffa, H. Decisions with Multiple Objectives: Preferences and Value Tradeoffs; Wiley and Sons: New York, NY, USA, 1976. [Google Scholar]
- Store, R.; Kangas, J. Integrating spatial multi-criteria evaluation and expert knowledge for GIS-based habitat suitability modeling. Landscape Urban Plan. 2001, 55, 79–93. [Google Scholar] [CrossRef]
- Wang, X.; Yu, S.; Huang, G.H. Land allocation based on integrated GIS optimization modeling at a watershed level. Landscape Urban Plan. 2004, 66, 61–74. [Google Scholar] [CrossRef]
- Ianni, E.; Geneletti, D. Applying the ecosystem approach to select priority areas for forest landscape restoration in the yungas, northwestern Argentina. Environ. Manage. 2010, 46, 748–760. [Google Scholar] [CrossRef]
- Oliver, I. An expert panel-based approach to the assessment of vegetation condition within the context of biodiversity conservation: Stage 1: The identification of condition indicators. Ecol. Indic. 2002, 2, 223–237. [Google Scholar] [CrossRef]
© 2014 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/3.0/).