Sp

Ca

b

c

a

ARRA

KUSML

I

tDprptPtoB

cpsb

e

0h

Land Use Policy 30 (2013) 519– 527

Contents lists available at SciVerse ScienceDirect

Land Use Policy

jou rn al h om epa g e: www.elsev ier .com/ locate / landusepol

trategic Environmental Assessment in Latin America: A methodologicalroposal for urban planning in the Metropolitan Area of Concepción (Chile)

arolina Rojasa, Joan Pinob,c,∗, Edilia Jaquea

Department of Geography, Architecture, Urbanism and Geography Faculty, University of Concepción, Victor Lamas 1290, Casilla 160-C, Concepción, ChileCREAF, Cerdanyola del Vallès 08193, SpainUniv Autònoma Barcelona, Cerdanyola del Vallès 08193, Spain

r t i c l e i n f o

rticle history:eceived 8 September 2010eceived in revised form 18 April 2012ccepted 22 April 2012

eywords:rban planningtrategic Environmental Assessmentetropolitan areas

andscape ecology

a b s t r a c t

This work describes a methodology for Strategic Environmental Assessment of urban areas in Latin Amer-ica based on the recently approved European Planning Directive, and applies it to the Metropolitan Areaof Concepción (Chile). The method is based on the Land Suitability Index (LSI), a cartographic GIS-basedindex originally developed for the region of Barcelona (Spain) and aimed at determining the suitability ofeach point in a region for urban development, considering three sub-indexes: (i) Naturalness, (ii) Ecolog-ical Connectivity and (iii) Natural Risk. Using the LSI we evaluated the already approved urban plans ofthe municipalities in the region, considering two scenarios: the initial land use or baseline scenario (S0)and the designated land use or planned scenario (S1). The results show that overall the planned scenariowill result in a loss of around 16% of naturalness, with particularly negative effects on brushwood and

wetland areas. Connectivity will decrease by around 17%, and urban areas exposed to many types of nat-ural risks will increase considerably, from approximately 49% to 92% of the total urban surface. Finally theLSI shows that around 252 ha are suitable for new urbanization in the extension area. This correspondsto around 0.7% of the total extension area (37.381 ha), which represents 12% of the region (271.398 ha).We propose this methodology can be a valuable contribution to the design of Strategic EnvironmentalAssessment applications and indicators for land planning in Latin America.

ntroduction

Strategic Environmental Assessment (SEA) has been a manda-ory land planning procedure in Europe since the approval of theirective 2001/42/EC, on the assessment of the effects of plans andrograms on the environment. SEA is the process by which envi-onmental considerations are required to be integrated into thereparation and adoption of these plans and programs in ordero promote environmentally sustainable development (Jiricka andröbstl, 2008). In consequence, SEA will contribute to the reduc-ion (or avoidance) of the environmental, social and economic costsften associated to excessive or chaotic urban growth (Portal andéjar, 2005; Botequilha-Leitão and Ahern, 2002).

SEA is inspired by two objectives: (1) to overcome the insuffi-iencies of Environmental Impact Assessment (EIA) by evaluating

rojects earlier in the decision making process and (2) to empha-ize the importance of a territory’s limitations and opportunitiesy defining the options of sustainable development. While being a

∗ Corresponding author. Tel.: +34 935814665; fax: +34 935814151.E-mail addresses: crojasq@udec.cl (C. Rojas), joan.pino@uab.es (J. Pino),

djaque@udec.cl (E. Jaque).

264-8377/$ – see front matter © 2012 Elsevier Ltd. All rights reserved.ttp://dx.doi.org/10.1016/j.landusepol.2012.04.018

© 2012 Elsevier Ltd. All rights reserved.

more general procedure than project-specific assessment instru-ments such as EIA, it poses significant challenges for decisionmaking (Unalan and Cowell, 2009). A particular challenge of SEAis how to adequately integrate all the dimensions of sustainabledevelopment so that it becomes an achieveable, practical objec-tive, which can thereby incorporate the environment into policies,plans, and programs (Onate et al., 2002).

A particular focus of application for SEA is urban plans formetropolitan areas. Currently, more than 3 billion people are livingin urban areas worldwide, and this figure will have increased to6.4 billion by 2050 (United Nations, 2009). A critical feature of thisprojection is that the largest population growth expected in urbanareas will be concentrated in the cities and towns of developingregions (United Nations, 2009), such as Latin America. However,these overpopulated metropolitan regions still house importantnatural areas featuring considerable ecological diversity, and pro-viding ecological services to the population. There is, therefore, anurgent need to rethink the appraisal of urban plans and projectsto consider metropolitan areas as a mosaic for natural systems

and population. It is paramount to combine traditional urban plan-ning that emphasizes quality of life, with conservationist planningthat focuses on the conservation of ecosystems and biodiversity(Forman, 2004).

5 se Poli

peTILpAmshabt

etlrnwtanwiqaao

paig(haO2aoLhocofrt

t(meiecHeweaoto

20 C. Rojas et al. / Land U

In Latin America, a number of countries are progressively incor-orating a SEA rationale via EIA, environmental frameworks andnvironmental mainstreaming in policies and plans (Sánchez-riana and Quintero, 2003). In the case of Chile, the Environmentalmpact Evaluation System (SEIA), detailed by the Environmentalaw (1994), states that land planning instruments have to incor-orate environmental impact studies in their approval process.dditionally, EIA regulations state that environmental assessmentust be present from the project’s beginning and the predicted

cenario must be monitored to see if the environmental changesave been effectively minimized or avoided. However, these evalu-tions are frequently performed considering the impacts producedy each project or initiative and not with respect to the overallerritorial model as required by SEA (Suazo et al., 2009).

In consequence, the EIA process in Chile does not sufficientlynsure the conservation of natural systems in metropolitan areashat are subjected to strong urbanizing pressures (83% of the popu-ation lives in urban areas; MINVU, 2008). In southern Chile, urbanegulating plans for the coastal cities of Tomé, Penco and Coro-el approved by the SEIA promote urban development in areasith a high risk of flooding or landslide (Suazo et al., 2009). Thus,

he current urban growth regulation poses serious environmentalnd security challenges in metropolitan areas of Chile, because theorms are focused towards specific projects, while SEA embraces aider view trying to protect and enhance the natural environment,

ntegrating social and economic factors alongside environmentalualities (Wallington et al., 2007). Thus, we consider SEA is a moredequate assessment tool that favors the integration of impactssessment and sustainable urban growth in metropolitan areasf Chile.

One of the main difficulties of applying SEA is that many regionallans frequently fail to take environmental factors properly intoccount (Marull et al., 2007). Quantitative socio-environmentalndices, already in use for aquatic systems (Paul, 2003), may be aood choice to assess the impact on land of diverse alternative plansLugeri et al., 2000). Indeed, a number of interesting applicationsave been developed on the interface between landscape ecologynd urban planning principles (Botequilha-Leitão and Ahern, 2002;pdam et al., 2002; Corry and Nassauer, 2005; Termorshuizen et al.,007). However, such approaches have not provided significantdvances in perhaps the most important constraint of these meth-ds: the lack of standard methodologies. The recently developedand Suitability Index (LSI) aims to overcome this challenge andas been used to evaluate the adequacy of land for urban devel-pment in the region of Barcelona (Spain) in the SEA context. Itombines three main sub-indices concerning (i) the vulnerabilityf the biosphere, lithosphere, and hydrosphere to impacts arisingrom implementing the predevelopment proposals; (ii) the natu-al heritage value of the target area; and (iii) its contribution toerrestrial ecological connectivity (Marull et al., 2007).

The development of these methodologies and their applicationo SEA might be, however, affected by data quality and availabilityMarull et al., 2007; Desmond, 2007). In particular, spatial and the-

atic accuracy of cartographic datasets might have non-negligibleffects on the SEA results and their consequences on decision mak-ng. Therefore, it might be desirable to generate statistically validstimates of the accuracy of these maps, e.g. describing their mis-lassification errors (Nusser and Klaas, 2003; Serra et al., 2003).owever, the effects of data quality on uncertainty have beenxplored only recently within SEA context (João, 2007), togetherith other uncertainty sources affecting it (João, 2007). SEA is gen-

rally seen as a process dealing with highly diverse data sources

nd societal values, and supporting great flexibility in its expectedutcomes (Partidário, 1996, 2007; OECD, 2006). Thus, it is assumedhat data quality should not preclude the application of SEA butnly modulate its expectative.

cy 30 (2013) 519– 527

This paper aims at incorporating SEA principles into the assess-ment of the Plan for the Metropolitan Area of Concepción (CentralChile) based on the Land Suitability Index developed for theMetropolitan Area of Barcelona (Marull et al., 2007). The originalmethodology has been adapted to the general Latin-American sce-nario, characterized by low availability of reliable GIS data. Wepresent the method developed and tested in the Metropolitan Areaof Concepción (Chile), which focused on the evaluation of ecologi-cal functionality and natural risk for the population, and was basedon a limited number of GIS layers. Using this methodology, we havedetermined (i) the expected changes in environmental dimensions(natural heritage, natural risk, and ecological function) under theplan application and (ii) the suitability for urban development ofthe planned areas, based on these dimensions.

Materials and methods

Study area

The Concepción Metropolitan Area (CMA thereafter) is locatedin south central Chile, between 36◦ 35′ and 37◦ 00′ south latitudeand 72◦ 45′ to 73◦ 15′ west, comprising 2830.40 km2. It spreadsover 60 km of coastline from the township of Tomé (the north-ern limit) to that of Lota (the southern limit), and also comprisesthe townships of Chiguayante, Concepción, Coronel, Hualqui, Lota,Penco, San Pedro de la Paz, Santa Juana, Talcahuano, and Tomé.The municipality of Concepción is the capital of the Bio Bio Regionand corresponds to 7.6% of the Region. The population lives princi-pally on the western branch of the coastal mountains, the tertiarymarine platforms and the deltas emerging from the Andean BioBio tributaries. According to the Population and Housing Census(2002), 902,712 people live in the CMA. This corresponds to 48.49%of the total population of the Bio Bio Region, with a density of 318.9inhabitants per km2. In the CMA, 97% of the population lives inurban areas, representing 57.31% of the urban population of BioBio Region (INE, 2002).

The Metropolitan Urban Plan of Concepción

The Metropolitan Urban Plan of Concepción (MUPC thereafter)was approved in 2003 and its principal goals are (i) to regulate thephysical development of the urban and rural areas of the inter-county system, (ii) to guarantee the integrity and continuity ofdifferent elements that structure neighboring counties, (iii) to pro-vide a stable reference frame to evaluate and invest in projectsthat are of community interest, and (iv) to move towards sus-tainable development in the CMA. The counties included in thePlan are: Concepción, Chiguayante, Hualpén, Hualqui Talcahuano,Penco, San Pedro, Tomé, Coronel, Lota and Santa Juana. The planalso regulates the administrative structuring of this particularlydynamic area. For example, it was responsible for the creation ofa new county Hualpen out of a previously existing one in March2004.

The zoning map of MUPC is based on instrumental objec-tives that should strengthen the territory’s physical potential andsocioeconomic specialization by facilitating integral functioningand specifically differentiating between inhabitable and protectedspaces. The habitable area is divided between urban and ruralareas within the urban limits of the CMA. The urban area issubdivided into the consolidated urban area (22,504 ha; 8.47%)

that receives population growth and the urban extension area(31,381 ha; 11.82%) for future growth which, in turn, is dividedinto areas of: (i) mixed housing, (ii) preferential housing, (iii) con-ditioned development, and (iv) extension in sloped terrain.

C. Rojas et al. / Land Use Policy 30 (2013) 519– 527 521

he Me

4ff2i1trrti(

Fig. 1. Main zoning categories considered in t

The MUPC proposes a huge urban growth from the current.2–20.4% of the CMA surface, which is poorly justified by theorecasted population growth. This has been estimated to riserom the current 902,712 inhabitants to 1083.128 by 2020 (INE,008). On the other hand, a set of protected areas where urban-

zation is proscribed have been included in the plan (43,981 ha;6.57% of the CMA). These areas are devoted to the conserva-ion of some remains of the original natural heritage in steepocky marine, drainage, and beach areas with natural value. The

ural area (167,533 ha; 63.12%) includes land especially devotedo agriculture, forestry, mining and tourist activities, and its characterized by small towns and agricultural communitiesFig. 1).

tropolitan Urban Plan of Concepción (MUPC).

Developing the LSI

The method is based on the Land Suitability Index (LSI), whichwas developed to evaluate the adequacy of land for urban devel-opment in the Metropolitan Region of Barcelona in the SEA context(Marull et al., 2007). The LSI focuses on environmental dimensions,considering that the new strategic planning based on sustainabil-ity principles requires a balance between population needs andnature conservation (Forman, 2004). This approach assumes that

new urban areas should be located in the areas most apt for urbangrowth, favoring urban extensions that minimize impact on nat-ural coverage, conserve the landscape with an integral approachwithout isolating protected natural spaces, and avoid locating

522 C. Rojas et al. / Land Use Poli

Table 1Correspondence matrix between land cover categories and values of the NaturalnessIndex (NI).

Value Naturalness Land use

0 Null Urban – infrastructure1 Very low Bare soil2 Low Agricultural land – lumber plantations3 Medium Grassland – water bodies – beach and dunes

p(iomciLatl(

uiatNC1pucT2tFsWd1

itr2rna

TM

4 High Brushwood areas – wetlands5 Very high Native forest – renewal

opulation settlements in areas vulnerable to natural hazardsMarull et al., 2007). The LSI formally combines three axes assess-ng the suitability for urban development based on (i) the impactsn physical environment and the associated natural risk, (ii) theaintenance of the functional environment focused on ecologi-

al connectivity, and (iii) the biological environment according tots natural heritage value (Marull et al., 2007). Our version of theSI has been adapted to the socio-environmental context and datavailability of the CMA, by combining three sub-indexes measuringhree environmental dimensions that are similar to those estab-ished in the original LSI: Naturalness (NI), Ecological ConnectivityECI) and Natural Risk (NRI).

The Naturalness Index (NI) has substituted the original LSI’s Nat-ral Heritage Value Index (NHVI), which was supported by detailed

nformation on species and habitat conservation. Since these datare not available for the CMA, the NI was conceived as a proxy forhe degree of human disturbance in habitats (Pino et al., 2009). TheI was based on a land cover map of the CMA, generated by thehilean Forestry Services at 1:50.000 scale using Landsat images of998 (CONAF et al., 1999). Despite this map was relatively contem-orary with the start of the MUPC (2003), it was updated with therban areas of 2001 through standard methods of Landsat imagelassification and further on-screen digitizing (Rojas et al., 2009).he overall accuracy of the classification of these urban areas of001 was moderate (Kappa index = 70%), while no information onhat of the map generated by CONAF et al. (1999) was available.inally, the land cover categories of the resulting map were reclas-ified to the following: 1, Urban; 2, Native forest; 3, Brushwood; 4,

ater bodies; 5, Bare soil; 6, Lumber plantations; 7, Beaches andunes; 8, Grassland; 9, Renewals; 10, Agricultural; 11, Wetlands;2, Infrastructures; and 13, Other.

The 13 categories of the resulting land cover map were scoredn five categories from 0 (very low) to 5 (very high) degree of dis-urbance produced by human intervention (Table 1). Vegetationeferences from the Santiago Metropolitan Area (Rodríguez et al.,

000) and Spain (Sastre et al., 2002) were used to set up the cor-espondences between vegetation and naturalness classes. Thus,ative forests, renewals and brushwood areas were considereds the most natural classes. Several studies, indicating that these

able 2atrix of impedances used to calculate cost distances in the Ecological Connectivity Inde

Class 1 2 3 4 5 6

1. Urban 1 100 75 – –

2. Native forest 1 50 75 75 53. Brushwood 1 75 50 74. Water bodies 1 –

5. Bare soil 1

6. Lumber plantations

7. Beaches and dunes

8. Grassland

9. Renewals

10. Agricultural11. Wetlands12. Infrastructures

13. Other

cy 30 (2013) 519– 527

vegetation types show the greatest plant species richness in theRegion, confirm this categorization (Cavieres et al., 2005). In con-trast, lumber plantations were assigned to the low naturalness classbecause they are anthropogenic and constitute one of the principalcauses of landscape alteration that threatens the conservation offlora in the study area.

The NI was calculated with the following equation:

I =n∑

i=1

ni × si

S(1)

where ni are the different ordinal categories of the correspond-ing index and si their area in the CMA; S corresponds to the CMAarea.

The Ecological Connectivity Index (ECI) aims at measuringthe extent to which organism movement between habitats isimpeded by landscape features in general, and by urban areasand infrastructures in particular (Adriaensen et al., 2003). Thereare a myriad of metrics for assessing connectivity (Beier andNoss, 1998; Brandt, 1995; Kubes, 1996; Marull and Mallarach,2005; Sepp et al., 1999). The first version of the LSI inMarull et al. (2007) included a refined version of the Eco-logical Connectivity Index (ECI; Marull and Mallarach, 2005),which has been successfully used in land planning assess-ment (Pino and Marull, 2012). This index essentially takes intoaccount the distance between different functional ecological areas,the affinity of their habitats and the impact of human-madebarriers.

In accordance with previous works (Adriaensen et al., 2003;Sastre et al., 2002), we developed a simplified ecological con-nectivity index based on the cost distance between the pixels ofeach habitat or land cover category (Foltête et al., 2008). Cost dis-tance is a commonly used proxy for ecological connectivity andit is included in the Spatial Analyst extension of the ArcGIS 9.2Program. This method calculates the distance between the pixelsof each land cover class modulated by the impedance – or fric-tion – of the surrounding land cover matrix that organisms moveacross. In the majority of cases, impedance values are assumed tobe inversely associated with the affinity between the land covercategories.

We constructed an impedance matrix for the categories of theabovementioned land cover map (Table 2). Values were assignedconsidering previous studies on the ecological value of the vegeta-tion types in the Metropolitan Area of Santiago de Chile (Rodríguez

et al., 2000) and taking as reference the impedance values usedin other connectivity indexes for Madrid (Sastre et al., 2002) andBelgium (Adriaensen et al., 2003). These values were then usedto calculate the cost distances for the native forest, renewals,

x (ECI).

7 8 9 10 11 12 13

– – 50 100 – – – –0 75 75 25 75 75 100 755 75 25 50 50 75 75 75– – 75 75 – – – –– – 50 75 – – – –1 – 75 50 – – – –

1 75 75 – – – –1 75 25 75 50 50

1 75 75 100 751 – – –

1 – –1 –

1

C. Rojas et al. / Land Use Policy 30 (2013) 519– 527 523

Table 3Categories considered in the Natural Risk Index (NRI) and their description.

Value Risk level Type of risk

0 Null Without risk1 Very low Wind deflation2 Low Low risk of landslide3 Medium Medium risk of flooding, landslide

bi

E

Tfiwlc

nIps(rmAa(oto

tttimwtcsts(me(r(riT

TCst

Table 5Correspondence between the impact values resulting from combining the ECI, NIand NRI, and the suitability categories summarized in the LSI.

Combined index Suitability

Value Impact Value Suitability

5 Very high 1 Not suitable4 High 2 Very low3 Medium 3 Low

4 High High risk of flooding or landslide5 Very high Active landslide – tsunami

rushwood and prairie. Thus, the ECI is calculated for each pixeln the study region as the sum of these four cost distances (CDi)

CI =n∑

i=1

CDi (2)

he map created by the sum of the costs of each ECI was reclassi-ed into six ordinal categories (natural breaks methods or groupith similar values), assigning the value 5 to the group with the

owest displacement cost values and representative of the highestonnectivity, and 0 to the highest cost.

The Natural Risk Index (NRI): The third environmental compo-ent of the former LSI corresponded to the Territorial Vulnerability

ndex (TVI), which summarized the ecosystem’s resilience to theotential impacts of urban and/or infrastructural plans based on aet of variables regarding biosphere, lithosphere, and hydrosphereMarull et al., 2007). In the case of the CMA, this concept wasestricted to hydro-geological risk, which was considered to be theost dangerous agent due to the location of new urbanizations.

first map of hydro-geological risk including flooding, landslide,nd wind deflation was generated for the coastal sector of the CMAMardones and Vidal, 2001). This map was then classified into 6rdinal categories, assigning the value 0 to areas without risk andhe value 5 to those with the greatest danger according to the devel-ped classification system (Table 3).

Unfortunately, the rest of the CMA namely the inland rural sec-or did not account for comparable cartography on risk factors orhe essential information for its generation. In consequence, we hado obtain a simplified risk map for the rural sector only consider-ng landslide risk. Hill slope was the main factor considered in this

ap due to its influence on almost all morphogenetic processes,hile being combined with vegetation that modulates exposure

o erosive agents such as rainfall. Soil and lithology could not beonsidered as there was no reliable cartography on soil types andhallow lithological formations for the inland CMA. Then, followinghe risk matrix proposed by Mardones and Vidal (2001), a land-lide risk map was obtained by combining a digital elevation modelDEM-SRTM) reclassified in three slope classes with the land cover

ap mentioned above, reclassified into woody vegetation (for-st, brushwood, renewal and plantations), herbaceous vegetationincluding prairies), and bare soil. The resulting map was finallyeclassified into three landslide categories: low, medium and highsee equivalencies in Table 4). The risk categories of this map were

elatively comparable to those of the coastal map as both werenspired by the same risk assessment (Mardones and Vidal, 2001).hen, a joint map was obtained for the entire CMA through GIS

able 4orrespondence matrix between the vegetation types in the land cover map and thelope categories of the DEM map used for generating the natural risk index map ofhe inland sector of the CMA.

Slope Tree vegetation Herbaceous vegetation Without vegetation

<10◦ Low Low Medium10–20◦ Low Medium High>20◦ Medium High High

2 Low 4 Moderate1 Very low 5 Suitable0 Without impact 6 Very suitable

procedures, by substituting the non-data areas in the coastal mapwith the values of the inland rural map.

Using this joint map, we calculated the Natural Risk Index (NRI)as:

NRI =∑ URi × 100

US(3)

where URi represents the urban surface for each type of risk and USthe total urban surface.

The Land Suitability Index (LSI): the NI, ECI and NRI were com-bined into the Land Suitability Index (LSI), in order to evaluate theland’s aptitude to support urban and infrastructure developmentin the CMA. First, we scaled the NI, ECI and NRI indices to six ordi-nal categories (0–5). Then, we assigned the maximum value of thethree indices to each pixel in the study area

CI = max(ECI, NI, NRI) (4)

and finally, we obtained the LSI by calculating the opposite ofthe CI

LSI = 6 − CI (5)

Thus, the lowest value of the combination represents the areas withless impact when urbanized, i.e. less ecological connectivity, natu-ralness, and hydro-geological risks, and they correspond to thosemost suitable for urban use (Table 5).

Measuring the sustainability of the MUPC

A second aim of our study was to assess the effect of the MUPCon the environment, based on changes in the developed indicesdue to the application of this plan. This corresponds to one ofthe SEA functions, namely, to analyze potential losses and accu-mulative deterioration in the environment (Onate et al., 2002). Toapply the Land Suitability Index (LSI), we considered two scenar-ios. The first (S0) corresponded to the baseline situation (prior tocarrying out the CMA plan) that was inferred from the land covermap of 1998 updated with the urban areas of 2001 (see “Devel-oping the LSI”). The second corresponded to the planned scenario(S1), which was obtained by superposing the MUPC (Fig. 1) tothe land use map used in S0. Thus, we assumed that all the landdevoted to urbanization has been developed in S1 according tothe MUPC.

S0 and S1 were used to assess the changes in natural risk, nat-uralness, ecological connectivity, and land suitability in the CMA.We calculated the mean value of the corresponding indices for S0and S1. Then, the amount of change (in %) for each index in the CMAwas calculated as

�I = (I1 − I0) × 100I0

(6)

where I0 and I1 correspond to the index values in S0 and S1 respec-tively.

In the case of ECI, two measures of �I were calculated: the firstfor all the CMA, aimed at assessing the overall connectivity loss in

524 C. Rojas et al. / Land Use Policy 30 (2013) 519– 527

Fr

tacs

R

E

tad(attaio

Saca0c7cfsw

Fno

risk areas will increase respectively from approximately 255 ha to18,744 ha and from around 1284 ha to 18,066 ha (Fig. 4).

ig. 2. Percentages of changes (% of the initial area) in the categories of the Natu-alness Index (NI) in the CMA, between S0 (2001) and S1 (2020) scenarios.

he study area; the second was calculated only for the protectedreas included in the MUPC, in order to assess the importance ofonnectivity loss strictly for the most valuable areas for biodiver-ity.

esults

xpected changes in environmental dimensions

The application of the MUPC will cause noticeable changes inhe environmental dimensions studied. Natural heritage will show

considerable loss, since the average value of NI in the CMA willecrease by around 16% after the application of S1 compared to S0Fig. 2). This is principally due to the loss of bare soil (around 66%),gricultural land and lumber plantations (around 18%). However,he losses also affected natural land cover categories, includinghe most natural brushwood and wetland areas that decreased bypproximately 19%. Losses in native forests and renewals are lowern percentage (around 7%), yet they have a noticeable impact on theverall CMA naturalness.

Changes in the ECI across the CMA show that the application of1 will also decrease landscape functionality: On average, S1 shows

drop of approximately 17% in ECI values, basically due to theonnectivity loss in brushwood areas (28%), and in both renewalnd native forest (19%). Prairies are expected to be affected only by.30% connectivity loss. In protected areas, the average loss of ECIonsidering all the natural coverage types studied will be around%, which indicates that these areas will suffer a noticeable loss of

onnectivity with S1 application, even though they are kept apartrom the urbanizing process. The habitats with high natural valuehowing the most connectivity loss in protected spaces are brush-ood areas (around −17%), followed by renewal (approximately

ig. 3. Percentages of losses and gains (% of the initial area) in the Ecological Con-ectivity Index (ECI) for each vegetation type in the CMA and in protected areasnly, from S0 (2001) to S1 (2020) scenarios.

Fig. 4. Frequency distribution of the categories of the Natural Risk Index (NRI) inthe CMA, under S0 (2001) and S1 (2020) scenarios.

−11%), and native forest (around −6.5%). In contrast, the most ben-efited habitats in terms of connectivity are the prairies with anincrease of approximately 4.5% (Fig. 3).

The Natural Risk Index (NRI) shows that, formerly (S0), the CMAwas principally exposed to a landslide risk: approximately 51% ofthe territory had low landslide risk, 33% was medium risk, 3% washigh risk, and 7.5% corresponded to active landslide areas. In thenew scenario (S1), the amount of consolidated and extension urbanareas that will be exposed to natural risks will increase from around5943 ha to 49,618 ha. In the S0, around 49% of the urbanized land(more than 12,000 ha) was exposed to natural risk, but this valuewill increase to 90% under the S1, where urbanized land will gobeyond 50,000 ha. The area of urbanized land in low and medium

Fig. 5. Map resulting from the application of the Land Suitability Index (LSI) to theMetropolitan Area of Concepción (Santa Maria Island was not included).

C. Rojas et al. / Land Use Poli

FU

E

T(aMm5a(0pmsa5(

D

paiaMvlthab(

ahztpiaaUsbWu

ig. 6. Frequency distribution of the Land Suitability Index (LSI) categories in therban Extension Zones in the CMA.

nvironmental suitability of the planned urban development

Fig. 5 shows the distribution of the LSI categories in the CMA.here are around 5376 ha located in suitable areas for developmentLSI 5–6), 90% of which are consolidated urban area and only 4.6%re urban extension areas designated for new settlements by theUPC. These areas correspond to 12% of the metropolitan area withore than 30,000 ha, and are primarily located (approximately

2%; 19,626.66 ha) in low to moderate suitable areas (LSI = 3–4)nd secondarily (around 47%; 17,502.66 ha) in not suitable areasLSI = 1–2). Thus, suitable areas (LSI 5–6) approximately represent.7% of extension areas (252.09 ha). These areas are dominated byreferential extension areas (45.4%; 114.48 ha) and secondarily byixed extension areas (22.2%; 55.98 ha; Fig. 6). Not suitable exten-

ion areas correspond to approximately 17,500 ha (9.3%) of the totalrea with LSI 1–2 (around 187,000 ha). Of these extension areas,6.5% (9875.79 ha) correspond to extension in slope, and 15.2%2662.65 ha) correspond to mixed or preferential housing.

iscussion and conclusions

This paper provides a preliminary approach to the appraisal oflanning instruments in Chile, based on the SEA rationale through

modification of the LSI proposed by Marull et al. (2007). The LSIs related to the concept of hosting capacity, generally understoods a territory’s aptness to support distinct land use (Marull andallarach, 2005; Otero et al., 2006). The application of an ad hoc

ersion of this index in the CMA thus provided a first evaluation theand’s aptitude to support urban and infrastructure development inhe CMA based on physical and ecological parameters such naturaleritage, geotechnical risk, and ecological functionality. This evalu-tion could help to improve the dominating land planning criteriaased on population demography, socioeconomics, and mobilityMalczewski, 2004).

Despite Chile having a specific plan appraisal tool (SEIA), there is growing consensus that the application of the SEA rationale mightelp to improve the sustainability of some derivatives of Chileanoning plans and programs, e.g. in urbanization processes. Indeed,he contribution to sustainable development of regional and locallanning instruments in Chile is, nowadays, very limited. The MUPC

s a general normative framework for local planning that provides restricting approach especially as regards as zoning natural valuereas. Indeed, in their Articles 2.1.17 and 2.1.18, the General Law ofrbanism and Construction (Chile Government, 2008) only con-

iders as protected land those zones that are already protectedy the competing institution, the National System of Protectedild Areas (Sistema Nacional de Áreas Silvestres Protegidas). Local

rban planning frequently has to restrict residential use in areas

cy 30 (2013) 519– 527 525

not necessarily designed for environmental conservation, and thusplanning instruments have acquired more competencies than thoseassigned by the Law.

A key limitation of plan and project appraisal in Chile is thelack of clearly identifiable indicators and comprehensive method-ological frameworks for their use (Portal and Béjar, 2005). Thedevelopment of SEA indicators might be, in turn, constrained by theavailability and quality of environmental cartography. Our modi-fied version of LSI was aimed at overcoming these difficulties usinggeneric environmental mapping that might be obtained with rela-tively limited resource, in order to facilitate the application of themethod across the Chilean metropolis. However, even so, someconstraints remained in the case of the CMA. First, land cover car-tography was not up-to-date enough, despite the most dynamiccategories (i.e. urban areas) were updated to 2001 and thus werequite contemporary with the start of the MUPC (2003). Second,information about its accuracy (e.g. Kappa indices) was availablefor our classification of urban areas in 2001 but not for the mainland use map by CONAF et al. (1999). As no error values could beprovided when assessing the distribution and changes of our SEAindexes, figures have to be considered approximate and mostlyindicative for the average status and trends in the region, espe-cially for the most quantitative ones (ECI). However, it should betaken into account that the MUPC is not completed, and how it willbe carried out in the near future might add a greater bias in ourresults than that arising from data inaccuracy. Moreover, the use offew land cover categories might help to reduce the misclassificationeffects in these indices. In the case of LSI, the scaling of its compo-nents into ordinal categories might help to reduce these effects.We then conclude that the potential constraints in land cover dataaccuracy and updating would not invalidate our SEA results butmodulate their expectative.

Still more important are the constraints derived from dataavailability regarding natural risks. These risks have only been con-sidered in their whole for the coastal sector of the CMA, which isthe most populated in the region and it is potentially affected byrecurrent flooding and tsunami, especially in the ocean shorelineand in the Bio Bio margins (Mardones and Vidal, 2001). Coastal flatareas also are locally affected by wind deflation, which might posein risk several human settlements close to dune areas. In contrast,the risk map for the inland rural sector only considered landslide,through a very simple approach due to the lack of key informationon soil and surface lithological formations. While this approachis supported by the landslide risk matrix suggested by Mardonesand Vidal (2001), some field assessment of the resulting risk mapthrough ground control areas should be done in the future. Thisrisk map did not include the flooding or the tsunami risks of ruralareas, which are concentrated in the coastal area but that mightaffect the Bio Bio margins in some in land areas. It neither includedthe wind deflation risk that is a typically coastal process. Therefore,the assessment of natural risks in the CMA needs to be improvedwith more reliable methods and more complete and up-to-datecartography, as recently discussed in the context of the earthquakeof 2010 (Romero et al., 2010).

In any case, the methodology provides a first quantitativeassessment of the application of the MUPC in SEA context. Our studybased on the LSI identified negative trends in the three dimensionsanalyzed in the CMA: (i) high overall impact, (ii) planned urbandevelopment in highly valuable and risky areas and (iii) impact onthe ecological functionality of protected areas due to land transfor-mation in their surroundings. These trends are in agreement withthe local increases in large irrigation areas in the coastal cities of the

CMA (Suazo et al., 2009), suggesting that the planned urban growthis highly unsustainable since the new urban areas will necessarilyoccupy lands that require protection because they are susceptibleor sensitive to natural risks. In short, the proposed land use plan

5 se Poli

dfafa

citdnaa(nup2

ct2thnhrAdnuutrrs

icmabdndtpauccaocioloe

iidatb

26 C. Rojas et al. / Land U

oes not correspond to the territory’s physical capacities. There-ore the majority of unsuitable areas were classified as rural zonesnd these are still considered extension areas (i.e., reserves foruture growth) for the city and will be used once the extensionreas neighboring the built up areas become urbanized.

Human population in the CMA is concentrated in medium-sizedities located on the coast, which is associated to a natural dynamicnfluenced by conditioning factors of steep slopes and river mouthshat are very vulnerable to urbanization urban expansion. In recentecades, the CMA has been characterized by an intense, simulta-eous urban sprawl, dominated by growth in the periphery with

disperse pattern through road infrastructures (Rojas et al., 2009)nd with principally negative impacts on the natural ecosystemPauchard et al., 2006). The future scenario model or urban plan-ing model will not change much even when it does not permitrban settlement in the higher valued natural areas, which atresent are especially affected by urbanization (Pauchard et al.,006; Villagran-Mella et al., 2006).

Additionally, the CMA is overflowing into areas that are diffi-ult to handle and are exposed to strong natural risks, increasinghe probability of environmental hazard (Mardones and Vidal,001). The impacts on connectivity are principally reflected inhe fragmentation of natural areas and, especially, in habitats ofigh conservation concern. Paradoxically, the connectivity of semi-atural habitats like prairies, which are the outcome of secularuman intervention and have lower conservation concern than theemaining natural habitats, might increase in the future scenario.lso, it should be noted that the MUPC contemplates extensive landevelopment, which corresponds to more than 20 times the growtheeds estimated by Baeriswyl (2007). Consequently, this extensiverban development is not justified, even considering the 20% pop-lation increase estimated for the year 2020 (INE, 2008). Moreover,his high urban development will cause important losses in natu-alness and ecological connectivity; however the highest losses areelatively acceptable due to the low naturalness of categories (bareoil and lumber plantations).

The protected areas included in the plan will become naturalslands in an urbanized matrix because no ecological corridors areontemplated in the MUPC. If the proposed scenario (S1) is imple-ented, the CMA will face serious impacts on its ecological function

nd mechanisms must also be implemented to diminish the possi-le negative tendencies produced by the planned urbanization. Theecrease in ecological connectivity within protected areas whereo urban development is planned is an unexpected result thateserves special attention. It can be interpreted as an extrapola-ion to the land planning scale of the statement by Janzen (1983) ‘noark is an island’. Clearly, the ecological connectivity of protectedreas will be highly affected by changes in the surrounding landse and land cover categories. Thus, ecological networks should beonsidered by selecting corridor areas that maintain or increaseonnectivity between protected areas, although the inclusion ofreas and corridors cannot be used to mitigate unsustainable devel-pment in the surrounding land matrix. Agricultural areas play aomplementary role to natural parks in maintaining biodiversityn metropolitan regions (Santos et al., 2008). Sustainable devel-pment will, then, require a regional policy aimed at increasingandscape heterogeneity and evaluating rural landscapes in termsf ecological services (such as biodiversity maintenance) as well asconomic value.

It should be noted that the proposed methodology for adapt-ng SEA to the Chilean case cannot replace the environmentalmpact evaluation required by Chilean Law (SEIA). However, it

oes provide a complementary assessment using a prospectivepproach that considers the sustainability of future land use. Also,he indices put forward do not disqualify the Plan’s proposed zoningecause their objective is evaluative and not normative. Rather, this

cy 30 (2013) 519– 527

evaluation aims at promoting discussion and reflection withrespect to how planning can be improved in order to revert negativetendencies of unsustainable urban development. The incorpora-tion of these findings corresponds to political decisions and policymakers.

Acknowledgements

This research was financed by the Project FONDECYT No.11090163 “Valoración del Territorio Metropolitano. Aproximacionesdesde su Sostenibilidad y Evaluación Ambiental Estratégica (Evalu-ation of the metropolitan territory using sustainable approachesand strategic environmental assessment)”. Land use and cover mapwas provided by the Secretary of Bio Bio Region of the Housing andUrbanism Ministry of Chile. Risk maps were elaborated within theFondecyt Regular Project 1992 “Zonificación y Evaluación de Ries-gos Naturales en el Área Metropolitana de Concepción y Talcahuano(Zoning and assessing the natural risks in the Concepción and Talc-ahuano Metropolitan Area)”.

References

Adriaensen, F., Chardon, J.P., De Blust, G., Swinnen, E., Villalba, S., Gulinck, H., et al.,2003. The application of ‘least-cost’ modelling as a functional landscape model.Landscape and Urban Planning 64, 233–247.

Baeriswyl, S., 2007. Patrones de Ocupación del Gran Concepción, A través del modelode consumo de suelo. In: Hidalgo, R., de Mattos, C., Arenas, F. (Eds.), Chile: Delpaís urbano al país metropolitano. GeoLibros Pontificia Universidad Católica deChile, Santiago de Chile, pp. 241–325 (in Spanish).

Beier, P., Noss, R., 1998. Do habitat corridors provide connectivity? ConservationBiology 12, 1241–1252.

Botequilha-Leitão, A., Ahern, J., 2002. Applying landscape ecological concepts andmetrics in sustainable landscape planning. Landscape and Urban Planning 59,65–93.

Brandt, J., 1995. Ecological networks in Danish planning. Landscharp 12,63–76.

Cavieres, L., Mihoc, M., Marticorena, A., Marticorena, C., Baeza, C., Arroyo, M., 2005.Flora vascular de la Cordillera de la Costa en la región del Bío Bío: riqueza deespecies, géneros, familias y endemismos. In: Smith, C., Armesto, J., Valdovinos,C. (Eds.), Historia, biodiversidad y ecología de los bosques costeros de Chile.Editorial Universitaria, pp. 245–252 (in Spanish).

Chile Government, 2008. Ley General de Urbanismo y Construcción. Retrieved fromhttp://www.minvu.cl (in Spanish).

CONAF, CONAMA, BIRF, Universidad Austral de Chile, Pontificia Universidad Católicade Chile, Universidad Católica de Temuco, 1999. Catastro y evaluación de losrecursos vegetacionales nativos de Chile. Informe regional Octava Región, San-tiago, Chile (in Spanish).

Corry, R.C., Nassauer, J.I., 2005. Limitations of using landscape pattern indices to eval-uate the ecological consequences of alternative plans and designs. Landscapeand Urban Planning 72, 265–280.

Desmond, M., 2007. Strategic Environmental Assessment (SEA): a tool for environ-mental decision-making. Irish Geography 40, 63–78.

Foltête, J., Berthier, K., Cosson, J., 2008. Cost distance defined by a topological functionof landscape. Ecological Modelling 210, 104–114.

Forman, R.T.T., 2004. Mosaico Territorial para la región metropolitana de Barcelona.Gustavo Gili, Barcelona, 150 pp. (in Spanish).

INE, 2002. Censo Nacional de Población y Vivienda 2002. Retrieved fromhttp://www.ine.cl, September 2010.

INE, 2008. Proyecciones de Población. Retrieved from http://www.ine.cl, September2010.

Janzen, D.H., 1983. No park is an island: increase in interference from outside as parksize decreases. Oikos 41, 402–410.

Jiricka, A., Pröbstl, U., 2008. SEA in local land use planning – first experience in theAlpine States. Environmental Impact Assessment Review 28, 328–337.

João, E., 2007. A research agenda on data and scale issues for SEA. EnvironmentalImpact Assessment Review 27, 479–491.

Kubes, J., 1996. Biocentres and corridors in a cultural landscape. A critical assessmentof the territorial system of ecological stability. Landscape and Urban Planning35, 231–240.

Lugeri, N., Amadei, M., Bagnaia, R., Dragan, M., Fernetti, M., Laureti, L.,et al., 2000. Environmental quality and territorial vulnerability assess-ment through the GIS of Landscape Units of Italy: the experience of theMap of Nature project. Retrieved from http://proceedings.esri.com/library/userconf/proc00/professional/papers/pap522/p522.htm, September 2010.

Malczewski, J., 2004. GIS-based land-used suitability analysis: a critical overview.Progress in Planning 62, 3–65.

Mardones, M., Vidal, C., 2001. La zonificación y evaluación de los riesgos naturales detipo geomorfológico: un instrumento para la planificación urbana en la ciudadde Concepción (The zoning and evaluation of natural hazards of geomorfological

e Poli

M

M

M

N

O

O

O

O

P

P

P

P

P

P

P

R

C. Rojas et al. / Land Us

type: a tool for urban planning in the city of Concepción). EURE 27, 97–122 (inSpanish: English Abstract).

arull, J., Mallarach, J., 2005. A GIS methodology for assessing ecological connec-tivity: application to the Barcelona Metropolitan Area. Landscape and UrbanPlanning 71, 243–262.

arull, J., Pino, J., Mallarach, J., Cordobilla, M., 2007. A Land Suitability Indexfor Strategic Environmental Assessment in metropolitan areas. Landscape andUrban Planning 81, 200–212.

INVU, 2008. Diagnóstico Urbano 1990–2007. Retrieved from http://www.observatoriourbano.cl/docs, September 2010.

usser, S.M., Klaas, E.E., 2003. Survey methods for assessing land cover map accu-racy. Environmental and Ecological Statistics 10, 309–331.

ECD, 2006. Applying Strategic Environmental Assessment. OECD Publishing, Mas-sachusetts, 164 pp.

pdam, P., Foppen, R., Vos, C., 2002. Bridging the gap between ecology and spatialplanning in landscape ecology. Landscape Ecology 16, 767–779.

nate, J., Pereira, D., Suárez, F., Rodríguez, J.J., Cachón, J., 2002. Evaluación AmbientalEstratégica. La Evaluación Ambiental de Políticas, Planes y Programas. EdicionesMundi Prensa, Madrid, 382 pp. (in Spanish).

tero, I., Varela, E., Mancebo, S., García, M., 2006. Modelo de Localización de áreaspotencialmente urbanizables: aplicación al Corredor del Henares. Ciudad yT.erritorio Estudios Territoriales 148, 331–351 (in Spanish).

artidário, M.R., 1996. Strategic environmental assessment: key issues emergingfrom recent practice. Environmental Impact Assessment Review 16, 31–55.

artidário, M.R., 2007. Scales and associated data – what is enough for SEA needs?Environmental Impact Assessment Review 27, 460–478.

auchard, A., Aguayo, M., Pena, E., Urrutia, R., 2006. Multiple effects of urbaniza-tion on the biodiversity of developing countries: the case of a fast-growingmetropolitan area (Concepción, Chile). Biological Conservation 127, 272–281.

aul, J., 2003. Developing and applying an index of environmental integrity for theUS Mid-Atlantic region. Journal of Environment Management 67, 175–185.

ino, J., Marull, J., 2012. Ecological networks: are they enough for connectivity con-servation? A case study in the Barcelona Metropolitan Region (NE Spain). LandUse Policy 29, 684–690.

ino, J., Vilà, M., Alvárez, N., Seguí, J., Guerrero, C., 2009. Niche breadth rather thanreproductive traits explains the response of wetland monocotyledons to landcover change. Applied Vegetation Science 12, 9–21.

ortal, B., Béjar, C., 2005. Proposición Metodológica para la Evaluación del Impacto

Ambiental de los Planes Reguladores Comunales. Contribuciones científicas ytecnológicas 33, 56–68 (in Spanish).

odríguez, M., Pedernera, P., Castillo, M., 2000. Descripción del paisaje en la regiónmetropolitana: Una primera aproximación. Ciencias Forestales 14–15, 29–37 (inSpanish).

cy 30 (2013) 519– 527 527

Rojas, C., Opazo, S., Jaque, E., 2009. Dinámica de Crecimiento Urbano del ÁreaMetropolitana de Concepción. Tendencias de las últimas décadas. In: Hidalgo,R., de Mattos, C., Arenas, F. (Eds.), Chile: Del país urbano al país metropoli-tano. GeoLibros Pontificia Universidad Católica de Chile, Santiago de Chile, pp.257–269 (in Spanish).

Romero, H., Fuentes, C., Smith, P., 2010. Ecología política de los riesgos naturales yde la contaminación ambiental en Santiago de Chile: necesidad de justicia ambi-ental. Scripta Nova 331, http://www.ub.es/geocrit/sn/sn-331/sn-331-52.htm.

Sánchez-Triana, E., Quintero, J., 2003. Experiences with Strategic Environmen-tal Assessment in Latin America (LAC): the case of Colombia. Wash-ington, DC. Retrieved from http://info.worldbank.org/etools/docs/library/136045/SEA%5FM3%5FErnesto%20Sanchez.pdf, September 2010.

Santos, K.C., Pino, J., Rodà, F., Guirado, M., Ribas, J., 2008. Beyond the reserves: the roleof non-protected rural areas for avifauna conservation in the area of Barcelona(NE of Spain). Landscape and Urban Planning 84, 140–151.

Sastre, P., de Lucio, J., Martínez, C., 2002. Modelos de conectividad del paisaje a dis-tintas escalas. Ejemplos de aplicación en la comunidad de Madrid. Ecosistemas11. Retrieved from http://www.revistaecosistemas.net/articulo.asp?Id=287,September 2010 (in Spanish).

Sepp, K., Palang, H., Mander, U., Kaasik, A., 1999. Prospects for nature and landscapeprotection in Estonia. Landscape and Urban Planning 46, 161–167.

Serra, P., Pons, X., Saurí, D., 2003. Post-classification change detection with datafrom different sensors. Some accuracy considerations. International Journal ofRemote Sensing 24, 3311–3340.

Suazo, B., Jaque, E., Rojas, C., Utz, R., 2009. Assessment of sustainability in urbanplanning of a coastal city: the city of Coronel, southern Chile. In: Brebbia, C.A.,Neophytou, M., Beriatos, E., Ioannou, I., Kungolos, A.G. (Eds.), Sustainable Devel-opment and Planning, vol. IV. WIT Press, pp. 665–679.

Termorshuizen, J., Opdam, P., van den Brink, A., 2007. Incorporating ecologi-cal sustainability into landscape planning. Landscape and Urban Planning 79,374–384.

Unalan, D., Cowell, R., 2009. Adoption of the EU SEA Directive in Turkey. Environ-mental Impact Assessment Review 29, 243–251.

United Nations, 2009. World urbanization prospects: Department of Economicand Social Affairs. Retrieved from http://esa.un.org/unpd/wup/index.htm,September 2010.

Villagran-Mella, R., Aguayo, M., Parra, L., González, A., 2006. Relación entre car-acterísticas del hábitat y estructura del ensamble de insectos en humedales

palustres urbanos del centro-sur de Chile. Revista Chilena de Historia Natural79, 195–211 (in Spanish: English abstract).

Wallington, T., Bina, O., Thissen, W., 2007. Theorising strategic environmentalassessment: fresh perspectives and future challenges. Environmental ImpactAssessment Review 27, 569–584.