A municipality-based pressure index developed in the Campania Region as a geo-stratification tool for human and environmental monitoring studies

In response to the complex social, environmental and economic situation caused by the "Terra dei Fuochi" phenomenon, the IZSM collaborated closely with the Terra dei Fuochi working group, carrying out additional monitoring surveys both on food (QR Code Campania project[30]), and on the environment (Campania Trasparente project[31]). This model was developed in the context of the experience accumulated in the field of environmental and food monitoring, and represents an innovative tool aimed at increasing knowledge of the environmental context of the Campania Region through an objective, integrated and organic synthesis of complex environmental phe-nomena and territorial dynamics. The model proposed here is useful for the global and synthetic assessment of environmental pressure on a municipal basis. As shown, it can also be applied to aggregations of municipalities. Furthermore, it can be used in the context of institutional actions for the planning and monitoring of improvements on a local or regional scale. Finally, the pro-posed municipality-based environmental pressure index represents the basis for geo-stratification of the sample in the context of population biomonitoring studies on a regional scale, as in the described biomonitoring study design applicable to the Campania Region.


Introduction
Since the 1980s, the organized crime has been responsible for the continued illegal trafficking of industrial waste and toxic materials in the so-called "Terra dei Fuochi" (Land of Fire), a territory mostly located in the provinces of Naples and Caserta of the Campania region in Southern Italy. The term "Terra dei Fuochi" was introduced by the Italian environmental association Legambiente and refers to the fact that waste was abandoned and illicitly disposed through uncontrolled combustion [1].
In the early 1990s, the Campania region suffered from a prolonged "waste crisis", which lasted roughly 15 years, caused by the inability of the Institutions to provide for the proper mangament of urban solid waste.
Waste that accumulated in municipal areas was often set on fire by citizens exasperated by the nauseating smell [2], which generated fears of being exposed to dioxins among indewelling citizens [3].
Public concern about the threats posed to human health by environmental contamination grew in 2004, when Mazza and Senior [4] used the expression "Triangle of Death" to indicate a geographical area comprised within the municipalities of Acerra, Nola and Marigliano of the province of Naples. The authors concluded that the area was characterised by an unexpected high incidence of some forms of malignant neoplasms, which they assumed was the result of exposure to toxic waste. While the report was extensively covered by the media, its methodological limits highlighted by other researchers [5] were largely ignored [3].
While a growing body of scientific evidence suggested that citizens indwelling in the provinces of Naples and Caserta could be affected by an increased risk of death, cancer-related mortality, cancer incidence and congenital abnormalities [6] [7] [8], the "bad reputation" of the Land of Fires severely harmed the local economy over the years and especially in 2013 and 2014, among widespread fears of consumers that the food produced in the Campania region was contaminated. As an example, in 2014 revenues of one typical product of the Campania Region, such as the water-buffalo mozzarella cheese, dropped by approximately 57 4 million Euros [9]. In order to tackle the social, economic and environmental emergency situation, a "Terra dei Fuochi Working Group" was established by the law 6 / 2014 [10]. In an area of 92 hectares assessed in the Region, 21 were identified as unsuitable for agri-food production by the Working Group, although none of the agricultural products analysed were found to be non-compliant with regulatory limits for toxic subsatnces [11]. The activities carried out by the Working Group sure had merits, but also suffered from several weaknesses, such as: 1) soil was the only environmental matrix analyzed (no air or water samples were assessed); 2) not all municipalities were included in the environmental monitoring plan; and 3) no human biomonitoring survey was conducted.
Although human biomonitoring studies play a key role in assessing the threats posed by environmental pollution, only few biomonitoring studies have been conducted in the Campania Region [12] [13]. In a territory as vast and densely populated as that of the Campania region, which presents a surface of 13,590 km² and has over 5.5 million inhabitants residing in 550 municipalities divided in five provinces [14] [15], a systematic biomonitoring survey can be effectively carried out at a regional level if the recrutiment plan is wisely designed.
In this original work, we constructed a mathematical model that computes a synthetic index of enviromental pressure at a municipality level (Municipality Environmental Pressure Index -MEPI). We computed the MEPI for all municipalities of the Campania Region and we used it as a geo-stratification tool for the recruitment plan of a human biomonitoring survey at a regional level [16].

Development of the environmental pressure index applicable on a municipal basis
The Municipal Environmental Pressure Index (MEPI) is defined based on a pairwise comparison process between variables (Table 1) to which scores of relative significance are assigned through a multi-criteria 5 approach based on the Analytic Hierarchy Process (AHP) method [14,15]. With this approach it was possible to move from a qualitative to a quantitative assessment of environmental sensitivity and to locally (at the municipal scale) establish the value of each variable in terms of its contribution to MEPI, according to the semantic classification proposed by Saaty [16] (Table 2). Figure 1 shows the block diagram of the algorithm for calculating MEPI. In the process, Ai is the single source of contamination considered and aij is the numerical value resulting from the comparison between criteria i and j, which can vary from 1 to 9, where each value of the scale is assigned according to criterias proposed in 2) if the variable Ai is judged to be of equal intensity relative to Aj, then aij = aji= 1.
The last row in matrix A shows the sum of the individual elements that make up each column.
Matrix A was normalised, dividing each element aij by the sum relative to the j-th column. Subsequently, the average value of each i-th row of the matrix was calculated, defining the "priority vector" as shown in Table   4.
For each source of contamination Ai, the model gave its percentage weight and in table 5 sources are sorted in descending order.
To evaluate whether matrix A was consistent, or that the requirements of consistency and significance in the judgments expressed by the "preference indices" were met, all the cells belonging to the i-th row of the nonnormalised matrix were added together and multiplied vectorially by the sum of the priority vector and divided by the weight of the criterion relating to that row. In this way it was possible to quantify the 6 consistency of each priority as shown in Table 6.
The consistency index (CI) of the entire matrix A was calculated using the following relation, where λ represents the maximum eigenvalue of matrix A and n the dimension of the matrix itself (Eq. 1): In equation 1, if the value of CI is equal to 0 then the matrix is consistent; if it deviates from n, then the matrix is not perfectly consistent, although the methodology used accepts a low degree of inconsistency because this does not affect the validity of the result obtained. As a first approximation, the maximum eigenvalue of matrix A can be evaluated by referring to the average of the consistencies relating to the individual variables; the result is a maximum eigenvalue equal to 8.70 which is close to the dimension n of matrix A.
Once the CI was known, it was possible to define the Random Consistency Index (RCI); for matrix A (with a n value equal to 8) the RCI value is equal to 1.41 (Table 7). For matrix A to be consistent, the value of CR must be less than 0.1. In the specific case, Eq. 2 gave a result of 0.07 stating the consistency of the matrix.
Once the "Pi" weights to be assigned to each pressure variable were determined, the MEPI values were determined. Specifically, for each municipality, MEPI was calculated by a linear combination of the set of pressure variables considered, multiplied, in turn, by specific amplification coefficients as function of the number, type, extent, hazard, environmental status, and impact of the variable itself. These coefficients were introduced in such a way as to be able to define the model on the environmental and territorial characteristics of each municipality in Campania region. MEPI relating to the i-th municipality of Campania region is expressed by the following relationship: In order to make a comparison between the environmental pressure indices determined, the variable was normalised in such a way as to have values between 0 and 100. The normalisation operation was carried out through application of the following relationship: to Terra dei Fuochi. Following investigations by the Campania Regional Environmental Protection Agency (ARPAC), the plots of land defined by decree were further classified as reported in Table 8.
In determining the environmental pressure index, only types B, NC and D of class 3, 4 and 5 plots of land were considered. Table 9 shows the p' scores assigned: the criterion adopted in this case was to attribute a significant significance, in terms of hazard, to contaminated landfills.
A further p'' score was attributed according to the specific spatial extension of these sites, assuming that the degree of pressure is directly proportional to the extent of the site as reported in Table 10.
The pressure index relating to the variable in question is expressed by the following mathematical relationship, where the sum is extended to all the contaminated sites present in a municipality:

"Areas of particular interest" variable
This variable includes sites of national (SNI) and regional (SRI) interest present in the area, illegal landfills awaiting investigation, as well as potentially contaminated sites investigated. For each area of interest considered, a preliminary score p' is assigned on the basis of the different degree of presumed pressure. In the specific case, sites of national interest are assigned the highest score, followed by illegal landfills, areas awaiting characterisation and, finally, potentially contaminated sites. The potential risk index relating to the "Areas of particular interest" variable was evaluated by assigning the score to the single element of the variable as shown in Table 11. From an analytical point of view, the environmental pressure index for the considered variable is expressed by the following relationship: ,( .

"Zoning" variable
The Zoning variable considers the different impacts, both direct and indirect, exerted by different land uses (urban, agricultural and commercial/industrial areas). To define the environmental pressure index relating to the "zoning" variable, scores (p') were assigned to the different land use destinations present in the 2012 edition of the "Corine Land Cover" map, based on qualitative assessments of pressures exerted in each of them. Based on the available data, the criterion followed consisted in attributing a greater weight, in terms of environmental hazard, to residential areas, followed by industrial areas and then agricultural and wooded areas (Table 12 ).
To represent the relative extension of each intended use over the entire municipal area, a specific p" indicator was introduced, divided into the following parameters:  Tables 13 and 14 show the bands considered and the relative scores assigned.
The environmental pressure index associated with the zoning variable is expressed analytically by the following expression: ,( where the sum is extended to the n land uses present in a municipality.

"Illegal waste spills and fires" variable
The"Illegal waste spills and fires" variable indicates the presence of abandoned waste and uncontrolled fires.
To determine the pressure index relating to this variable, the number of waste spills detected in the municipalities by the monitoring activity carried out by SMACampania was taken into consideration. The environmental pressure index relating to the "illegal waste spills and fires" variable coincides with the number n of spills detected on a municipal basis. It can be mathematically formalised by the following formula:

"Waste management plants" variable
This variable includes incineration, storage, composting, selection, purification, recovery, scrapping plants and controlled landfills. A p' score is assigned to the specific type of plant. Table 15 shows the types of plants considered and the relative scores assigned.
In assigning the weights, it was decided to attribute the same significance to the types of plants which, on the basis of the waste treated, present the same level of environmental hazard. The environmental pressure index relating to the variable was mathematically formalised by the following relationship:

"Terra dei Fuochi plots of land decree" variable
As regards the Plots of land of the TdF Decree, defined in the Directive of 23/12/2013, all the plots for which a site-specific investigation has not yet been carried out (2.a and 2.b) and those of class 5, 4 and 3 which, following investigations, are not contaminated (Table 16) were taken into consideration.
The environmental pressure index relating to the variable "TdF decree plots" was assessed by assigning the score p' to the single element of the variable according to the formula presented below.

"Potential hazard -soil analysis" variable
The potential hazard variable was created starting from analysis of the spatial distribution of the concentration values of contaminants using spatial statistics models, which made it possible to reconstruct continuous concentration areas on the entire regional territory and to estimate the probability of exceeding the legal limits or reference values in areas not covered by sampling [17][18][19]. The Potential Hazard map is very useful insofar as, in addition to enabling the identification of areas potentially at risk, it serves to define the background/baseline values of the various geochemical elements investigated, according to the various types present in the substrate. On the basis of this cartography, the following indicator was taken into consideration for each municipality: -Potential Hazard area/Municipal area.
The values of this ratio were divided into five classes defined on the basis of a classification of a natural breaks [20] type. A p' score was applied to each of the intervals thus defined in Table 17.
The environmental pressure index relating to this variable is determined through the relationship shown below, where a represents the number of analytes the concentrations of which have exceeded regulatory limits and the sum present is extended to the n areas potentially at risk, with reference to the municipal territory:

"Water bodies status -water analysis" variable
This variable takes into account the quality status of groundwater bodies. To fully define the environmental pressure index relating to this variable, a series of attributes were introduced that indicate the qualitative status (Chemical Status of Groundwater -CSG) and a series of indicators that take into account the percentage municipal coverage of the underground aquifer. In fact, the CSG is an index that summarises the qualitative state of groundwater based on the comparison of the average annual concentrations of the chemical parameters analysed with the relative quality standards and threshold values defined at national level by Legislative Decree 30/09 [21], also taking into account natural background values. Based on this, a p' score was assigned to the qualitative status of the groundwater body. The highest score was assigned to the "poor" status, insofar as this condition presupposes exceeding of the reference values (standard and threshold), even for a single parameter. The assigned score took into consideration the anthropogenic or natural origin of the aforementioned exceedances. Table 18 shows the scores assigned.
Subsequently, for each municipality, the following indicator was introduced to take into account the extension of the groundwater body in relation to the municipal area: -Groundwater body area/Municipal area The values of these indicators were divided into a series of intervals (bands), established according to a classification of the "natural breaks"type. A second p'' score was then assigned to each interval thus defined. Tables 13 and 14 show the bands considered and the relative scores assigned. The environmental pressure index related to the variable was evaluated through the following relationship: where the sum is extended to all groundwater bodies within a municipality.

Calculation of the environmental pressure index on a municipal basis
Following the application of multicriteria analysis, the contaminated sites variable assumes greater significance than the others considered as health risk has been ascertained (exceeding Risk Threshold Concentrations -RTC) for the potentially exposed population. With a difference of about 9 percentage points, it follows the Areas of particular interest variable, which includes all those territorial circumstances in which

Design of a human biomonitoring study in the Campania Region
For the design of a human biomonitoring study to be conducted at a regional level, a total of 174 Municipalities of the Campania Region, representing 80% of the regional population, were chosen on the basis of geographical contiguity and logistical constraints. First, the Municipalities were grouped into 3 areas, as described in Tables 21 and 22, based on geographical contiguity and classified at high, medium and low environmental pressure based on the average MEPI weighted per municipality residents. We then grouped municipalities within the same area into "clusters", which represents the actual tool for geo-stratification to be used for the biomonitoring study, following the "Natural Breaks" approach [20], wit hthe expetion of municipalities of the Sabato and Irno Valleys, which were included into 3 separate clusters (Valle dell'Irno 1,

Discussion
Existing pressures, making it possible to identify a geo-stratification unit on which to perform population sampling, with significant resource savings and faster recruitment [13].
In conclusion, the model proposed here is useful for the global and synthetic assessment of environmental pressure on a municipal basis. As shown, it can also be applied to aggregations of municipalities. Furthermore, it can be used in the context of institutional actions for the planning and monitoring of improvements on a local or regional scale. Finally, the proposed municipality-based environmental pressure index represents the basis for geo-stratification of the sample in the context of population biomonitoring studies on a regional scale, as in the described biomonitoring study design applicable to the Campania Region. Http://Www.Arpacampania.It/Impianti-Autorizzati-Alla-Gestione-Dei-Rifiuti-in-Regione-Campania.

Values aij Interpretation
1 i and j are equally important 3 i is slightly more important than j 5 i is much more important than j 7 i is very much more important than j 9 i is extremely more important than j

Sites of national interest 9
Illegal landfills 7 Areas awaiting characterisation 5 Potentially contaminated sites 3