application of spatial multi-criteria analysis to site selection for a …sd1229/690h/application of...

ARTICLE IN PRESS

0301-4797/$ - se

doi:10.1016/j.je

�CorrespondE-mail addr

[email protected] Regio

Journal of Environmental Management 88 (2008) 752–769

www.elsevier.com/locate/jenvman

Application of spatial multi-criteria analysis to site selection for a localpark: A case study in the Bergamo Province, Italy

Antonella Zuccaa,�, Ali M. Sharifib, Andrea G. Fabbria,c

aDISAT, Department of Environmental Sciences, University of Milano Bicocca, Piazza della Scienza 1, 20126 Milan, ItalybDepartment of Urban Regional Planning and Geo-Information Management, ITC-Hengelosestraat 99, P.O. Box 6, 7500 AA Enschede, The Netherlands

cSPINlab, IVM, Institute of Environmental Sciences, Vrije Universiteit, De Boelelaan 1087, 1081 HV Amsterdam, Netherlands

Received 1 September 2005; received in revised form 14 January 2007; accepted 9 April 2007

Available online 20 September 2007

Abstract

This contribution discusses a site selection process for establishing a local park. It was supported by a value-focused approach and

spatial multi-criteria evaluation techniques. A first set of spatial criteria was used to design a number of potential sites. Next, a new set of

spatial and non-spatial criteria was employed, including the social functions and the financial costs, together with the degree of suitability

for the park to evaluate the potential sites and to recommend the most acceptable one. The whole process was facilitated by a new

software tool that supports spatial multiple criteria evaluation, or SMCE. The application of this tool, combined with a continual

feedback by the public administration, has provided an effective methodology to solve complex decisional problem in land-use and urban

planning.

r 2007 Elsevier Ltd. All rights reserved.

Keywords: Decision making; GIS; Site selection; Spatial multi-criteria analysis

1. Introduction

This contribution discusses an application of spatialmultiple criteria decision analysis (SMCDA). An introduc-tion to this field can be found in Malczewski (1999), whocontributed to bridging the gap between geographicalinformation systems, GIS, and multi-criteria decisionanalysis, MCDA. SMCDA was applied here in supportof a real management problem of the Provincial Admin-istration of the City of Bergamo in northern Italy’sLombardy Region. Within a part of its administrativeterritory, The Park Office of the Province of Bergamoneeded to select the most suitable sites to build a local parkof extra-municipal character (Parco Locale di InteresseSovracomunale, or PLIS). This particular type of park,introduced in the Lombardy Region by a regional law(L.R. 30 novembre 1983, n. 86),1 often does not have the

e front matter r 2007 Elsevier Ltd. All rights reserved.

nvman.2007.04.026

ing author. Tel.: +393471570042.

esses: [email protected] (A. Zucca),

l (A.M. Sharifi), [email protected] (A.G. Fabbri).

nale (Regional law).

same naturalistic value as do other types of protected areas.It has a strategic role, however, in protecting andimproving the environmental quality of a territory.Usually the establishment of a PLIS is promoted by the

local community and put into practice by the ProvincialAuthority (L.R. 5 gennaio 2000, n. 1). In this particularcase, however, the Provincial Authority (represented by theProvincial Park Office) wished to promote the establish-ment of a new park and therefore needed to select the mostsuitable location and design so that the decision would bejustified by the municipalities involved.Formalized methodologies to select suitable areas to be

included in a park did not exist in that Province. Theselection was mainly based on expert’s knowledge andstrongly influenced by the existing City Master Plans.Owing to the fact that there was no formal method for thistype of site selection process, the Park Office would haveliked to benefit from a more analytical and transparentapproach. This situation made it feasible to develop theapplication discussed here.The process considered was a complex decision problem,

which involved multiple criteria and dealt with a large

www.elsevier.com/locate/jenvman

dx.doi.org/10.1016/j.jenvman.2007.04.026

mailto:[email protected]



ARTICLE IN PRESSA. Zucca et al. / Journal of Environmental Management 88 (2008) 752–769 753

number of environmental factors and socio-economicalconstraints. Such factors and constraints were mainly sitespecific; therefore the geographic attribute of variouslocations was to play a major role in site selection.Moreover, the process that led to the final decision hadto be as clear and transparent as possible to the municipalCouncils and to the relevant stakeholders.

This contribution briefly presents the procedure devel-oped and applied in support of the design and theevaluation phases of this type of park. The methodologyused benefited from multi criteria decision analysis andspatial multi-criteria evaluation (or SMCE) and made theprocess more rational and transparent. The process wasintended to be used by the Park Authorities for design andevaluation of proper sites in supporting decision makingand dialog with the Municipal Councils.

In the following section the study area and its relatedproblems are introduced first. Then the applied methodol-ogy is illustrated, followed by development and implemen-tation of a conceptual framework for the design andevaluation of alternative locations for a new park. Next,the potential sites are evaluated and the most suitable oneis selected. Finally a short discussion is made on themethodology and tools applied.

1.1. Study area

The Serio-Oglio study area, shown in Fig. 1, is located inthe Province of Bergamo in a plane between two rivers, theSerio River to the west and the Oglio River to the east. The

Fig. 1. The Serio-Oglio study area in the

size of the area is approximately 153 km2 and includes13 municipalities, with a total of 49,650 inhabitants.The main land use is agricultural and the area fallsbetween two parks of regional relevance, the Serio RiverRegional Park and the Oglio River Regional Park. A greatvariety of cultural, historic and architectural assets aredisseminated over the entire area. In the southern part it isstill possible to find many active fountainheads. These arewater resurgences springing in the transitional areabetween the higher part and the lower part of the PoRiver Plain, where the terrain porosity decreases and thegroundwater gets spontaneously to the surface creating wetzones with flora and fauna typical of marshlands. Thisnatural phenomenon has been largely utilized in the past tosupply irrigation canals and to grow fodder plants even inwinter, because the temperature of the water is persistentlybetween 101 and 151. The fountainheads are an importantcultural heritage to be preserved and even if they are notcompletely natural, they represent a unique ecologicalenvironment.Human activities, such as the construction of a

new motorway (the Bre.Be.Mi project of the Societa diProgetto Brebemi SpA, 2004) and an increasing urbangrowth, are threatening this area. Therefore, it has becomenecessary to preserve the environmental quality of theregion in general, and specifically to avoid inappropriatedevelopment of human activities as a consequence of theconstruction of new infrastructures. The establishment of aPLIS should represent an important step towards thisgeneral objective.

Province of Bergamo, northern Italy.

ARTICLE IN PRESS

Table 1

Sequence of activities performed in this study

Phase Activities

Intelligence a. Development of a conceptual framework including:

� identification of the main objective, the sub-objectives

and the contents of the local park;

� definition of the criteria structure: development of a

criteria tree that can be used to assess the satisfaction

of each sub-objective;

� Definition of constraints or characteristics of areas

unsuitable for the park.

b. Verification of the criteria tree through:

� analysis of the existing parks;

� verification through field visits and discussions with

park authorities.

Design c. Design of proper locations for the park by:

� performing a spatial multi-criteria evaluation using the

criteria structure and the set of constraints to produce

a suitability map;

� designing primary park elements (PPE) for the park,

using the suitability map together with the topographic

map, the aerial photographs and the expert’s

knowledge;

� designing alternative parks based on the PPEs and the

field work.

Choice d. Evaluation and ranking of the designed alternatives:

� definition of a new criteria structure for the evaluation

phase, including environmental, social and economic

criteria;

� performing spatial multi-criteria evaluation with

ILWISs SMCE using the new criteria tree;

� identification of the most appropriate locations for the

park.

The value-focused approach applied was guided by the framework for

planning and decision making developed by Sharifi et al. (2004b).

A. Zucca et al. / Journal of Environmental Management 88 (2008) 752–769754

2. Methodology

Decision making is a process involving a sequence ofactivities that starts with recognition of a decision problemand ends with recommendation for a decision. The qualityof the decision depends on the sequence and quality of theactivities that are carried out. Depending on the situation,there is a number of ways in which the sequence ofactivities can be organized. According to Keeney (1992),two major approaches can be distinguished: alternative-focused, and value-focused. The alternative-focused ap-proach starts with development of alternative options,proceeds with the specification of values and criteria andthen ends with evaluation and recommendation of anoption. The value-focused approach on the other hand,considers the values as the fundamental element in thedecision analysis. Therefore, it first focuses on thespecification of values (value structure), then consideringthe values, it develops feasible options to be evaluatedaccording to the predefined value and criteria structure.This implies that decision alternatives are to be generatedso that the values specified for a decision situation are bestachieved. In other words, the order of thinking is focusedon what is desired, rather than on the evaluation ofalternatives. In fact alternatives are considered as means toachieve the more fundamental values, rather than being anend to themselves. Naturally in decision problems, in whichalternative options have to be developed first and thenevaluated, the value-focused approach can be much moreeffective. However, if the decision problem starts with achoice of options, the alternative-focused approach seemsmore relevant.

In this study a value-focused approach was applied thatwas guided by the framework for planning and decisionmaking developed by Sharifi et al. (2002, 2004). Theapproach included the sequence of activities shown inTable 1.

The problem formulation phase (intelligence) has led tothe definition and specification of the park. The designphase has led to the identification of few potential sitesbased on the specified characteristics. This was obtained inthree steps: first, using a criteria structure, a suitability mapwas generated; then primary park elements (PPE) wereidentified using the suitability map, expert’s knowledge

together with other relevant ancillary information such astopographic maps, land-ownership and land-use maps,aerial photographs, and satellite images. Finally in aparticipatory process with the Park Officer of the Province,potential parks were derived through various combinationsof PPEs.

The choice phase has led to the final location for thepark. The costs and benefits of each site were analyzed andthe potential parks were compared and ranked on the basisof their overall attractiveness (utilities). In the implementa-tion phase, not yet completed to date, the boundaries of theselected site should be further adjusted based on groundsurveys and negotiation with the stakeholders involved.

In a spatial decision problem like the one consideredhere, options can be described by a defined set of mapsproviding information on each criterion. Therefore, thespatial decision problem can be visualized as a ‘‘table ofmaps’’, or ‘‘map of tables’’ (Sharifi et al., 2006), which hasto be transformed into one final ranking of alternatives.Thus it is necessary to aggregate not only thematically butalso spatially. The aggregation function can be simplifiedby distinguishing two operations: (i) aggregation in thespatial dimension, and (ii) aggregation in the thematicspace. These two operations can be performed eithersimultaneously or in successions. In the first case,represented as Path 3 in Fig. 2, all the information isprocessed by the decision maker alone and converted in aranking. The other two options shown in Fig. 2, are Path 1and Path 2. Following Path 1, the spatial information isfirstly aggregated to a non-spatial value for each themeseparately. Then, traditional techniques of multi-criteriaevaluation (MCE) can be used to derive the final utility foreach alternative. Following Path 2 instead, first the theme

ARTICLE IN PRESS

gc1

gc2

gc3

gc4

c1

a1

a1

a2

a2

a3

a3

a1

a2

a3

a1

a2

a3

c2

c3

c4

c1

c2

c3

c4

g

f f

Path 1

Path 3

Path 2

Fig. 2. The decision paths in spatial decision problems described as ‘‘table

of maps’’ or ‘‘map of tables’’ in van Herwijnen (1999).

A. Zucca et al. / Journal of Environmental Management 88 (2008) 752–769 755

maps are combined through MCA techniques to obtain asuitability map for each alternative, thus reflecting theperformance of the alternative across the space. Theneach map is aggregated to a single non-spatial value.Advantages and disadvantages of the two paths aredescribed in detail by van Herwijnen (1999) and Sharifiet al. (2006). In general, the final result might be the sameor it might be different, depending on the aggregationmethods applied. Nevertheless, following Path 2 it ispossible to perform a multi-criteria evaluation usingspatial criteria along with non-spatial criteria, and that isdone without losing the spatial dimension. In this study wehave preferred to follow Path 2. This was in order topreserve as long as possible throughout the decisionprocess the spatial information given by the spatialdistribution of the values.

3. Problem formulation

3.1. Development of a conceptual framework

To define the criteria structure for the design of suitablesites in the study area and establish a PLIS, it was firstnecessary to identify the main objectives of the park.Consideration was given to the laws issued (L.R. 30novembre 1983, n. 86, L.R. 5 gennaio 2000, n. 1), theliterature (Di Fidio et al., 2001; Mauri, 2000; Provincia diMilano, 2002) and the previously established PLIS. Theprotection and improvement of the environmental qualityof the territory appeared to be the main goal of the park.The goal could be met through the following objectives: (i)protection of valuable areas and (ii) restoration ofdegraded situations.

(i)
The degree of satisfaction of the first objective,‘‘protection of valuable areas’’, was measured throughthe following criteria:� Protection of agricultural areas: from a sustainable-
development point of view, it was not enough toprotect single assets. It was also necessary to reduce

pressures on the environment related to the humanactivities, and to enhance the regeneration capabilityof natural resources, favouring more sustainableforms of utilization. Agricultural areas could play animportant role in this. Moreover, they could have aconnective function between natural areas.� Protection and enhancement of rural, historical and

architectural assets: in a territory characterized by astrong human influence, the presence of someheritage elements could contribute to the environ-mental quality.� Protection of assets of high natural value: to protect

biodiversity it was important to preserve the existingnatural values, e.g., forests, geologic and geomor-phologic assets, by including them in the park. Itwas also important to preserve the ecologicalcorridors, and in some cases the park as a wholecould function as a connection belt between valuablenatural areas of higher order.

(ii)
Satisfaction of the second objective, ‘‘restoration ofdegraded situations’’, was measured through thefollowing criteria:� Restoration of degraded urban areas: the establish-
ment of a PLIS could offer a good opportunity torestore degraded situations that are resulting fromsome particular human activities, such as quarryareas or former industrial areas.� Mitigation of and environmental compensation for

impacts of new infrastructures: the park could be agood opportunity to mitigate or to compensate for anew unavoidable human impact, such as a newmotorway.

The formulation of objectives and their related criterialed to the identification of the characteristics and, as aresult, to the definition of the park. The criteria structurecould then be used for assessing the suitability of each areafor the park. Objectives and criteria were organized in thestructure termed criteria tree, shown in Fig. 3.In the same way, it was also necessary to define the

constraints that vetoed the suitability of an area for a park.These were:

�
According with the guidelines of the Regional Govern-ment, a local park could only be built in areas with aparticular classification (‘‘E’’ or ‘‘F’’) in the City’sMaster Plan. Corresponding to this classification weremainly green areas, agricultural areas, woodlands, waterbodies, and water meadows. � According to the same guidelines, the local park could
not be built inside an existing regional park. Areas beingpart of the two regional parks (the Serio River Park andthe Oglio River Park) had to be excluded.
� The park was not to include waste-disposal sites nor be
in their immediate vicinity.
� The park was not to include active quarries nor be in
their immediate vicinity.

ARTICLE IN PRESS

Protection and improvement of the

environmental qualityof the territory

Protection of high natural value areas and

assets

Protection and valorisation of rural,

historic and architectural assets

Protection and valorisation of historic-architectural assets

Recovery of degraded situations

Protection of agriculturalareas

Protection of valuable areas

Protection of single natural assets

Creation of ecological corridors

Restoration of degradedurban areas

Mitigation and environmental

compensation of impacts of new infrastructures

Protection and restoring of typical landscape elements

Protection of natural areas

Preserving rivers and water basins

Fig. 3. The criteria tree constructed for assessing the suitability of each area considered for the local park in the Serio-Oglio study area.


�
The park was not to include contaminated sites nor be intheir immediate vicinity.
3.2. Verification of the criteria tree

In order to test the functionality of the criteria treeconstructed, 32 previously established PLIS were analysedto verify whether the framework developed is capable ofdescribing them. According to the description of theexisting local parks available in Di Fidio et al. (2001) andafter some field visits, this was measured through thecounting of the number of objectives fulfilled by each park.The result is summarizsed in Fig. 4.

All the parks satisfied one criterion or more criteriarelated to the protection of valuable areas. In particular, 29parks (90%) aimed to the protection of natural areas, 20parks (60%) included rural, historical and architecturalassets, and 19 parks (about 60%) intended to preservevaluable agricultural areas.

Moreover, 20 parks satisfied one or more criteria relatedto the recovery of degraded areas. In particular, 19 parksincluded former degraded urban areas and 12 parks (40%)provided a mitigation/compensation function.

Looking at the above information it can be seen that,although no formal method has been applied in theselection of proper sites for the 29 existing parks, implicitlymost of the criteria suggested were considered. In this sensethe analysis confirmed the validity of the conceptualframework developed. Further verification of the frame-work has been obtained by consulting directly with the

technical personnel working in the Park Office of theProvince of Bergamo.

4. Design of potential locations for the park

The design of potential parks was carried out throughthe following 2 steps:

�
development of a suitability map; � design of alternative locations for the park.
4.1. Development of suitability map

To develop the suitability map, the theoretical frame-work described in Sections 2 and 3, was directly convertedinto the branches and leaves of a criteria tree. For this theinteraction and processing capabilities were exploited of arelatively new version of the software package ILWISs 3.3(ILWIS 3.3, 2005), the SMCE module for spatial multi-criteria evaluation. The ILWIS SMCE criteria treeconstructed is shown in Fig. 5. It included the conceptsthat follow:

�
Constraints: A constraint is a hard criterion thatdetermines which areas should be excluded from orincluded in the suitability analysis. The excluded areaswill get a nil (0) performance value in the compositeindex map, while the remaining areas will obtain a valuebetween 0 and 1.

ARTICLE IN PRESS

0 5 10 15 20 25 30

Protection and valorization of rural,

historical, architectural assets

Protection of valuable agricultural

areas

Protection of areas and single

elements with high natural value

Total: Protection of valuable areas

Restoration of degraded urban areas

Mitigation and environmental

compensation

Total: Recovery of degraded areas

Ob

jec

tiv

es

Number of parks

Fig. 4. Histogram synthesizing the result of comparing 32 established PLIS parks according to 7 groups of objectives fulfilled by each.

Fig. 5. The criteria tree used in the analysis of the Serio-Oglio study area. On the left are constraints, factors and associated weights with descriptors. On

the right the corresponding file names of the digital maps spatially representing constraints and factors. The interaction structure is from the ILWISs

SMCE module (ILWIS 3.3, 2005), that introduces multi-criteria evaluation in a GIS environment. See text for explanation.


ARTICLE IN PRESS

Table 2

Maps and standardization methods used to represent the constraints

Constraint Source data Derived map Standardization

E and F areasa City Master Plans – E and F areas are standardized to 1, all the others to 0

Existing parks Map of protected areas – Park areas are standardized to 0, all the others to 1

Waste disposal Map of waste disposals Distance mapb Distance p200 is standardized to 0; all other values are standardized to 1

Active quarries Map of active quarries Distance map Distance p200 is standardized to 0; all other values are standardized to 1

Contaminated sites Map of contaminated sites Distance map Distance p200 is standardized to 0; all other values are standardized to 1

aE and F denominations in the City’s Master Plan correspond to green areas, agricultural areas, woodlands, water bodies and water meadows.bA distance map is a derived raster map where each pixel has a value correspondent to the distance from source pixels in the original map.


�

2

to t

to

Factors: A factor is a soft criterion that contributes to acertain degree to the output (suitability). There are twotypes of factors, (i) benefit criteria and (ii) cost criteria.A benefit criterion contributes positively to the output(the higher are the values, the better it is), while a costcriterion contributes negatively to the output (the lowerare the values, the better it is). As opposed toconstraints, which cannot be compensated, poor per-formance of a factor can be compensated by goodperformance of another factor. Using compensatorydecision rules, such as a weighted sum,2 this can still leadto good overall performance in the composite indexmap.
� Group of factors: A group of factors defines an
intermediate or a partial goal, given by a combinationof factors.

The constraints applied in this study were as follows (seealso Section 3.1): (1) E or F area, (2) no waste disposal, (3)no active quarries, (4) no contaminated sites, and (5) noexisting park. Table 2 lists these constraints, the availablesource maps used to represent them, the derived maps fromthe source maps, and the standardization methods appliedto the derived maps or to the source maps. It is importantto stress that the goal achievement of the local parkestablishment was assessed by several criteria referring totwo main objectives: the protection of valuable areas andthe restoration of degraded areas. Table 3 presents thehierarchical structure of objectives and criteria and thefactors which indicate the achievement of each criterion.For each factor, Table 3 shows the related criterion, theavailable source maps used to represent the factor, thederived maps, and the standardization methods applied toderived or source maps. All factors represented wereconsidered as costs, except for the presence of agriculturalareas, which was assumed to be a benefit.

To each factor or group of factors a weight was assigned,expressing its relative relevance in determining the finalscores representing the overall attractiveness or utility of apoint for belonging to the park. Three different methods of

Using a weighted sum algorithm, the normalized value corresponding

he performance of each factor is first multiplied by the weight assigned

that factor; then all the weighted values are summed.

weight assignment were available in the ILWISs SMCEmodule: direct method, pair-wise comparison, and rank

ordering. For this analysis, the Rank Order method hasbeen applied due to its relevance and simplicity ofapplication from the perspective of the decision makers.This method simply consisted of arranging the criteria inorder of importance and then translating that order into aquantitative ranking using different techniques.In this case, the cooperating Park Officer, supported by a

group of experts, was first requested to rank the criteria inorder of importance; thus, different sets of weights werecalculated using the available algorithms: Expected Value(Rietveld, 1980) and Rank Sum (Janssen and VanHerwijnen, 1994). Finally, between the two resulting sets,the Park Officer selected the set calculated on the basis ofthe Expected Value Method. The result of this process isshown in Table 4. A higher weight was attributed to theprotection of the valuable areas with respect to therecovery of degraded situations. Further, to all branches

and leaves, equal weights were assigned as no furtherinformation was available to differentiate them.Once all the maps were converted to partial suitability,

and standardized to the same value range, their corre-sponding relative importance weights were assigned. Thepartial suitability maps were then combined applying theweighted sum method to derive the overall suitability index.This was facilitated by combining the various attributes ofeach pixel to derive the suitability index map as shown inFig. 6.

4.2. Design of alternative sites for the park

The suitability map was then used, together with thetopographic base map, aerial photographs, and otherrelated information, to manually design PPEs. In thisprocess, considering the physical (natural and artificial)boundaries, the areas with the highest suitability wereincluded in a PPE; as specifically dictated by law (L.R. 5gennaio 2000, n. 1), existing physical (natural and artificial)boundaries were considered in delimiting the PPE. After-wards, in a participatory process with the Park Officer ofthe Province, the 11 PPEs were combined into fouralternative options for the park. The boundaries of the

ARTICLE IN PRESS

Table 3

Criteria and related factors

Criterion Group Factor Source data Derived

map

Standardization

Protection of agricultural

areas

– Presence of agricultural

areas

Landscape mapa – Attribute tableb

Protection of rural-historic-

architectural assets

– Distance from typical

landscape elements

Map of natural and

landscape assets

Distance

map

Linear standardization (the

lower the distance, the higher

the score)

Protection of

historic-

architectural

assets

Distance from historic-

architectural assets

protected by lawc

Map of assets of

historic-artistic

interest

Distance

map



the score)

Distance from other

historic-architectural

assets

Map of assets of

historic-artistic

interest

Distance

map



the score)

Distance from urban

centres

Map of historic urban

centres

Distance

map



the score)

Distance from

archaeological areas

Map of archaeological

areas

Distance

map



the score)

Protection of highly natural

areas and assets

– Distance from natural

assets (points)

Map of natural and

landscape assets

Distance

map



the score)


assets (lines)

Map of natural and

landscape assets

Distance

map



the score)


assets (areas)

Map of natural and

landscape assets

Distance

map



the score)

– Distance from features

with ecological

connection function

Map of features with

ecological connection

function

Distance

map



the score)

– Distance from river and

water basins

Land use map Distance

map



the score)

Restoration of degraded

urban areas

– Distance from degraded

areas

Map of degraded

areas

Distance

map



the score)

Mitigation/compensation

function with respect to new

infrastructures

– Distance from new

infrastructures

Map of new roads Distance

map



the score)

Digital maps and standardization methods used to represent the different factors.aThe landscape map was derived from the land use map of the Provincial Administration of Bergamo.bDifferent scores were assigned to agricultural areas with a high landscape value (score ¼ 1), to other agricultural areas (score ¼ 0.5), and to other

classes (score ¼ 0).cD. Lgs. 29 ottobre 1999 n. 490.


alternatives designed are presented in Figs. 7 and 8. Theycan be briefly described as follows:

Alternative 1 is located close to the Serio River RegionalPark, as a protection buffer. It extends between threeinhabited centres, limiting the urban sprawl and the risk ofconurbation. The southern part of the park couldalso represent a mitigation band for the new highwayand the new railway that will cross the area from the Westto the East.

Alternative 2 is located in the southern part of the studyarea, characterized by extensive agricultural activities and

small scattered settlements; in this area there still are a lotof active fountainheads.

Alternative 3 is located in the southern-central partof the study area. It represents a mainly agricultural area,with scattered small settlements and some active fountain-heads. This is a bit smaller than the other alternatives butincludes five municipalities. It is positioned in a strategiclocation.

Alternative 4 is located in the eastern part of thestudy area, very close to the Oglio River RegionalPark, for which it could represent a protection buffer. It

ARTICLE IN PRESS

Table 4

Weights assigned to each factor and group in the design phase

Fig. 6. The composite suitability index map of the Serio-Oglio study area,

obtained combining the attributes of each pixel using SMCE. The higher

are the values of overall performance, the higher is the suitability.


is mainly an agricultural area, with some activefountainheads and many rural and historic-architecturalassets.

5. Evaluation and ranking of alternative sites

5.1. Definition of a new criteria structure

The last phase of decision making is the evaluation andchoice of alternative options. The four alternative sites forthe park were evaluated using a different criteria structure.Besides the suitability index, resulting from the first part ofthe analysis, environmental-ecological, social and econom-ic criteria were used for this part of the study. Differentlyfrom the indicators used to build the suitability map, theindicators selected for the evaluation phase are dependingon the boundaries of the park. Fig. 9 shows the conceptualframework used for the evaluation. For this last phase thefollowing aspects have been considered: (i) degree ofsuitability, (ii) environmental-ecological effects, (iii) socialeffects, and (iv) economic effects.

5.1.1. Degree of suitability

The area selected for the park has to have a highsuitability Index. The higher is the value of this Index, thehigher is the performance of the potential park.

5.1.2. Environmental-ecological effects

Considering the environmental-ecological function ofthe park, the following criteria were used for the evaluationof the alternatives: (a) the fragmentation due to thepresence of roads inside the park, (b) the capability ofthe park to limit the urban sprawl, or the risk ofconurbation between two or more municipalities, (c) therole of connection between the existing regional parks, and(d) the protection role with respect to the existing regionalparks.To calculate the fragmentation of each potential park an

Infrastructural Fragmentation Index (IFI) was used,developed by the University of l’Aquila, Department ofArchitecture and Urban Planning (Di Ludovico and

ARTICLE IN PRESS

Fig. 7. The boundaries of the four alternative options of the local park

overlaid on the suitability index map of the Serio-Oglio study area. See

text and Fig. 8 for explanation.


Romano, 2000; Romano and Di Ludovico, 2000). Thisindex appeared effective in describing the fragmentation ofthe study area due to the presence of infrastructures even ifit is computationally simple and does not require a specificdata set. The mathematical expression of the index is asfollows:

IFI ¼ ½SiðLioiÞ�½N=A�P, (1)

where, Li is the length of the infrastructure i in theterritorial unit; oi the weight assigned to each category ofinfrastructure (o ¼ 1 for highway and railways; o ¼ 0.5 fornational, regional, and provincial roads; 0 ¼ 0.3 for localand municipal roads); N the number of patches in whichthe territorial unit is divided; A the area of the territorialunit; and P the perimeter of the territorial unit.

To estimate the park role on limiting the urban sprawl, apair-wise comparison technique (Saaty, 1980) was used.Looking at their spatial location, pairs of alternatives werecompared with respect to their capability to limit the urbangrowth. Next, verbal comparisons were converted to anumerical Saaty scale, as shown in Table 5. The nume-rical values were then transformed into final scores bymultiplying the elements in each row and taking the Nth

root of the product, where N is the number of rowelements.Concerning the functions of connection and protection

with respect to the existing regional parks, two distancemaps have been used with a different standardization. Theresult is shown in Fig. 10. In practice, the standardizedpixel scores decrease between 0 and 1500m of distancefrom the existing regional parks and they are nil between1500 and 2500m (left diagram), or they increase againbetween 2500 and 4000m of distance (right diagram). Theaverage distance is in fact 4000m.

5.1.3. Social effects

The social function of the park was assessed by two typesof indicators: the educational and the recreational indica-tor. For the educational indicator, the same factors wereused as considered in constructing the suitability map, suchas the presence of cultural values, of historic-architecturalvalues, of archaeological areas and of historic urbancentres. For the recreational aspect, the presence offacilities such as cycling paths or agri-tourism activities,and other factors such as the dimension of the park and thepopulation served were considered for the evaluation.

5.1.4. Economic effects

Many authors dealt with the problem, referred to theItalian situation, of the economic value of the protectedareas (Diviacco, 1994), of the economic impacts on thelocal community of a park establishment (Fava, 1994;Bernetti and Marone, 2000) and of the development ofeconomic activities in the national parks (Bernetti, 2000).However, all those authors refer to National or Regionalparks that are subjected to specific and detailed regula-tions. In the case of PLIS, owing to their strictly localcharacter, general regulations and prescriptions did notexist, a fact that has resulted in the difficulty in estimatingthe economic effects of the park. Many problems wereencountered in defining precisely the economic gains andthe losses that resulted from the establishment of the park.A set of generic criteria have been laid down, after someinterviews with a number of technical personnel working inthe existing parks. For costs, the financial costs wereconsidered—split in fixed costs (personnel and officemaintenance costs) and variable costs related to theactivities carried on in the park—and the losses causedby the regulation imposed on the land use after theestablishment of the park. For benefits, the possiblepositive consequences were considered of the park estab-lishment on the neighbouring areas on the areas inside thepark, and on the value of the agricultural and handicraftproducts coming from the area inside the park.The largest component of the financial costs of the park

is related to the maintenance costs, such as reforestation,restoration of degraded habitats, provision of pathwaysand educational activities, and other actions to promotethe park. These costs can vary considerably from park topark, with respect to the state of the environment or of the

ARTICLE IN PRESS

Fig. 8. The boundaries of the four alternative options identified in Fig. 7 overlaid on a false colour composite representation of an aster image (June 2003).

Red, light blue and grey represent all cultivated fields, with different cultures and irrigation practices. It is also possible to recognize towns, main roads,

and, in dark blue, water bodies.


single natural or historic-architectural assets inside thepark. This has been simplified by considering the linearrelationships between these costs and the size of the park.Moreover, the financial contribution of each municipalityfor the establishment of park maintenance was estimatedusing a formula applied in practice by the park adminis-trators of the existing local parks.3 Actually, eachmunicipality has to pay for the park proportionally tothe portion of park in its own territory and to the number

3Contribution of Municipality X ¼ 20*Ix%+50*Cx%+30*Px, where

Ix is equal to the inhabitants of the municipality X divided by the total of

inhabitants; Cx is equal to the area of municipal territory X in the park

divided by the total area of the park; Px is equal to the portion of the

municipal territory X in the park divided by the sum of the portions of all

municipal territories in the park.

of inhabitants. Then the standard deviation of theseestimated contributions was considered as indicator forthe equitable distribution of the financial costs of the park.The costs of the park could be more bearable andacceptable, if more municipalities were involved in themanagement and the balanced distribution of the costsbetween the municipalities involved. Thus, if the standarddeviation is high, the cost is not equally distributed and thepark is less sustainable.To assess the gains and losses deriving from the

hypothetical variations in the land value and in theeconomic activities inside the park, the pair-wise compar-ison method was used. In accordance with the City’sMaster Plan, the land-use destinations that could beaffected by some positive or negative changes inside the

ARTICLE IN PRESS

Environmental

and ecological

effects

Social effects

Limitation of urban growth

Connection of existing regional parks

Creation of a protection buffer for existing regional parks

Education/sightseeing

Recreation

Economic effects

Financial costs

Low Fragmentation

Decrease of land value

Benefits lossIncome loss for restrictions in agricultural and breeding

activities

Gains

Losses

Increase of land value

Increase of value of products

Fixed costs

Costs for park activities

Presence of historical-architectural values

Distance from archaeological areas

Population served

Dimension

Presence of cycle paths

Distance from agriturism activities

Distance from historic urban centers

Presence of cultural values

Suitability

EV

AL

UA

TIO

N C

RIT

ER

IA

Fig. 9. The criteria tree structure used for the evaluation and choice phase of alternative options in the Serio-Oglio study area.

Table 5

Pair-wise comparison matrix, expressing the capability of each alternative

in limiting the urban sprawl

Alt1 Alt3 Alt2 Alt4 Score

Alt1 1.00 5.00 7.00 9.00 4.21

Alt3 0.20 1.00 6.00 7.00 1.70

Alt2 0.14 0.17 1.00 6.00 0.61

Alt4 0.11 0.14 0.17 1.00 0.23


park were first selected. Then, those land uses werecompared according with the relevance of such changes.After the park establishment, the commercial value of landwill likely decrease for certain type of land-use destination,and the relative loss of value will be different, for instance,for areas that are classified as ‘‘residential’’ and for areasthat are classified as ‘‘industrial.’’ On the other hand, theland value will increase inside the park for other types ofland uses, such as ‘‘agricultural’’ areas. In addition, theeconomic activities related to a given land use will beaffected by the park establishment. In some cases the effectwill be positive, for instance for handicraft or specializedagriculture activities, in other cases it will be negative, asfor manufacturing activities.

According with these assumptions and after collectingthe experience of technicians working in existing localparks through a number of interviews, four decisionmatrices were obtained, related to: (i) decreasing landvalue inside the park; (ii) increasing land value insidethe park; (iii) increasing product value inside the park;(iv) generating income loss due to restrictions. Table 6shows the normalized scores that were used to prepare thefour attribute tables to be linked to the land-use destinationmap and perform the multi-criteria evaluation.For the land-value variation outside the park, just the

residential areas were considered. The total residential areawithin a 1 km buffer around the park was used as anindicator. The values of this indicator and of the other non-spatial indicators are listed in Table 7.It is worth specifying that the non-spatial indicators

are in practice considered in the spatial multi-criteriaevaluation as maps having a unique pixel value all overthe map.

5.2. Weight assignment

As previously discussed for the suitability map, theweights were assigned, using the rank order method, by the

ARTICLE IN PRESS

Table 6

Normalized scores assigned to the land use destinations inside the park to estimate gains and losses deriving from hypothetical variations in the land value

and in the economic activities carried out inside the park

Decreasing land value

inside the park

Increasing land value

inside the park

Increasing product value

inside the park

Income loss for

restrictions

Industrial-consolidated 0.33 – – 0.56

Handicraft-consolidated 0.18 – 0.13 0.19

Handicraft-expansion 0.16 – 0.05 0.19

Residential-expansion 0.11 – – –

Residential-

transformation

0.10 0.04 – –

Residential-consolidated 0.06 0.04 – –

Residential-restoring 0.03 0.17 – –

Agricultural-consolidated 0.02 0.29 0.25 0.04

Specialized cultures-

consolidated

0.01 0.46 0.57 0.02

0 1000 2000 3000 4000 5000 6000

distance (m)

0

0.8

0.6

0.4

0.2

1

valu

e

0 1000 2000 3000 4000 5000 6000

distance (m)

0

0.8

0.6

0.4

0.2

1

valu

e

Fig. 10. Standardization functions of two distance maps expressing the protection (left) and the connection (right) utility of potential PLIS for the existing

regional parks.

Table 7

Values of non-spatial indicators

Effects Indicators Alt 1 Alt 2 Alt 3 Alt 4

Environmental-ecological Infrastructure Fragmentation Index (IFI) 429.46 15.91 41.55 13.66

Limitation of urban sprawl 4.21 0.61 1.70 0.23

Social Number of cultural values inside the park 30 24 29 22

Number of historic-architectural values inside the park 4 1 1 2

Total length of cycle paths (m) 15 609.10 0.00 0.00 0.00

Social/economic Size (km2) 11.64 12.68 8.59 10.05

Economic Number of municipalities involved 6 3 5 3

Number of inhabitants served 23 560 6 603 13 346 7 350

Standard deviation of municipalities’ contributions for the park (%) 12.99 26.27 12.21 3.82

Residential areas in 1 km buffer around the park (m2) 22 688.00 49 976.00 15 745.00 5752.00


Park Officer supported by a group of experts. The result isshown in Table 8.

Equal priority was given to the four main branches ofthe criteria tree: the suitability value, the environmental-ecological effects, the social effects, and the economic

effects related to the park establishment. Regarding theenvironmental-ecological effects, a higher importance wasattached to the role played by the potential park in limitingthe urban sprawl and in connecting the existing regionalparks, rather than to the low fragmentation and to the

ARTICLE IN PRESS

Table 8

Weights assigned to each factor and group of factors in the evaluation phase


protection function of the park. These preferences reflectthe main concerns of the Park Office about the con-sequences on the study area of the increasing humanpressures. As concerns the social effects, a higher pre-ference was attributed to the presence of cultural andhistoric-architectural values, to the amount of populationserved and to the size of the park. In fact to the Park Officethese criteria appeared more relevant than the other ones inselecting the best alternative for the park. As to theeconomic effects, the assigned weights reflect both theprovincial Park Office view and the impressions collectedby talking with the technical personnel working in theexisting local parks. The financial costs of the park, and inparticular the costs related to the park activities, seemed tobe the higher economic charge in managing a local park,while the loss of benefits is considered less important.The gains deriving from the park establishment might berelated mainly to the increasing land value inside thepark, secondarily to the higher value gained by theproducts of the economic activities carried out insidethe park and just in minimal part to the increasing landvalue around the park.

5.3. Evaluation results

Once all criteria maps were in a form representing thedegree of suitability of each picture element or pixel of thepark from different aspects and perspectives, the evalua-tion of alternatives could proceed. As shown in Fig. 11, forevery alternative the result was a map (utility map) in which

the value of each pixel represented the overall suitability ofthat pixel for the park. To compare the alternatives it wasnecessary to aggregate the pixel values. One of the simplestand often-used non-spatial aggregation methods is takingthe average or the sum of pixel values. Combined withother descriptive statistical parameters, such as minimumand maximum values, the average and the sum were usedto describe the overall performance of each alternative. Theresults are shown in Table 9. As it is presented in the table,alternative 1 appears to be the most attractive site for thepark. In fact, this park alternative outranks the others in allparameters.In order to evaluate the consistency of the results, a

simple sensitivity analysis has been performed. Fortechnical and administrative reasons, the discussion withmain stakeholders (citizens, environmentalists, municipalCouncils, farmers, etc.) was not possible at the time thispart of the research was carried out. Nevertheless theconsultation activities were supposed to take placeimmediately after the conclusion of the study.Carrying out a complete sensitivity analysis is quite a

complex and difficult process to implement in the spatialMCE, especially with respect to error margins in eachscore, which corresponds basically to a pixel in the map.The available software module ILWIS-SMCE did notsupport such a type of analysis. Moreover a completesensitivity analysis was not considered essential, given thatthe result of the first MCE, which was used to design theprimary park elements, was supported by the MCE andwas carried out manually. In the second round of SMCE

ARTICLE IN PRESS

Fig. 11. Comparison of the four alternative choices of the Serio-Oglio study area. The colours indicate utility values. The overall performance of each

alternative was obtained as a composite index that is expressed in Table 9 where Alternative 1 appears the most attractive.

Table 9

Results of the evaluation phase

Area (km2) Perimeter (km) Number of pixel Average Sum Maximum value Minimum value

Alt1 11.64 37.49 29 080 0.68 17 036.09 0.78 0.50

Alt2 12.68 31.84 31 686 0.62 15 836.43 0.71 0.44

Alt3 8.59 19.94 21 504 0.58 10 457.41 0.73 0.39

Alt4 10.05 25.01 25 112 0.54 11 379.98 0.64 0.37


for choosing the more acceptable park, sensitivity analysiswas made only to explore the robustness of the finalrecommendation with respect to the weights. Again, as this

was not feasible to implement through a sort of MonteCarlo type analysis, it was decided to introduce insteadthree different sets of weights to represent three different


view points on the main criteria and to analyse the resultsof such introductions.

The results are shown in Fig. 12. The set of weights ‘‘A’’corresponds to the view of the Park Officer, illustrated

0 5000 10000 15000 20000

Set A

Set B

Set C

Set D

Se

t o

f W

eig

hts

Set

of

Weig

hts

Set

of

Weig

hts

Sum

alt1 alt2 alt3 alt4

0 0.2 0.4 0.6 0.8 1

0 0.2 0.4 0.6 0.8 1

Set A

Set B

Set C

Set D

Average

Set A

Set B

Set C

Set D

Minimum Value

Set

of

Weig

hts

0 0.2 0.4 0.6 0.8 1

Set A

Set B

Set C

Set D

Maximum Value

above. In a first alternative view, set of weights ‘‘B’’, adifferent ranking of the criteria was made: a higherimportance was given to the economical aspects and tothe suitability, while social and economical effects wereconsidered less important. Such a position would likely beundertaken by the Municipal Councils, whose usual mainconcerns were the effective costs of the park establishment.In a second potential view, set of weights ‘‘C’’, the

environmental-ecological aspects and the suitability wereassumed to have the higher importance, while social andeconomical aspects were considered less important. Thiswould likely happen to be the position of environmentalassociations and of part of the citizens.In a third hypothetical view, set of weights ‘‘D’’, which

might correspond to the farmers and to part of the citizens,higher importance was given to social and economicalaspects, while a lower weight was assigned to the suitabilityand to environmental-ecological aspects.Evaluation of the parks considering the above extreme

set of weights shows that alternative 1 is consistently themost favourable option.

6. Discussion and conclusion

The methodology that was developed and applied in thisstudy combined a value-focused approach with spatialmulti-criteria evaluation techniques in supporting a landmanagement problem: the selection of the most suitablesite for the establishment of a new local park.In the last decades a variety of applications of multi-

criteria evaluation and spatial analysis to site selection havebeen made. Many of them concerned the search of the mostsuitable location where to build things, such as wastedisposal sites (Sharifi and Retsios, 2004; Keeney and vonWinterfeld, 1994). Some applications regarded habitatsuitability modelling (Store and Kangas, 2001) and otherenvironmental management issues (Giupponi et al., 1999;Phua and Minora, 2004). Few applications concerned apark planning problem. Some instances mainly concernedthe establishment (Keisler and Sundell, 1997) or theboundary relocation (Sharifi et al., 2002) of NationalParks. Moreover, spatial multi-criteria evaluation has beenused either for design purposes or for the evaluation ofdifferent alternatives.In this study a framework is provided to effectively

design and evaluate alternative sites for a specific type of

Fig. 12. Results of the evaluation when assigning different sets of weights

to the main branches of the criteria tree: suitability, environmental-

ecological effects, social effects, and economical effects. The set of weights

‘‘A’’ corresponds to the view of the Park Officer. In a first alternative view,

set of weights ‘‘B’’, higher importance was given to the economical aspects

and to the suitability; in a second potential view, set of weights ‘‘C’’, the

environmental-ecological aspects and the suitability were assumed to have

higher importance; in a third hypothetical view, set of weights ‘‘D’’, higher

importance was given to social and economical aspects.

ARTICLE IN PRESSA. Zucca et al. / Journal of Environmental Management 88 (2008) 752–769768

parks, with local relevance. For this kind of park a specificmethodology or a set of specific criteria to select the areasto include in the park did not exist in the Province ofBergamo, where the study area is located, as well as inItaly. The selection was mainly based on experts’ knowl-

edge and strongly influenced by the existing City MasterPlans. This means that there is always the risk that a localpark would mainly include residual areas that are notsuitable for other uses. The methodology applied allowsenhancing the role of this type of park as an opportunityfor a more sustainable land planning.

In fact, following the value-focused approach used inthis study, first the values were specified, then consideringthe values, feasible options were developed and evaluated.

According to the need of the Decision Maker to upholdthe choice with population and Municipal Councils, apowerful model was provided to facilitate a rationaldecision-making process in a transparent and participatorymanner. In this process the role of geo-information wasenhanced so that large amounts of detailed information ofvarious types and formats were used in a planning anddecision-making process. Following sustainable develop-ment concept, besides ‘‘physical’’ criteria of suitability,several environmental-ecological, social and economicalcriteria were considered in determining the ‘‘best’’ or mostacceptable location and design for the park.

The processing was supported by the ILWISs SMCEmodule, a GIS software that demonstrated to be aneffective tool for managing and combining a large amountof spatial and non spatial information. Further itfacilitated processing ‘‘tables of maps’’, as well as ‘‘mapof tables’’ (Sharifi et al., 2006). In this study the spatialdimension has been maintained until the end of theevaluation process. Therefore the spatial distribution ofthe values could be obtained along with the ranking of thealternatives. Moreover, fast and efficient processing oflarge data sets allowed completing several analyses in arelatively short time. To test the robustness and sensitivityof the final result with respect to the potential errors in theanalysis, three different weight-sets representing threedifferent viewpoints on the main criteria were selectedand applied in a multiple criteria evaluation process. Amore exhaustive sensitivity analysis could have beenattempted, however, that process was found too complexto implement in the context of spatial multiple criteriaevaluation for the Serio-Oglio study area. This aspect willbe taken under consideration in a further development ofthis study. Nevertheless the role of the sensitivity analysisin this case was considered less useful than in othercircumstances, owing to the fact that there was a narrowuncertainty on the decision maker’s preferences and arestricted room for the participation of stakeholders. Thislack of participation, mainly due to the social and politicalsituation at the time of the study, represented a short-coming, which should be avoided in future applications.

The main advantages of the methodology used in thisstudy are the efficient structuring of geo-information, and

the processing large data sets to support managementactivities. This process proposed permits the efficientcombination of multi-criteria evaluation with spatial dataanalysis tools that support a sustainable land managementand provide a logical and scientific foundation into whichthe values of decision makers and stakeholders can beintegrated.

Acknowledgements

This study has been carried out during an internshipperiod at ITC, in The Netherlands, within the frameworkof a Ph.D. degree program at the University of Milano-Bicocca in Italy. The first author is grateful to ITC for theacademic support and to professors Angelo Cavallin andAndrea Zanchi, of the University of Milano-Bicocca, forthe financial support. Moreover, a special thank goes toDr. Karin Pfeffer (ITC) for her kind and precioussuggestions and to Drs. Dick van der Zee and LucBoerboom (ITC) for the useful discussions. It is worthnoting that this study would have not materialized withoutthe kind cooperation of the Lombardy Region and of theProvince of Bergamo that provided the data set. Specialthanks have to be extended to Architects Moris Lorenziand Anna Nicotera of the Province of Bergamo. Finally, asincere gratitude goes to the technical personnel working inthe Local Parks of the Lombardy Region for the usefulinformation provided about their own ‘‘pioneer’’ experi-ence, and in particular to Architect Luisa De Carli.

References

Bernetti, I., 2000. Valutazione e sviluppo delle attivita economiche nei

parchi nazionali di nuova Istituzione. In: Atti del XXX Incontro di

studio del CeSET ‘‘Gestione delle risorse naturali nei territori rurali e

nelle aree protette: aspetti ecomici, giuridici ed estimativi,’’ Potenza,

5–6 ottobre 2000.

Bernetti, I., and E. Marone., 2000. La valutazione dell’impatto

sull’economia locale derivante dall’istituzione di un’area protetta.

Una metodologia di analisi ex-ante. In: Atti del XXXVII Convegno

SIDEA, Bologna, 14–16 settembre 2000.

Di Fidio, M., Ferrari, A., Lazzeri, O. (Eds.), 2001. I Parchi Locali di

Interesse Sovracomunale in Lombardia. Fondazione Lombardia per

l’Ambiente, Milano.

Di Ludovico, D., Romano, B., 2000. The evaluation of environmental

fragmentation using GIS techniques. PLANECO Newsletter no. 5/

2000. Available at /http://dau.ing.univaq.it/plan/newsletter.htmS.

Diviacco, G., 1994. Il valore delle aree protette. Rivista del Coordina-

mento Nazionale dei parchi e delle Riserve Naturali no. 11/1994.

Available at /http://www.parks.it/federparchi/rivista/index.htmS.

D. Lgs. 29 ottobre 1999. n. 490 ‘‘Testo unico delle disposizioni legislative

in materia di beni culturali e ambientali, a norma dell’art. 1 della legge

8 ottobre 1997, n. 352’’. G.U. n. 302 del 27/12/1999.

Fava, U., 1994. La valutazione dell’impatto socio-economico nei parchi

naturali. Rivista del Coordinamento Nazionale dei parchi e delle

Riserve Naturali no. 11/1994. Available at /http://www.parks.it/

federparchi/rivista/index.htmS.

Giupponi, C., Eiselt, B., Ghetti, P.F., 1999. A multicriteria approach for

mapping risks of agricultural pollution for water resources: the Venice

Lagoon watershed case study. Journal of Environmental Management

56, 259–269.

http://dau.ing.univaq.it/plan/newsletter.htm

http://www.parks.it/federparchi/rivista/index.htm




van Herwijnen, M., 1999. Spatial Decision Support for Environmental

Management. Faculty of Economical Sciences and Econometrics, Free

University of Amsterdam, Amsterdam.

ILWIS 3.3, 2005. The Integrated Land and Water Information System.

ITC, Enschede, /http://www.itc.nl/ilwis/S.

Janssen, R., Van Herwijnen, M., 1994. Multiobjective decision support for

environmental management+DEFINITE DEcisions on an FINITE set

of alternatives: demonstration disks and instruction. Kluwer Academic

Publishers, Dordrecht (Netherlands), 232pp., 16 p.+two 3,500disks.

Keeney, R.L., 1992. Value-focused Thinking. Harvard University Press,

Cambridge.

Keeney, R.L., von Winterfeld, D., 1994. Managing nuclear waste from

power plants. Risk Analysis 14, 3–19.

Keisler, J.M., Sundell, R.C., 1997. Combining multi-attribute utility and

geographic information for boundary decisions: an application to park

planning. Journal of Geographic Information and Decision Analysis 1

(2), 100–119.

L.R. 30 novembre, 1983. n. 86 ‘‘Piano generale delle aree regionali

protette. Norme per l’istituzione e la gestione delle riserve, dei parchi e

dei monumenti naturali nonche delle aree di particolare rilevanza

naturale ed ambientale’’. B.U. 2/12/1983, n. 48, 2 suppl. ord.

L.R. 5 gennaio, 2000. n. 1 ‘‘Riordino del sistema delle autonomie in

Lombardia’’. B.U. 10/1/2000, n. 2, 1 suppl. ord.

Malczewski, J., 1999. GIS and Multicriteria Decision Analysis. Wiley,

New York, 392pp.

Mauri, M. (Eds.), 2000. Atti del seminario: Parchi locali d’interesse

sovracomunale: nuovi protagonisti nelle strategie di valorizzazione del

territorio. Regione Lombardia, Consorzio Parco del Molgora.

Phua, M., Minora, M., 2004. A GIS-based multi-criteria decision making

approach to forest conservation planning at a landscape scale: a case

study in the Kinabalu Area, Sabah, Malaysia. Landscape and Urban

Planning 71, 207–222.

Provincia di Milano, 2002. Criteri e modalita di pianificazione e gestione

dei parchi locali di interesse sovracomunale in provincia di Milano

(Regolamento approvato con D.G.P. 941/02 del 20 dicembre 2002).

Available at /http://www.provincia.mi.it/agricolturaeparchi/parchi.

htmS.

Rietveld, P., 1980. Multiple Objective Decision Methods and Regional

Planning. North-Holland Pub. Co., New York.

Romano, B. and Di Ludovico, D., 2000. Methodology hypothesis for

defragmentation of the urban organism. Available at /http://www.

lifeeconet.com/appendices/chapter4/appendix11/Methodologies.docS.

Saaty, T.L., 1980. The Analytic Hierarchy Process. McGraw-Hill, Inc.,

New York.

Sharifi, M.A., Retsios, V., 2004. Site selection for waste disposal through

spatial multiple criteria decision analysis. Journal of Telecommunica-

tions and Information Technology 3, 28–38.

Sharifi, M.A., van den Toorn, W., Rico, A., Emmanuel, M., 2002.

Application of GIS and multicriteria evaluation in locating sustainable

boundary between the Tunari National Park and Cochabamba City

(Bolivia). Journal of Multi-Criteria Decision Analysis 11, 151–164.

Sharifi, M.A, Boerboom, L., Shamsudin, K., 2004. Evaluating rail

network options using multicriteria decision analysis. Case study

Klang Valley Malaysia. In: Application of Planning and Decision

Support Systems, International Islamic University of Malaysia, pp.

21–60.

Sharifi, M.A., van Herwijnen, M., van den Toorn, W., 2006. Spatial

Decision Support Systems: Theory and Practice. ISPRS WG II/IV, to

be published 2007.

Societa di Progetto Brebemi SpA, 2004. Raccordo Autostradale Diretto

Brescia-Milano (Bre.Be.Mi.), /http://www.brebemi.itS.

Store, R., Kangas, J., 2001. Integrating spatial multi-criteria evaluation

and expert knowledge for GIS-based habitat suitability modelling.

Landscape and Urban Planning 55, 79–93.

http://www.itc.nl/ilwis/

http://www.provincia.mi.it/agricolturaeparchi/parchi.htm

http://www.provincia.mi.it/agricolturaeparchi/parchi.htm

http://www.lifeeconet.com/appendices/chapter4/appendix11/Methodologies.doc

http://www.lifeeconet.com/appendices/chapter4/appendix11/Methodologies.doc

http://www.brebemi.it

application of spatial multi-criteria analysis to site selection for a …sd1229/690h/application of...

Documents