Modeling Environmental Susceptibility of Municipal Solid Waste Disposal Sites : A Case Study in São Paulo State , Brazil

The large excess of solid waste generated in cities is a result of population growth and economic development. Properly managing this municipal solid waste (MSW) is a challenge, mainly in underdeveloped and developing countries where financial concerns are an added problem. From the environmental point of view, a major issue is properly disposing MSW taking into consideration a wide range of factors, and working with different spatial data. In this study, we used geographic information system (GIS) to perform multi-criteria decision analysis (MCDA) conducted by analytical hierarchy process (AHP). The development of the environmental impact susceptibility model (EISM) for municipal solid waste disposal sites (MSWDS) applied to the state of São Paulo, Brazil considered factors such as geology, pedology, geomorphology, water resources, and climate represented by fifteen associated sub-factors. The results indicated that more than 82% of São Paulo’s territory is situated in areas with very low, low, and medium environmental impact susceptibility categories. However, in the remaining 18% of the state land area, 85 landfills are located in areas with high and very high environmental impact susceptibility categories. These results are alarming because these 85 landfills receive approximately 17,886 tons of MSW on a daily basis, which corresponds to 46% of all municipal solid waste disposed in São Paulo state. Therefore, decision makers, urban planners and policymakers could use the findings of the EISM towards mitigating the environmental impacts caused by MSWDS.


Introduction
The rapid world population growth and economic development are causing changes in terrestrial systems that can have serious and lasting consequences.
The large amount of municipal solid waste (MSW) generated exceeds the capacity of the environment to decompose and recycle these wastes through natural processes [1]. The lack of proper waste management is a major environmental problem [2]. Sustainable management of MSW is required to achieve low environmental impact. One essential part in this process is to properly dispose waste, since disposal sites are permanent facilities that pose risks to the environment and population as they need to be monitored for extended periods of time [3].
Among many methods to dispose MSW in underdeveloped countries, the most common are open dumps and landfills. Open dumps are uncontrolled facilities where waste is directly disposed in the ground without any control causing several impacts. In contrast, sanitary landfills use techniques and methods to better control environmental impacts and are commonly used around the world, particularly in developed countries [4]. Although the number of sanitary landfills is increasing in the last decades in Brazil [5], the nation is through an inadequate MSW disposal scenario, with more than 60% of its cities still disposing MSW in open dumps [6] [7].
It is important take into consideration the environmental impact caused by municipal solid waste disposal (MSWD) in São Paulo, Brazil due to several factors. First, São Paulo is the most populous state in America and Western Hemisphere. Second, São Paulo is the biggest producer of MSW among all Brazilian states. Third, the per capita waste generation rate in São Paulo state is the biggest rate in Brazil with 1.4 kg/habitant/day with a growing trend over the years [8].
Finally, various São Paulo cities still dispose MSW improperly [9]. Therefore, all these factors together lead to the occurrence of negative environmental impacts.
Among diverse kinds of environmental impacts caused by humans, the MSWD is one of the most impactful, because solid wastes are retained in the same place where they are deposited even though they may undergo chemical and physical transformations over the years [10] [11] [12].
For this reason, assessing the environmental impact caused by MSWD sites must consider different parameters, to avoid potential negative effects. Developing a model for assessing environmental impact susceptibility must take into consideration multiple issues, values, scales and degrees of uncertainty, as well as assist stakeholder engagement. In this process, the models are usually built to satisfy one or more of five main purposes: 1) prediction, 2) forecasting, 3) management and decision-making under uncertainty, 4) social learning, and 5) de-veloping system understanding and experimentation [24].
In this study to develop an environmental impact susceptibility model (EISM) for municipal solid waste disposal sites (MSWDS), we used a multi criteria decision analysis (MCDA) approach via an analytic hierarchic process (AHP) coupled with geographic information system (GIS). This paper is organized as follows: Section 2 discusses the literature review of GIS, MCDA and AHP applied to environmental studies. Section 3 describes the methods used to develop the EISM and describes the study area. Section 4 presents the model results for the state of São Paulo and the MSWDS assessment. Finally, the conclusions are presented in Section 5.

Background Literature Review
In this section, the literature review is divided into four parts: Section 2.1 includes the advantages of GIS in environmental studies, Section 2.2 demonstrates the importance of MCDA applied to municipal solid waste issues, and Section 2.3 explains the use of AHP.

Geographic Information System (GIS)
The use of geographic information system is one of the most promising approaches to investigate complex spatial phenomena, because GIS has the advantage of storing, retrieving and analyzing a considerable amount of disaggregated data from various sources and displaying the results spatially, which helps decision makers solve several problems [25] [26].
GIS has also been used in numerous studies to improve municipal solid waste management (MSWM

Multi Criteria Decision Analysis (MCDA)
Multi criteria decision analysis is a method to structure a problem through the action concepts and intelligible criterion group to facilitate the communication in decision process, forming a conviction rather than determining an optimum [61]. Combining MCDA with spatial decision problems usually contains a large set of feasible alternatives and conflicts with an incommensurate evaluation criteria [62].
The MCDA applied to environmental studies had a significant growth over the last decade [63] [64]. The integration of spatial analysis using GIS to MCDA has been used in different environmental studies. Examples include, analyzing the possibility to convert pastures to croplands in Brazil [65], mapping the landslide susceptibility [31], and identifying geotechnical land suitability [66].
Spatial analyses associated with MCDA is considered one of the main application for GIS [67], and have also been used in several studies related to municipal solid waste issues [26] [75]. Because MSWM involves multiple factors such as environmental, economic, political and social [37], combining MCDA with GIS increases the analysis effectiveness and accuracy [50] helping to understand the complexity of the problem, ensuring the robustness and reliability of the final decision.

Analytical Hierarchy Process (AHP)
In this study, we use the Analytical Hierarchy Process (AHP), which is a component of the Multi Criteria Decision Analysis method. The AHP was developed by Saaty in the 1970s [76] and consists of an assessment theory through pairwise comparison to help decision makers set priorities and choose the best decision [37] [59]. The AHP in combination with GIS has been widely used in the field of natural resources and environmental management [77], first because the combined approaches are easy to implement using map algebra operations and cartographic models, and second because the approaches are intuitively appealing to decision makers [62]. The comparisons are made using a scale of absolute judgments ranging from one to nine, where one represents equal importance and nine represents the highest importance from one element to another (Table  1). In addition, a reciprocal value is used to express the inverse comparison [78]. The process of AHP determination involves these subsequent steps: 1) compute sum of values in each column of pairwise matrix, 2) normalize the matrix by dividing each element by its column total and, 3) compute the mean of the elements in each row of the normalized matrix [60].
Afterwards to determinate the consistency of the AHP judgment, a consistency index (CI) (Equation (1)) is determined [76].
Subsequent the determination of CI, a consistency ratio (CR) needs to be calculated (Equation (2)) [76].  [78]. The CR is acceptable if its value is less than 10%. However, if this number is higher than 10%, the judgments may be inconsistent and should be re-evaluated [53].

Methods
To develop the environmental impact susceptibility model for municipal solid waste disposal sites, we considered six major steps: 1) selection of environmental decision factors and sub-factors; 2) data acquisition and integration into a GIS database; 3) definition of classes and assignment of ratings; 4) data standardization to a common scale of measurement; 5) calculation of relative weights using the AHP technique; and 6) derivation of the final model map using weighted linear combination (WLC) aggregation method ( Figure 1). Each step is described as follows.

Selection of Environmental Decision Factors and Sub-Factors
In this study, the selection of environmental factors and sub-factors was based on the literature that takes into account the environmental impact susceptibility associated with disposal of municipal solid waste e.g. [ [83]. We also took into consideration guidelines, relevant legislation and regulations, experts' opinions, and available data. Overall, a total of five factors including geology, pedology, geomorphology, water resources, and climate, with fifteen associated sub-factors were used in the model ( Figure 2). This list is not exhaustive; we only considered what the literature included as the most important criteria to develop the environmental impact susceptibility model for municipal solid waste sites.

Geology
Geological features influence the environmental susceptibility of municipal solid waste disposal sites because they can cause land instability in an earthquake region [66]. They can also, influence water infiltration if the rock formations are porous or have faults [84]. For this reason, when municipal solid waste is disposed above susceptible rocks, the process of waste landslide and water contamination may occur. Some geological aspects are considered in previous studies, including [43] [53] [54] [83] [85]. However, these studies did not consider simultaneously the four geological sub-factors used in this model, which are 1) distance to faults, 2) porosity of rocks, 3) distance to seismic areas, and 4) distance to caves.

Pedology
Soil parameters, such as depth and physical characteristics, could interfere in environmental susceptibility related to siting municipal solid waste facilities, mainly for two reasons. First, strength characteristics of the soil are important to support the overlying load from the waste mass. Second, the soil permeability can interfere in infiltration process, which in turn can cause contamination of water bodies. Multiple studies in MSWDS issues included pedologic aspects in their assessments, e.g., [3] [46] [82] [83]. In particular, we used the pedology sub-factors of 1) type of soil and 2) infiltration rate.

Geomorphology
Geomorphology is mainly related to terrain features and the influence of these characteristics on the topography and runoff process. For example, flat areas influence leachate infiltration, while steep areas influence terrain instability.

Water Resources
Another aspect that affects environmental susceptibility is associated with surface and underground water resources. It is not appropriate to have MSWDS close to surface water sources or in areas where the water table level is shallow due to the higher contamination risk. Several studies took into consideration these aspects, e.g., [ [82]. In this study, we used surface water resources sub-factors of 1) distance to rivers and lakes and 2) flood risk, while, for underground water resources, we used the sub-factors of 1) distance to wells, 2) aquifer flow, and 3) aquifer vulnerability to pollution.

Climate
Climate factors need to be used in modeling the environmental impact susceptibility for municipal solid waste disposal, mainly because they can interfere in the decomposition process of solid waste and in the volume of leachate generated, due to the water balance as well as the amount of landfill gas generated. Climate aspects also were considered in previous investigations, e.g., [44] [57] [59] [82]. In this study, we used the climatic sub-factors of 1) precipitation and 2) temperature.

Data Acquisition and Integration into a GIS Database
The spatial database used in the environmental impact susceptibility model for municipal solid waste disposal sites applied to the São Paulo state was created using a variety of sources including geologic, pedologic, geomorphologic, hydrologic and, climatologic data of different scales ( Table 2). The successful use of GIS depends on the accessibility of data, as well as its quality, representing the real world conditions through diverse layers [56].
In this study, all data layers were stored, manipulated, analyzed, and visualized using ArcGIS version 10.2 ModelBuilder as a starting point for a multi-criteria decision analysis. ModelBuilder is a GIS extension that encodes complex sequences of GIS operations into a simple graphic model from which the steps can be executed [86]. The data layers were georeferenced using the UTM System Datum SIRGAS 2000 (Zone 22 and 23 South).

Definition of Classes and Rating
Each of the fifteen sub-factors used in the environmental impact susceptibility model for municipal solid waste disposal sites was divided into classes. Each class was rated on a scale from one to ten, where one represents the lowest level of susceptibility and ten represents the highest level of susceptibility for environmental impact.
The rating intervals from one to ten was selected based on similar scales used by [43] [55] [87] [88], as well as based on the experience and judgment of the authors. Furthermore, the importance for each class could vary based on the region of interest and characteristics of the specific area [56]. In this study the classes were assigned considering the relevant conditions in the state of São Paulo (Table 3).

Data Standardization to a Common Scale of Measurement
In order to overlay the spatial information to calculate the environmental impact susceptibility, it is necessary to standardize the data into a common measurement scale. Therefore, the fifteen sub-factors were converted into raster grid

Weight Assignment Using AHP
The construction of a comparison matrix and the derivation of weights in our study uses the analytical hierarchic process web-based tool developed by [89].
First, the AHP methodology was applied to the factors (Table 4) and sub-factors (Table 5). Then, by multiplying these two results, the global weighting for each sub-factors was obtained (Table 6).

Weight Linear Combination (WLC) Method
After checking the reliability of the pairwise comparisons for factors and sub-factors, the environmental impact susceptibility model for municipal solid waste disposal sites in the São Paulo state was built using a weighted linear combination method, following (Equation (3)).
In this equation, S is the EISM final score, i W is the sub-factor weight, and i X is the standardized class rating of factor i. As the sum of weight for factor i is a multiplication of i W and i X for each sub-factor, the i W is constrained to one, while i X varies from zero to ten, and the final combined estimate is presented on this scale.
Therefore, the EISM final score was obtained for each raster cell as a sum of   the products of ratings assigned for each class (Table 3) and global weights obtained by AHP (Table 6) (Figure 4). The results were grouped into five categories of environmental impact susceptibility for municipal solid waste disposal sites: Very Low (S1), Low (S2), Medium (S3), High (S4) and Very High (S5) ( Table 7).

Environmental Impact Susceptibility Model for Municipal Solid Waste Disposal Sites
The results of the environmental impact susceptibility model for municipal solid waste disposal sites in the state of São Paulo are presented in ( Figure 5). The area for each susceptibility category indicate that most part of São Paulo state, 77.3% have medium environmental impact susceptibility category (S3), 16.8% has high category (S4), 4.8% has low category (S2), 1.1% has very high category (S5) and there is no representative areas for the very low category (S1) ( Table 8).
The high and very high categories (S4 and S5, respectively) in the state of São Paulo extend are located near the surface water resources, which is correlated to the EISM global weights that has the sub-factor distance to rivers and lakes as the most important contributor. There is also a concentration of the higher categories near the Atlantic Ocean mainly in the southeast of the state of São Paulo, which can be explained by a combination of geographical variables. For example, there is a mountain range in this area formed by the Serra do Mar and Serra da Mantiqueira, which has a concentration of steep areas. In addition, these mountain range stop the humidity that comes from the ocean to the continent, which makes the precipitation near the coast very high in comparison to the rest of the state of São Paulo.
The EISM for MSWDS in the state of São Paulo was progressed well due to availability and reliability of spatial data and the findings in this study provide V. F. Nascimento et al.    an advancement to previous models due to three main reasons: 1) a higher number of factors used, 2) a higher number of sub-factors used, and 3) a more extensive set of combined factors and sub-factors used. Therefore, the EISM results for MSWDS indicate a decent environmental impact susceptibility representation of the state of São Paulo.

Analysis of Susceptibility for MSW Disposal Sites in São Paulo State
In order to evaluate the environmental impact susceptibility for each municipal solid waste disposal site in São Paulo state we developed a spatial analysis ( Figure 6) and statistical study ( Table 9).
The geographical coordinates of municipal solid waste disposal sites for the 645 municipalities in São Paulo state were obtained from spreadsheets used to assess the waste quality index developed by the Environmental Company of São Paulo State (CETESB) [90]. Because some of São Paulo's cities use consortia to dispose solid waste, there are currently 420 municipal solid waste disposal sites   When a separate analysis was performed for sanitary and ditch landfills, a total of six ditch landfills were in the lower susceptibility categories (S1 and S2) and just one sanitary landfill was in the very high susceptibility category (S5) (  (Table 10).

Conclusions
Through the development of the environmental impact susceptibility model for municipal solid waste disposal sites using multi criteria decision analysis and analytical hierarchic processes coupled with geographic information system, it was possible to identify the most and least environmentally susceptible areas using five environmental factors associated with fifteen sub-factors. With the application of the EISM, it was also possible to assess the current susceptibility of municipal solid waste disposal sites in São Paulo state, Brazil.
In this study, the results of the environmental impact susceptibility model in- The development of this model took three main modeling purposes into consideration, including prediction, management decision-making under uncertainty, and developing system understanding and experimentation. This type of spatial analysis can help stakeholders promote the mitigation of environmental  For future studies, to improve the environmental impact susceptibility assessment for MSWDS the authors suggest adding 1) forecasting, using different climate scenarios that influence leachate generation and emission of greenhouse gases, and 2) social learning, coupling a social model with the EISM, which could result in a greater understanding of global susceptibility.