Landscape improvement impacts on roadside safety in Texas · Landscape and Urban Planning 78 (2006)...

A

rbpaw

aarsatra©

K

f

f

hj

0

Landscape and Urban Planning 78 (2006) 263–274

Landscape improvement impacts on roadside safety in Texas

Jeong-Hun Mok a,∗, Harlow C. Landphair b,∗∗, Jody R. Naderi c,1

a Seoul Development Institute, 391 Seocho-dong Seocho-gu, Seoul 137-071, Republic of Koreab Texas Transportation Institute, 2929 Gilchrist Building, Room 128A, Texas A&M University, College Station, TX 77843-3135, USA

c Department of Landscape Architecture and Urban Planning, Texas A&M University, College Station, TX 77843-3135, USA

Received 12 August 2003; received in revised form 4 March 2005; accepted 16 September 2005Available online 28 November 2005

bstract

Environmental psychologists suggest that appropriately landscaped roadside scenes may have a reducing influence on travel-elated stress or may improve attention, yet there is very little data available that establishes the nature of the relationshipetween roadside landscaping and driver safety. Traditional transportation researchers suggest that aesthetic enhancements are aroblematic component of the roadside landscape because of the severity of vehicle/tree collisions and a perception that roadsideesthetics can distract the driver causing safety risk. Costly planning processes arise as members of the local communities debateith public utility and transportation management staff on the subject of appropriate roadside landscaping.To test the effect of landscape improvements on driver performance, this study used a comparison of before-and-after crashes

s a quantitative measure of roadside greening. Researchers examined 61 road sections in Texas that were landscape designeds either urban arterials or state highways. The hypothesis tested was to determine whether landscape-improved sections of theoadway were safer compared to the same road section before landscape improvements at 10 sites were very well controlled astudy sites. The findings of this study show a significant decrease in crash rate after landscape improvements were implemented
t the 95% confidence level on 10 urban arterial or highway sites in Texas. The contribution of this study is to further investigatehe effect that landscape features are having on driver behavior which appear to be associated with positive changes in safetyesult from design. However, these findings need further research to verify a relationship between driver’s visual perceptionccording to travelway corridor landscape treatments.
2005 Elsevier B.V. All rights reserved.

eywords: Roadside landscape; Streetscape enhancements; Visual percepti

∗ Corresponding author. Tel.: +82 02 2149 1047;ax: +82 02 2149 1060.∗∗ Co-corresponding author. Tel.: +1 979 845 7871;ax: +1 979 862 1759.

E-mail addresses: [email protected] (J.-H. Mok),[email protected] (H.C. Landphair),[email protected] (J.R. Naderi).

1 Tel.: +1 979 458 0103.

1

Tnmr

169-2046/$ – see front matter © 2005 Elsevier B.V. All rights reserved.doi:10.1016/j.landurbplan.2005.09.002

on; Crash rate; Safety

. Introduction

The American Association of State Highway and
ransportation Officials (AASHTO), has long recog-ized that the proper landscape and aesthetic develop-ent of urban streets provides a desirable touch of natu-
al beauty in a built environment. These improvements

2 d Urba

aoo

hvfiieo

AaUmvfsrtmbstdmeeao“iDaptpptad

mlpeTt

wir

2

ei(teOdwiatdian

2

rpOBiatubB

otffmw

64 J.-H. Mok et al. / Landscape an

re often the means of improving the economic valuesf the areas adjacent to the streets and creating a sensef community identity (AASHTO, 1970).

On the other hand, some AASHTO safety interestsave expressed concern about the potential hazards ofehicle tree collisions. Concerns about collisions withxed roadside objects must be balanced against the pos-

tive factors of visual preference, noise abatement, androsion control achieved by roadside landscape devel-pment (AASHTO, 1984).

The Intermodal Surface Transportation Efficiencyct of 1991 (ISTEA), mandated a more balanced

pproach to transportation system development in thenited States by including considerations of environ-ental, cultural, economic, and social conditions in

ehicular transportation projects. The Act also includedunding set asides for transportation enhancements, andcenic byways both within and outside transportationights-of-way. Landscape and aesthetic improvementso rights-of-way are a significant part of the enhance-

ents program. ISTEA emphasized that, in addition toeing safe and cost effective, projects must fit theirurrounding environments, especially in scenic, his-oric or culturally sensitive areas. Federal planning andesign guidelines published in 1995, stated that designsay take into account: the constructed and natural

nvironment of the area; impacts of the project uponnvironmental, scenic, aesthetic, historic, community,nd preservation interests; and access for other modesf transportation. Then in 1997, FHWA publishedFlexibility in Highway Design” which has evolvednto what has become known as Context Sensitiveesign/Context Sensitive Solutions (CSD/CSS). This

pproach to the project delivery process applies therinciples of flexibility in delivering new transporta-ion projects or reconstructing current transportationrojects. In 2001, Texas State Department of Trans-ortation (TxDOT) published its “Landscape and Aes-hetic Design Manual” which provided engineeringnd transportation staff with guidance on the planningesign and implementation of roadside landscapes.

This study attempts to compare the safety perfor-ance of urban arterial road sections before and after

andscape improvements. This before-and-after com-
arison study allows researchers to examine the differ-nce of safety performance by landscape treatments.his study contributes to further investigate the effect
hat landscape features are having on driver behavior

oaba

n Planning 78 (2006) 263–274

hich appear to be associated with positive changesn safety result from within and outside transportationights-of-way design.

. Background and broad concepts of study

Several field studies on the impact of landscapenhancements have demonstrated a variety of positivempacts on communities and traffic safety. Topp’s study1990) of German streetscape enhancements charac-erized by a landscaped center strip, found these to beffective in calming traffic and increased traffic safety.ver the period of Topp’s study (1990), overall acci-ents were reduced by 30%, the number of accidentsith injuries was cut by about 60%, and accidents

nvolving street crossing pedestrians were reduced bybout 80%. In Toronto, Bahar and Naderi (1997) foundhat the frequency and severity of mid-block acci-ents decreased after landscape improvements werenstalled. Mid-block accidents decreased significantlyt all the sites studied while there was an increase in theumber and severity of mid-block accidents city-wide.

.1. Research background of study

Environmental psychologists have developed theo-ies that attempt to explain the relationship betweeneople’s interest and attention to their environment.ne of the better known theories was advanced byerlyne who related attention to the visual complex-

ty of what was seen. Berlyne (1971) suggested thatttention was aroused as visual stimulus increases upo a level of complexity, at which point if visual stim-lus continues to become more complex, subjects willecome confused and lose interest. This is known aserlyne’s “Arousal Theory”.

In a 1976 study, Wohlwill applied Berlyne’s the-ry to landscape aesthetics. Wohlwill hypothesized thathere was an optimal level of stimulus or informationrom the landscape; too much information is stress-ul, and too little information is boring. With this inind, we look at the work of Taylor et al. (1987)ho demonstrated that driving information is mostly
btained from the outside environment. The landscapelong the roadside will contribute to the decisions madey drivers with either positive or negative results. Moknd Landphair’s research (2003) on parkways suggest

d Urba

tewthtiqis

itetpwsoc

speacthddtefhw

2

ssowoAtaf

6dtcbnteoihwnh

2

taradwho

tMslP

•

•

•

•

J.-H. Mok et al. / Landscape an

hat carefully landscaped roadside edges which havenough features to be interesting may be the featurehich makes the drive a more pleasant experience than

he interstate highway option and may also contribute toigher degrees of attentiveness. The landscape levels ofhe parkway are not so complex that they were confus-ng or oppressive. The speculation is that both increaseduality of visual aesthetic and increased attention mayndicate a positive influence on improvement in driverafety.

Nature scenes may have comparatively positivenfluence on driver behavior, since natural settings mayend to have an optimal level of complexity to be inter-sting (Wohlwill, 1976). Topp (1990) also indicatedhat appropriate tree planting and landscaping has asychological effect of reducing driving speed. In otherords, streets characterized by a landscaped center

trip or median planting may alter drivers’ perceptionf lane width and therefore reduce driving speeds asso-iated with increases in severity of accidents.

Our interest here is to measure the impact of road-ide landscape on the functionality of the driver usingrimarily crash results as an indicator of impact. Gen-rally, road capacity is identified by how many vehiclesspecific roadway design can carry in a day over a spe-ific period of time based on the speed and geometry ofhe road. Within this context, many aspects of the roadave an impact on safety and a series of design stan-ards used in roadway construction today have beeneveloped from research on road configuration, sightriangles, construction materials, pavement markings,tc. (AASHTO, 1996). The contribution this study is tourther investigate the effect that landscape features areaving on driver behavior which appear to be associatedith positive changes in safety results from design.

.2. Trees on the roadside

Trees are often cited as the most hazardous road-ide objects by researchers in the field of transportationafety. Trees account for more single-vehicle, fixed-bject fatalities than any other object along the road-ay (Turner and Mansfield, 1990). This study focusedn urban tree collisions in Michigan and Huntsville,
labama. According to the Michigan tree study review,
he major reason for fatal tree collisions was drinkingnd reckless driving. More than 60% of the drivers inatal crashes with trees had been drinking, and over

••

n Planning 78 (2006) 263–274 265

0% of the fatalities were under the age of 35. Malerivers outnumbered female drivers by more than twoo one. In addition, greater than two-thirds of theseollisions occurred on weekends, with the prime timeeing the extremely late hours of Friday and Saturdayights (Zeigler, 1986). The most frequently encoun-ered problem related to tree collisions results fromrrant motorists (AASHTO, 1996). However, the issuef trees on the roadside is often a political and socialssue among community residents, environmentalists,istorical preservationists, and traffic safety engineersho view roadside trees as integral to local commu-ity identity, contributing to economic revival, publicealth, and pedestrian safety.

.3. Roadside landscape design guidelines

It is generally assumed that modern freeway charac-eristics of paved shoulders, concrete median barriersnd extended vegetation clear zones represent safetyelated design elements. On the other hand, parkwaysre characterized by landscaped edges, grassed shoul-ers, vegetated medians, other landscape elementsithin 30 ft. of the edge of the driving lanes, and inarmony with existing nature and surrounding devel-pment.

In case of Texas, Texas Department of Transporta-ion (TxDOT) has a Landscape and Aesthetics Design

anual (TxDOT, 2001) for Texas system roads andtreets to incorporate many safety criteria in the guide-ines for roadside landscape and aesthetic treatments.lanting guidelines used by TxDOT are as follows:

Roadside vegetation should be designed or main-tained to accomplish specific goals of sight-distance,clear view of obstructions, erosion control, and aes-thetics.Plants must not be planted where they may obstructany signs, sightlines, or driver visibility.On frontage roads, allow a minimum of 3 ft. clearspace between the back of curb and any area to bemaintained for maintenance personnel.Plant use in intersection areas must be limited tolow-growing varieties.
Plants must not be placed near merging lanes.Landscape improvements must avoid the creationof unsafe conditions for motorists or maintenancepersonnel (see Fig. 1).

266 J.-H. Mok et al. / Landscape and Urban Planning 78 (2006) 263–274

motori

Toqstwrsctc

3

slaswa

3

olUeaa

pliletefmapf

•

•

•

Fig. 1. Photos of landscape edges for

here are planting guidelines which states have devel-ped that support safer driver performance. What is ofuestion here is whether proper application of plantingtandards is resulting in overall safety improvementshat are being unrecognized. Standards or guidelineshich assist in improving safety need to be identified in

oadside landscaping, Many of these guidelines avoidpecific dimensions because of liability issues asso-iated with discrete dimensions. Research is ongoingo develop more definitive standards for visibility andlearance.

. Methods of study

This research was based on comparisons of theafety performance of selected sites before and afterandscape improvements on major urban freeways andrterial streets. Sites with landscape development wereelected from across Texas and all types of crash dataere collected for the comparisons of sections before

nd after landscape improvements.

.1. Case selections

Before and after landscape improvement sites werebtained from various data source. Firstly obtainedandscape improvement projects were 61 in Texas,
SA. They consist of 39 landscaping/streetscape
nhancement projects along urban roads and city streetsnd 22 landscape development projects along urbanrterial roads in Texas.

ts

sts or maintenance personnel safety.

Thirty nine landscaping/streetscape enhancementrojects of the total 61 projects were as fol-ows: 30 TxDOT ‘Enhancement Projects (Landscap-ng/Streetscapes)’ between the year 1995 and 2000, fiveandscaping projects in City of Austin, three projectsxtracted from ‘Austin Metropolitan Area Transporta-ion Plan–2050 Roadway Plan’ and one streetscapenhancement project in City of Lubbock extractedrom ‘Visual Database of Transportation Enhance-ent’. Only two projects (Amarillo Route 6th street

nd Austin Airport Blvd.) from the 39 enhancementrojects were remained. Thirty-seven were removed asollowing reasons:

Absence of completion year or Control Section Job(CSJ) number of projects: several projects have noinformation to verify ‘landscape construction year’or ‘exact landscape construction location’.Bias in sample of sites: study sites should be treatedby only landscaping treatment but many projectswere treated by other treatments, such as pedestriansidewalk widening, expansion of existing shoulders,or installation of bicycling path.Unsuitable projects in terms of accident data: land-scape construction should be finished between theyear 1987 and 1996 for comparing 3–5 year periodsbefore and after landscape intervention because theTexas accident data is available from the year 1984
to 1999.
Twenty-two landscape development projects of theotal 61 projects were as follows: 12 TxDOT land-cape development projects (only landscaping) along

J.-H. Mok et al. / Landscape and Urban Planning 78 (2006) 263–274 267

Table 1Urban arterial highways and streets before and after landscape improvements in Texas

Location Section locations (average of averageannual daily traffic for 3–5 yearsbefore and after landscapeintervention)

Landscape treatments(section length)

Treatment year(year period)

Urban arterialroad/highway

1. Austin Interchange (Loop 1 & US 183)(before: 35,000/after 19,440)

Interchange landscaping(1 mile)

1993 (1988–1998)

2. Dallas SPUR 303 (from Co. line toMountain Creek Lake) (before:10,895/after 11,873)

Roadside landscaping(5.328 mile)

1992 (1987–1997)

3. Dallas US 75 North (from Loop 635 to Co.line) (before: 138,237/after 142,096)


1992 (1987–1997)

4. McKinney US 380 (from US 75 to SH 5)(before: 7893/after 11,244)

Median landscaping(2.053 mile)

1995 (1991–1999)

5. Plano US 75 (from Plano Pkwy to SpringCreek Dr.) (before: 49,348/after62,265)


1995 (1991–1999)

6. Houston Interchange (IH 10 & IH 45) (before:106,307/after 123,901)

Interchange landscaping(1 mile)

1996 (1993–1999)

7. Lubbock IH 27 (from 58th St. to 82nd St.)(before: 5150/after 5800)


1995 (1991–1999)

8. Odessa BI 20 (from Loop 338 to SH 349)(before: 17,681/after 13,880)


1988 (1984–1992)

City street 9. Austin Airport Blvd. (from IH 35 to ManorRd.) (before: 64,080/after 66,081)

Median landscaping(1.985 mile)

1988 (1984–1992)

10. Amarillo Amarillo Route 6th Street (SL 279)0,367)

Sidewalk improvement 1996 (1993–1999)

u1(Omr

•

•

ptl

dsObacor

3

Tsdid

(before: 9550/after 1

rban arterial roads in Texas between the year 1984 and999 and 10 TxDOT landscape development projectsonly landscaping) between the year 1993 and 1999.nly eight projects from the 22 landscape develop-ent projects were remained. Fourteen projects were

emoved as following reasons:

Thirteen projects were taken out because there wasno the CSJ number or landscape construction year.Projects along minor arterial roads in Texas cannotbe verified by the CSJ number because the CSJ num-ber was assigned for only major arterial roads inTexas.One project in Houston was also taken out becauseit was not coincidence with the location of majorlandscape construction indicated by the CSJ numberin the project file.

Ten sites for the study were selected from 61rojects through the verifying of landscape construc-ion year/location and crash date/location for control-ing variables (Table 1). These 10 sites represent 8

stSm

and roadside planting(4.026 mile)

ifferent cities, 8 of the 10 sites selected reflect inter-tate or major arterial roads and 2 were city streets.verall 5874 crashes at 10 sites for 3–5 year periodsefore and after landscape intervention were used fornalyzing crash rate in this research. It is an importantontribution to control sample bias and to get a numberf crashes from study sites to increase the reliability ofesearch findings.

.2. Data collection

Data collection involved state accident data andxDOT roadway inventory data. The accident data fortudy sections were extracted from the Texas accidentataset, for 3–5 year periods before and after landscapemprovements in each study section. The Texas acci-ent data contains all types of crashes that occur on
tate system roads and streets. These data are main-ained and reported by the Texas Department of Publicafety. Each site included a runoff zone of approxi-ately 1000 ft. on either end of the project site and

2 d Urba

twc

3

ah

•

•

Twcg

3

lubtbr

ricrItamdima(

w1

coAaaft(owam1

tctaiatttwbetff

Numb

× Sec

wVoiassps

sc


he section of improved landscape development. Sitesere selected to avoid major intersections which could

onfound the data.

.3. Research hypotheses

Based on the review of related theories and research,nd considering the research problems, the researchypotheses of this study are:

Crash rates significantly decreased after the land-scape improvement at study sites.A decrease in the number of tree collisions occurredafter landscape improvements.

he two hypotheses that were tested are based on theork of the environmental psychologists who indi-

ated that there may be a positive benefit from roadsidereening that current standards may not be sensitive to.

.4. Research methodology

To assess the crash rate reduction effectiveness ofandscape improvement projects, many different eval-ation methodologies and statistical techniques cane employed. However, it should be emphasized thathe methodology can be used in this research mighte referred to as a quasi-experimental design becauseesearchers cannot assign treatment to locations

andomly and cannot control conditions prior to themposition of treatment. Before-and-after design isommonly found in the highway research literature,eferenced sites will not be given (Council et al., 1980).n the simple before and after design, measurements areaken in time: one before the imposition of the treatmentnd one after the treatment has been put in place. Theost basic assumption underlying the before-and-after

esign is that if the treatment in question had not beenmposed, the after measurement would equal the before

easurement. Therefore, any difference in the beforend after measurement is attributable to the treatment

Crash rate (CR) =Traffic volume

Griffin, 1982).The before-and-after design is however relatively

eak in terms of internal validity (Cook and Campbell,979; Council et al., 1980; Griffin, 1997). To over-

cwTm

n Planning 78 (2006) 263–274

ome this weakness in internal validity, ‘the conceptf multiple cases’ was applied to this research design.ccording to Griffin (1997), by imposing the treatment

t multiple locations and aggregating the multiple casest different times, treatment effect would be separatedrom the ‘uniqueness’ of a particular treatment loca-ion, and the likelihood of falling to ‘an unknown threat’other events happened between the pretest and posttestr before pretest that also affect posttest observation)ould be reduced. In addition, increased number of

ccident data by combining data from several locationsay increase sample size and statistical power (Griffin,

997).Researchers collected crash data at ten study loca-

ions for 3–5 year periods to test for any change ofrash rate in sections with landscape treatments. Mul-iple cases at different times and different locationsre selected to enhance statistical validity and min-mize the unknown threats and the ‘uniqueness’ of

particular treatment location. Conceptual explana-ion of crash data collection and unknown researchhreats for the comparison period at each study loca-ion is delineated in Fig. 2. The number of crashesas converted into the crash rate based on the num-er of crashes per one million Vehicle Miles Trav-led (VMT) because in most cases traffic tendedo increase over the study period. The calculationormula of crash rate per one million VMT is asollows:

er of crashes × 1000000

tion length × Time period (years) × 365

here crash rate: number of crashes per one millionMT at a study road section in a period of time; numberf crashes: number of crashes at a study road sectionn a period of time; traffic volume: average of averagennual daily traffic (AADT) volume at a study roadection in a period of time; section length: length of atudy road section; time period (years): at least 3–5 yeareriods before and after landscape treatments at the 10tudy road sections between the years 1984 and 1999.

The difference in landscape characteristics attudy sites before and after landscape treatmentsonstitutes independent variable and the crash rate
onstitutes the dependent variable. Control variablesere constituted during data filtering and analyses.he traffic volume before and after landscape treat-ents at the study sections was controlled through

J.-H. Mok et al. / Landscape and Urban Planning 78 (2006) 263–274 269

tion be

tswodTcdwSrTct

cmd

tc

R

wlna

4

4

TN

L

1234567891

TA

i

Fig. 2. Concept diagram of data collec

he calculation of crash rates in order to bring theections to a standardized comparison study. Crashesithin construction/maintenance zones were ruledut to control construction zone bias. In addition,ata filtering controlled data recording change in theexas accident dataset. Property damage only (PDO)rashes were ruled out from the dataset to controlata recording bias because PDO recording thresholdas changed on July 1st, 1995. According to TxDOTtandard Specifications Item 192 (TxDOT, 1993),oadside planting or landscape establishment period inexas is 90–365 days. Crashes in the year of landscapeonstruction are ruled out to control for fluctuations inhe landscape form immediately post installation.

The number of tree collisions in the research wasompared by a reduction factor. The reduction factorethod is commonly used in before-and-after study

esign to compare treatment effects before and after

t

r

able 2umber of crashes and crash rates before and after landscape improvement

ocations Number of crashes

Before After

. Interchange landscaping 1 2

. Roadside landscaping 313 315


. Median landscaping 15 11


. Interchange landscaping 32 81



. Median landscaping 320 1280. Sidewalk improvement and roadside planting 64 64

otal 3807 2067verage

a The number of crashes were counted from the crash dataset selected frontervention from the year 1984 to 1999.

b The number of PDO crashes was ruled out to control PDO crash recordc Crashes within construction/maintenance zone were ruled out to controld The values are obtained by deducting crashes or crash rate (before) from

fore and after landscape intervention.

he treatment intervention (Al-Masaeid, 1997). The treeollision reduction factor formula is given by:

eduction factor = Xb − Xa

Xb× 100

here Xb is the number of tree collisions before theandscape treatments at the study sections, Xa is theumber of tree collisions after the landscape treatmentst the study sections.

. Results

.1. Crash rate before and after landscape
reatments
Eight of 10 study sites showed a decrease in crashate and two sites showed an increase in crash rate after

sa,b,c Crash rates

Differenced Before After Differenced

+1 0.0157 0.0564 +0.0407+2 2.9545 2.7285 −0.2260

−1492 1.3682 0.5939 −0.7743−4 0.6340 0.3264 −0.3076

−50 0.4822 0.2447 −0.2375+49 0.2749 0.5970 +0.3221

0 0.0876 0.0778 −0.0098−54 0.4296 0.4171 −0.0125

−192 1.7231 0.6684 −1.05470 2.2802 2.1005 −0.1797

−17401.0250 0.7811 −0.2439

m 10 study sections for 3–5 year periods before and after landscape

ing bias in Texas.construction zone bias.crashes or crash rate (after).

270 J.-H. Mok et al. / Landscape and Urban Planning 78 (2006) 263–274

change

tsao9iad−th

5bo

prccsi5s

tL

Fig. 3. Photos of inter

he landscape treatment (Table 2). Both of the sites thathowed increases in crash rate were complex grade sep-rated interchanges (see Fig. 3). There were also twoutstanding positive sites, study locations nos. 3 and(see Fig. 4) because both of the sites had significant

mprovement they were tested against the other sitesnd each of them showed to be within three standardeviations (3D) of the mean for the 10 sites (mean:0.2439, standard deviation: 0.4023). It suggests that

hese two are not an extreme or anomalous. This staticas shown an approximately normal distribution.

Difference in crash rate calculated by analyzing874 crashes at 10 study sites for 3–5 year periodsefore and after landscape intervention was tested byne-sided paired t-test because crashes were inde-

Fig. 4. Photos of two outstanding positive

pram

landscape treatments.

endent before and after landscape intervention. Theesults show that there was a significant decrease inrash rate after landscape improvements at the 95%onfidence level (p-value: 0.0437, N = 10). It could betatistically valid even if the number of sample sitess 10 but the crash rates were calculated by analyzing874 crashes based on the control of traffic volume andection length.

Interestingly, two negative sites reflect landscapereatments in interchanges. According to the TxDOTandscape and Aesthetic Design Manual (2001), the

sites (study location nos. 3 and 9).

rimary feature of an interchange is vertical grade sepa-ation of the intersecting routes. The grade separation ischieved using a series of ramps and bridges to accom-odate the various directional movements. The series

J.-H. Mok et al. / Landscape and Urba

Table 3Number of tree collisions before and after landscape improvements

Location number(treatment types)

Number of tree collisionsa,b,c

Before After Differenced

1. Interchange landscaping 0 0 02. Roadside landscaping 2 1 −13. Roadside landscaping 18 3 −154. Median landscaping 0 1 +15. Roadside landscaping 1 1 06. Interchange landscaping 0 0 07. Roadside landscaping 0 0 08. Roadside landscaping 1 0 −19. Median landscaping 0 0 010. Sidewalk improvement

and roadside planting2 1 −1

Total 24 7 −17

Tree collision reduction factor 24−724 × 100 = 70.83%

a The number of tree collisions was counted for 3–5 year periodsbefore and after the landscape improvements at 10 study sectionsbetween 1984 and 1999.

b The number of PDO crashes was ruled out to control PDO crashrecording bias in Texas.

c Tree collisions within construction/maintenance zone were ruledo

t

obt5rc

4

ss(ctfFaoaaibtia

5

whs

ut.d The value was obtained by deducting the crash rate (before) from

he crash rate (after).

f ramps and bridges in interchange areas need a num-er of bridge columns or roadside vertical objects. On
hese sites detailed data analysis revealed that about0% of vehicle crashes at this location were related tooadside fixed objects, such as median barriers, con-rete traffic barriers, guardrails, and sides of bridges.
Asac

Fig. 5. Number of tree collisions at study loc

n Planning 78 (2006) 263–274 271

.2. Tree collisions

The calculated reduction factor of tree collisionshows a decrease of about 70.83% of tree colli-ions after landscape treatments at ten study locationsTable 3). As shown in Table 3, there were no extremehanges in tree collisions before and after landscapereatments except for location no. 3. This site is on aour lane divided section of US 75 North in Dallas (seeig. 4). This section has grade separated interchangesnd is bounded by frontage roads. After the installationf roadside landscape improvement the site showedsignificant decrease in tree collisions. The change

ppears to be associated with the landscape treatmentn 1992 (see Fig. 5). The change may also be explainedy TxDOT landscape design guidelines which broughthe site into compliance with clear zone rules and plant-ng setback rules; trees should not be placed forward ofny light standard and retaining wall (TxDOT, 2001).

. Discussions

The findings in this research seem to support earlierork by Topp (1990), Bahar and Naderi (1997) whoave looked at the impact of environmental variables,uch as landscape on human performance and safety.
t the same time, the authors are very cautious about
uggesting that any precise conclusions can be drawnbout the degree to which landscape developmentould be used as a tool to improve the safety of

ation no. 3 (US 75 North in Dallas).

2 d Urba

tohdisar

5

mitotodbmse

TrawaaNjoamlslws

rpgcavc

triugs

6

samdrasdeosfimudsnbm

tmragrccrciTm


ransportation corridors in our urban centers. On thether hand, the fact that this and several other studiesave demonstrated a connection between landscapeevelopment on the roadside and improved safetyf measured by accident reduction warrants furthertudy and consideration. This single treatment mayccount for much of the improvement in accidentates.

.1. Need for further study

The before-and-after studies provided a simpleeans to test the hypothesis that urban landscape

mprovements had a positive impact on crash rates andherefore may contribute to better safety performancef our urban highways and streets. The findings fromhis research suggest that the setting of urban highwaysr streets characterized by landscape improvementsoes have impact on overall performance which maye a result of the effect of greening on human perfor-ance as suggested by environmental psychologists,

uch as Berlyne (1971), Wohlwill (1976), and Parsonst al. (1998).

However, the measures and the data are very coarse.hat is, it is very difficult due to the way the data is

ecorded to get an accurate fix on the exact location ofn accident if it did not occur in an intersection. Like-ise, the lack of information on property damage only

ccidents prevents developing a better understanding ofccident types in relation to landscape improvements.umbers of sites is also a concern because sites with

ust roadside improvements were relatively easy to findn state rights of way. However, municipalities do notlways keep records of their streetscape developmentaking it difficult to compare data between sites and

andscape types. Obtaining a more complete under-tanding of whether, and to what degree, the roadsideandscape contributes to overall safety performanceill require developing more complete data for study

ites.In the further research, crash data associated with

oadside treatments will be investigated to identifyroperties that result in serious injuries. Field investi-ation at each accident location within project sections
an give more specific information, which crash datand roadway inventory data do not have. Researcherserify often struck roadside objects at the project sites,heck the types of struck objects, and measure the dis-
ttac

n Planning 78 (2006) 263–274

ance of clear zone, struck object size, and slope ofoadside/embankment, etc. The data obtained from sitenvestigation and community survey will be used topdate or recommend urban roadside landscape designuidelines to enhance aesthetics and safety of land-cape project sites.

. Conclusions

The findings suggest that the use of roadside land-cape is having a positive affect on overall performancend can be used as a tool to improve the safety perfor-ance of urban streets if the specifics of the affect on

river performance can be identified. There is a cor-elation between certain types of landscape treatmentsnd reduction in crash rates, indicating that the land-cape along the roadside is having a positive affect onriver behavior and perception. Further study into theffect of the specific treatments on driver attentivenessr alertness is required to begin to develop more specificafety-outcome design guidelines. In this regard severalndings seem to be relevant to the design and develop-ent of urban highways and streets. When developing

rban corridors, consideration should be given to theevelopment of the landscape as an integral part of theafety management within the corridor. The landscapeot only contributes to greater aesthetic compatibilityetween the urban environment and the highway butay contribute to a safer street.One very important observation made from this par-

icular study was that all of the landscape improve-ents made on the sites studied adhered strictly to

ules governing setback of non-yielding obstructionsnd visibility considerations. Many of the standardsoverning road design have been developed from soundesearch in driver perception, behavior and response toues from the environment. While this research indi-ates that tree planting may not always be bad along aoad and may in fact have a positive effect on safety inertain contexts, standards which are restrictive regard-ng planting of trees need to applied site specifically.hese needed restrictions are often criticized by welleaning community leaders and other members of

he design professions. Because of the complexity ofhe transportation design problems, experience in thepplication of AASHTO guidelines and newer FHWAontext sensitive design recommendations are needed

d Urba

tpgeia

cabtu

R

A

A

A

A

B

B

C

C

F

G

G

M

P

T

T

T

T

T

W

Z

JoRdlPaNbHtrngmHA(mHCtAAfrr

J.-H. Mok et al. / Landscape an

o encourage effective use of landscape for safety pur-oses. It is the considered opinion of the authors thatood roadside design can be accomplished within thestablished criteria for geometry and safety while meet-ng the desires of the community for streets with moreesthetic appeal.

Clearly, more work will be required to develop spe-ific tools or recommendations that have direct designpplication. However, given the strong correlationsetween this study and the work in other disciplineshe potential benefits would appear to be worth contin-ed pursuit.

eferences

l-Masaeid, H.R., 1997. Performance of safety evaluation methods.J. Transportation Eng. 123 (5), 364–369.

merican Association of State Highway and Transportation Offi-cials (AASHTO), 1970. A Guide for Highway Landscape andEnvironmental Design. Washington, DC.

merican Association of State Highway and Transportation Officials(AASHTO), 1984. A Policy on Geometric Design of Highwaysand Streets. Washington, DC.

merican Association of State Highway and Transportation Officials(AASHTO), 1996. Roadside Design Guide. Washington, DC.

ahar, G.B., Naderi, J.R., 1997. An integrated approach to envi-ronmental impact mitigation and safety management-case stud-ies in the municipality of metropolitan Toronto. Session:roads/transportation and the environment. In: XIIIth World Meet-ing of the International Road Federation Toronto, Ont., Canada.

erlyne, D.E., 1971. Aesthetics and Psychobiology. Appleton Cen-tury Crofts, New York, pp. 1–48.

ook, T.D., Campbell, D.T., 1979. Quasi-Experimentation-Designand Analysis Issues for Field Settings. Rand McNally CollegePublishing, Chicago, IL, pp. 95–102.

ouncil, F.M., Reinfurt, D.W., Campbell, B.J., Roediger, F.L., Car-roll, C.L., Dutt, A.K., Dunham, J.R., 1980. Accident ResearchManual. Highway Safety Research Center, University of NorthCarolina, Chapel Hill, NC, pp. 42–60.

ederal Highway Administration (FHWA), 1997. Flexibility in High-way Design. U.S. Department of Transportation, Washington,DC.

riffin, L.I., 1982. Three procedures for evaluating highway safetyimprovement programs. TARE Technical Report No. 51. TexasTransportation Institute, Texas A&M University, College Station,TX.

riffin, L.I., 1997. A Preface to a Discussion of Six Proceduresfor Evaluating Highway Safety Projects. Research Report No.
FHWA-RD-08-033. Texas Transportation Institute, Texas A&MUniversity, College Station, TX.
ok, J., Landphair, H.C., 2003. Parkways and freeways: safetyperformance linked to corridor landscape type. In: CD-Rom:Transportation Research Board 82nd Annual Meeting.

bGApC

n Planning 78 (2006) 263–274 273

arsons, R., Tassinary, L.G., Ulrich, R.S., Hebl, M.R., Alexander,M.G., 1998. The view from the road: implication for stress recov-ery and immunization. J. Environ. Psychol. 18, 113–140.

aylor, J.G., Zube, E.H., Sell, J.L., 1987. Landscape assessment andperception research methods. In: Bechtel, R.B., Marans, R.W.,Michelson, W. (Eds.), Methods in Environmental and Behav-ioral Research. Van Nostrand Reinhold Company, New York,pp. 361–393.

exas Department of Transportation (TxDOT), 1993. StandardSpecification Book (English). Available at http://www.dot.state.tx.us/apps/specs/specs.htm.

exas Department of Transportation (TxDOT), 2001. Landscape andAesthetic Design Manual. Available at http://manuals.dot.state.tx.us/docs/coldesig/forms/lad.pdf.

opp, H.H., 1990. Traffic safety, usability and streetscape effects ofnew design principles for major urban roads. Transportation 16,297–310.

urner, D.S., Mansfield, E.R., 1990. Urban trees and roadside safety.J. Transportation Eng. 116 (1), 90–104.

ohlwill, J.F., 1976. Environmental aesthetics: the environment asa source of affect. In: Altman, I., Wohlwill, J.F. (Eds.), HumanBehavior and Environment. Plenum, New York, pp. 37–86.

eigler, A.J., 1986. Guide to Management of Roadside Trees. ReportNo. FHWA-IP-86-17, Michigan Department of Transportation,Lansing, MI.

eong-Hun Mok, PhD, is associate research fellow at Seoul Devel-pment Institute, Republic of Korea. He got his PhD in Urban andegional Science (1999–2003) from Texas A&M University; andissertation in delineating traffic safety benefits of travelway corridorandscape characteristics and landscape improvements. Dissertationrofessor is Harlow C. Landphair. He got his master of landscaperchitecture in Graduate School of Environmental Studies, Seoulational University from 1993 to 1995. His master’s thesis wasased on the study on the delivery plan of landscape information.e got his bachelor of agriculture in major of landscape archi-

ecture in 1989–1993 from Sung Kyun Kwan University. He wasesearch associate fellow (2003–2005) at Department of Urban Plan-ing & Design, Seoul Development Institute, Seoul Korea. He wasraduate research assistant (2000–2003) at Department of Environ-ental Management, Texas Transportation Institute, Texas, USA.e was landscape architect and urban designer (1996–1998) atRCHIPLAN Inc., Seoul Korea. He worked as associate researcher

1994–1995) at Graduate School of Environmental Studies, Environ-ental Planning Institute, Seoul National University, Seoul, Korea.e got affiliations membership registrations from Korea Landscapeouncil (general director, 2005–2007); Korea Architect Associa-

ion (registered engineer, current); a full member of Korea Plannerssociation; a full member of The Korean Institute of Landscaperchitecture; a full member of Urban Design Institute of Korea; a

ull member of Korea Institute of Ecological Architecture and Envi-onment. His professional activities include his work as a designeviewer of Turn-Key Bidding at Kyoung-Gi Province; a mem-
er of Task Force, Ministry of Culture & Tourism, South Koreaovernment; An Advisory Committee Member, Professional Urbanrchitecture and Design Forum at Samsung Co. Ltd. His area ofrofessional interest are: Urban Scenic View Protection Study andity Landscape Management Plan; Neighborhood Plan based on
http://www.dot.state.tx.us/apps/specs/specs.htm

http://manuals.dot.state.tx.us/docs/coldesig/forms/lad.pdf

2 d Urba

CfPHin1ticJmSFn

oTNHPm2MSH


ommunity Participation; Context Sensitive Designs and Solutionsor Sustainable Urban Design; Design of Walking Environments foredestrian Safety and Health. His publication work includes: Jeong-un Mok, “A Case Study on Urban Scenic View Protection Methods

n USA for the Management of Urban Scenic View in Korea”, Jour-al of the Korean Institute of Landscape Architecture. Vol. 33, No., 2005; Jeong-Hun Mok, “Case Study for Landscaping & Aes-hetic Enhancement Programs of Urban Transportation Corridorsn USA”, Seoul Studies, Vol. 4, No. 1, 2003. His conference pro-eedings includes: Jeonghun Mok and Harlow C. Landphair, and
ody R. Naderi, “Before-and-after Comparison of Safety Perfor-ance at Landscape improvement Sections”, Proceedings from theecond Urban Street Symposium, Transportation Research Board,ederal Highway Administration, American Society of Civil Engi-eers, 2003, 7; Jeonghun Mok and Harlow C. Landphair, “Parkways
tGES2

n Planning 78 (2006) 263–274

r Freeways: Safety Performance Linked to Corridor Landscapeypes”, Transportation Research Board 82nd Annual Meeting, Theational Academies, Washington DC, 2003, 1; Jeonghun Mok andarlow C. Landphair, “Comparisons of the Safety Performance ofarallel Sections of Parkway and Freeway, Sponsored by TRB Geo-etric Design Committee’s Context Sensitive Design Subcommittee,

003, 1. His published technical reports and media are: Jeong-Hunok and Seok Jung, “Seoul Landscape Management Plan 2005”

eoul Metropolitan Government, 2005; Young-Duk Kweon, Jeong-un Mok, and Hee-Jung Lee, “Community Participation and Main-

enance Methods in District Unit Plan in Seoul” Seoul Metropolitanovernment, 2004; In-Ju Song and Jeong-Hun Mok, “Study on thenvironmental Assessment Index and Design Guidelines of Greenpace for District Unit Plan in Seoul” Seoul Development Institute,004.

Landscape improvement impacts on roadside safety in Texas · Landscape and Urban Planning 78 (2006)...

Documents

Transcript of Landscape improvement impacts on roadside safety in Texas · Landscape and Urban Planning 78 (2006)...