landscape and urban planning, volume 94, issue 2, 28 february 2010, pages 105–115,

Landscape and Urban Planning 94 (2010) 105115

Contents lists available at ScienceDirect

Landscape and Urban Planning

journa l homepage: www.e lsev ier .com/ l

Creating disaster-resilient communities: Evaluatperformance of new urbanism

Mark R. SDepartment of ast BuNC 27599-3140

a r t i c l

Article history:Received 1 ApReceived in reAccepted 22 AAvailable onlin

Keywords:Disaster resilieNatural hazard mitigationNew UrbanismPlanning staff technical assistanceLand use planning

nt pzardsonveally aat rehazately

nal deUrbanist developments are incorporating hazard mitigation techniques at a greater rate than are con-ventional developments. We nd that New Urbanist design does not appear to make a difference in theuse of hazard mitigation techniques. While New Urbanist developments use more hazard mitigationtechniques on average than do conventional developments, this difference appears to stem not from thedifference in design type but rather from increased local government technical assistance in the review

1. Introdu

Natural h$25 and $30ing relative(Cutter, 200quakes arethey cause htern of deceof developm

New Urbpossibly cousprawling d2001; Talenintended to

CorresponE-mail add

pberke@email1 Present ad

Settlements, UBritish Columb

0169-2046/$ doi:10.1016/j.of New Urbanist relative to conventional developments. We recommend that New Urbanist designersand local governments engage in more proactive land use planning and take more responsibility to makesure that hazard mitigation techniques are integrated into New Urbanist project site designs.

2009 Elsevier B.V. All rights reserved.

ction

azards cause average annual economic losses betweenbillion in the United States, and losses have been ris-to increases in population and gross national product1). While hazardous events such as oods and earth-

naturally occurring phenomena, the amount of damageas been exacerbated by the conventional land use pat-ntralized sprawl, which has fostered a massive buildupent in areas subject to natural hazards (Burby, 2006).anist design has been promoted as an alternative tonter certain adverse societal outcomes of conventionalevelopment (Duany et al., 2000; Calthorpe and Fulton,, 2005). Based on a set of design principles that arefoster more intentional delineation of open space, a

ding author. Tel.: +1 604 347 5812; fax: +1 604 822 3787.resses: [email protected] (M.R. Stevens),.unc.edu (P.R. Berke), [email protected] (Y. Song).dress: School of Community andRegional Planning, Centre forHumanniversity of British Columbia, #223-1933 West Mall, Vancouver,ia V6T 1Z2, Canada.

better mixture of land uses built at relatively high densities, andpedestrian-oriented transportationnetworks,NewUrbanist designhas also drawn increasing attention for its potential to reduce nat-ural hazard vulnerability (Thompson, 2005; Miller, 2007).

Despite this potential, however, when a New Urbanist develop-ment locates in a hazardous area, its relatively high developmentdensities canmean thatmorepeople andproperty are placed at riskthan would have been the case with a low-density development onthe same parcel of land (Berke and Campanella, 2006; Berke et al.,2009). Song et al. (2009) show that more than one-third of all NewUrbanist developments in theUnited States thatwere completed orunder constructionasofDecember2003are located inareas subjectto oodhazards, and because of relatively high development densi-ties, these New Urbanist developments can pose a greater risk thanconventional low-density developments if ood hazards are notanticipated and ood hazard mitigation is not promoted in projectdesign.

Recent research has compared New Urbanist developmentswith conventional low-density developments to determinewhether local communities put forth more effort in reviewing pro-posals for New Urbanist developments, and whether that effortappears to translate into design that is more resilient to naturalhazards. Berke et al. (2009) found that, on average, in comparison

see front matter 2009 Elsevier B.V. All rights reserved.landurbplan.2009.08.004tevens ,1, Philip R. Berke, Yan SongCity and Regional Planning, University of North Carolina-Chapel Hill, UNC CH, New E, United States

e i n f o

ril 2009vised form 13 July 2009ugust 2009e 15 September 2009

nce

a b s t r a c t

Conventional low-density developmeing per capita losses from natural haUrbanist design as an alternative to cNewUrbanist designmake it theoreticNew Urbanist developments are builtthey locate in areas subject to naturalniques such that hazard risk is adequapairs of New Urbanist and conventioocate / landurbplan

ing the promise and

ilding, Campus box #3140, Chapel Hill,

atterns have been cited as a partial explanation for increas-in the United States. There is an emerging appeal for New

ntional low-density development, and particular features ofmenable to reducing natural hazard losses. However, becauselatively high densities, they can exacerbate hazard risk whenrds and do not incorporate sufcient hazard mitigation tech-reduced. We present a comparative evaluation of 33 matchedvelopments located in oodplains to evaluate whether New

106 M.R. Stevens et al. / Landscape and Urban Planning 94 (2010) 105115

with conventional developments, New Urbanist developments (1)were subject to stronger local government development manage-ment regulations, (2) involved greater levels of public participationand local government planning staff technical assistance during thedevelopmehazard miti

In this panalytical tfounding fadifferenceswith respecthat are himultiple redevelopmeconventionrelevant facthe potentiaborhoods a

To help aticular featufoster the ias additionato inuencement projemethods, wpairs of Neventional lothroughoutmodeling indesign for tconclude wfor plannin

2. Designinresilient to

Increasineffort on thScholars haa guiding pareas. Withhoods andhazard evenout needin1999; GodsCampanellamust beproand Gailus,intentionallthe use of dthe effects o

Becauserisk on anularly impohazard mitiexisting Nehazardmitifocus on boand the degtechniques.rivers, lakeof varying fCommitteements maywe choose

rence and the level of damage they cause. Floods accounted for90% of all natural hazards in the United States from 1992 to 2001(GAO, 2005), contributing to roughly 900 deaths and $55 billion inproperty damages (GAO, 2004).

mosandm d

III-IVoos thapmetemstualfour999;

onmlopmpro

ral dwat

ff thate pom chtreameyanturalbilityplain

e poion:

olarsfram

e Nepacee nunellas of

pacehe mnmene ovioussigne totionof pl09) dce th

netes thitivectiostor

scapeodiity.t netandlopmfor sed uots. Rnt review process, and (3) incorporated more naturalgation techniques.aper, we expand upon previous research by utilizingechniques that allow us to control for potential con-ctors that might explain some (or all) of the apparentbetweenNewUrbanist and conventional developmentst to the incorporation of hazard mitigation techniquesghlighted in previous research. In particular, we usegression analysis to examine whether New Urbanistnts incorporatemorehazardmitigation techniques thanal developments, while controlling for other potentiallytors. Answering this question can help assess whetherl of NewUrbanist design is being translated into neigh-

nd communities that are resilient to natural hazards.nswer this question, we rst identify and describe par-res of New Urbanist design that we hypothesize will

ncorporation of hazard mitigation techniques, as welll factors that have been found by previous researchersthe use of hazard mitigation techniques in develop-

cts. After describing our research design, variables, ande then present a comparative evaluation of 33 matchedw Urbanist developments and a control group of con-w-density developments located in ood-prone areasthe United States, including the results of regressiontended to evaluate the importance of New Urbanisthe incorporation of hazard mitigation techniques. Weith a discussion of our ndings and their implicationsg practice and natural hazard mitigation.

g neighborhoods and communities that arenatural hazards

g losses from natural hazards have inspired signicante part of researchers to identify potential solutions.ve recently emphasized the concept of resiliency asrinciple for designing new development in hazardousin the context of natural hazards, resilient neighbor-communities are those that can withstand naturalts without experiencing devastating losses and with-

g signicant assistance from external entities (Mileti,chalk et al., 2003; Mileti and Gailus, 2005; Berke and, 2006). To achieve this kind of resiliency, communitiesactive in controllinghowdevelopmentproceeds (Mileti2005). In particular, the concept of resiliency should bey built into the planning and building process throughesign and construction techniques that help to mitigatef natural hazards (Bosher et al., 2007).they are likely to place more people and property atequivalent land unit exposed to hazards, it is partic-rtant for New Urbanist developments to incorporategation techniques into their design. To gauge whetherw Urbanist developments actually incorporate moregation techniques than conventional developments, weth types of developments located in oodplain areasree to which they incorporate ood hazard mitigationFloodplain areas are the low-lying lands adjacent to

s, and oceans that are ooded periodically at intervalsrequency (Interagency Floodplain Management Review, 1994). While New Urbanist and conventional develop-be subject to other types of natural hazards as well,to focus on oods because of their pervasive occur-

Thepeoplefree fro1986:amongofcialdeveloecosysconceplowinget al., 1

Envirdevevicesnatu

Stormrunoltra

Strealize sconv

Strucneraood

2.1. Thmitigat

Schdesignbecausopen sing thCampaportionopen sfrom tenviroriverin

Presite de(relativprotecsenseal. (20enhan

Highspacsensprotetratelandnel mveloc

Streetion,deveneedis freing lt effective approach to reducing ood losses is to keepproperty out of harms way by keeping oodplains

evelopment (Federal Emergency Management Agency,). Furthermore, there has been an increasing awarenessdplainmanagers, planners, environmentalists, and localt protecting the natural functions of oodplains fromnt can help to reduce ood risk and preserve natural(Morris, 1997: 24).With this inmind, we draw on prior

izations of ood hazard mitigation that specify the fol-broad categories of mitigation techniques (GodschalkFederal Emergency Management Agency, 2002):

entally sensitive area protection involves preventingent in oodplains and protecting ood mitigation ser-vided by oodplain ecosystems, upland wetlands andrainage systems.er best management practices (BMPs) are used to storet reduceson-site anddownstreamooding, aswell as tollutants in runoff and inltrate runoff to groundwater.annel modication is used to clear, enlarge, and stabi-

channels in or near the development site to facilitatece of stormwater off the site quickly as possible.protection involves techniques to reduce structural vul-to oods if development is located in or near the.

tential advantages of New Urbanist design for hazardin concept

have argued that New Urbanism offers a model urbanework for creating resilient communities, in part

w Urbanist design affords opportunities to maximizewithin a development site without necessarily reduc-mber of dwelling units that can be built (Berke and, 2006). Increasing development densities on certaindevelopment sites while setting aside other portions ascan enable project designers to steer construction awayost hazardous areas while simultaneously protectingtally sensitive features (e.g. wetlands, sand dunes, and

odplains) that provide ood hazard mitigation services.researchers have used three particular New Urbanist

features to compare the effects of New Urbanist designconventional low-density development) on watershed(Berke et al., 2003), transportation (Crane, 1996), andace (Brown and Cropper, 2001), respectively. Berke etescribe these design features, which we expect to alsoe integration of ood hazard mitigation techniques:

density provides more opportunities to create openat can be used to avoid development in environmentallyareas, install BMPs, and reduce reliance on structuraln practices. Compact development patterns concen-mwater runoff rather than spreading runoff across the, which is likely to increase the need for stream chan-

cation to accommodate increased runoff volume and

work design (e.g. narrow streets, pedestrian orienta-on-street parking) offers more opportunities to avoident in sensitive areas and to use BMPs, and reduces thetructural protection because more development spacep due to reduced demand for wide driveways and park-educed project footprints can mean that stormwater

M.R. Stevens et al. / Landscape and Urban Planning 94 (2010) 105115 107

runoff is more concentrated, thus possibly increasing the needfor stream channel modication.

Mixed land uses offers more room for avoidance of sensitive areasand installation of BMPs, and reduces reliance on structural pro-tection because placement of businesses and civic uses next toresidential uses increases pedestrian accessibility which relievespressures for parking, and permits multipurpose trips whereina single parking space can serve several trip purposes whichdecreases demand for spaces. Further, because this design featureencourages compactness in urban forms, it may induce greateruse of streamchannelmodications due to higher concentrationsof runoff.

Table 1 shows the hypothesized impacts of each site designfeature onUrbanist de

2.2. Other fmitigation t

The chointo develothe developas a proceto determinupon particresearchersreview procment projec

1. Developmdevelopmimpact oprogramages (Olsintrusion

2. Planningthe deveassistancimplicatifound thalocal govDalton, 1plannerscan be ican increaddressegiven toclaricattoapplicaother puet al., 200

3. Public painuenceticular m

to particular features that place people andproperty at risk. Suchactivism is relevant because elected ofcials, planning staff, andother relevant actors often respond to citizens. The more activecitizen interest groups are in demanding attention to naturalhazards, the more hazard mitigation techniques local govern-ments tend to adopt (Burby et al., 1997). Burby (2003) found thatwhen a narrow set of groups is involved, mitigation is likely tobe ignored or minimized. Conversely, broad involvement cre-ates the potential for planners to expand their understandingof problems and to develop a stronger set of policies for dealingwith them.Godschalk et al. (2003) argue that increased involve-ment and understanding on the part of the public with respectto natural hazard mitigation should pay off in more widely sup-ported hazard mitigation and should lead to more resilient andsafer communities.

addited ationpmemenrbann timther

nal dpmeeve

levethatpmerowdtionviewes thst prpatiost pre woon inpmepmequesqualviousf comiti

mund toral h

erienwill

0; Brberent

Table 1New Urbanist hniqu

Category of

Sensitive areStormwaterStream chanStructural pr

(+) Means tech sed.the use of ood hazard mitigation techniques in Newvelopments.

actors that might inuence the use of ood hazardechniques

ice to incorporate ood hazard mitigation techniquespment project site designs is typically made duringment review process, described by Berke et al. (2006)ss in which planners and applicants directly interacte how the techniques should be applied, dependingular features of each development project. Previoushave identied certain aspects of the developmentess that help determine the extent to which develop-ts incorporate mitigation techniques, including:

ent management programs: Research suggests thatent management programs and policies can have ann the built environment. Development managements can help protect future development from ood dam-hansky and Kartez, 1998), and can be used to preventinto oodplains and wetlands (McElsh, 2004).staff capacity: The number of planners involved inlopment review process and the number of technicale services they provide to applicants have signicantons for hazard mitigation, as previous studies havet planning staff capacity has a positive inuence on theernment efforts to address natural hazards (Burby and994; Berke and Roenigk, 1996; Brody et al., 2004). Morecan mean that more effort, expertise, and perspectivesnjected into the development review process, whichase the likelihood of hazard issues being identied andd. Increased capacity can lead to more attention beingthe explanation of plan policies and associated rules,ion of issues to be addressed, provision of informationnts, conveyanceofpolicyadvice, andcoordinationwith

blic agencies charged with development review (Berke3; Brody et al., 2004).rticipation: Citizens can potentially participate in andthedevelopment reviewprocessbyadvocating forpar-itigation design features and/or expressing opposition

IndescribmitigadevelogovernNew Upoint iexpectventiodevelo

Howpationnotesdevelomore copposiand redensitiUrbaniparticiUrbanithen wattentidevelodevelotechnibeing e

Preistics ohazard

1. Comfounnatuexpthey200

2. Numopm

design features and hypothesized impacts on the use of ood hazard mitigation tec

New Urbanist design feature

mitigation technique High net density

a protection +BMPs +nel modication +otection niques are more likely to be used; and () means techniques are less likely to be uion to the three New Urbanist site design featuresbove, we expect that the incorporation of ood hazardtechniques into both New Urbanist and conventionalnts will depend in part upon these aspects of the localt development review process. Assuming that pairs ofist and conventional projects are reviewed at the samee by the same respective local government, we woulde to be little difference across New Urbanist and con-evelopmentswith respect to planning staff capacity andnt management practices.r, we do expect signicant differences in public partici-ls for the two types of projects. Downs (2005: 270271)existing community residents often fear that densent projects will lead to more trafc congestion anded schools and other facilities, and that strong publicfrequently surfaces when such projects are proposeded. Since New Urbanist developments have higheran conventional developments, we would expect Newojects to inspire a greater level of public reaction andn during development review. If it is true that Newojects inspire more public participation and scrutiny,uld expect that natural hazard issues will receive morethe development review process for New Urbanist

nts than for conventional, and thus that New Urbanistnts will incorporate more ood hazard mitigationthan will conventional developments, other things.researchers have also identied particular character-

mmunities and project sites that inuence the use ofgation techniques. These characteristics include:

ity ood history: Prior experiencewithoodinghas beenhaveasignicanteffectoncommunityefforts toaddressazards. In general, the more recently a community hasced natural hazard-related damages, the more effortput into mitigating future damages (Lindell and Prater,ody, 2003).of acres: Previous research has found that larger devel-sites (in terms of acres) offer project designers more

es.

Pedestrian orientation Mixed land uses

+ ++ ++ +


opportunity and exibility for protecting hydrologically sensi-tive areas (MacDonald, 2004). Thus, larger sitesmay also providemore room and exibility for incorporating hazard mitigationtechniques.

3. Numberexpectedtechniquthat needavoid enOn the otgreater tated withmay be tniques, a

4. Project logreenelers morMacDonawill haveniques.

5. FloodplaihavemixAs oodpneed foroodplaidesigner

3. Researc

3.1. Sample

To exammitigationopments, wtypes of devused the Nedevelopmewere comp2003. (Wethe plannina large degthus the inWe intersecoodplain bMaps (FIRMinterviewinresults, weexposed todevelopmeor under coderivation o

Next, wments by cgovernmenof the 114identify a cof all conveUrbanist deof oodplaihousing ungreeneldas an openurban devedevelopedexisting urbunderdevel

construction of all New Urbanist developments began duringthe mid-1980s or later, all conventional developments had to becompleted by 1985 or later. Paired-samples t-tests for mean dif-ferences across matched pairs revealed no signicant differences

0) fointe

lidityUr

chniqewh

extentmee (erbanticstionticswa mast deners

ir jurthatovernnal da preovered pst deed coesigstionvelopterised patopat proandareheckred cple

rbanich wthe

0-yet wasfor rell/redinedrbanelopmta forwereeld lin S

epend

depd haareaaterl moctura sepmitiof dwelling units: The number of dwelling units can betohavemixed inuenceon theuseofhazardmitigationes. On one hand, the greater number of dwelling unitss to be accommodated, the more difcult it will be to

vironmentally sensitive portions of development sites.her hand, the greater the number of dwelling units, thehe stormwater impacts and the greater the risk associ-the project, and thus the more likely project designers

o incorporate BMPs, stream channel modication tech-nd structural protection techniques.cation: Compared to redevelopment/inll locations,d locations have been found to provide project design-e exibility in project design (Berke et al., 2003;ld, 2004). This may suggest that greeneld projectsmore exibility to incorporate hazard mitigation tech-

n exposure: Floodplain exposure can be expected toed inuenceon theuseofhazardmitigation techniques.lain exposure increases, so might ood risk and theood hazard mitigation techniques. However, largern exposure may also make it more difcult for projects to avoid environmentally sensitive areas

h design, variables, and methods

and data collection

ine and compare the incorporation of ood hazardtechniques into New Urbanist and conventional devel-e developed a national sample of matched pairs of bothelopments that are at risk from ood hazards. We rstw Urban News (December 2003) list of New Urbanist

nts to identify 318 New Urbanist developments thatleted or well into the construction stage as of Decemberomitted 329 New Urbanist developments that were ing or groundbreaking stage, as these projects involveree of uncertainty regarding nal project design andcorporation of ood hazard mitigation techniques).ted the boundaries of the 318 projects with 100-yearoundaries identied on digitized Flood Insurance Rates). Through this geocoding procedure and telephoneg with local government planning staff to verifydetermined that 114 New Urbanist developments areood hazards (or roughly 36% of the 318 New Urbanistnts throughout the United States that were completednstruction) (see Song et al., 2009 for further details onf the sample).

e identied a control group of conventional develop-onducting telephone interviews with the lead localt planner charged with development review for eachNew Urbanist developments. We asked planners toonventional development that was most comparablentional developments in their community to the Newvelopment based on several criteria, including percentn exposure of the entire site, acreage, and number ofits. We also matched in terms of project location (i.e.or inll/redevelopment, where greeneld is denedspace site located adjacent to or outside an existing

lopment boundary, inll as an open space or previouslysite that was redeveloped and is located within anan development boundary, and redevelopment as anoped space in a developed zone). Moreover, because

(p


project. Each variable is thus measured in the form of an inte-ger, corresponding to the number of mitigation techniques usedby the development project (see Table A.1 for a description of themitigation techniques included in each category).

3.3. Site des

In orderincorporatedevelopmevariable thaUrbanist (g0). We usedesign featopment, instreet netwactual presUrbanist deNew Urbanclassied as

3.4. Develo

Weutilizprocess varplanning stseparate dregressionpractices inmitigationvariable (foing the presmitigation ttally sensitprogram vaopment magovernmen

We userst variabinvolved inlation. Theassistance scant, rangin

To measthat indicatto natural hdescription

3.5. Commu

Wemeasite level. Tood histornity had beeood damadiscusses thnity when ain establisha natural dihazards modo not staytime periodterized by adisasters, thand Atkissoa communi

The four site level variables relate to project size, location, andoodplain exposure. We measure projects size in terms of numberof acres and number of dwelling units (both divided by 100 to aidinterpretation). We measure project location as either greeneld

l/redrcentar oe-sp

alyti

anaent hpmeressiodepeommapp

a lineepensed a, the uan ph (orstanthatant

ip act. Beces (aionountoccuptiongression m

ssonn a shis aauses inith pctionampquestion tnce.obse

odelshat tons oIn shthatnship

ding

es Ne?

rstd hast antes cst anitigatign variable

to determine whether New Urbanist developmentsmore hazard mitigation techniques than conventional

nts, we include in our regression models a dichotomoust indicates whether the development project is Newiven a value of 1) or conventional (given a value ofthis variable as a proxy for the presence of certainures that are fundamental to New Urbanist devel-cluding relatively high densities, pedestrian-orientedorks and mixed land uses. (We do not measure theence and extent of such design features in the Newvelopments in our sample: rather, we have relied onNews to determine whether each project should beNew Urbanist or not).

pment review process variables

e the following three categories of development reviewiables: (1) development management program; (2)aff capacity; and (3) public participation. We use aevelopment management program variable in eachmodel, to ensure that the development managementcluded in each regression model relate to the hazardtechniques associated with the respective dependentr example, development management practices requir-ervation of environmentally sensitive areas and hazardechniques relating to the preservation of environmen-ive areas). We create the development managementriables by counting the number of relevant devel-nagement practices ofcially adopted by the localt that relate to the respective dependent variable.two variables to measure planning staff capacity. Thele is the number of planning staff members typicallydevelopment review, per 100,000 community popu-second variable is a count of the number of technicalervices provided by the planning agency to the appli-g from zero to six.ure public participation, we use a dichotomous variablees whether or not citizens raised issues with respectazards during development review (see Table A.2 for aof development review process variables).

nity and site-specic variables

sure one variable at the community level and four at thehe community level variable relates to a communitysy, measured in terms of whether or not the commu-n declared eligible for federal disaster aid as a result of

ges between the years 1995 and 2005. Birkland (1997)e notion of windows,which aremoments of opportu-problem has become urgent enough to inspire changesed societal practices. Such windows can open up aftersaster and encourage communities to address naturalre proactively in the future, but such windows typicallyopen for long (Berke and Campanella, 2006). While theimmediately following a hazard event is often charac-relatively high level of public interest in averting futureis level of interest tends to decline with time (Petakn, 1982). For this reason, we limit our measurement oftys recent ood history to a period of 10 years.

or inlthe pe100-yeand sit

3.6. An

Theto presdeveloof regof ourmost cis onlyUsingical) dare biawordsables ctoo higand/orabilitysignictionshdrawnvariablregresseling ceventassumson reregress

Poitions i219). Tple becfeaturemon wjurisdi(for extechnimitigadiffereamongsionmsense tviolatip. 86).abilityrelatio

4. Fin

4.1. Dopractic

Ourof ooUrbaniillustraUrbaniard mevelopment. Lastly, we measure oodplain exposure asage of the project footprint that is located inside theodplain (see Table A.3 for a description of community

ecic variables).

cal techniques

lytical techniques we use include descriptive statisticsazard mitigation scores for individual matched pairs ofnt projects and multiple regression analysis. Our choicen model techniques is based on the measurementndent variables. Linear regression, while perhaps theonly used regression technique in the social sciences,ropriate for dependent variables that are continuous.ar regression model with a non-continuous (categor-dent variable can produce regression coefcients thatnd inconsistent (Long and Freese, 2006, p. 4). In otherse of linear regressionwith categorical dependent vari-roduce regression coefcients that are systematicallytoo low) in relation to the population parameter valuedard errors that are inated, which increases the prob-a researcher will fail to reject a null hypothesis of norelationship between two variables when such a rela-ually exists in the population from which a sample wasause our dependent variables are measured as countnd are thus categorical), we choose to use the Poissontechnique, which is an appropriate technique for mod-variables that measure how many times a particular

rs. Poisson regression does not require that all of thes of linear regression be met. Among other things, Pois-ion does not require that the dependent variable in aodel be continuous.

regression models do assume, however, that observa-ample are independent of one another (Long, 1997, p.ssumption may not hold for observations in our sam-development projects within a matched pair may have

common with each other that they do not have in com-rojects from other matched pairs, reviewed in others. This means that projects within a matched pair mayle) incorporate similar numbers of hazard mitigation, with the intra-pair difference in numbers of hazardechniques being smaller on average than the inter-pairTo address this issue of possible lack of independencervations, we use robust standard errors in our regres-. Robust standards errors are considered robust in thehey produce correct standard errors in the presence off the assumptions of the model (Long and Freese, 2006,ort, the use of robust standard errors increases the prob-researcherswill draw correct conclusions regarding thes between variables.

s

ew Urbanist design support ood hazard mitigation in

set of ndings involves a comparison of the numberzard mitigation techniques incorporated into the Newd conventional developments in our sample. Table 2omparisons of the scores for each matched pair of Newd conventional developments for each of the four haz-ion categories. The table shows that the mean score


Table 2Comparison of hazard mitigation scores between matched pairs of New Urbanist and conventional developments.a.

NUD name Sensitive areaprotection (max=10)

BMPs (max=5) Stream channelmodication (max=3)

Structuralprotection (max=5)

ArizonaCivano 2 (

FloridaCagan Cro 5 (Carillon To 4 (Lakeside V 4 (Village of 5 (Northlake 4 (Pembroke 3 (Post Harbo 5 (Summervi 4 (

GeorgiaSmyrna To 2 (

MarylandClarksburgSunset IslaWaterView

MichiganCherry Hil

North CaroliAyersleyCarpenterMeadowmMorrisonSouthern VSpring Bro

OregonFairview VTwin Cree

PennsylvaniLantern HWeatherst

South CaroliBaxterIOn

TennesseeWesthave

TexasCraig RancHighland PPlum CreeTurtle Cre

VirginiaPrince Wil

WisconsinSmiths CrDifference

a The rst v* Signicant

** Signicant

for New Urconventioning BMPs, sThere is notally sensitaverage, Neard mitigat

4.2. Whichmitigation t

The ndUrbanist de4 (6)

ssings 6 (6)wn Center 1 (5)illage 4 (3)Bridgewater 4 (3)Park 8 (3)Neighborhood 6 (0)ur Place 6 (6)lle 6 (6)

wn Center 5 (9)Town Center 8 (9) 2 (nd 4 (5) 4 (

0 (5) 1 (

l Village 3 (3) 5 (

na5 (7) 2 (

Village 7 (7) 2 (ont 9 (9) 2 (

Plantation 3 (6) 3 (illage 9 (9) 2 (ok Meadows 3 (4) 2 (

illage 10 (8) 5 (ks 9 (6) 5 (

aill 9 (0) 2one 8 (8) 2 (

na9 (9) 2 (8 (8) 1 (

n 8 (8) 2 (

h 5 (8) 4 (ark 9 (4) 3 (k 2 (9) 1 (ek Village 6 (9) 2 (

liam County Center 4 (6) 2 (

ossing 8 (7) 4 (in mean scores for New Urbanist and conventional (0.2)

alue is the New Urbanist score; the second value in parenthesis is the matched conventioat p


Table 3Regression analyses that explain variation in the use of hazard mitigation techniques.

Variable name Sensitive area protection BMPs Stream channel modication Structural protection

IRRa IRR IRR IRR

Site designDevelopmenNumber of pTechnical asCitizen issueCommunityNumber of aNumber of dProject locatFloodplain econstant

2

(df)N

a An inciden -unit ithe dependent n thevariable count endenindicate a neg

b The develo mentvariable.

* Signicant** Signicant

*** Signicant

The mosthat site deof any cateother factogory of miticonventionthe uniquecally amenaare not outto the incorour ndingNew Urbanmore attentprocess thaopments. GUrbanist dement manalocal governal., 2009), aNew Urbanof hazard mpublic partimunity andthat the dibetween Nearenot the rist design fe

The proning agencthat explaitional devetechniques.associatedtion techniqincreases wvided. Oneassistance mknowledgetechniquesinterpretati

ancefor aand Mremresica

quesrs inensitBMPremes. Cnd sthe

pected byt bePs o0.96 1.03t managementb 1.01 1.09lanners 1.03* 0.94*

sistance services 1.09 1.26**

s 1.10 1.02ood history 0.90 0.81**

cres 0.99 0.99welling units 1.00 1.01***

ion 1.30 0.97xposure 1.00 1.01**

3.83*** 1.12

32.33*** 138.85***

10 1061 61

ce rate ratio (IRR) of 1.50, for example, should be interpreted to mean that a onevariable count by 50%. Conversely, an IRR of 0.50 means that a one-unit increase iby 50%. IRRs greater than 1.00 indicate a positive relationship between the indep

ative relationship; IRRs equal to 1.00 indicate no relationship.pment management variable is different for each model, and includes only develop

at p


explained in terms of risk perception and response. As oodplainexposure increases, so does the likelihood that built structures onthe site will be damaged by future ooding. Project designers maybe responding to this elevated risk by employing additional mea-sures to pro

5. Discussi

5.1. Implicamitigation

In priordesign onthat local cUrbanist deNew Urbantechniques.examiningist and convthe developcharacterist

Our ndand convenin the incoring that thehas not necshow that pand convenof hazard mproject siteof effort puparticular,themselveshave not gement plannduring the dthat hazardpartuponsuassociated wmore critica

One poshave typicaance of locNew Urbanards. Whileprinciples scharacter, ssupported(Berkeet al.attention ofguidelines wand mitigatings suggesbeen unlikedo so by lowelfare of t

Recent rever, thatNto natural h2005, Duanbe the rstment entitl9.0 (the momain thrusard areas fr

Urbanist plans (i.e. Regional Scale, NewCommunity Scale, and InllCommunityScale), the standards call forprojectdesigners to relyonoodplain maps, zoning information, and site conditions to desig-nate areas in the project site that arewithin the 100-year oodplain

at shated, forZybeendian pst defrom

) reqemence Pelevardmlly bted indditiand

lievemit

f Neore noodise placan pnces. Lonce pistanng strkinrevioresoBrodcreasnd m

mitat

orewitat

reseaourts lopmeur s

velopt to ee Neto be. Wht sigtionnev

t is porbanedever outhat

ts.ond,elopmng wtect structures from ood damages.

on

tions for planning practice and natural hazard

comparative analyses of the effects of New Urbanistnatural hazard mitigation, Berke et al. (2009) foundommunities put forth more effort into reviewing Newvelopments than conventional developments, and thatist developments incorporated more hazard mitigationIn this paper, we expanded upon previous ndings by

the use of hazard mitigation techniques in New Urban-entional developments while controlling for factors inment review process and community and site-specicics.ings suggest that the distinction between New Urbanisttional developments has not made a direct differenceporation of ood hazard mitigation techniques, mean-potential of New Urbanist design for hazard mitigationessarily translated into increased hazard resilience. Wereviously observed differences between New Urbanisttional developments with respect to the incorporationitigation techniques stem not from the difference indesign per s, but rather from differences in the levelt forth into reviewing the development projects. In

we show that New Urbanist design features that lendto the incorporation of hazard mitigation techniquesnerally been taken advantage of unless local govern-ers provide technical assistance to project designersevelopment review process. While it may also be truemitigation in conventional projectswill depend in largechbehaviors fromplanners, theheightened levelof riskith dense New Urbanist developments makes it evenl that natural hazard risks be addressed.sible explanation for why New Urbanist developmentslly not responded to hazard risks without the guid-al government planners is the lack of attention thatist design codes have historically paid to natural haz-the widely publicized New Urbanist model codes andupport goals relating to such concerns as communityense of place, and pedestrian movement, they have notland use and design standards for hazard mitigation, 2009).As a result, these codeshaveneitherdirected theproject designers to natural hazard issues nor providedith respect to how natural hazards can be addressed

ed through project design. It follows then, as our nd-t, that designers of New Urbanist projects may havely to address natural hazard issues unless directed tocal government entities responsible for protecting thehe public.evisions to New Urbanist design codes suggest, how-ewUrbanist designersmaybeginpayingmore attentionazard issues than they have in the past. In October ofy Plater-Zyberk and Company released what appear toNew Urbanist hazard mitigation standards, in a docu-ed SmartCode Modules that supplements SmartCodedel development code for New Urbanist design). The

t of the standards relates to keeping certain ood haz-ee from development. In three different scales of New

and thdesignSpacePlater-

Depdards cUrbaniresults(FEMAmanagInsuration isthehazoreticais loca

In awrittenalso behazardning othatmfrom land umentsassistaprocesassistacal assplanniand wouponpity and1994;that inplans a

5.2. Li

Beftain limfuturewhichopmendevelonot in othe derespecthat thlikelysamplewas nomitigashouldSince iNew Uinll/rwhethmentsprojec

Secist devindictiould be kept free from development. When areas arein thismanner, they are to bedeclaredpermanent Civicuse only as parks, open space, and/or recreation (Duanyrk and Company, 2008).ng on whether and how they are used, these stan-otentially provide an additional level of safety for Newvelopments located inside the oodplain beyond what

minimum Federal Emergency Management Agencyuirements. Whereas FEMA regulations for oodplainnt in communities that participate in the National Floodrogram primarily focus on ensuring that new construc-ated to the elevation associated with a 100-year ood,

itigation standards included in theSmartCode can the-e used to reduce the extent to which new constructionside the oodplain in the rst place.on to making sure that New Urbanist design codes areimplemented so as to prioritize hazard mitigation, wethat local governments should take a stronger role in

igation by paying greater attention to proactive plan-w Urbanist developments. Prior research has shownon-structuralmitigation techniques are used and lossesng are lower in communities that adopt comprehensivens (Burby, 2006). Our ndings show that local govern-romote hazard mitigation through providing technical

to project developers during the development reviewcal governments can possibly improve their technicalrovision by standardizing the process by which techni-ce services are provided and by ensuring that sufcientaff resources exist for reviewing development projectsg with project applicants. This recommendation buildsus studies on the importance of planning agency capac-urces for landuse planning activities (Burby andDalton,y et al., 2006). In general, previous studies have founded planning agency capacity fosters stronger land useore successful plan implementation.

ions and suggestions for future research

e conclude,we consider it prudent to acknowledge cer-ions in our study and to provide suggestions for howrchers can build upon our ndings. First, the extent to

ndings can be generalized to all New Urbanist devel-cated in oodplains is limited to the extent that thents in our sample are representative of developmentsample.Wewere not able tomake comparisons betweenments in our sample and those not in our sample withvery potentially relevant characteristic, but we did ndw Urbanist developments in our sample were morelocated in greeneld locations than those not in ourile our regression models showed that project locationnicantly associated with any category of ood hazardtechniques, our ndings and the conclusions we drawertheless be interpreted with this difference in mind.ssible that ourndings aremoregenerally applicable toist developments located in greeneld locations than inelopment locations, future researchers should examiner ndings hold in samples of New Urbanist develop-include a greater percentage of inll/redevelopment

the site design variable we use to compare New Urban-entswith conventional developments is dichotomous,

hether a project is New Urbanist or conventional. In


reality, there is likely to be variation in design features within oursample of New Urbanist developments, such that some may morefaithfully represent New Urbanist design principles than others.(Theremay also be variation indesign features among conventionaldevelopments, such that some may be more or less conventionalthanothers.) Future research canpossibly improveourunderstand-ing of the differences between New Urbanist and conventionaldevelopments in the use of hazard mitigation techniques by utiliz-ingmore detailedmeasures of site design that account for variationin design within each of the two site design types. Among otherthings, this approachwould enable for the effects of particular NewUrbanist design features to be isolated and examined individu-ally, in order to determine which features (if any) actually makea positive contribution to the incorporation of hazard mitigationtechniques.

Future research should also evaluate whether and how therecently developed New Urbanist hazard mitigation standards areused in New Urbanist developments and whether they make apositive contribution to public safety by reducing the incidenceof development inside the oodplain. The New Urbanist develop-ments in our sample were all designed prior to 2004, meaning thatthe New Urbanist hazard mitigation standards that were adoptedin 2005 did not inuence the use of hazard mitigation techniquesin the projects in our sample. Future studies should focus on NewUrbanist developments located inside the oodplain that weredesigned after the hazard mitigation standards were adopted, toexamine the inuence that the standards have on project designwith respec

Lastly, fmeasuremeof whetherinside a 100site that wacondent tsome degrelimitationssuch limitatyear oodpooding thaand Doyle,in our sampmay nevertinside and o

even be accurate, as the precise location of the oodplain bound-ary is ultimately uncertain (due to data limitations, the stochasticnature of oods, and possible errors in model assumptions) (ibid).Furthermore, the 100-year oodplain designation does not distin-guish among different types of ooding (e.g. uvial, pluvial, andcoastal), despite the possibility that different types of ooding canhave varying associated risk levels. Future studies can thus improveupon our research by utilizing measures of ood risk that accountfor these deciencies in our measures.

6. Conclusion

Relative to conventional low-density development, NewUrban-ist design presents exciting opportunities for reducing naturalhazard-related losses. Yet, New Urbanist design also raises con-cerns pertaining to the placement of high-density development inhazardous areas. Our ndings suggest that New Urbanist design-ers are generally not taking full advantage of the potential for NewUrbanist design to reduce hazard risk. In the absence of techni-cal assistance from local government planning staff that mightdirect attention to natural hazard issues and the use of ood hazardmitigation techniques, the design of New Urbanist developmentsappears to be no different from that of conventional low-densitydevelopments with respect to ood hazard mitigation.

We believe, however, that this outcome is not inevitable. IfNewUrbanist designers can continue to improve upon natural haz-ard mitigation standards in their design codes and ensure that

idelin, andare ierest dehat a

wled

ndNatipiniosedarilye graents.t to ood hazard mitigation.uture studies can possibly utilize more sophisticatednts of ood risk. We measured ood risk both in termsa development project was located at least partially-year oodplain area, and the percentage of the projects located inside the 100-year oodplain. While we arehat these measures do in fact reect ood risks withe of accuracy, we acknowledge that they have certainthat can possibly be addressed in future research. Oneion relates to the fact that land slightly outside the 100-lain boundary is only slightly less likely to experiencen landslightly inside the100-yearoodplain (Patterson2009). Thus, portions of development sites for projectsle that are technically outside the 100-year oodplainheless be at risk. In addition, the distinction betweenutside the 100-year oodplain may in some cases not

the gumentsissuesthen thUrbaninities t

Ackno

Theby theAny oexpresnecessWe arcommes are consistently implemented into future develop-if local communities can make sure that natural hazarddentied and addressed during development review,are reasons to be optimistic that the promise of Newsign can be translated into neighborhoods and commu-re more resilient to natural disasters.

gements

ings reported in this paper are part of a study supportedonal Science Foundation (NSF Grant # CMS-0407720).ns, ndings, and conclusions or recommendationsin this paper are those of the authors and do notreect the views of the National Science Foundation.teful to four anonymous reviewers for their helpful


Appendix A.

Tables A.1A.3.

Table A.1Categories of h

Technique c

Sensitive are

Stormwater

Stream chan

Structural pr

Table A.2Development

Variable

Developmenarea prote

Developmenstormwate

Developmenchannel m

Developmenstructural

Number of p

Technical as

Citizen issue

a Percent Yazard mitigation techniques.

ategory Techniques Mean Standard deviation Min Max n

a protection 1. Minimized ll in wetlands 6.0 2.5 0 9 652. Minimized grading in wetlands3. Maintained wetland vegetation buffers4. Minimized ll in oodplain5. Minimized grading in oodplain6. Maintained oodplain vegetation buffers7. Protected topsoil during construction8. Preserved natural drainage systems9. Restored natural contours on site after construction10. Reforestation to stabilize landslide prone slopes

BMPs 1. Excavated ponds to provide ood storage 2.8 1.3 0 5 652. Constructed wetlands3. Used detention/retention basins4. Provided compensatory ood storage5. Used erosion/sediment control devices

nel modication 1. Deepened, widened, and/or lined streams 0.9 1.0 0 3 652. Stabilized banks3. Cleared debris and/or obstructions in streams

otection 1. Raised elevation of buildings 1.3 1.1 0 3 652. Added ll to raise elevation of roads3. Additional structural strengthening for buildings4. Built ood control dams on streams5. Built oodwalls and/or levees along streams

review process variables.

Description Mean Standarddeviation

Min Max n

t management: sensitivection

1. Prohibition of all development in oodways 6.5 2.3 2 12 63

2. Prohibition of residential development in oodplains3. Low-density zoning in oodplain areas4. Prohibition on extending water and sewer to serve development inoodplain areas5. Land acquisition program to acquire oodplain areas6. Land bank program for oodplain areas7. Accept dedication of conservation easements8. Require river, stream, oodway, wetland, and/or oodplain buffers9. Require greenways10. Require setbacks11. Reduced taxation on undeveloped land to maintain open space12. Government policy not to locate public facilities in oodplain areas13. Require dedication and/or preservation of open space

t management:r BMPs

1. Require compensatory ood storage for oodwaters displaced bydevelopment

1.4 0.6 0 2 63

2. Require on-site stormwater retention

t management: streamodication

1. Prohibition of all development in oodways 1.3 0.7 0 2 63

2. Require river, stream, oodway, wetland, and/or oodplain buffers

t management:protection

1. Mandatory oodproong of nonresidential structures in oodplains 1.2 0.8 0 2 63

2. Require additional freeboard beyond elevation required by NationalFlood Insurance Program

lanners The number of planning staff members typically involved in thedevelopment review process, per 100,000 community population

2.2 2.3 0.2 9.6 63

sistance services 1. One-on-one technical assistance during plan reviews 3.1 0.9 0 5 652. Predevelopment conference3. Checklist of items to be included on site plan4. Workshops to explain code provisions5. Newsletters/bulletins6. Audio/video tapes

s Whether or not citizens raised issues with respect to natural hazardsduring development review

21.5a 65

es responses.


Table A.3Community and site-specic variables.

Variable Description Mean Standarddeviation

Min Max n

Community ood history Whether the community had been declared eligible for federal disaster aidas a result of ooding from 1995 to 2005

37.7a 61

Number of acres Size of development site in acres, in hundreds 5.4 7.4 0.1 40.0 64Number of dwelling units Number of dwelling units approved in development project, in hundreds 12.3 17.2 0.3 89.0 65Project location Location of development site within local community: greeneld or 87.7b 65

Floodplain e

a Percent Yb Percent g

References

Berke, P.R., Ba2006.Whaand imple

Berke, P.R., CaAcad. Polit

Berke, P.R., McdevelopmJ. Am. Plan

Berke, P.R., Roplanning m7996.

Berke, P.R., SoUrban and

Birkland, T., 19Georgetow

Bosher, L., Carrtainable buDisasters 3

Brody, S.D., 20sis of plan191201.

Brody, S.D., Csprawl red

Brody, S.D., Hiof local jurPlann. 69,

Brown, B., Crouating psy

Burby, R.J., 20action. J. A

Burby, R.J., 200icy: bringiAcad. Polit

Burby, R.J., Dastate plannAdmin. Re

Burby, R.J., Mathe comm

Burby, R.J., MaGovernmeUniversity

Burby, R.J.,Mayand centra

Burby, R.J., Machoices an

Calthorpe, P.,Island Pres

Campbell, D.T.Research.

Crane, R., 1996tional plan

Cutter, S., 2001ters. Josep

Downs, A., 200Assoc. 71,

Duany, A., Platthe Declin

Duany Plater-and Comp

Federal Emergplain man

Emergplainronme04. Nerties05. N. GAO

lk, D.Rgation

lk, D.Rgationn. Mancy Fe: oo.K., P

pariso., 1997E Publ., Freeg Statald, J.At Siteel Hil, J.M.hingto.S., 19ed Sta.S., Gnitedge 10

J., 20Towngn for

., 19. Amerky, R.BEd.), Cning f202.n, L.Augh sur. As.J., Atcipati, Berks: a reasters 2., 2005tledgeon, R.,

Stevivers

w an aniversinll/redevelopmentxposure Percentage of development site located inside the oodplain

es responses.reeneld.

ckhurst, M., Day, M., Ericksen, N., Laurian, L., Crawford, J., Dixon, J.,tmakes plan implementation successful? An evaluation of local plansmentation practices in New Zealand. Environ. Plann. B 33, 581600.mpanella, T.J., 2006. Planning for postdisaster resiliency. Ann. Am.. SS 604, 192207.Donald, J., White, N., Holmes, M., Oury, K., Ryznar, R., 2003. Greeningent forwatershed protection: does NewUrbanismmake a difference?n. Assoc. 69, 397413.enigk, D.J., 1996. Enhancing plan quality: evaluating the role of stateandates for natural hazard mitigation. J. Environ. Plann. Manage. 39,

ng, Y., Stevens, M.R., 2009. Integrating hazard mitigation into Newconventional developments. J. Plann. Educ. Res. 28, 441455.97. After Disaster: Agenda Setting, Public Policy, and Focusing Events.n University Press, Washington, DC.illo, P., Dainty, A., Glass, J., Price, A., 2007. Realising a resilient and sus-ilt environment: towards a strategic agenda for the United Kingdom.1, 236255.03. Are we learning to make better plans? A longitudinal analy-quality associated with natural hazards. J. Plann. Educ. Res. 23,

arrasco, V., Higheld, W.E., 2006. Measuring the adoption of localuction planning policies in Florida. J. Plann. Educ. Res. 25, 294310.gheld, W., Carrasco, V., 2004. Measuring the collective capabilitiesisdictions to manage ecosystems in southern Florida. Landsc. Urban3350.pper, V., 2001. New urban and standard suburban subdivisions: eval-chological and social goals. J. Am. Plann. Assoc. 67, 402419.03. Making plans that matter: citizen involvement and governmentm. Plann. Assoc. 69, 3349.6. Hurricane Katrina and the paradoxes of government disaster pol-ng about wise government decisions for hazardous areas. Ann. Am.. SS 604, 171191.lton, L.C., 1994. Plans can matter! The role of land use plans anding mandates in limiting the development of hazardous areas. Publicv. 54, 229238.y, P.J., 1998. Intergovernmental environmental planning: addressingitment conundrum. J. Environ. Plann. Manage. 41, 95110.y, P.J., Berke, P.R., Dalton, L.C., French, S.P., Kaiser, E.J., 1997. Makingnts Plan: State Experiments inManaging LandUse. The JohnsHopkinsPress, Baltimore/London., P.J.,Malizia, E.E., Levine, J., 2000. Building codeenforcementburdensl city decline. J. Am. Plann. Assoc. 66, 143161.y, P.J., Paterson, R.C., 1998. Improving compliance with regulations:d outcomes for local government. J. Am. Plann. Assoc. 64, 324334.Fulton, W., 2001. The Regional City: Planning for the End of Sprawl.s, Washington, DC., Stanley, J.C., 1963. Experimental andQuasi-Experimental Designs forRand McNalley College Publishing Company, Chicago.. Cars and drivers in the suburbs: linking access to travel in neotradi-ning. J. Am. Plann. Assoc. 62, 5165.. American Hazardscapes: The Regionalization of Hazards and Disas-h Henry Press, Washington, DC.5. Smart growth: why we discuss it more than we do it. J. Am. Plann.

Federaloodenvi

GAO, 20prop

GAO, 20claim

GodschaMitiDC.

GodschamitiPlan

Interageleng

Lindell, Mcom

Long, J.SSAG

Long, J.SUsin

MacDonmenChap

McElshWas

Mileti, DUnit

Mileti, Dthe UChan

Miller,Thedesi

Morris, Mvice

OlshansR.J. (Plan167

PattersothroReso

Petak, WAnti

Song, Y.tionDisa

Talen, ERou

Thomps

Dr.Markat the UnHe is noat the U367378.er-Zyberk, E., Speck, J., 2000. SuburbanNation: The Rise of Sprawl ande of the American Dream. North Point Press, New York.Zyberk and Company, 2008. SmartCode v 9.0. Duany Plater-Zyberkany.encyManagementAgency, 1986.Auniednationalprogramforood-agement. FEMA Report #100. FEMA, Washington, DC.

Dr. Phil Berkefessor in the DCarolina-Chap

Dr. Yan SongPlanning at th27.1 33.8 0.2 100.0 65

ency Management Agency, 2002. The natural benecial functions ofs: reducing ood losses by protecting and restoring the oodplainnt. FEMA Report #409. FEMA, Washington, DC.

ational Flood Insurance Program: actions to address repetitive loss. GAO Report GAO-04-401T. Washington, DC.ational Flood Insurance Program: oversight of policy issuance andReport Gao-05-532T. Washington, DC.., Beatley, T., Berke, P.R., Brower, D., Kaiser, E., 1999. Natural Hazard: Recasting Disaster Policy and Planning. Island Press, Washington,

., Brody, S., Burby, R.J., 2003. Public participation in natural hazardpolicy formation: challenges for comprehensive planning. J. Environ.

nage. 46, 733754.loodplain Management Review Committee, 1994. Sharing the chal-dplain management into the 21st century. Washington, DC.rater, C., 2000. Household adoption of seismic hazard adjustments: an of residents in two states. Int. J. Mass Emerg. Disasters 18, 317338.. RegressionModels for Categorical and LimitedDependent Variables.ications, Thousand Oaks, London/New Delhi.se, J., 2006. Regression Models for Categorical Dependent Variablesa. Stata Press, College Station, TX.., 2004. Predicting Watershed Protectiveness of Approved Develop-Plans. Department of City and Regional Planning, UNC-Chapel Hill,l, Doctor of Philosophy., 2004. Nature-Friendly Ordinances. Environmental Law Institute,n, DC.99. Disasters by Design: A Reassessment of Natural Hazards in thetes. Joseph Henry Press, Washington, DC.ailus, J.L., 2005. Sustainable development and hazards mitigation inStates: disasters by design revisited. Mitigation Adapt. Strat. Global, 491504.07. Design for disaster: Is the New Urbanism indestructible?Paper, 9, MarchApril (http://www.tndtownpaper.com/Volume9/disaster.htm, accessed November 5, 2008).

97. Subdivision Design in Flood Hazard Areas. Planning Advisory Ser-ican Planning Association, Chicago.., Kartez, J.D., 1998. Managing land use to build resilience. In: Burby,ooperating with Nature: Confronting Natural Hazards with Land-Useor Sustainable Communities. JosephHenry Press,Washington, DC, pp.

., Doyle, M.W., 2009. Assessing effectiveness of national ood policypatiotemporal monitoring of socioeconomic exposure. J. Am. Waterssoc. 45, 237252.kisson, A.A., 1982. Natural Hazard Risk Assessment and Public Policy:ng the Unexpected. Springer-Verlag, New York/Heidelberg/Berlin.e, P.R., Stevens, M.R., 2009. Smart developments in dangerous loca-lity check of existing New Urbanist developments. Int. J. Mass Emerg.7, 124.. New Urbanism and American Planning: The Conict of Cultures.

, London.2005. Mississippi tackles a tough one. Planning Magazine, 611.

enswas an instructor in theDepartment of City andRegional Planningity of North Carolina-Chapel Hill at the time this studywas conducted.ssistant professor in the School of Community and Regional Planningity of British Columbia.is the Deputy Director of the Institute for the Environment and pro-epartment of City and Regional Planning at the University of Northel Hill.

is an associate professor in the Department of City and Regionale University of North Carolina-Chapel Hill.

Creating disaster-resilient communities: Evaluating the promise and performance of new urbanismIntroductionDesigning neighborhoods and communities that are resilient to natural hazardsThe potential advantages of New Urbanist design for hazard mitigation: in conceptOther factors that might influence the use of flood hazard mitigation techniques

Research design, variables, and methodsSample and data collectionDependent variablesSite design variableDevelopment review process variablesCommunity and site-specific variablesAnalytical techniques

FindingsDoes New Urbanist design support flood hazard mitigation in practice?Which factors explain differences in the use of flood hazard mitigation techniques?

DiscussionImplications for planning practice and natural hazard mitigationLimitations and suggestions for future research

ConclusionAcknowledgementsAppendix AReferences

landscape and urban planning, volume 94, issue 2, 28 february 2010, pages 105–115,

Documents