UBC Faculty Research and Publications

Proximity of public elementary schools to major roads in Canadian urban areas Amram, Ofer; Abernethy, Rebecca; Brauer, Michael; Davies, Hugh; Allen, Ryan W Dec 21, 2011

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-12942_2011_Article_464.pdf [ 455.62kB ]
JSON: 52383-1.0223760.json
JSON-LD: 52383-1.0223760-ld.json
RDF/XML (Pretty): 52383-1.0223760-rdf.xml
RDF/JSON: 52383-1.0223760-rdf.json
Turtle: 52383-1.0223760-turtle.txt
N-Triples: 52383-1.0223760-rdf-ntriples.txt
Original Record: 52383-1.0223760-source.json
Full Text

Full Text

RESEARCH Open AccessProximity of public elementary schools to majorroads in Canadian urban areasOfer Amram1, Rebecca Abernethy2, Michael Brauer2, Hugh Davies2 and Ryan W Allen3*AbstractBackground: Epidemiologic studies have linked exposure to traffic-generated air and noise pollution with a widerange of adverse health effects in children. Children spend a large portion of time at school, and both air pollutionand noise are elevated in close proximity to roads, so school location may be an important determinant ofexposure. No studies have yet examined the proximity of schools to major roads in Canadian cities.Methods: Data on public elementary schools in Canada’s 10 most populous cities were obtained from onlinedatabases. School addresses were geocoded and proximity to the nearest major road, defined using a standardizednational road classification scheme, was calculated for each school. Based on measurements of nitrogen oxideconcentrations, ultrafine particle counts, and noise levels in three Canadian cities we conservatively defineddistances < 75 m from major roads as the zone of primary interest. Census data at the city and neighborhoodlevels were used to evaluate relationships between school proximity to major roads, urban density, and indicatorsof socioeconomic status.Results: Addresses were obtained for 1,556 public elementary schools, 95% of which were successfully geocoded.Across all 10 cities, 16.3% of schools were located within 75 m of a major road, with wide variability between cities.Schools in neighborhoods with higher median income were less likely to be near major roads (OR per $20,000increase: 0.81; 95% CI: 0.65, 1.00), while schools in densely populated neighborhoods were more frequently close tomajor roads (OR per 1,000 dwellings/km2: 1.07; 95% CI: 1.00, 1.16). Over 22% of schools in the lowestneighborhood income quintile were close to major roads, compared to 13% of schools in the highest incomequintile.Conclusions: A substantial fraction of students at public elementary schools in Canada, particularly studentsattending schools in low income neighborhoods, may be exposed to elevated levels of air pollution and noisewhile at school. As a result, the locations of schools may negatively impact the healthy development andacademic performance of a large number of Canadian children.IntroductionMotor vehicles are a major source of both air and noisepollution in communities. Epidemiologic studies havelinked exposure to traffic-generated air pollution with awide range of adverse effects in children including reducedlung function [1], decrements in lung growth [2], incidentasthma [3], otitis media [4], and decreased cognitive func-tion [5]. Chronic exposure to traffic noise among childrenhas been linked with increased blood pressure [6], reducedsleep quality [7], and cognitive deficits [8].For children, school is an important environment forexposure to traffic-related pollution due to the amount oftime spent there [9]. According to the Canadian HumanActivity Pattern Survey, children 11-17 years spend anaverage of 12% of time of their time at school, making itthe second most common microenvironment, while forchildren < 11 years school is the 3rd most importantmicroenvironment, accounting for 6% of time on average[10]. Both noise and traffic-generated air pollutants suchas diesel soot, ultrafine particles, oxides of nitrogen (NOx),and carbon monoxide are elevated within approximately100-500 meters of major roadways [11-15], so the proxi-mity of schools to major roads may be an important deter-minant of exposure. A study in the Netherlands found that* Correspondence: allenr@sfu.ca3Faculty of Health Sciences, Simon Fraser University, Burnaby, BC, CanadaFull list of author information is available at the end of the articleAmram et al. International Journal of Health Geographics 2011, 10:68http://www.ij-healthgeographics.com/content/10/1/68INTERNATIONAL JOURNAL OF HEALTH GEOGRAPHICS© 2011 Amram et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative CommonsAttribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction inany medium, provided the original work is properly cited.for children attending a school within 100 m of a freeway,soot exposure was 30% higher and NOx exposure was 37%higher than among children attending a school in a back-ground location [16]. Students attending schools close tomajor roads can be exposed to traffic-related air pollutioneven while indoors because outdoor pollution infiltratesinto classrooms [17,18]. Inverse correlations between con-centrations of traffic-related air pollution inside class-rooms and distance to the nearest major road have beenreported [19,20].Several studies have quantified the distances fromschools to major roads in the US [21-24], but no resultsfor schools in other countries have been published. Herewe present the results from an investigation of the road-way proximities of public elementary schools in the tenlargest Canadian cities. Our objectives were to: 1) validatethe use of roadway proximity as a surrogate for outdoorconcentrations of traffic-generated air and noise pollutionusing measurements from previous field studies; 2) deter-mine the proximities of public schools to major roads; and3) explore urban characteristics and socio-economic indi-cators as correlates of schools’ proximities to major roads.MethodsSchools DataPublic elementary schools for the 10 most populated citiesin Canada were chosen for this analysis as children in thisage group may be particularly susceptible to the effects ofenvironmental pollutants. Elementary schools weredefined as those with students in kindergarten throughgrade 5 (schools with students in other grades wereincluded if their enrollment also included students ingrades of interest). We included only public schools dueprimarily to concerns about the quality and completenessof private school data in provincial databases. In addition,only about 6% of Canadian students attend private schools[25]. Cities from five provinces were included in the analy-sis. In Ontario: Toronto, Hamilton, Mississauga andOttawa; in Quebec: Montreal and Quebec City; in Alberta:Calgary and Edmonton; in British Columbia: Vancouver;and in Manitoba: Winnipeg. For each city we onlyincluded schools within the municipality as defined by thecensus subdivision (i.e., we did not include schools in sub-urban communities). We included only urban areas pri-marily due to their higher levels of traffic-related airpollution and noise and concerns about geocoding accu-racy in low density communities [26,27]. In addition, themajority of pollution measurements used to validate roadproximity as a surrogate for pollution levels (describedbelow) were collected in urban areas. To evaluate the sen-sitivity of our results to the exclusion of suburban commu-nities, we randomly selected one census subdivisionadjacent to each of five of our cities and geocoded thelocations of schools in those five suburbs (the adjacentcensus subdivisions were Burlington, Brampton, Laval,Markham, and Richmond; these are adjacent to Hamilton,Mississagua, Montreal, Toronto, and Vancouver,respectively).Relevant school attributes included the address, gradelevels and type of school (public, private or other). Asinformation regarding Canadian educational institutions isnot centrally collected, the majority of this informationresides with the provincial Ministries of Education (MoE).As a result, the availability and format of these data differby province. Public school locations were collected usingdata available from the MoE websites for each province[see Additional file 1].School Geocoding and Road Proximity CalculationsWe used the commercially available DMTI CanMap roadnetwork to identify road locations and attributes (DMTISpatial, Markham, ON). This product covers Canada anddivides roads into 6 categories. In our primary analysis wecombined DMTI road categories 1 (expressways, usuallyfour lanes), 2 (principal highways, which are multi-laneconduits for intracity traffic), 3 (secondary highways,which are typically thoroughfares with multiple lanes andlarge traffic capacity), and 4 (major roads, used for shortertrips within the city) into a single layer (henceforth “majorroads”) for analysis of school proximities [28]. GeoPin-Point (GPP) software, a product of DMTI, was used togeocode school addresses into latitude/longitude coordi-nates using a 10 m offset from the road’s centerline.Designed for use within Canada, GPP uses the DMTI roadnetwork and allows for the geocoding of French-languageaddresses. In addition, it provides a summary output thatdetails the number of schools successfully geocoded.After geocoding, we calculated the distance from eachschool to the nearest major road using ArcGIS 9.2 (ESRI,Redland, CA). As a secondary analysis, we also quantifiedthe distance from each school to the nearest expresswayor principal highway (DMTI road categories 1 and 2) toallow for comparisons with previous studies in the US[21-23].Assessing Geocoding AccuracyBecause school buildings and grounds cover large areas,and because geocoding of schools can produce substantiallocation errors [27], we manually determined the locationsof a subset of schools for comparison with our automatedgeocoding results. First, we selected schools with geocodedlocations within 200 m of a major road (N = 533). Fromthese we randomly selected 148 schools (10% of the 1,476schools in the analysis) while requiring that at least fiveschools from each city be included. For each of these 148schools, we then used satellite images from Google Mapsto manually determine the coordinates of the point of theschool building nearest to a major road and calculated theAmram et al. International Journal of Health Geographics 2011, 10:68http://www.ij-healthgeographics.com/content/10/1/68Page 2 of 11distance from that location to the nearest major roadusing ArcGIS. The major road proximities assessed by thismanual method were considered the “true” distances forcomparison with the distances estimated from the GPPgeocoding procedure.Pollution DataWe used measurements from previous field samplingcampaigns in Edmonton, Vancouver, and Winnipeg tovalidate the assumption that roadway proximity acts as asurrogate for concentrations of traffic-generated air pollu-tion and noise. Nitrogen oxide (NO) concentrations weremeasured on a 2-week basis using passive Ogawa samplersat 50 locations each in Edmonton and Winnipeg [29] and105 locations in Vancouver [30]. Locations were selectedto cover the study areas and to capture a wide range ofroad proximities. Each location was monitored twice indifferent seasons, and the two measurements at each loca-tion were combined to estimate the long-term averageconcentration. Abernethy et al. [31] measured concentra-tions of ultrafine (< 0.1 μm diameter) particles, anothertraffic-generated air pollutant, for 1-hour periods at 80 ofthe NO monitoring locations in Vancouver using conden-sation particle counters (TSI CPC 3007, Shoreview, MN).Measurements were adjusted to account for temporal var-iation in ultrafine particle concentrations, and compari-sons of measurements collected in different seasonssuggest that 1-hour measurements represent long-termconditions. Davies et al. [32] measured 5-min equivalentcontinuous sound pressure levels (Leq) at the NO monitor-ing locations in Vancouver using a Larson Davis 870Bsound level meter (Larson Davis, Depew, NY). We havepreviously shown a strong correlation between 5-minnoise levels measured in different seasons [11], suggestingthat these measurements are indicative of long-term noiselevels. The locations of these NO, ultrafine particle, andnoise measurements were recorded by field techniciansusing GPS. We used ArcGIS to calculate the distancefrom the measurement locations to the nearest majorroad, defined using the same DMTI data and road classifi-cation scheme as in the school proximity calculations.Correlates of Road ProximityData from the 2006 Canadian census were used to evalu-ate the relationship between schools’ proximities to majorroads and both dwelling density and socio-economic vari-ables. Because we hypothesized that both city- and neigh-borhood-level characteristics might be correlated withproximity, we obtained data for the census subdivision(CSD) and census tract (CT) in which each school islocated. CSD areas correspond to city boundaries, whileCTs typically have populations between 2,500 and 8,000and are useful proxies for neighborhoods in Canada [33].Specific variables included dwelling density at both thecity (CSD) and neighborhood (CT) levels as well as med-ian income and percent of population without a highschool diploma or equivalent at the neighborhood level[23]. Socio-economic variables were not considered at thecity level because these variables are assumed to be mean-ingful primarily in the local context and may not bedirectly comparable between cities due to differences incosts of living or other factors. To account for clusteringwithin cities and neighborhoods, we used mixed modelswith random intercepts at the city and neighborhoodlevels. We modeled school proximity as a binary proximityvariable with 75 m cut point (PROC GLIMMIX, SAS v9.2)and also as a continuous variable (PROC MIXED).Although the influence of traffic-generated air pollutionand noise extends well beyond 75 m [14], we chose thisdistance to be conservative, given geocoding errors andthe relatively large areas of schools and playgrounds. Toevaluate the sensitivity of model results to the choice ofbinary distance, we also modeled school proximity as abinary variable using 200 m cut point. Contrasts in predic-tor variables were scaled to roughly correspond to inter-quartile ranges (IQR) to allow for comparisons of effectsizes between variables.ResultsAddresses were obtained for a total of 1,556 public ele-mentary schools, 1,476 (94.9%) of which were successfullygeocoded into a latitude/longitude location (Table 1). Thegeocoding success rate in individual cities ranged between75.7% (in Calgary) and 100% (in four cities). Variablesaffecting geocoding success included addresses with nomatch in the road network and use of post office boxes asmailing addresses. The populations in the 10 citiesincluded in this analysis ranged between approximately490,000 in Quebec City and 2.5 million in Toronto [34].The combined population of these 10 cities was approxi-mately 9.5 million, or nearly one third of the total Cana-dian population.Pollution measurements in Edmonton, Vancouver, andWinnipeg were inversely correlated with the natural loga-rithm of distance to the nearest major road, with strongercorrelations in Winnipeg (r = -0.44; p < 0.01) and Van-couver (r = -0.50 to -0.61; p < 0.01) than in Edmonton(r = -0.27; p = 0.06). Similar correlations were foundwhen including only measurements within 200 m of amajor road. Based on these measurements we conserva-tively defined ‘near roads’ as < 75 m (Figure 1). Mean (±SD) NO concentrations measured < 75 m from the near-est major road were greater than those measured ≥ 75 min both Winnipeg (14.4 ± 7.3 ppb vs. 9.5 ± 4.2 ppb; 2-sam-ple t-test p-value: < 0.01) and Vancouver (48.1 ± 20.3 ppbvs. 23.4 ± 11.3 ppb; p < 0.01). In Edmonton the differencewas less pronounced (15.6 ± 7.4 ppb vs. 12.6 ± 4.1 ppb;p = 0.17). Ultrafine particles (26,000 ± 18,200 p/cc vs.Amram et al. International Journal of Health Geographics 2011, 10:68http://www.ij-healthgeographics.com/content/10/1/68Page 3 of 11Table1CitycharacteristicsandproximitiesofpublicelementaryschoolstomajorroadsCityCityPopulationaCityArea(km2)aCityPopulationDensity(persons/km2)aCityDwellingDensity(dwellings/km2)aMedian(IQR)SchoolNeighborhoodDwellingDensity(dwellings/km2)bMedian(IQR)SchoolNeighborhoodIncome($10,000)bMedian(IQR)SchoolNeighborhood%PopulationWithoutHSDiploma(%)b#ofSchoolsSuccessfullyGeocodedGeocodingSuccessRate(%)DistancetotheNearestHighwayorMajorRoadMean(±SD)MedianToronto2,503,2816303,9721,5541,805(1,774)5.46(2.11)7.0(6.5)487100.0265±197240Montreal1,620,6933654,4392,0364,058(3,780)3.61(0.86)8.5(6.1)16995.5181±156156Calgary988,1937271,360530973(527)6.42(2.37)4.7(4.6)8475.7431±275366Ottawa812,1292,778292116910(892)7.76(3.81)3.5(2.8)116100.0346±298278Edmonton730,3726841,067435987(498)5.71(2.22)7.9(6.3)13787.3362±193346Mississauga668,5492892,3177451,185(1,012)7.51(2.70)5.6(3.2)103100.0445±233397Winnipeg633,4514641,3655631,162(736)4.93(2.36)7.8(5.7)16397.0402±290368Vancouver578,0411145,0392,2091,761(1,185)5.17(1.23)6.6(7.6)9386.9212±153191Hamilton504,5591,1174521741,152(1,045)5.87(3.01)7.7(4.0)98100.0278±217265Quebec491,1424541,0815023,166(3,582)3.37(2.22)6.5(6.0)2681.3257±198217All9,530,4107,6231,2505091,385(1,515)5.42(2.73)6.7(6.2)1,47694.9318±221282a 2006censussubdivisionstatistics.b2006censustractstatistics.IQR=interquartilerange.Amram et al. International Journal of Health Geographics 2011, 10:68http://www.ij-healthgeographics.com/content/10/1/68Page 4 of 1112,000 ± 6,200 p/cc; p < 0.01) and noise (70.2 ± 5.7 dBAvs. 57.9 ± 6.5 dBA; p < 0.01) in Vancouver were also sig-nificantly elevated within 75 m of major roads. In fact, inVancouver the influence of major roads extended toapproximately 200 m (Figure 1). Concentrations within200 m of a major road were significantly higher than those≥ 200 m for NO (39.7 ± 20.1 ppb vs. 20.6 ± 9.1 ppb; 2-sample t-test p-value: < 0.01), ultrafine particles (21,300 ±16,300 p/cc vs. 11,700 ± 6,400 p/cc; p < 0.01), and noise(65.1 ± 8.2 dBA vs. 57.2 ± 6.6 dBA; p < 0.01).Across all 10 cities, 16.3% of schools were locatedwithin 75 m of a major road (Figure 2). There was con-siderable variability between cities, ranging between 2.9%of schools in Mississauga, Ontario and 33.7% in Mon-treal, Quebec. Using a less conservative cut-off distanceof 200 m, 36.1% of schools were located close to a majorroad, ranging between 11.7% of schools in Mississaugaand 58.0% in Montreal (Figure 2). There was not a strongeast-west gradient in school proximities by city. Whenconsidering only expressways or principal highways(DMTI road categories 1 and 2) to allow for comparisonswith previous studies in the US, we found that 4.7% ofschools were located within 200 m, ranging between 0%in both Calgary and Hamilton and 16.0% in Montreal.Based on comparisons between five cities and commu-nities adjacent to each, we found that a larger percentageof schools included in our analysis were located nearmajor roads than schools in adjacent communities. InHamilton, Mississauga, Montreal, Toronto, and Vancouver18.7% of 950 schools were within 75 m of a major road,while in the five selected adjacent communities 9.3% of236 schools were within 75 m. The cities all had higherpercentages of proximate schools than their adjacent com-munities, with the exception of Mississauga, where thepercentage of schools within 75 m of a major road (2.9%)was lower than the adjacent community of Brampton(4.1%).When modeling schools’ proximities to major roads as abinary variable (< 75 m or ≥ 75 m) we found that bothhigher neighborhood dwelling density (OR per 1,000 dwell-ings/km2 increase: 1.07; 95% CI: 1.00, 1.16) and lowerneighborhood median income (OR per $20,000 increase:0.81; 95% CI: 0.65, 1.00) were associated with closer schoolproximity to the nearest major road (Table 2). Similarresults were observed when schools were categorized as <200 m or 200 m, and when school proximity was modeledas a continuous variable (Table 2). For example, for each$20,000 increase in median neighborhood income, schoolswere an average of 47 m (95% CI: 33, 61) further frommajor roads. The relationship between neighborhood med-ian income and school proximity to roads is summarizedFigure 2 Percent of public elementary schools that are locatedclose to a highway or major road by city.Figure 1 Measured nitrogen oxide, ultrafine particles, and noise vs. distance to the nearest major road in three Canadian cities. Linesin the upper plots are locally weighted regression curves fit to the data. Solid lines in boxplots represent medians; dashed lines represent means.Amram et al. International Journal of Health Geographics 2011, 10:68http://www.ij-healthgeographics.com/content/10/1/68Page 5 of 11Table2Resultsofmulti-levelmodelsofschoolproximitytomajorroadwaysModelOutcomeVariableLevelVariableContrastaEffectEstimate(95%CI)bSchoolwithin75mofahighwayormajorroad(binary)City(CensusSubdivision)DwellingDensity1000dwellings/km21.26(0.71,2.24)Neighborhood(CensusTract)DwellingDensity1000dwellings/km21.07(1.00,1.16)MedianHouseholdIncome$20,0000.81(0.65,1.00)%ofPopulationWithoutHSDiploma5%0.95(0.77,1.18)Schoolwithin200mofahighwayormajorroad(binary)City(CensusSubdivision)DwellingDensity1000dwellings/km21.34(0.75,2.38)Neighborhood(CensusTract)DwellingDensity1000dwellings/km21.19(1.09,1.29)MedianHouseholdIncome$20,0000.74(0.63,0.89)%ofPopulationWithoutHSDiploma5%1.08(0.91,1.28)Distancefromschooltothenearesthighwayormajorroad(m,continuous)City(CensusSubdivision)DwellingDensity1000dwellings/km2-51.6(-102,-1.3)Neighborhood(CensusTract)DwellingDensity1000dwellings/km2-10.3(-17.0,-3.6)MedianHouseholdIncome$20,00047.1(32.8,61.3)%ofPopulationWithoutHSDiploma5%4.7(-11.8,21.2)a Contrastsroughlycorrespondtointerquartileranges(seeTable1)toallowforcomparisonsofeffectsizesacrossvariables.bForthebinaryoutcomemodelstheeffectestimatesaretheoddsratiospervariablecontrast.Forthecontinuousmodeltheeffectestimatesarethechangesinaverageschooldistancetothenearesthighwayormajorroad(inmeters)pervariablecontrast.Amram et al. International Journal of Health Geographics 2011, 10:68http://www.ij-healthgeographics.com/content/10/1/68Page 6 of 11in Figure 3, which shows the percent of schools in closeproximity to major roads in each city-specific neighbor-hood-level income quintile. Of the schools located in thepoorest neighborhoods in each city, more than 22% are <75 m from a major road. In the highest income quintileonly 13% of schools are within 75 m of a major road. Thesame relationship with neighborhood income was observedwhen close proximity was defined < 200 m (Figure 3).For the 148 schools that were manually located, themedian absolute value difference in estimated majorroad proximity between the GPP geocoded locationsand the manually determined locations was 26 m(range: 0 - 244 m). In general, the automated geocodingprocedure resulted in similar road proximity (mediandistance: 89 m; range: 1 - 200 m) as the manual proce-dure (median distance: 81 m; range: 6 - 301 m). Whendichotomizing the 148 schools as < 75 m or ≥ 75 mfrom the nearest major road, 119 schools (80%) wereplaced in the same category by both methods (Table 3).Of the 71 schools automatically geolocated within 75 mof a major road by GPP, 79% were actually within 75 mof a major road, 87% were actually within 100 m, 90%were actually within 150 m, and 97% were actuallywithin 200 m.DiscussionWe found that 16.3% of schools in Canada’s 10 largestcities were located within 75 m of a highway or majorroad. To our knowledge this is the first such study outsideof the US. Unlike previous studies of schools’ proximitiesto major roads, we used measurement data to demonstratea clear relationship between proximity to major roads andelevated levels of traffic-related noise and air pollution,and defined close proximity based on those measure-ments. A growing body of epidemiologic evidence linkschronic exposure to traffic-related air pollution and noisewith a wide range of health effects in children [1-6,13,35].Thus, poorly sited schools may be placing a sizable frac-tion of Canadian public elementary school students atincreased risk of adverse health effects. In addition, thereis evidence that both noise [36,37] and air pollution [38] atschools may negatively affect academic performance.Several studies have examined schools’ proximities tomajor roads in the US. A study in California found thatapproximately 7.2% and 2.3% of public schools werewithin 150 m of medium (25,000-49,999 vehicles/day) andhigh traffic (≥ 50,000 vehicles/day) roads, respectively [23].A similar study in Detroit found that 4.9% of schools werewithin 150 m of high traffic (≥ 50,000 vehicles/day) roads[22]. These results are generally consistent with our find-ing that 4.7% of schools were within 200 m of an express-way or principal highway. Most recently, Appatova et al.calculated roadway proximities for public schools in 9major US cities and reported that 33% of schools werewithin 400 m of a major roadway (defined as federal inter-state, US highway, or state highway) and nearly 12% werewithin 100 m [21].Instead of using traffic volumes we defined major roadsusing the DMTI road classification scheme, which is stan-dardized for roads across Canada. Although road cate-gories are imperfect surrogates for pollutionconcentrations due to variability in traffic flows, vehicletypes, and pollution emissions, our decision to definemajor roads using DMTI categories was supported byclear relationships with measured NO, ultrafine particles,and noise in 3 Canadian cities. A study in Vancouverfound that roads in DMTI categories 1-4 (our definition of“major road”) had mean daily traffic counts of 114,000,21,000, 18,000, and 15,000 vehicles/day, respectively [28].Our choice to define close proximity as only < 75 m froma major road for our primary analysis was more conserva-tive than previous studies.Our finding that neighborhood-level income correlatedwith school proximity to major roads has important envir-onmental justice implications and is consistent with sev-eral previous studies indicating a relationship betweensocioeconomic status and environmental quality aroundschools. For example, Green et. al. [23] found that severalindicators of lower socioeconomic status - including per-centage of students receiving reduced-price meals atschool, percentage of census tract population with incomebelow poverty level, and percentage of census tract popu-lation with no high school diploma - were positively asso-ciated with traffic within 150 meters of schools inCalifornia. Similarly, Wu et al. [22] found that studentsattending schools near high traffic roads in and aroundDetroit were more likely to be ethnic minorities and toreside in a low-income area. A study in Sweden reportedan inverse correlation between NO2 concentrationsFigure 3 Percent of public elementary schools that are locatedclose to a highway or major road by city-specific quintile ofmedian neighborhood-level income at the school location.Amram et al. International Journal of Health Geographics 2011, 10:68http://www.ij-healthgeographics.com/content/10/1/68Page 7 of 11outside schools and neighborhood income [39]. Houstonet al. reported that child care facilities in disadvantagedareas in California were more likely to be situated nearbusy roads than facilities in more affluent areas [40].Unfortunately, our data did not enable us to investigatethe chronology of school and road construction andneighborhood-level income changes, and more research isneeded to understand the underlying causes of our find-ings. For example, it might be useful to explore whetherlow-income residents are drawn to neighborhoods withschools close to roads (e.g., due to lower housing prices),or if low-income neighborhoods are more likely to haveschools and roads constructed in close proximity to oneanother (e.g., due to low-income residents having lessinfluence on community decision-making) [41].The relationship between neighborhood dwelling densityand proximity demonstrates the challenge in balancing thehealth risks of environmental pollution with the potentialbenefits of urban living. Dwelling density and other indica-tors of urban “compactness” are often seen as desirabledue to associations with increased physical activity [42,43]and decreased risks of obesity and associated morbidities[44-47]. However, our results and others’ suggest that den-sity may also lead to increases in exposure to environmen-tal pollution. For example, Marshall et al. [48] found thatVancouver neighborhoods with a high walkability score(based on residential density, intersection density, retailfloor area ratio, and land use mix) tended to also haverelatively high levels of NO.We did not find a strong east-west gradient in the frac-tion of schools located close to busy roads. This findingdiffers from the results of Appatova et al. [21] in the US.Their finding of a strong east-west gradient was drivenprimarily by schools on the “urban fringe” and they didnot find a clear gradient for schools in urban centers.Thus, the lack of a clear gradient in our study may be due,in part, to our exclusion of suburban communities.While this study provides the first assessment of schools’proximities to major roads outside of the US, some limita-tions should be noted. A 2007 study estimated that themedian error for geocoded school addresses was 41 m,with larger errors in rural locations [27]. Location errorscan be exacerbated by the large footprint of school build-ings and surrounding playgrounds. In our assessmentbased on manually locating 10% of schools we found thatthe median error in estimated major road proximity was26 m. However, the influence of location errors on ourconclusions is minimized by our conservative choice of <75 m as the distance of primary interest. In reality, thearea of impact for vehicle emissions may extend out to500 m depending on the specific pollutant [14]. We wereencouraged that 87% of schools geocoded within 75 m amajor road were actually within 100 m, while 90% wereactually within 150 m. An additional benefit of our conser-vative definition of close proximity was that it reduced theinfluence of representing schools as points. As schools’sizes and layouts vary, the placement of classrooms andplaygrounds in relation to roads can also affect students’exposures to traffic-related pollutants. Ideally, we wouldhave overlaid the road network on a layer representingschool footprints as polygons to calculate the portion ofeach school that is located in close proximity to a majorroad.An additional limitation is that since schools data werenot available from a single provider we relied on publiclyavailable online data for this analysis, and this may havenot captured all schools. In addition, data for privateschools were not available or were incomplete for severalcities, so this analysis included only public schools. How-ever, while we may be missing some elementary schoolsin these cities, it seems unlikely that the roadway proxi-mities of the missing schools would be systematically dif-ferent from the included schools, and thus it is doubtfulthat our main findings and conclusions would be alteredsubstantially by missing data. We only included schoolswithin the census subdivision boundaries for majorCanadian cities. Our sensitivity analysis suggested that, ingeneral, schools in suburban communities were less fre-quently located in close proximity to major roads. There-fore, our results cannot be extrapolated outside of theTable 3 Comparison between GeoPinPoint geocoding and manual locating for a random subset of schoolsDistances Based on Google Maps GeocodesDistances Based on GeoPinPointGeocodesNumber (%) of Schools <75 mNumber (%) of Schools 75 -100 mNumber (%) of Schools >100 mTotalsNumber (%) of Schools < 75 m 56 6 9 71(38%) (4%) (6%) (48%)Number (%) of Schools 75 - 100 m 5 1 2 8(3%) (1%) (1%) (5%)Number (%) of Schools > 100 m 9 9 51 69(6%) (6%) (34%) (47%)Totals 70 16 62 148(47%) (11%) (42%) (100%)Amram et al. International Journal of Health Geographics 2011, 10:68http://www.ij-healthgeographics.com/content/10/1/68Page 8 of 11cities included in this analysis. Nevertheless, since weincluded the 10 largest Canadian cities, which accountfor nearly one third of the Canadian population, ourresults apply to a large proportion of Canadian elemen-tary students. Finally, we only considered the schoolenvironment, but other microenvironments and activitiesmay make substantial contributions to the air pollutionand noise exposures of school-aged children. For exam-ple, children can receive high exposures to some pollu-tants while commuting on diesel school buses [49],although exposures depend on a wide range of factorsincluding emissions controls [50], fuels [51,52], androutes [52]. The relationship between school locationand students’ exposures is complicated by the fact thatschool location may affect accessibility and the amountof time that students spend in transit. In addition, attend-ing a school located near a major road may also influencehealth by discouraging walking and cycling to school[53].There are several possibilities for minimizing students’air pollution and noise exposures in and around schools.Concentrations of traffic-related air pollution can bereduced both by technical improvements that reduceper-vehicle emissions, such as improved engine effi-ciency, and urban planning/policy efforts that reduceautomobile use, such as public transit enhancements andimprovements in cycling infrastructure. New schoolscould be set back from major traffic corridors, and it mayalso be beneficial to orient the school facilities such thatthe outdoor playgrounds are located as far as possiblefrom major roads [54]. For example, California State Bill352 requires health risk assessments to be conducted forproposed school sites that are within 150 m of a busyroadway [55], while legislation in New Jersey (AssemblyBill 856, which was motivated by safety concerns and notenvironmental pollution) forbids the construction of newhighway ramps within 300 m of an existing school [56].The importance of traffic-related pollution in school sit-ing decisions is also gaining recognition in Canada. Forexample, the British Columbia Ministry of Environmentrecommends that schools and other sensitive facilities beplaced at least 150 m from roads with over 15,000 vehi-cles/day [57]. Given the small spatial scales over whichtraffic-related air pollutants and noise vary, shiftingschool locations by relatively small distances could resultin substantial reductions in students’ exposures, healthrisks, and impacts on academic performance.There are also several potential strategies for reducingexposures at existing schools. As part of New YorkCity’s Asthma Free School Zone Project, Richmond-Bry-ant et al. [58] evaluated relationships between pollutionconcentrations and vehicle traffic and idling duringschool dismissal periods. They concluded that programsfocused on school bus idling and redirecting school bustraffic could have small but measurable effects on dieselsoot concentrations near schools. Some communitieshave implemented programs that limit outdoor activitiesduring high outdoor air pollution days [59], but theseprograms are likely to be more effective for highly tem-porally variable pollutants like ozone than for trafficrelated pollutants, which are consistently elevated nearroads. Another possible strategy is to modify schoolfacilities. For example, the Port of Long Beach in Cali-fornia has created a “Schools and Related Sites GrantProgram” in which schools and daycare facilities in closeproximity to the Port may apply for funding to mitigateair pollution and noise impacts through improvementssuch as installing high efficiency particulate air (HEPA)filters in ventilation systems, replacing window and doorseals, constructing sound barriers, and installing doubleglazed windows [60].ConclusionWe conducted the first assessment of schools’ proximi-ties to major roads outside of the US and found that16% of public elementary schools were located within75 m of highways or major roads. We conservativelychose 75 m as the distance of interest based on mea-surements of traffic-related air and noise pollution in 3Canadian cities with different characteristics. There wasconsiderable variability between cities in the percentageof schools located near roads, and distance to the near-est highway or major road was correlated with neighbor-hood income and inversely correlated withneighborhood dwelling density. In the lowest quintile ofneighborhood income, 22% of schools were locatedwithin 75 m of a highway or major road. A substantialfraction of students at public elementary schools inCanada, particularly students attending schools in lowincome neighborhoods, may be exposed to elevatedlevels of air pollution and noise while at school. As aresult, the school environment may negatively impactthe academic performance and healthy development of alarge number of Canadian children.Additional materialAdditional file 1: Data sources for school locations andcharacteristics. The table provides the websites used to obtain schooladdresses.AcknowledgementsWe would like to thank those who collected the air pollution and noisedata. We are also grateful to Dr. Winnie Chu and her staff for analyzing NOxsamples.Competing interestsThe authors declare that they have no competing interests.Amram et al. International Journal of Health Geographics 2011, 10:68http://www.ij-healthgeographics.com/content/10/1/68Page 9 of 11Authors’ contributionsOA and RWA designed the study and drafted the manuscript. RA, MB, andHD designed the pollution measurement studies, and RA supervisedcollection of ultrafine particle data. All authors contributed to theinterpretation of data, and read and approved the final manuscript.Author details1Department of Geography, Simon Fraser University, Burnaby, BC, Canada.2School of Population and Public Health, The University of British Columbia,Vancouver, BC, Canada. 3Faculty of Health Sciences, Simon Fraser University,Burnaby, BC, Canada.Received: 29 September 2011 Accepted: 21 December 2011Published: 21 December 2011References1. Brunekreef B, Janssen NAH, deHartog J, Harssema H, Knape M, vanVliet P:Air pollution from truck traffic and lung function in children living nearmotorways. Epidemiology 1997, 8:298-303.2. Gauderman WJ, Vora H, McConnell R, Berhane K, Gilliland F, Thomas D,Lurmann F, Avoli E, Kunzli N, Jerrett M, Peters J: Effect of exposure totraffic on lung development from 10 to 18 years of age: a cohort study.Lancet 2007, 369:571-577.3. Clark NA, Demers PA, Karr CJ, Koehoorn M, Lencar C, Tamburic L, Brauer M:Effect of early life exposure to air pollution on development ofchildhood asthma. Environ Health Perspect 2010, 118:284-290.4. Brauer M, Gehring U, Brunekreef B, de Jongste J, Gerritsen J, Rovers M,Wichmann HE, Wijga A, Heinrich J: Traffic-related air pollution and otitismedia. Environ Health Perspect 2006, 114:1414-1418.5. Suglia SF, Gryparis A, Wright RO, Schwartz J, Wright RJ: Association ofblack carbon with cognition among children in a prospective birthcohort study. Am J Epidemiol 2008, 167:280-286.6. Paunovic K, Stansfeld S, Clark C, Belojevic G: Epidemiological studies onnoise and blood pressure in children: Observations and suggestions.Environment International 2011, 37:1030-1041.7. Ohrstrom E, Hadzibajramovic E, Holmes M, Svensson H: Effects of roadtraffic noise on sleep: Studies on children and adults. Journal OfEnvironmental Psychology 2006, 26:116-126.8. Clark C, Martin R, van Kempen E, Alfred T, Head J, Davies HW, Haines MM,Barrio IL, Matheson M, Stansfeld SA: Exposure-effect relations betweenaircraft and road traffic noise exposure at school and readingcomprehension - The RANCH project. Am J Epidemiol 2006, 163:27-37.9. Klepeis NE, Nelson WC, Ott WR, Robinson JP, Tsang AM, Switzer P, Behar JV,Hern SC, Engelmann WH: The National Human Activity Pattern Survey(NHAPS): a resource for assessing exposure to environmental pollutants.J Expo Anal Environ Epidemiol 2001, 11:231-252.10. Leech JA, Nelson WC, Burnett RT, Aaron S, Raizenne ME: It’s about time: Acomparison of Canadian and American time-activity patterns. J Expo AnalEnviron Epidemiol 2002, 12:427-432.11. Allen RW, Davies H, Cohen MA, Mallach G, Kaufman JD, Adar SD: Thespatial relationship between traffic-generated air pollution and noise in2 US cities. Environ Res 2009, 109:334-342.12. Zhu YF, Hinds WC, Kim S, Sioutas C: Concentration and size distribution ofultrafine particles near a major highway. J Air Waste Manag Assoc 2002,52:1032-1042.13. HEI: Health Effects Institute Special Report 17, Traffic Related AirPollution: A Critical Review of the Literature. 2010 .14. Zhou Y, Levy JI: Factors influencing the spatial extent of mobile sourceair pollution impacts: a meta-analysis. BMC Public Health 2007, 7.15. Beckerman B, Jerrett M, Brook JR, Verma DK, Arain MA, Finkelstein MM:Correlation of nitrogen dioxide with other traffic pollutants near a majorexpressway. Atmos Environ 2008, 42:275-290.16. Van Roosbroeck S, Jacobs J, Janssen NAH, Oldenwening M, Hoek G,Brunekreef B: Long-term personal exposure to PM2.5, soot and NOx inchildren attending schools located near busy roads, a validation study.Atmos Environ 2007, 41:3381-3394.17. Diapouli E, Chaloulakou A, Spyrellis N: Levels of ultrafine particles indifferent microenvironments - Implications to children exposure. SciTotal Environ 2007, 388:128-136.18. Guo H, Morawska L, He CR, Zhang YLL, Ayoko G, Cao M: Characterizationof particle number concentrations and PM(2.5) in a school: influence ofoutdoor air pollution on indoor air. Environmental Science And PollutionResearch 2010, 17:1268-1278.19. Roorda-Knape MC, Janssen NAH, de Hartog JJ, van Vliet PHN, Harssema H,Brunekreef B: Air pollution from traffic in city districts near majormotorways. Atmos Environ 1998, 32:1921-1930.20. Janssen NAH, van Vliet PHN, Aarts F, Harssema H, Brunekreef B: Assessmentof exposure to traffic related air pollution of children attending schoolsnear motorways. Atmos Environ 2001, 35:3875-3884.21. Appatova AS, Ryan PH, LeMasters G, Grinshpun SA: Proximal exposure ofpublic schools and students to major roadways: a nationwide US survey.Journal of Environmental Planning and Management 2008, 51:631-646.22. Wu Y-C, Batterman SA: Proximity of schools in Detroit, Michigan toautomobile and truck traffic. Journal Of Exposure Science AndEnvironmental Epidemiology 2006, 16:457-470.23. Green RS, Smorodinsky S, Kim JJ, McLaughlin R, Ostro B: Proximity ofCalifornia public schools to busy roads. Environ Health Perspect 2004,112:61-66.24. Chakraborty J, Zandbergen PA: Children at risk: measuring racial/ethnicdisparities in potential exposure to air pollution at school and home. JEpidemiol Community Health 2007, 61:1074-1079.25. Aurini J, Davies S: The transformation of private tutoring: Education in afranchise form. Canadian Journal Of Sociology-Cahiers Canadiens DeSociologie 2004, 29:419-438.26. Bonner MR, Han D, Nie J, Rogerson P, Vena JE, Freudenheim AL: Positionalaccuracy of geocoded addresses in epidemiologic research. Epidemiology2003, 14:408-412.27. Zandbergen PA, Green JW: Error and bias in determining exposurepotential of children at school locations using proximity-based GIStechniques. Environ Health Perspect 2007, 115:1363-1370.28. Setton EM, Hystad PW, Keller PC: Road Classification Schemes - GoodIndicators of Traffic Volume? University of Victoria Spatial SciencesLaboratories Working Paper 2005, 05-014.29. Allen RW, Amram O, Wheeler AJ, Brauer M: The Transferability of NO andNO2 Land Use Regression Models Between Cities and Pollutants. AtmosEnviron 2011, 45:369-378.30. Henderson SB, Beckerman B, Jerrett M, Brauer M: Application of land useregression to estimate long-term concentrations of traffic-relatednitrogen oxides and fine particulate matter. Environmental Science &Technology 2007, 41:2422-2428.31. Abernethy R, McKendry I, Allen R, Brauer M: Developing a Land UseRegression Model for Ultrafine Particle Concentrations in Vancouver,Canada. Environ Health Perspect 2011, Available online, http://ehp03.niehs.nih.gov/article/fetchArticle.action?articleURI=info:doi/10.1289/ehp.isee2011.32. Davies HW, Vlaanderen JJ, Henderson SB, Brauer M: Correlation betweencoexposures to noise and air pollution from traffic sources. OccupEnviron Med 2009, 66:347-350.33. Ross NA, Tremblay S, Khan S, Crouse D, Tremblay M, Berthelot J-M: Bodymass index in urban Canada: Neighborhood and metropolitan areaeffects. Am J Public Health 2007, 97:500-508.34. Census Canada: Population and dwelling counts for census metropolitanareas and census agglomerations.[http://dc1.chass.utoronto.ca/census/2006/index.html].35. Babisch W: Transportation noise and cardiovascular risk: Updated Reviewand synthesis of epidemiological studies indicate that evidence hasincreased. Noise & Health 2006, 8:1-29.36. Shield BM, Dockrell JE: The effects of environmental and classroom noiseon the academic attainments of primary school children. J Acoust SocAm 2008, 123:133-144.37. Sanz SA, Garcia AM, Garcia A: Road Traffic Noise Around Schools - A RiskFor Pupils Performance. Int Arch Occup Environ Health 1993, 65:205-207.38. Mohai P, Kweon B-S, Lee S, Ard K: Air Pollution Around Schools Is LinkedTo Poorer Student Health And Academic Performance. Health Aff(Millwood) 2011, 30:852-862.39. Chaix B, Gustafsson S, Jerrett M, Kristersson H, Lithman T, Boalt A, Merlo J:Children’s exposure to nitrogen dioxide in Sweden: investigatingenvironmental injustice in an egalitarian country. J Epidemiol CommunityHealth 2006, 60:234-241.40. Houston D, Ong P, Wu J, Winer A: Proximity of licensed child carefacilities to near-roadway vehicle pollution. Am J Public Health 2006,96:1611-1617.Amram et al. International Journal of Health Geographics 2011, 10:68http://www.ij-healthgeographics.com/content/10/1/68Page 10 of 1141. Schweitzer L, Valenzuela A: Environmental injustice and transportation:The claims and the evidence. Journal Of Planning Literature 2004,18:383-398.42. Saelens BE, Sallis JF, Frank LD: Environmental correlates of walking andcycling: Findings from the transportation, urban design, and planningliteratures. Ann Behav Med 2003, 25:80-91.43. Frank LD, Schmid TL, Sallis JF, Chapman J, Saelens BE: Linking objectivelymeasured physical activity with objectively measured urban form -Findings from SMARTRAQ. Am J Prev Med 2005, 28:117-125.44. Zhao ZX, Kaestner R: Effects of urban sprawl on obesity. J Health Econ2011, 29:779-787.45. Ewing R, Schmid T, Killingsworth R, Zlot A, Raudenbush S: Relationshipbetween urban sprawl and physical activity, obesity, and morbidity. AmJ Health Promot 2003, 18:47-57.46. Ewing R, Brownson RC, Berrigan D: Relationship between urban sprawland weight of United States youth. Am J Prev Med 2006, 31:464-474.47. Lopez R: Urban sprawl and risk for being overweight or obese. Am JPublic Health 2004, 94:1574-1579.48. Marshall JD, Brauer M, Frank LD: Healthy Neighborhoods: Walkability andAir Pollution. Environ Health Perspect 2009, 117:1752-1759.49. Behrentz E, Sabin LD, Winer AM, Fitz DR, Pankratz DV, Colome SD, Fruin SA:Relative importance of school bus-related microenvironments tochildren’s pollutant exposure. J Air Waste Manag Assoc 2005, 55:1418-1430.50. Trenbath K, Hannigan MP, Milford JB: Evaluation of retrofit crankcaseventilation controls and diesel oxidation catalysts for reducing airpollution in school buses. Atmos Environ 2009, 43:5916-5922.51. Mazzoleni C, Kuhns HD, Moosm ller H, Witt J, Nussbaum NJ, Chang MCO,Parthasarathy G, Nathagoundenpalayam SKK, Nikolich G, Watson JG: A casestudy of real-world tailpipe emissions for school buses using a 20%biodiesel blend. Sci Total Environ 2007, 385:146-159.52. Sabin LD, Kozawa K, Behrentz E, Winer AM, Fitz DR, Pankratz DV,Colome SD, Fruin SA: Analysis of real-time variables affecting children’sexposure to diesel-related pollutants during school bus commutes inLos Angeles. Atmos Environ 2005, 39:5243-5254.53. O’Loghlen S, Pickett W, Janssen I: Active Transportation EnvironmentsSurrounding Canadian Schools. Canadian Journal Of Public Health-RevueCanadienne De Sante Publique 2011, 102:364-368.54. Rundell KW, Caviston R, Hollenbach AM, Murphy K: Vehicular air pollution,playgrounds, and youth athletic fields. Inhal Toxicol 2006, 18:541-547.55. State of California: State Bill Number 352.[http://info.sen.ca.gov/pub/03-04/bill/sen/sb_0351-0400/sb_352_bill_20031003_chaptered.html].56. State of New Jersey: Assembly Bill No. 856.[http://www.njleg.state.nj.us/2006/Bills/PL07/308_.PDF].57. British Columbia Ministry of Environment: Environmental BestManagement Practices for Urban and Rural Land Development in BritishColumia.[http://www.bcairquality.ca/reports/pdfs/aqbmps_feb16_06.pdf].58. Richmond-Bryant J, Saganich C, Bukiewicz L, Kalin R: Associations of PM(2.5) and black carbon concentrations with traffic, idling, backgroundpollution, and meteorology during school dismissals. Sci Total Environ2009, 407:3357-3364.59. Shendell DG, Rawling MM, Foster C, Bohlke A, Edwards B, Rico SA, Felix J,Eaton S, Moen S, Roberts EM, Love MB: The outdoor air quality flagprogram in central california: A school-based educational intervention topotentially help reduce children’s exposure to environmental asthmatriggers. J Environ Health 2007, 70:28-31.60. Port of Long Beach: Schools and Related Sites Grant Program.[http://www.polb.com/environment/grants/schools.asp].doi:10.1186/1476-072X-10-68Cite this article as: Amram et al.: Proximity of public elementary schoolsto major roads in Canadian urban areas. International Journal of HealthGeographics 2011 10:68.Submit your next manuscript to BioMed Centraland take full advantage of: • Convenient online submission• Thorough peer review• No space constraints or color figure charges• Immediate publication on acceptance• Inclusion in PubMed, CAS, Scopus and Google Scholar• Research which is freely available for redistributionSubmit your manuscript at www.biomedcentral.com/submitAmram et al. International Journal of Health Geographics 2011, 10:68http://www.ij-healthgeographics.com/content/10/1/68Page 11 of 11


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items