SEARCH

SEARCH BY CITATION

Keywords:

  • Remediation technologies comparison;
  • Multicriteria decision analysis;
  • Contaminated sites;
  • Decision support system

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. BACKGROUND
  5. PROPOSED METHODOLOGY FOR THE SELECTION OF REMEDIATION TECHNOLOGIES
  6. APPLICATION OF THE PROPOSED METHODOLOGY AND DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. References

A methodology for selecting remediation technologies is presented as part of a decision support system for the rehabilitation of contaminated sites. It includes 2 steps: In the 1st step, a pool of suitable technologies is selected within a technologies database according to their applicability to site-specific conditions; in the 2nd step, the selected technologies are ranked according to a multicriteria decision analysis (MCDA) approach. The MCDA was applied to allow for a transparent procedure and for the integration of expert analyses. The methodology was implemented in a previously developed georeferenced information system–based decision support system for the rehabilitation of contaminated sites and then applied to a case study (Porto Marghera, Venice, Italy). On the basis of the obtained results, the proposed methodology appeared suitable to select remediation technologies according to both technical features and requirements of available technologies, as well as site-specific environmental conditions of the site of concern, such as chemical contamination levels and remediation objectives.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. BACKGROUND
  5. PROPOSED METHODOLOGY FOR THE SELECTION OF REMEDIATION TECHNOLOGIES
  6. APPLICATION OF THE PROPOSED METHODOLOGY AND DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. References

A large number of contaminated sites are disseminated in over the postindustrialized countries (estimated European contaminated sites vary from 300,000 to 1.5 million [EC 2002]; on average, 5 contaminated sites per 1,000 inhabitants) with expected high financial consequences (e.g., overall costs for the remediation of European contaminated sites range between 59 and 109 billion Euros, according to the European Environmental Agency [EC 1999]). Criteria and methodologies for the rehabilitation of contaminated sites are urgently needed.

In recent years, new regulatory frameworks have been proposed (Ferguson and Kasamas 1999) that highlight the fundamental role played by the environmental risk assessment and call for the implementation of procedures allowing the comparison of remediation technologies within the decision process leading to the rehabilitation of a contaminated site (Ferguson et al. 1998; Vik and Bardos 2003). The best available technology at sustainable cost is a well-established criterion for the selection of reclamation technologies. The choice of best available technologies, however, should not be based just on scientific and technical considerations because economic, social, and political aspects need to be taken into account. The socioeconomic implications behind site requalification, according to well-defined land end uses, heavily influence the identification of remediation objectives and, therefore, the choice of most suitable technologies.

The integration of environmental risk assessment and best available technologies according to land end uses is a difficult task, especially in the case of so-called megasites, areas of notable extension (hundreds or thousands of ha) in which several different sources of contamination are present simultaneously. Rehabilitation of megasites requires a set of technologies to be identified, taking into consideration spatial, temporal, and logistic aspects. Moreover, megasites usually occupy strategic portions of land, triggering a multidisciplinary approach in which risk analysis, socioeconomic analysis, and the comparison of remediation technologies need to be integrated according to a transparent framework.

This work will focus on the issues behind the selection and comparison of remediation technologies. The main purpose is to define a site-specific methodology for selecting a set of remediation technologies applicable to a contaminated site and to develop a comparative system for these technologies. A “set of technologies” is regarded as a combination of technologies, possibly organized in train technologies and extended over space and time, able to achieve the remedial goals defined for the site under examination. Multicriteria decision analysis (MCDA) is applied to rank remediation technologies and to develop alternative remediation scenarios to be offered to decision makers (i.e., stakeholders).

The proposed methodology is included in the decision support system “decision support system for rehabilitation of contaminated sites” (DESYRE), software designed to assist experts in the rehabilitation of large contaminated sites (i.e., megasites; Carlon et al. 2003). A decision support system is designed to assist the decision maker involved with contaminated lands in carrying out specialized analyses and ensuring reproducible and transparent processes that consider all the variables involved (Bardos et al. 2003).

Very few decision support systems exist in the field of remediation technologies selection. Among them, the Cost–Benefit Analysis for Remediation of Land Contamination, by the UK Environment Agency (UKEA), guides in the selection of a short list of potential remediation techniques through MCDA and cost–benefit analysis (UKEA 1999).

All the other available resources are mainly searchable databases, such as the US Environmental Protection Agency ReachIT (USEPA 2004) or the Remediation Technologies Screening Matrix by the US Federal Remediation Technologies Roundtable (FRTR 2002).

BACKGROUND

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. BACKGROUND
  5. PROPOSED METHODOLOGY FOR THE SELECTION OF REMEDIATION TECHNOLOGIES
  6. APPLICATION OF THE PROPOSED METHODOLOGY AND DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. References

The DESYRE system

The structure of DESYRE (Carlon et al. 2003) consists of 5 modules: 1) characterization, 2) socioeconomic analysis, 3) remediation technologies comparison, 4) risk analysis (RA), and 5) decision making.

The characterization module, developed according to the requirement of environmental risk assessment procedures (ASTM 1998), allows users to define the conceptual model of the site and to provide information regarding contaminants distribution and transport through the different environmental media, with georeferenced information system tools for handling spatial data. The socioeconomic module supports the selection of the optimal future use of the site. Socio-economic parameters are inputs of a fuzzy expert system that generates a composite indicator of suitability for alternative land uses. The remediation technologies comparison and risk analysis modules integrate with each other to produce an effective framework for the selection of cleanup techniques. The stepwise structure of this framework (Figure 1) allows the definition of different “remediation scenarios”: Each scenario refers to a suitable solution for the rehabilitation of the contaminated site, encompassing the final land use, the socioeconomic benefits, a technologies set (with reference to costs, intervention time, and environmental impacts), and the associated residual risk.

The final decision-making module provides the description of alternative remediation scenarios. Technological, environmental, and socioeconomic aspects are described by macro-indexes that are derived directly from the technological, risk assessment, and socioeconomic modules, respectively (Carlon et al. 2003).

Review of remedial technologies

The database created by the FRTR (2002) is universally recognized (Vik and Bardos 2003) as an exhaustive and up-to-date source of remedial technologies. It collects information from studies and remediation actions within the Superfund Program, and it is continuously updated. Currently, the FRTR matrix includes 64 technologies divided according to the treated environmental matrix: solid (soil, sediment, bedrock, and sludge), liquid (groundwater, surface water, and leachate), or gaseous (air emission or off-gas treatment). Moreover, the techniques are considered according to where the operations take place—in situ or ex situ (i.e., ex situ treatments require the removal of contaminated matrix, in situ treatments do not)—and the adopted containment systems—solutions meant to prevent or reduce the migration of contaminants through the ground or subsurface water.

Finally, the FRTR matrix groups the technologies according to their target pollutants, which are split into 6 categories of contaminants: Nonhalogenated volatile organic compounds, halogenated volatile organic compounds, nonhalogenated semivolatile organic compounds, halogenated semivolatile organic compounds, fuels, and inorganics.

Review of criteria for the comparison of remediation technologies

The comparison of remediation technologies was conducted by selecting internationally recognized evaluation criteria proposed by the FRTR (2002), the Organization of the United Nations (UN 1997), and UKEA (1999).

The set of criteria proposed by FRTR and related to the technologies evaluation screening matrix introduced in the previous paragraph includes a series of parameters that allow a detailed description of the different techniques without selection or ranking intent.

The UN study indicates criteria similar to those of FRTR (e.g., overall cost, cleanup time, system reliability and maintainability), and provides a ranking of the technologies, which is based on a fixed, non–site-specific scoring system.

Finally, the UKEA defines a selection procedure based on a cost–benefits analysis approach. It includes analysis of all the aspects that characterize remediation before and after the intervention actions. Each selected aspect is scored as a result of expert judgment on site-specific conditions.

Review of MCDA methods

The MCDA approach is normally used for problems in which a decision maker classifies or chooses among some alternatives that are measured by a finite number of criteria or attributes (Norris and Marshall 1995). For cases in which no single alternative exhibits the most preferred available value for all attributes, MCDA is especially useful (Saaty 1980; Norris and Marshall 1995). This method corresponds to what happens in the planning phase of a remediation intervention when the experts have to select 1 or few remediation options among several cleanup technologies.

The MCDA method can be classified as a single- or multiple-person decision process. The latter involves a group of experts or decision makers and is considered a member of group decision theory. In this case, the MCDA algorithms have to include suitable consensus measures that show how much the group of decision makers agree or disagree about the alternative ranking (Carlsson et al. 1992).

In the DESYRE tool, the comparison of remediation technologies was designed as a single-person activity, involving the presence of only 1 expert. In fact, this choice was suggested by the most probable usage of the software tool, in which only 1 expert or user at a time has an active role in the decision process; if more than 1 participates, unanimous agreement with all judgments is expected. Thus, hereafter, only single MCDA algorithms will be considered.

According to the specialized literature (Vincke 1992), single MCDA tools are classified as multiattribute utility theory (MAUT), outranking and interactive. However, as discussed in a previous paper (Giove et al. 2006) and reported by Linkov et al. (2004), the applications of MCDA to environmental problems are mainly based on MAUT, the analytic hierarchic process (AHP), and some outranking methodology (in particular, the Elimination et Choix Traduisant la Realité [ELECTRE] method).

In MAUT methods, the attribute values of each alternative are aggregated by means of a suitable “utility” function (or “value” functions) to obtain the score of the investigated alternatives. This approach is based on the hypothesis of a rational and consistent decision maker (Bridges et al. 2004) and implies the existence of both the value functions and a suitable aggregation operator. Many methods exist to define the value functions (Keeney 1976), but the description of those methodologies is beyond the aim of this paper. Even the aggregation operator needs to be carefully selected, and as discussed in a previous paper (Giove et al. 2006), the simplest and most widely used aggregation function in the MAUT context is given by the weighted averaging operator. However, it should be stressed that the choice of a particular aggregation operator depends on several elements, such as the kind of the problem, available information, and subjective reasoning of the expert.

thumbnail image

Figure Figure 1.. Framework of the remediation technologies (RT) comparison and risk assessment (RA) module of the DESYRE decision support system.

Download figure to PowerPoint

As far as the AHP is concerned, it was 1st developed by Saaty (1980) and widely reviewed and applied in the literature (Saaty 2000; Ramanathan 2001). The AHP aggregation algorithm simply computes a linear combination of the criteria values as weighted averaging operators, making use of pairwise comparison of alternatives (i.e., for each pair of attributes, the expert specifies a judgment of “how much more important” 1 attribute is compared with another), allowing a numerical value to be obtained even from intangible criteria. The pairwise comparison is a robust method, giving local and global weights that, as verified in many applications, respect the hidden preference structure of the user. A main characteristic of the consistency property is that it implies that a user is coherent if the pairwise comparisons satisfy the transitivity property (Saaty 1980, 2000). However, given the uncertain characteristics of human thinking, a limited amount of inconsistency can be accepted.

thumbnail image

Figure Figure 2.. Hydrogeological model of the Porto Marghera (Venice, Italy) contaminated megasite. a.s.l. = above sea level.

Download figure to PowerPoint

As discussed in a previous paper (Giove et al. 2006), despite the great popularity and the large number of AHP applications, some criticisms are warranted (Barzilai 2001): primarily, the rank reversal phenomenon and the great number of comparisons required as the number of alternatives increase. Methods are available to solve these criticisms, including the supermatrix approach and the AHP network, originally suggested by Saaty (1980), and the adoption of the absolute mode (Ramanathan 2001). In the absolute mode, the pairwise comparison is limited only to the 1st level of the hierarchy to obtain the global weights, whereas for the lower levels, a direct judgment is required by the user. In so doing, the amount of information required decreases by 1 order of magnitude (i.e., from N2 to N, where N is the number of data points). Moreover, rank reversal is avoided because deleting an alternative, or inserting a new one, does not change the scoring of those remaining (because they were assigned from a direct judgment, unchangeable by the insertion or deletion); thus, the ranking will not change. This method will be described in the methodology section with particular reference for application to remediation technologies comparison.

Study area

The proposed methodology was applied to the Porto Marghera industrial zone, a contaminated megasite bordering a complex and fragile ecological–naturalistic and hydrogeological transitional environment such as the Venice lagoon.

The Porto Marghera site covers a surface of 3,595 ha, out of which 479 ha are occupied by canals. The site was originally a mudflat (so-called barene) that has been raised up ˜2 m above sea level by fill, with material from the dredging of lagoon canals (Gatto and Carbognin 1981) and waste production residues, including industrial toxic waste. The area started to operate at the beginning of the last century.

As for hydrogeology, the site is located over a coastal multiple-aquifer system subject to tidal effects. By assembling previously available information (Carbognin et al. 1972; Critto et al. 2004) implemented by further experimental investigations, the hydrogeological conceptual model of the site is shown in Figure 2.

The topsoil is covered by a layer of fill material contaminated by industrial waste, including red bauxitic mud, black organic sludge, or both. Below the topsoil, a 1st impermeable layer is found, consisting of Holocene deposits of lagoon mudflats and an overconsolidated silt clay layer called caranto. The 1st impermeable layer is followed by a semiconfined aquifer, delimited at the bottom by clayey Pleistocene sediments that constitute the deepest impermeable layer (Carbognin et al. 1972).

Chemical contamination was characterized extensively on a 100-m sampling grid. In the topsoil (i.e., the fill material layer), several classes of pollutants were found: amines, chlorobenzenes, chloronitrobenzenes, chlorophenols, dioxins, aliphatic hydrocarbons, polynuclear aromatic hydrocarbons, metals, metalloids, and inorganic anions (Venice City Council 2001). In the semiconfined aquifer, the same classes of pollutants were found.

Metals and metalloids showed the highest concentration levels and the widest spread of contamination. The soil samples displayed high concentrations of arsenic (hot spots of 900 mg/kg dry wt at 0–1 m depth), chromium (hot spots of 4,200 mg/kg dry wt at 0–4 m depth), cadmium (hot spots of 900 mg/kg dry wt at 3–4 m depth), copper (hot spots of 3,000 mg/kg dry wt at 3–4 m depth), mercury (hot spots of 130 mg/kg dry wt at 1–2 m depth), and lead (hot spots of 26,000 mg/kg at 3–4 m depth). Moreover, on the basis of the number of samples that exceeded Italian law (Law Decree 471/99 1999) for acceptable concentration limit (ACL) for industrial use, widespread contamination was found, especially for arsenic (117 of 1,392 samples exceeding limits; ACL = 20 mg/kg dry wt), cadmium (127 of 1,389; ACL = 2 mg/kg dry wt), and mercury (105 of 1,389; ACL = 1 mg/kg dry wt). Lead contamination was less widespread (39 of 1,393 samples exceeding regulatory limits; ACL = 100 mg/kg dry wt); finally, for copper (15 of 1,384; ACL = 120 mg/kg dry wt) and chromium (4 of 1,354; ACL = 150 mg/kg dry wt), the contamination appeared to be confined to only 1 hot spot.

PROPOSED METHODOLOGY FOR THE SELECTION OF REMEDIATION TECHNOLOGIES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. BACKGROUND
  5. PROPOSED METHODOLOGY FOR THE SELECTION OF REMEDIATION TECHNOLOGIES
  6. APPLICATION OF THE PROPOSED METHODOLOGY AND DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. References

The proposed methodology consists of 3 main steps: 1) selecting the remediation technologies, 2) setting comparative criteria, and 3) ranking the selected remediation technologies with a comparative procedure.

The 1st step provides a pool of remediation technologies suitable to the case study; the 2nd step assesses the previously selected techniques according to several evaluation criteria. The last step develops a ranking algorithm by MCDA that allows a comparison and classification of the selected technologies.

Selection of remediation technologies

To select the remediation technologies, 2 subsequent selection filters were applied to the input database of technologies provided by the FRTR matrix (Figure 3). The 1st filter was based on 2 basic parameters, commercial availability of the considered technology and the technology target contaminants overlapping those found in the site under study. The 2nd filter was concerned with site-specific parameters (i.e., hydrogeological and physicochemical characteristics of the investigated environmental matrix) affecting the feasibility of remediation technologies.

At the end of the selection, a pool of remediation technologies was chosen, including all technologies applicable simultaneously or subsequently to the study site.

This selection procedure was implemented into the DESYRE software as a hypertext database document developed in Microsoft® Word through Visual Basic Macro. It included 3 interactive tables (A, B, and C) and a characterization database, which drive the potential expert in the aforementioned selection pathway.

thumbnail image

Figure Figure 3.. Steps applied to the selection of a suitable site-specific technologies pool.

Download figure to PowerPoint

The A table (Figure 4) includes the input database of technologies and groups the cleanup technologies according to the treated contaminated matrix (i.e., soil, surface water, or groundwater, emitted off-gas).

Each technology is characterized by target contaminants, commercial availability, general site characteristics required for applicability, main benefits and disadvantages or drawbacks, and the target pollutants found at the site. Moreover, the last column of the table contains a synthetic judgment or evaluation made by the expert (FA = the technology is fully applicable and the remediation action does not imply any impediment; AR = applicable with reserve [i.e., the cleanup actions have notable troubles in 1 of the listed parameters]; NA = not applicable [i.e., the technology shows a specific impediment]).

thumbnail image

Figure Figure 4.. Structures of the A, B, and C tables used sequentially in the selection of remediation technologies.

Download figure to PowerPoint

The B table (Figure 4) includes only the technologies selected by the expert in the previous table A after application of the 1st filter (i.e., the technologies marked FA or AR). Table B is only descriptive and provides the characterization of the selected technologies according to the additional criteria of remediation strategy (i.e., extraction, removal and retrieval, biodegradation, immobilization, destruction, and chemical transformation), specific cleanup effectiveness for the considered contaminant classes, and capability to be included in train technology treatments.

The C table (Figure 4) lists site-specific parameters that affect the applicability of the selected technologies: either related to the treated matrix (e.g., pH, total organic carbon, hydraulic conductivity, and soil cation exchange capacity) or related to the target contaminants (e.g., vapor pressure, solubility). The applicability range values of these parameters for each technology were obtained by published technical documents (Los Alamos National Laboratory 1996; NATO/CCMS 2001; Vik and Bardos 2003).

Finally, the characterization database (not reported here) includes all the information concerning the geotechnical (e.g., soil granulometry) and physicochemical (e.g., soil organic carbon content) characteristics of investigated environmental media obtained during the characterization process.

With the use of the characterization database and table C, the expert can apply the 2nd site-specific filter by identifying the fit between hydrogeological and physicochemical characteristics of the site, summarized in the characterization database, with the aforementioned applicability range values summarized in the 2nd column of table C (labeled “Site-specific criteria that affect the applicability”). The software then prompts the expert for a final applicability judgment or assessment for all technologies, which is inserted in the 3rd column of table C (labeled “Expert applicability judgment/assessment”). Only the technology marked “Applicable” will be selected, which leads to a pool of remediation technologies actually applicable to the case study.

Setting of comparative criteria set: Macrocriteria and evaluation matrix

To compare the pool of selected remediation technologies, the following 6 comparative macrocriteria were defined: reliability, course of action (i.e., intervention condition), “hazardousness,” community acceptability/impacts, effectiveness, and cost (Figure 5). Each macrocriterion is able to describe a specific aspect of a cleanup action.

The reliability macrocriterion considers the maintenance aspects and results obtained by the technology application to other case studies. The course of action (i.e., intervention condition) macrocriterion identifies the logistic and technical aspects related to a remediation action by differentiating between in situ, ex situ, and off-site technologies and by considering the possibility of creating a train technology. The macrocriterion measuring hazard allows assessment of the potential effects for human health resulting from the technology application (e.g., effects related to the use of hazardous reagents or the emission of dust and volatile substances). With the use of the community acceptability/impacts macrocriteria, the negative effects on the environment, as well as the main factors on which depend public judgment of a particular remediation technology, are evaluated. The effectiveness macrocriterion helps the expert to assess technology performance, which depends on cleanup time and removal rates. Finally, the cost macrocriterion points out the parameters on which depend the actual or real costs of a remediation action (e.g., time, installation and maintenance cost, need of waste disposal).

thumbnail image

Figure Figure 5.. Macrocriteria and associated evaluative criteria used for the comparison of remediation technologies.

Download figure to PowerPoint

Table Table 1.. Qualitative and quantitative rating of the evaluation criteria selected for comparing the remediation technologies
Thumbnail image of

Figure 5 reports the evaluation criteria associated with each macrocriterion. These evaluation criteria were identified from the review of international approaches (UN 1997; UKEA 1999; FRTR 2002), and each can contribute to the definition of more than 1 macrocriterion (e.g., cleanup time influences both effectiveness and cost), as shown in Figure 5. Some of the selected evaluation criteria are strictly correlated with technical aspects (i.e., costs, cleanup time, performance, reliability and maintenance, technology development status, cleanup operation locations, train technology, hazardous reagents use, contaminated matrix removal, and residuals production), whereas other criteria refer to the potential effects on human health and the environment (i.e., dust and volatile substances emission, effects on water, and consequences to soil and community acceptability).

Performance, cost, and cleanup time are the most important criteria in the description of a remediation technology according to cost–benefit analysis. Performance is the removal ratio, expressed as the ratio between the residual (i.e., after treatment) pollutant concentration in a given matrix and the initial concentration in the same matrix. Cost gives information on the overall cost per treated matrix unit (US$/t of soil matrix wet weight and US$/1,000 L of water matrix), and it is strictly related to the removal rates. The costs applied to the study case were obtained from reviewing case studies with features similar to those of the analyzed site, so that site-specific criteria could be introduced (Los Alamos National Laboratory 1996; NATO/CCMS 2001; OCETA 2001; USEPA 2001a, 2001b; Vik and Bardos 2003). Finally, cleanup time is the average time required to clean a site with a specific technology. According to the USEPA (2001a) approach, all cleanup times were estimated by referring to standard conditions, namely 20,000 t of soil and 3,785,000 L of water.

For each evaluation criterion, a qualitative or quantitative rating was defined as shown in Table 1. This rating scheme allowed the so-called evaluation matrix to be obtained—1 for each contaminant class found in the site analyzed (see Table 2, in which the evaluation matrix for inorganics is presented). In this matrix, evaluation criteria were reported in the columns, and the selected technologies suitable to treat specific pollutants were in the rows. The evaluation matrix provides the input data for the technologies ranking step, described in the following paragraph.

Comparative procedure and ranking of the selected remediation technologies

The 3rd step of the proposed methodology, the comparative procedure, aims at obtaining a ranking of remediation technologies according to the macrocriteria previously presented. The final classification will allow the experts to create remediation technologies sets that act as possible remediation solutions.

The comparative process is based on a ranking algorithm that was developed by single-person MCDA tools. Specifically, the selected MCDA method uses a decision matrix as the database (Table 7) that summarizes the available raw data for the decision maker. Each row of the matrix corresponds to a specific alternative (in the current case, a specific remediation technology), whereas each column corresponds to a proposed macrocriterion. Accordingly, each element ij of the matrix represents the judgment of the alternative i with respect to the macrocriterion j (i.e., the jth attribute value for the alternative i) and can be expressed in different ways (i.e., symbolic, numerical, Boolean).

Table Table 2.. Evaluation matrix for inorganic contaminants and the remediation technologies resulting from the selection step. For legend, see Table 1a
    Performance (% removal)           
Soil treatmentTechnologyCleanup timeOverall cost ($/t or $/1,000 L)AsCrCuCdHgNiPbZnCleanup operation locationTrain technologyEffects on watersEffects on soilsDust and volatile substances emissionsContaminated matrix removalResiduals productionCommunity acceptabilityHazardous reagents useReliability/maintenanceTechnology development statusb
  1. a As = arsenic; Cr = chromium; Cu = copper; Hg = mercury; Ni = nickel; Pb = lead; Zn = zinc; NA = not applicable; NAV = not available.

  2. b I = innovative; C = consolidated.

In situPhytoremediationAverage25NAV38NAVNAV42NAVNAV98In situYesYesNoNoNoYesHighNoWorseI
 Electrokinetic separationAverage6288908593NAV908973In situYesNoYesYesNoYesHighNoAverageC
 Solidification/stabilizationBetter90NAVNAVNAV60NAVNAV7460In situNoYesYesYesNoNoAverageNoBetterC
Ex situSeparationBetterNaVNANANANANANANANAEx situYesNoYesYesYesYesAverageNoBetterC
 Soil washingBetter200969387NAVNAV929198Ex situYesNoYesYesYesYesLowYesBetterC
 Solidification/stabilizationBetter1409999979898979897Off-siteNoNoYesYesYesYesLowYesBetterC
OtherLandfill capWorseNANANANANANANANANAIn situNoYesYesYesNoYesLowNoBetterC
 Landfill cap enhancementWorseNANANANANANANANANAIn situNoYesYesYesNoYesLowNoBetterC
 Excavation, retrieval and off-site disposalBetterNANANANANANANANANAOff-siteNoYesYesYesYesNALowNoBetterC

Moreover, according to the discussion in the Background section, the selected MCDA method is based on the weighted averaging operator associated with the absolute AHP to structure the problem into a suitable hierarchy and to determine the criterion weights. This choice was motivated from the aforementioned characteristics, including easy interpretability of its linear form, user-friendly capability, probability of low or null interaction among the criteria, and limited amount of information required.

According to the standard AHP, the absolute AHP is based on 3 fundamental steps: 1) Structuring the problem with hierarchy, which allows a complex problem to be divided into a series of levels of analysis so that each attribute is a member of a small set of attributes on the same level, all attributes are related to a single attribute on the level immediately above them, and the last level is formed by the available alternatives; 2) comparison of judgments, which allows the relative importance of the variables (attributes or alternatives) belonging to the same level and relative to each of the associated variables belonging to the upper level to be calculated with the use of a pairwise comparison method (i.e., for each pair of attributes, the expert specifies a judgment of “how much more important” 1 attribute is compared with another) with a predefined (and limited) scale, usually the natural scale of 1, 2, …, 9 points; 3) analysis of priorities, which leads, through suitable aggregation tools, to a final ranking of the alternatives (Saaty 1980; Norris and Marshall 1995).

Moreover, for specific application in the remediation technologies comparison, in the absolute AHP mode, the relative importance (i.e., the weight) was calculated with the pairwise comparison only for the hierarchic level in which the macrocriteria were included, not for all the levels, as the relative AHP provides for (i.e., overall goal concerning the environmental requalification, macrocriteria, evaluation criteria, and technological alternatives). Therefore, for the overall goal, the relative importance is intended to be the importance of each macrocriterion relative to the 1st hierarchic level (i.e., obtaining a ranking of the selected remediation technologies that facilitates the definition of a remediation plan for the study case).

In accordance with the general considerations for AHP reported in the Background section, the absolute AHP was adopted to avoid both the great number of required comparisons of the alternatives and the undesired rank reversal phenomenon. Thus, the values of the macroattributes for each alternative were directly assigned by the expert. This process allows the expert knowledge to be captured and implemented in a numeric form. In this way, the proposed process can be considered, at the same time, a support and check tool for the experts who have to assess the several selected technologies to obtain different technologies sets.

The overall defined comparative process finally comprised 1) structuring the problem with hierarchy, 2) computing macrocriterion weights, 3) assessing the technologies with the use of a judgment matrix, and 4) computing the final score for each technology.

Structuring the problem in hierarchic levels—The proposed structure include 4 hierarchic levels, as shown in Figure 6. The 1st level consists of the final goal (i.e., obtaining a ranking of the selected remediation technologies to define a remediation plan for the study case); the 2nd level consists of the proposed macrocriteria, the 3rd level concerns the selected evaluation criteria, and the 4th level concerns the different available alternatives (i.e., the selected remediation technologies).

Table Table 3.. Stepwise selection of remedial technologies for inorganic contaminants
Technologies for inorganics from technologies' database (Table A)Technologies resulting from the 1st selectionPool of technologies applicable to the study case for inorganics category
Treatments for soil sediments and sludge in situ
1Phytoremediation1Phytoremediation1Phytoremediation
2Electrokinetic separation2Electrokinetic separation2Electrokinetic separation
3Fracturing3Fracturing  
4Soil flushing4Soil flushing  
5Solidification/stabilization5Solidification/stabilization3Solidification/stabilization
6Vitrification    
Treatments for soil sediments and sludge ex situ
7Chemical extraction    
8Oxidation/reduction6Oxidation/reduction  
9Separation7Separation4Separation
10Soil washing8Soil washing5Soil washing
11Solidification/stabilization9Solidification/stabilization6Solidification/stabilization
Containment and other treatments
12Landfill cap10Landfill cap7Landfill cap
13Landfill cap alternatives11Landfill cap alternatives8Landfill cap alternatives
14Excavation, retrieval, and off-site disposal12Excavation, retrieval, and off-site disposal9Excavation, retrieval, and off-site disposal
Table Table 4.. Saaty's (1980) numerical scale applied in the matrix of pairwise comparison (reported in Table 5)
Intensity of importanceDefinition
1Equally as important
3Moderately more important
5Strongly more important
7Very strongly more important
9Extremely more important
2, 4, 6, 8Intermediate values between defined rankings

Computing macrocriteria weights—The 2nd step of the comparison process is the computing of macrocriteria weights. According to the AHP method (Saaty 1980), this is obtained with a pairwise comparison matrix (i.e., Table 5), in which rows and columns list the proposed macrocriteria. For each pair of macrocriteria, the expert specifies a judgment of “how much more important” 1 macrocriterion is than another. To specify the pairwise comparison judgments, a numerical approach was adopted (i.e., the expert answers each question with a number, as in “Attribute A is 3 times as important as Attribute B”) on the basis of the Saaty numeric scale (Saaty 1980) reported in Table 4. With the use of this scale, the macrocriteria weights were successively estimated by applying the eigenvector method (Saaty 1980). Moreover, the developed procedure allows both ordinal (maximum–minimum transitivity) and cardinal consistency analysis to be applied, with the aim to avoid intransitive cycles (Kwiesielewicz and Van Uden 2004) and too-low cardinal consistency (Saaty 1980, 2000).

Evaluating remediation technologies with the judgment matrix—This step evaluates the selected remediation technologies (i.e., the available alternatives) through the so-called judgment matrix (i.e., the aforementioned decision matrix). A specific judgment matrix has to be developed for each contaminant class. In this matrix (Table 7), the rows correspond to the selected remediation technologies that are able to treat a particular contaminant class, whereas the columns correspond to the proposed macrocriteria, whose weights were previously calculated. Each element ij of the matrix represents the score (i.e., expert judgment) of the technology i with respect to the macrocriterion j. The expert assigns this score by assessing the rating that the evaluation criteria, correlated to a specific macrocriteria, assumes for a selected remediation technology in the aforementioned evaluation matrix (see Table 2 for inorganics). These judgments (i.e., scores) are expressed according to a numerical scale (1 = sufficient, 2 = rather good, 3 = satisfactory, 4 = good, and 5 = excellent) that allows several judgment levels to be explained, while avoiding too-heavy computational efforts. Also in this step, a consistency check is performed: if an alternative is dominated by any others, the criteria judgment has to be coherent, otherwise an alarm is sent to the user.

Table Table 6.. Macrocriteria weights calculated by the eigenvector method (Saaty 1980)
MacrocriterionWeights
Effectivenessβ1 = 0.4750
Community acceptability/impactsβ2 = 0.0288
Reliabilityβ3 = 0.0644
Intervention conditionsβ4 = 0.0434
Hazardousnessβ5 = 0.1472
Costβ6 = 0.2412

Ranking algorithm—The last step of the comparative procedure concerns the definition of the ranking algorithm, which estimates the final score for each selected remediation technology. The proposed algorithm is the sum of the products of the numerical judgment (expressed for each macrocriterion through the judgment matrix) with the corresponding macrocriterion weight according to the Equation 1,

  • equation image(1)

where Pi is the total scoring associated with the ith remediation technology, β′j is the normalized weight associated with the jth macrocriterion, vij is the numerical judgment assigned to the ith alternative (technology) and correlated with the jth macrocriterion, and L is the number of macrocriteria considered.

Table Table 5.. Matrix of pairwise comparison applied to the calculation of macrocriteria weights
 EffectivenessCommunity acceptability/impactsReliabilityIntervention conditionsHazardousnessCost
Effectiveness175755
Community acceptability/impacts1/711/31/31/51/7
Reliability1/53131/51/7
Intervention conditions1/731/311/51/5
Hazardousness1/555511/3
Cost1/577531
Table Table 7.. Judgment matrix for inorganic contaminants and the remediation technologies selected in the stepwise procedurea
 EffectivenessCommunity acceptability/impactsReliabilityIntervention conditionsHazardousnessCost
  1. a 1 = sufficient; 2 = rather good; 3 = satisfactory; 4 = good; 5 = excellent.

Electrokinetic separation434533
Solidification/stabilization in situ415433
Phytoremediation141554
Separation425422
Soil washing525432
Solidification/stabilization ex situ415122
Landfill cap115342
Landfill cap alternatives114342
Excavation, retrieval and off-site disposal114112

APPLICATION OF THE PROPOSED METHODOLOGY AND DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. BACKGROUND
  5. PROPOSED METHODOLOGY FOR THE SELECTION OF REMEDIATION TECHNOLOGIES
  6. APPLICATION OF THE PROPOSED METHODOLOGY AND DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. References

The proposed methodology was applied to the contaminated megasite of Porto Marghera for the selection and comparison of treatment technologies for the removal of inorganic contaminants present in the surface fill material layer (Figure 2).

Remediation technologies selection

To obtain a pool of remediation technologies suitable for the case study, the characterization database was used, and the A, B, and C table were applied (Figure 4).

The 1st selection filter, following the application of table A to the 14 remedial technologies proposed by the FRTR matrix for soil matrix and inorganic pollutants (Table 3, 1st column), allowed the 12 technologies reported in Table 3, 2nd column, to be selected. Vitrification was eliminated because it could not assure any homogeneity of the intervention over the whole study area; chemical extraction was discarded because of the high cost of chemical reagents.

thumbnail image

Figure Figure 6.. Hierarchic structure applied in the proposed absolute AHP that was adopted for the comparison procedure. Gi = macrocriteria, βi = macrocriteria weight, Ci = evaluative criteria, Ti = selected remediation technologies.

Download figure to PowerPoint

The 12 selected technologies were characterized by the additional criteria reported in table B (Figure 4) and then were submitted to the 2nd selection filter resulting from the application of table C (Figure 4). By the matching of site-specific hydrogeological and physicochemical characteristics of the surface soil, summarized in the characterization database, with the applicability range values presented in the 2nd column of table C (Figure 4), a final remediation technologies pool was selected, which included 9 technologies applicable to the case study (Table 3, 3rd column). The selected pool consisted of both ex situ treatments (i.e., separation, soil washing, and solidification/stabilization) and in situ treatments (e.g., electrokinetic separation, solidification/stabilization). Moreover, it included the technologies suitable for areas with widespread, near-surface, and low-level contamination (i.e., phytoremediation). Finally, containment systems (i.e., landfill cap, landfill cap alternatives and excavation, retrieval, and off-site disposal) were also included in the pool because they can be applied whenever the contamination levels, volumes, or both are so high that no other remediation actions are feasible.

Comparative criteria set: Evaluation matrix

The selected remediation technologies were next subjected to comparison steps. For the comparative procedure and ranking algorithm, the qualitative or quantitative ratings defined for each evaluation criterion (Table 1) were used to implement the evaluation matrix for the inorganic pollutants (Table 2) through a specific and suitable software interface. All the matrix data were inserted according to the most conservative option (i.e., the worst case, for example, the average of the highest costs and the lowest performance obtained by the reviewed study cases with features similar to those of the analyzed site; Los Alamos National Laboratory 1996; NATO/CCMS 2001; OCETA 2001; USEPA 2001a, 2001b; Vik and Bardos 2003).

Comparative procedure and ranking

As previously illustrated in the Methods section, the comparative procedure consisted of several steps. The results obtained by the application of each step to all selected technologies are reported in the following paragraphs.

Computing macrocriteria weights—The pairwise comparison matrix was filled in (Table 5) and the macrocriteria weights were estimated (Table 6) with the use of the Saaty numerical scale (Table 4) and a specific software interface included in the DESYRE decision support system. According to Tables 5 and 6, the macrocriteria “cost” and “effectiveness” obtained the highest preference judgments and weights, followed by the macrocriterion “hazardousness.” This highlights the relative importance of technical, economic, and risk considerations within the process for the definition of remedial scenarios. The macrocriterion “community acceptability/impacts” obtained the lowest preference judgment and weight because of the low relevance given by the expert to the environmental and social consequences that the remediation process could cause to the already strongly compromised context of the industrial area of Porto Marghera.

Technologies evaluation with the judgment matrix—On the basis of the evaluation criteria rating (Table 1), the evaluation matrix (Table 2), and the judgment matrix (Table 7), the selected technologies were evaluated. The assigned scores (i.e., expert judgments) for each technology and for each macrocriterion (i.e., attribute) are reported in Table 7. For instance, soil washing obtained an excellent judgment for effectiveness and reliability because of the high performance and reliability and the low maintenance and cleanup time. However the rather good judgment obtained by soil washing for community acceptability/impacts and cost reflects the rather low community acceptability (normally correlated with ex situ treatment) and the high cost.

Ranking algorithm—Using the macrocriteria weights reported in Table 6 and the assigned scores reported in Table 7, Equation 1 (in the Methods section) was applied and a final scoring for each technology was estimated, thus obtaining the remediation technologies ranking (Table 8).

Soil washing was the technology with the highest ranking: It is an effective solution to treat soil contaminated by metals and metalloids (even for high concentration levels), and it is able to treat large amounts of soil. The 2nd option, electrokinetic separation, is a suitable in situ solution for hot spots (with metal concentrations from a few parts per million to tens of thousands of parts per million). The lowest score, assigned to landfill cap alternatives and excavation, retrieval, and off-site disposal, reflects the common features of these remedial options: no reduction, only a containment, of contaminated soil is obtained. In addition, they present a high hazardousness level because of possible percolation into groundwater or release of volatile compounds into the atmosphere. They usually are adopted when no other remediation technologies are technically or economically feasible.

CONCLUSIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. BACKGROUND
  5. PROPOSED METHODOLOGY FOR THE SELECTION OF REMEDIATION TECHNOLOGIES
  6. APPLICATION OF THE PROPOSED METHODOLOGY AND DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. References

A methodology for selecting and comparing a set of remediation technologies and for supporting the definition of suitable site-specific remediation scenarios was developed.

The proposed selection and comparative procedure is user-friendly (i.e., organized on specific tables and matrices integrated into a suitable software interface, which interacts with the experts), transparent (i.e., has stepwise traceability of any selection and comparison), and interdisciplinary (i.e., the framework integrates technological, environmental, and socioeconomic knowledge). The integration of guidelines and criteria shared by the international scientific community gives solidity and reliability to the entire process.

Finally, the proposed procedure, through the integration of software modules, expert knowledge, and professional judgment, plays a key role within the DESYRE software application.

The integration of technologies selection with other aspects of the remediation process in a single tool is the distinctive feature of DESYRE compared with other available tools. In fact, the other few decision support systems and existing databases provide the user with the selection of suitable remediation technologies, but DESYRE includes this functionality in a more general framework for remediation of large contaminated sites while considering socioeconomic, risk, and environmental impact aspects.

Moreover, DESYRE technologies selection through MCDA methodologies is based on a wide variety of criteria, encompassing not only cleanup time and cost, but also community acceptability/impacts, performance, reliability/maintenance, location of cleanup operations, and effects on waters and soils. In this respect, DESYRE supports a fully integrated and complete decision process for site remediation.

Table Table 8.. Ranking of the remediation technologies selected for soil matrix and inorganic contaminants
RankTechnologyScoring
1Soil washingP1 = 3.85
2Electrokinetics separationP2 = 3.64
3Solidification/stabilization in situP3 = 3.58
4SeparationP4 = 3.23
5Solidification/stabilization ex situP5 = 3.07
6PhytoremediationP6 = 2.57
7Landfill capP7 = 2.02
8Landfill cap alternativesP8 = 1.96
9Excavation, retrieval, and off-site disposalP9 = 1.43

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. BACKGROUND
  5. PROPOSED METHODOLOGY FOR THE SELECTION OF REMEDIATION TECHNOLOGIES
  6. APPLICATION OF THE PROPOSED METHODOLOGY AND DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. References

We thank Manuela Samiolo and Stefano Silvoni for assistance in analyzing data. This work and the DESYRE project were funded by the Italian Ministry for University and Scientific Research.

References

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. BACKGROUND
  5. PROPOSED METHODOLOGY FOR THE SELECTION OF REMEDIATION TECHNOLOGIES
  6. APPLICATION OF THE PROPOSED METHODOLOGY AND DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. References
  • [ASTM] American Society for Testing and Materials. 1998. Standard provisional guide for risk-based corrective action. Philadelphia (PA): ASTM. PS 104–98. p 100.
  • Bardos P, Lewis A, Nortcliff S, Mariotti C, Marot F, Sullivan. 2003 Review of decision support tools for contaminated land management and their use in Europe. A report from the Contaminated Land Rehabilitation Network for Environmental Technologies (CLARINET) by the Decision Support Tools Working Group. Vienna (AT): Federal Environment Agency.
  • Barzilai J. 2001. Notes on the analytical hierarchy process. In: Proceedings of the NSF Design and Manufacturing Research Conference; Tampa (FL). p 16.
  • Bridges T, Kiker G, Cura J, Apul D, Linkov I. 2004. Towards using comparative risk assessment to manage contaminated sites. In: LevnerE, LinkovI, ProthJM, editors, Strategic management of marine ecosystems. Amsterdam (NL): Kluwer. p 202210.
  • Carbognin L, Gatto P, Mozzi G. 1972. Hydrogeologic aspects of Venice Lagoon. Venice, Italy. Venice (Italy): National Research Council (CNR. Lab St.Din.Gr.-Ma.). TR32 Final Report. (In Italian.)
  • Carlon C, Critto A, Nadal N, Samiolo M, Marcomini A, Petruzzelli GA. 2003. Desyre—Decision support system for rehabilitation of contaminated sites. In: Proceedings of Consoil 2003 8th International Conference on Contaminated Soil. Gent, Belgium. Karlsruhe (DE): Forschungszentrun Karlsruhe GmbH. p 29872993.
  • Carlsson C, Ehrenberg D, Eklund P, Fedrizzi M, Gustafsson P, Lindholrn P, Merkuryeva G, Riissanen T, Ventre A. 1992. Consensus in distributed soft environments. Eur J Oper Res 61: 165185.
  • Critto A, Zuppi GM, Carlon C, Marcomini A. 2004. Effect of a contaminated site (the San Giuliano landfill, Venice, Italy) on the interaction between water bodies in a coastal aquifer system. Ann Chim 94: 303314.
  • [EC] European Commission. 1999. Directive 1999/31/EC concerning waste landfills. Gaz Ufficiale 182. July 16, 1999. p 19.
  • [EC] European Commission. 2002. Towards a schematic strategy for soil protection. Brussels (Belgium): Commission of the European Communities. Communication from the Commission to the Council, the European Parliament, the Economic and Social Committee and the Committee of the regions. COM(2002)179.
  • FergusonC, DarmendrailD, FreierK, JensenBK, JensenJ, KasamasH, UrzelaiA, VegterJ. editors. 1998. Risk assessment for contaminated sites in Europe. Vol 1: Scientific basis. Nottingham (UK): LQM Press. A report prepared as part of the Concerted Action on Risk Assessment for Contaminated Sites in the European Union (CARACAS). p 165.
  • FergusonC, KasamasH, editors. 1999. Risk assessment for contaminated sites in Europe. Vol 2: Policy frameworks. Nottingham (UK): LQM Press. Concerted Action for Risk Assessment for Contaminated Sites in the Europe Union (CARACAS). p 223.
  • [FRTR] Federal Remediation Technology Roundtable. 2002. Remediation technologies screening matrix and reference guide, 4th ed. www.frtr.gov/matrix2/ Accessed 1 February 2005.
  • Gatto P, Carbognin L. 1981. The lagoon of Venice: Natural environmental trend and man-induced modification. Hydrol Sci Bull 26: 379391.
  • Giove S, Critto A, Agostini P, Semenzin E, Marcomini A. 2006. Decision support system for the management of contaminated sites: A multi-criteria approach. In: KikerG, LinkovI, editors. Environmental security and environmental management: Decision analytical and risk-based approaches. Amsterdam (NL): Springer-Verlag. (forthcoming)
  • Golden BL. 1989. The analytic hierarchy process: Applications and studies. New York (NY): Springer-Verlag. p 265.
  • Keeney R, Raiffa H. 1976. Decision with multiple objectives: Preferences and value trade-off. New York (NY): Wiley. p 589.
  • Kwiesielewicz M, Van Uden E. 2004. Inconsistent and contradictory judgements in pairwise comparison method in the AHP. Comput Oper Res 31: 713719.
  • Law decree 471/99. 15 December 1999. Remediation of contaminated sites. Gaz Ufficiale 293. (In Italian.)
  • Linkov I, Sahay S, Kiker G, Bridges T, Seager TP. 2004. Multi-criteria decision analysis: A framework for managing contaminated sites. In: LevnerE, LinkovI, ProthJM, editors. Strategic management of marine ecosystems. Amsterdam (NL): Kluwer. p 151167.
  • Los Alamos National Laboratory. 1996. A compendium of cost data for environmental remediation technologies. Los Alamos (NM): LANL. LA-UR-96–2205.
  • [NATO/CCMS] North Atlantic Treaty Organization/Committee on the Challenges of Modern Science. 2001. Evaluation of demonstrated and emerging technologies for the treatment of contaminated groundwater (phase III). Washington DC: USEPA. EPA-542-R-01–002. Final Report.
  • Norris GA, Marshall HE. 1995. Multiattribute decision analysis method for evaluation buildings and building systems. Gaithersburg (MD): National Institute of Standards and Technology, Building and Fire Research Laboratory. NISTIR 5663. Final Report.
  • [OCETA] Ontario Centre for Environmental Technology Advancement. 2001. About remediation: Technology directory case studies. www.aboutremediation.com Accessed 01 December 2004.
  • Ramanathan R. 2001. A note on the use of the analytic hierarchy process for environmental impact assessment. J Environ Manag 63: 2735.
  • Saaty TL. 1980. The analytic hierarchic process. New York (NY): McGraw Hill. p 287.
  • Saaty TL. 2000. Fundamentals of decision making and priority theory with the analytic hierarchy process. Pittsburg (PA): RWS. p 477.
  • Triantaphyllou E. 1999. Multi-criteria decision making methods: A comparative study. London (UK): Kluwer Academic.
  • [UKEA] UK Environmental Agency. 1999. Cost & benefits analysis for remediation of land contamination. West Midlands (UK): UKEA, National Library & Information Service, Midlands Region, P 316. R&D technical report.
  • [UN] United Nations. 1997. Compendium of soil cleanup technologies and remediation companies. New York (NY) and Geneva (CH): United Nations. ISBN-92–1–116677–2. p 98.
  • [USEPA] US Environmental Protection Agency. 2001a. FRTR cost and performance remediation case studies and related information. Washington DC: USEPA. EPA-542-C-01–003. Technical report.
  • [USEPA] US Environmental Protection Agency. 2001b. Remediation technologies cost compendium. Washington DC: EPA-542-R-01–009. Technical report.
  • [USEPA] US Environmental Protection Agency. 2004. Remediation and characterization innovation technologies (EPA REACH IT). www.epareachit.org Accessed 01 July 2004.
  • Venice City Council. 2001. Knowledge framework to write out the master plan for the rehabilitation of Porto Marghera contaminated site, according to the DPCM 12.02.99. Venice (IT): Council of Venice, Environmental Planning Department. (In Italian.) p 20.
  • Vik EA, Bardos P. 2003. Remediation of contaminated land technology implementation in Europe. A report from the Contaminated Land Rehabilitation Network for Environmental Technologies (CLARINET) by the Remediation Technologies Working Group. Vienna (AT): Federal Environment Agency.
  • Vincke P. 1992. Multicriteria decision-aid. Chichester (UK): Wiley & Sons. p 152.