parkinglotlightingbasedupon ......to the photopic luminous efficiency function, v( ), also shown in...

Parking lot lighting based uponpredictions of scene brightness andpersonal safetyMS Rea PhD, JD Bullough PhD and JA Brons MScLighting Research Center, Rensselaer Polytechnic Institute, Troy, NY, USA

Received 9 July 2015; Revised 7 August 2015; Accepted 11 August 2015

Providing subjective impressions of security is central to outdoor lighting design.Current parking lot lighting recommendations are based upon photopic illumin-ances, regardless of spectrum. Scene brightness perception is directly related toimpressions of security, and depends upon both light level and spectrum.A provisional model was used to predict scene brightness for three parking lots,each illuminated to different levels by different light sources. Observers judgedscene brightness, security and other factors for each lot. The provisional modelaccurately predicted both scene brightness and security judgements. The lightingassociated with the best subjective ratings also had the lowest power density.A design method using ‘brightness illuminance’ is presented, which can lowersystem costs while maintaining a sense of security by users.

1. Introduction

Higher photopic illuminances are typicallyrecommended for outdoor spaces where goodvisual performance is needed.1 The NorthAmerican industry standard2 for parking lotlighting design, for example, recommendshigher illuminance levels for potential vehicleconflict points, such as vehicle entrances andexits. In addition to supporting good visualperformance, an important lighting designcriterion to consider for parking lot lighting isproviding pedestrians with a sense of personalsecurity and safety.2–4 Dimly illuminatedparking lots are perceived as less safe andsecure relative to those illuminated to higherlight levels.5,6

A variety of studies have been publishedindicating that both light level and lightspectrum affect how bright a scene

appears.5,7–12 Scenes illuminated to higherlight levels will appear brighter and, for thesame photopic illuminance, outdoor scenesilluminated by cooler light sources will alsoappear brighter.9 Despite this research, cur-rent recommendations for parking lot lightingare based upon photopic illuminances with-out regard to the spectral characteristics ofthe light source.

A provisional model of the spectral sensi-tivity of scene brightness perception has beenpublished.9 Research continues to providegreater insight into the phenomenon as wellas the neurophysiology underlying scenebrightness perception.10 Figure 1 shows two,very much simplified, luminous efficiencyfunctions based upon early developments ofthat evolving provisional model: one, denotedVB3(�), for ‘high’ levels like those found incommercial applications such as schools andoffices, and one, denoted VB2(�), for ‘low’levels that might be found for outdoorapplications such as parking lots.13 Theseluminous efficiency functions are comparable

Address for correspondence: MS Rea, Lighting ResearchCenter, Rensselaer Polytechnic Institute, 21 Union Street,Troy, NY 12180, USA.Email: [email protected]

Lighting Res. Technol. 2015; 0: 1–12

� The Chartered Institution of Building Services Engineers 2015 10.1177/1477153515603758

by guest on September 3, 2015lrt.sagepub.comDownloaded from

http://lrt.sagepub.com/

to the photopic luminous efficiency function,V(�), also shown in Figure 1, whereby thespectral power distribution of any light sourcecan be used to calculate photometric quan-tities such as illuminance and luminance.

Power density (W/m2) is also an importantdesign consideration for parking lot applica-tions. (Of course, other factors such as initial/life-cycle costs and maintenance requirementsare also important for light source selection.)It is common practice to compare powerdensities required by different lighting sys-tems for a given lighting installation where aspecific photopic illuminance is recommendedor prescribed.14 Using conventional photom-etry based upon the photopic luminous effi-ciency function [V(�)], high pressure sodium(HPS) has often been the source of choice foroutdoor lighting applications because, for agiven recommended photopic illuminance, thepower density would be less than thatrequired by other light source types.However, light-emitting diodes (LEDs) haverecently become a viable source for outdoorlighting applications because the photopicluminous efficacy of these sources hasbecome comparable to HPS. The first costsof LED systems are greater than those ofHPS, limiting their widespread application. Acommon estimator’s guide for construction

costs,15 for example, put LED cobra headluminaires ($1800; all costs are in US dollars)at about three times the cost of comparableHPS cobra heads ($580).

The present study was primarily designed tovalidate predictions of scene brightness basedupon the provisional model by Rea et al.9 Forexample, scenes illuminated by sources thatproduce ‘white’ light, like LEDs, should beperceived as brighter by a predictable amountthan ‘yellow’ HPS at the same photopicilluminance. Since a formal link has previouslybeen established between scene brightness andperceived sense of security,16 a given perceivedbrightness level should also provide a predict-able impression of personal safety and security.A second study goal was to determine howlighting design based upon a prescribed bright-ness criterion, and consequently a predictableimpression of personal safety and security,might be used to systematically reduce powerdensities for parking lot applications. Byreducing power densities to meet a prescribedcriterion of perceived safety and security, thelife-cycle cost of an LED system becomesmoreattractive for wider application.

The present study was conducted on theUniversity of Washington campus in Seattle,Washington, USA. Photometric measurementsand subjective evaluations were collected byresearchers at Rensselaer’s Lighting ResearchCenter (LRC) in collaboration with those atthe Seattle Lighting Design Lab (SLDL).

2. Method

2.1. Measurement sites

Three parking lots on the campus of theUniversity of Washington were selected forphotometric and subjective evaluations. Thefirst site was a large parking lot (about110m� 73m) and served as a reference con-dition; it was illuminated by recently installedLED sources. The two test sites had high-intensity discharge (HID) lighting systemsinstalled. One lot (about 67m� 33m) was

1

0.9

0.8

0.7

0.6

0.5

0.4

Effe

cien

cy

0.3

0.2

0.1

0

400 450 500 550 600 700

V(λ)

VB2(λ)VB3(λ)

Wavelength (nm)650

–0.1

Figure 1 Three luminous efficiency functions, the pho-topic luminous efficiency function [V(�)], and two bright-ness luminous efficiency functions [VB2(�) and VB3(�)]from Rea.13

2 MS Rea et al.




illuminated with HPS sources and the other(about 98m� 33m) was illuminated by metalhalide (MH). Table 1 describes the lightingsystems used to illuminate the three parkinglots. Information about lamp replacementand cleaning schedule was unavailable.

2.2. Photometric and radiometric

measurements

Researchers from SLDL performed illu-minance measurements (Minolta, T10) at allthree sites. They measured horizontal

photopic illuminance on the ground atapproximately 2.7m� 2.7m spacing, in atypical ‘bay’ (defined as the rectangular areain which the light level distribution would berepeated underneath a regular grid of lumin-aires) in each lot. Spaces containing parkedcars were omitted from the field measure-ments. Table 2 provides the descriptive stat-istics for the measured photopic illuminancelevels in three parking lots.

LRC researchers sampled spectral irradi-ance distributions (SIDs) at the three parking

Table 1 Lighting systems in the three parking lots.

Parking lot Luminaire description CIExa CIEya CCTa CRIa GAIa Power perluminaire (W)

Powerdensityb

(W/m2)

Mountingheightc

MHd Cobrahead-typeluminaire withdropped lens andexternal shielding

0.38 0.39 4055 43 46 458e 1.51 8.3 m

HPSd Cobrahead-typeluminaire with flatlens

0.52 0.41 1977 16 19 465e 0.97 8.3 m

LEDf Fully shielded IESType V mediumsquare distribution

0.38 0.37 4101 65 74 101g 0.32 11.3 m

Note: MH: metal halide; HPS: high pressure sodium; LED: light-emitting diode; CIEx and CIEy: CommissionInternationale de l’Eclairage 1931 chromaticity coordinates; CCT: correlated colour temperature; CRI: colour renderingindex; GAI: gamut area index (see Reference 13).aBased upon on-site spectroradiometric site measurements.bPower density assumptions: 101 W per LED head, two staggered heads per ‘bay’ (defined in Section 2.2), with an areaof approximately 621 m2 per bay; 465 W per HPS head, with an area of approximately 472 m2 per head; 458 W per MHhead, with an area of approximately 301 m2 per head.cAs reported by facilities personnel at University of Washington.dLuminaire distribution types for existing MH and HPS luminaires were unavailable.eEstimated according to Leslie and Rodgers.6fReference condition.gAs reported by LED luminaire manufacturer.

Table 2 Photometric measurement summary of (a) photopic illuminance; (b) brightness illuminance; and (c) uniformityratios in the three parking lots.

Photopic illuminance (lx) Brightness illuminance (B2-lx) Ratios

Source Mean Median SD Maximum Minimum Mean Median SD Maximum Mininum Maximum:Mininum

Avg:Mininum

MH 5.6 4.6 3.4 14.0 1.6 25.6 21.1 15.6 63.7 7.3 8.7 3.5HPS 45.9 37.1 33.8 157.1 7.0 53.7 43.4 39.5 183.6 8.2 22.5 6.6LED 14.5 15.1 2.5 19.4 8.6 62.6 65.0 10.8 83.6 37.2 2.3 1.7

Note: MH: metal halide; HPS: high pressure sodium; LED: light-emitting diode.

Parking lot lighting, brightness and safety 3




lot sites using a spectroradiometer with aremote sensor (Ocean Optics, USB650); alaptop computer was used to store andprocess the spectral data. Two steps weretaken to determine the ‘brightness illumin-ance’ provided by the different lighting sys-tems. First, the raw SID values wereconverted into photopic illuminance usingthe conventional photopic luminous efficacyfunction (Figure 2) and then scaled to equalthe measured photopic illuminance valuesfrom Table 2. Figure 3 shows the SIDvalues, scaled to the average photopic illu-minances from Table 2. From the scaled SIDvalues, the VB2(�) luminous efficiency func-tion from Rea13 was applied using the con-ventional unit conversion of 683 lm/W at555 nm (Figure 2). An equivalent methodfor determining brightness illuminance wouldbe multiplying the wavelength-specific

photopic illuminance values by a brightness(B2)/photopic (P) ratio for each light source,whereby the ratio of VB2(�) to V(�) ismultiplied by the scaled SIDs; the B2/Pratios are 4.32 for the LED source, 1.17 forthe HPS source and 4.55 for the MH source.Table 2 provides the descriptive statistics forthe three parking lots based upon brightnessilluminance (B2-lx). Uniformity ratios are alsoprovided in Table 2; these ratios are identicalfor both of the photopic and brightnessilluminance distributions.

2.3. Procedures

In December 2014, researchers at theSLDL recruited 18 (7 female) adults (meanage¼ 46 years) to provide subjective assess-ments of brightness during one night in thethree parking lots on the University ofWashington campus. All three parking lotswere mostly devoid of parked vehicles andother pedestrians during the experimentalsessions. Subjects met near the parking lotilluminated by the LED sources. Weatherconditions consisted of occasional light mist.The parking lot surfaces were initially dry,but towards the end of the nighttime session,they were beginning to become wet.

Instructions were given to all subjects at thesame time. Each was provided with a pen anda clipboard containing five copies of a ques-tionnaire, one to be completed at the firstparking lot and two each to be completed atthe other parking lots. The questionnairecomprised 10 Likert-type scale ratings, ran-ging, qualitatively, from ‘completely agree’ to‘completely disagree’ and, numerically, fromþ2 to �2 (see Appendix). Each rating scalewas associated with a different perceptualaspect of the parking lot lighting, rangingfrom a sensation of glare to a psychologicalsense of personal safety. All subjects saw theLED lot first and together as one group. Theywere told that the brightness of the LED lotshould be rated as a 10 and that the subjectivebrightness levels of the other two lots that

Lum

ens

per

wat

t

683

0400 500 600 700

Photopic Brightness

Wavelength (nm)

Figure 2 Luminous efficacy functions based upon thephotopic luminous efficiency function [V(�)] and the B2

luminous efficiency function [VB2(�)] from Rea.13

0.0025

0.0015

0.0005

0400 500 600 700

Wavelength (nm)

MH HPS LED

Irra

dian

ce (

W/m

2 /nm

)

0.002

0.001

Figure 3 Average measured spectral irradiance distribu-tions (SID) from the three parking lots, each illuminated todifferent levels by different light sources (MH, HPS andLED).

4 MS Rea et al.




they would be seeing would be rated relativeto this value; if a subsequently seen parkinglot appeared brighter than the first parkinglot, then it should be given a number greaterthan 10 proportional to its relative brightness.Similarly, if another parking lot appeareddimmer than the first parking lot, it should begiven a number proportionally less than 10.Both magnitude estimations of the parkinglot brightness and responses to a Likert-typescale rating about the brightness of the lotwere used as independent assessments of theperceived brightness of each parking lot.

A bus transported the subjects to the othertwo parking lots, located near each other. Forlogistical convenience, the subjects weredivided into two groups when they arrivedat the drop-off location near the two parkinglots; one group walked to the parking lotilluminated by the HPS sources, while theother walked to the parking lot illuminated bythe MH sources. Two viewing locations ateach parking lot had been previously identi-fied by the researchers, one view was acrossthe shorter side of the parking lot and theother was along the longer side. Participantsprovided subjective assessments at each of thetwo viewing locations for both HID lots. Thetwo groups of subjects moved to the otherparking lot and repeated their subjectiveassessments at both viewing locations. Aftercompleting the final questionnaire, subjects

were transported back to the LED site tosubmit their completed questionnaires andcollect a small incentive gift card. The proto-col was approved by Rensselaer’sInstitutional Review Board (IRB).

3. Results

Overall, most subjects judged the LED park-ing lot to be neither too bright nor too dark(Figure 4), indicating that the light levels inthis lot, which served as a reference condition,were not considered extremely high norextremely low.

As shown in Figure 5, subjects rated theHPS site as about the same brightness as theLED site, which was defined in the study ashaving a brightness of 10. The MH site wasrated as about half as bright as the LED site.Figure 6 compares the photometric charac-terisations of the three parking lots fromTable 2 to the subjective brightness judgmentsin Figure 5.

The results from each Likert-type scale (seeAppendix 1) are given in Figure 7. Table 3provides descriptive statistics for each scaletogether with the results of the inferentialstatistics (Student’s t-test) comparing subject-ive responses from the HPS and the MHparking lots to those from the reference, LEDparking lot. For all questions, subjects rated

100%

80%

60%

40%

20%

0%0%

Toodark

Slightlytoo dark

Neithertoo bright

nor too dark

Slightlytoo bright

Toobright

11%

83%

LED lot“Please judge the overall brightness of this lot.” (n=18)

6%0%P

erce

nt o

f sub

ject

s

Figure 4 Subjective brightness assessment of the reference (LED) parking lot.





the LED lot as statistically equal to or betterthan the HPS lot. Both the LED lot and theHPS lot were always rated more positivelythan the MH lot. In addition, the meanbrightness ratings for each parking lot(Figure 5) were very strongly correlated(r2¼ 0.998) with the mean Likert-type scaleresponses to the statement ‘This parking lotlooks bright’ (Figure 7).

Several of the Likert-type scale responses inFigure 7 specific to the MH parking lot canprobably be explained by the substantiallylower light level in that lot compared to theothers (Table 2). For example, the MH lotwas rated as least glaring although the MHluminaires used dropped lenses, while theHPS and LED lots used flat-lensed lumin-aires. The lower light level also resulted in

poorer colour identification under MH thanunder HPS, despite the superior colour-rendering characteristics of MH (Table 1).And, the MH lot was judged to have moreshadows despite being more uniformly lightedthan the HPS. These findings reinforce thenotion that overall light level, as well asspectral and spatial distributions, affects per-ceptions of lighted exteriors.

As shown in Figure 8, the brightnessfunction is also more predictive of impres-sions of safety than photopic illuminance.

4. Discussion

Outdoor lighting should provide pedestrianswith an acceptable level of perceived safetyand security.2–4 Perceptions of safety andsecurity are well correlated with perceptionsof scene brightness;16 dimly illuminated park-ing lots are perceived as less safe and securethan brightly illuminated parking lots.Obviously increasing light levels will makeoutdoor spaces appear brighter, and thereforethey will also be perceived as safer and moresecure. Increasing light levels, however, hasdirect and obvious implications for capitaland operational costs. Consequently lightingengineers and designers should optimise theirdesigns to provide the benefit of lighting forperceived safety and security while minimis-ing cost.

15

12

9

6

3

0

15

12

9

6

3

00 10

MH HPS LED MH HPS LED

r2 = 0.404 r2 = 0.963

20 30 40 50Photopic illuminance (lx)

0 10 20 30 40 50 60 70Brightness illuminance (B2-lx)

Brig

htne

ss r

atin

g

Brig

htne

ss r

atin

g

Figure 6 Functional relationships between subjective ratings of brightness (�SEM) and photopic illuminance (left,�SEM) and brightness illuminance (right, �SEM) for the three parking lots.

Brightness comparison* (n=17-18)15

10

5

0MH HPS LED

* “From where you are standing, please judge the overallbrightness of this lot, compared to the first lot you saw.”

Brig

htne

ss r

atin

g va

lue

Figure 5 Subjective brightness assessments of the HIDparking lots (HPS and MH) compared to those of thereference (LED) parking lot.

6 MS Rea et al.




One aspect of outdoor lighting that affectsapparent brightness, but is unrecognised incurrent lighting recommendations, is thespectral composition of the source of illumin-ation. At the same photopic illuminance,outdoor scenes illuminated by sources withrelatively more short-wavelength radiationwill appear brighter. This phenomenon

would have significant implications for out-door lighting practice if it could be reliablydemonstrated that by taking into accountboth light level and spectrum, an acceptablelevel of perceived safety and security could beachieved at lower capital and operationalcosts than could be achieved with currentpractice where only light level is considered.

MH HPS

–2 –1 0 1 2 –2 –1 0 1 2

–2 –1 0 1 2

–2 –1 0 1 2

–2 –1 0 1 2

–2 –1 0 1 2–2 –1 0 1 2

–2 –1 0 1 2

–2 –1 0 1 2

–2 –1 0 1 2

LED

“This parking lot looks bright.”

“I would feel comfortable driving in this parking lot.”

“This parking lot looks gloomy.”

“There are shadows where people could hide.” “I would be able to see other pedestrians easily.”

“I could identify colors of clothing and cars easily.”

“I could see potential hazards in this parking lot.”

“I could find coins or keys I dropped on the ground.”

“I would feel safe walking here.”

“The lighting is glaring.”

Disagree completely–2

Disagree somewhat–1

Agree somewhat1

Agree completely2

Neither agree/disagree0

Figure 7 Questionnaire results for the three parking lots, average responses (�SEM) for the two viewing locations atthe two high-intensity discharge sites; the star (?) highlights the scale used to assess perceptions of safety.





Systematic investigations of scene bright-ness perceptions have been published overseveral decades,17 with growing interest in thelast few years.5,7–12 A provisional model ofapparent scene brightness has been recently

published so that for any SID (both spectrumand amount), the apparent brightness of ascene can be predicted.9–12,18 Moreover, adirect link has been shown between apparentbrightness and perceived sense of safety and

Table 3 Questionnaire rating results.

HPS LED MH

Compared toreference

Referencecondition

Compared toreference

MeanRating

SEM t17 p MeanRating

SEM MeanRating

SEM t17 p

The lighting is glaring 0.09 0.23 0.10 ns 0.06 0.19 �1.31 0.16 5.04 50.01I would be able to see

other pedestrianseasily

1.00 0.14 1.16 ns 1.28 0.15 �0.83 0.18 7.82 50.01

There are shadows wherepeople could hide

�0.14 0.19 1.34 ns �0.67 0.21 1.58 0.12 6.94 50.01

The parking lot looksbright

0.50 0.17 0.00 ns 0.50 0.19 �1.54 0.13 7.85 50.01

I would feel safe walkinghere

0.78 0.16 1.73 ns 1.28 0.16 �0.72 0.19 8.03 50.01

I would feel comfortabledriving in this parkinglot

1.36 0.12 1.24 ns 1.67 0.10 �0.19 0.21 5.63 50.01

I could see potential haz-ards in this parking lot

0.78 0.17 0.68 ns 1.00 0.17 �0.39 0.21 3.46 50.01

I could identify colors ofclothing and carseasily

0.00 0.22 1.41 ns 0.61 0.22 �1.26 0.13 6.48 50.01

I could find coins or keys Idropped on theground

0.47 0.17 2.04 ns 1.11 0.14 �1.08 0.14 8.79 50.01

This parking lot looksgloomy

�0.44 0.20 3.41 50.01 �1.33 0.14 1.14 0.20 6.28 50.01

Note: MH: metal halide; HPS: high pressure sodium; LED: light-emitting diode.HPS and MH were significantly different on all rating scales at p50.01. ns¼not significant.

–2

–1

0

1

2

0 10

MH HPS LED MH HPS LED

r2 = 0.224 r2 = 0.999

20 30 40 50Photopic illuminance (lx)

0 10 20 30 40 50 60 70Brightness illuminance (B2-lx)

Saf

tey

ratin

g

–2

–1

0

1

2

Saf

tey

ratin

g

Figure 8 Safety question averages (�SEM) excerpted from Figure 7 (?); the brightness function is more predictive(r2¼ 0.999) of ratings of safety than the photopic spectral sensitivity function (r2

¼0.224).

8 MS Rea et al.




security,16 and the present results confirm thislink. In principle then, the body of researchmakes it possible to optimise the benefits andcosts of outdoor lighting by taking intoaccount the SID of illumination.19

The present study was aimed at validatingthe promise of this research for the design ofparking lot illumination. As shown inFigures 6 and 8, by taking into account theSID and an appropriate brightness luminousefficiency function [VB2(�)], it was possible topredict both perceptions of scene brightnessand perceptions of safety and security. Theseresults are consistent with the provisionalmodel predictions by Rea et al.9 and, there-fore, provide a justification for engineeringoutdoor lighting based upon apparent scenebrightness as it affects perceptions of safetyand security.

The costs associated with lighting also needto be considered when engineering outdoorlighting. Once a design goal is established,such as ‘to provide an acceptable level ofsecurity in the parking lot’, it is then possibleto specify a criterion level of brightnessilluminance needed to meet the design goal.For the present study, it was assumed that theLikert-type scale rating of þ1.5 in Figure 7(?), corresponding to a qualitative ratingmidway between ‘agree somewhat that Iwould be safe walking here’ (þ1.0) and‘agree completely that I would be safewalking here’ (þ2.0), would be the design

goal. Thus, it was assumed that a Likert-typevalue of þ1.5 would ‘provide an acceptablelevel of security in the parking lot’. Basedupon Figure 8, a Likert-type value of þ1.5corresponds to a brightness illuminance valueof 66.8 B2-lux.

From this criterion brightness illuminancevalue, it is then possible to compute thephotopic illuminances needed by the differentlight sources to equal 66.8 B2-lux (Table 4,third column). More importantly, it is pos-sible to compute the incremental energy costsof the three different lighting systems existingat the University of Washington parkingfacilities, HPS, MH and LED, needed tomeet the criterion brightness illuminancevalue and, thus the design goal. Two ofthese lighting systems, HPS and MH, werenot new, thus providing lower illuminancesthan would be provided when new. At theexisting lighting system efficacies, the powerdensity needed to provide the criterion bright-ness illuminance value of 66.8 B2-lux wasgreatest for the MH system and the least forthe (newer) LED system (Table 4, fourthcolumn). If these three types of sources wereinstalled new to meet the criterion brightnessilluminance, the HPS would have the highestpower density, three times that of new MHand LED systems (Table 4, fifth column). Ifthe design criterion was based upon lightoutput after five years of operation, takinginto account lumen depreciation, the LED

Table 4 The brightness illuminance and the associated photopic illuminances needed to achieve a þ1.5 Likert-scale(‘agree’) rating of security under the HPS, MH and LED sources.

Lightsource

Brightnessilluminance (B2-lx)

Photopicilluminance (lx)

Power density:existing (W/m2)

Power density:new (W/m2)

Power density:after five years (W/m2)

HPS 66.8 57.1 0.97 0.97 1.18MH 66.8 14.7 1.51 0.33 0.54LED 66.8 15.5 0.33 0.33 0.43

Note: MH: metal halide; HPS: high pressure sodium; LED: light-emitting diode.Brightness illuminance was calculated using the VB2(�) luminous efficiency function from Rea.13 Also listed are powerdensities required to provide these illuminances under the existing lighting conditions, when the same sources wouldbe new according to manufacturer reported data, and after five years of operation according to the US Department ofEnergy.20





system would have the lowest installed powerdensity (last column).

The initial costs of LED lighting systemsare greater than those of comparable HPSlighting systems.15 To be cost-effective then, itis important that the LED lighting system savesufficient energy over the lighting systemlifetime to justify its higher initial cost. Theamount of light needed to deliver an ‘accept-able’ (þ1.5 on the Likert-type scale) level oflight would be 66.8 B2-lux for any light source.From Table 4, the incremental power densityneeded by the HPS lighting system over theLED lighting system after five years, takinginto account lumen depreciation,20 is 0.75W/m2, or 0.00075 kW/m2. To determine thebreak-even cost after five years, it is necessaryto calculate the annual hours of use as well asthe incremental energy (power� time) costs ofthe HPS relative to the LED systems to deliver66.8 B2-lux after five years. The annualelectricity use (12 hour/day� 365 day/year)is 4380 hour/year or 21,900 hour/five years.The incremental annual power density savingsper year is 3.3 kWh/m2/year or 16.5 kWh/m2/five years. At an estimated utility rate,including use and demand charges, of $0.10/kWh, the annual energy cost savings per yearfrom the LED lighting system would be $0.33/m2/year or $1.65/m2/five years.

As an example, for a 4650m2 parking lot,the incremental annual savings would beapproximately $1500/year. So for the five-year payback, the initial cost of the LEDsystem over the HPS system should not bemore than $7500. Illumination of this parkinglot might require seven luminaires, based onthe number of luminaires needed to illuminatea proportionally larger lot.19 An initial systemcost increment of $7500 would correspond toabout $1070 per luminaire. As describedabove, the current estimated cost differencebetween LED and HPS luminaires is greater,approximately $1220,15 so LED system priceswould need to be somewhat lower to achievea five-year payback. Importantly, if the

parking lot were illuminated by the LEDlighting system to the same photopic illumin-ance as the HPS system needed to deliver an‘adequate’ light level, the LED parking lotwould appear much brighter, but there wouldbe an energy penalty as well as an initial costpenalty for the LED lighting system.

5. Conclusions

Outdoor lighting serves multiple objectivesand critical among these is providing animpression of personal safety and security tothe occupants of a space. Meeting thisobjective is practical with evolving lightsource technologies such as LEDs, which arebeginning to outpace traditional outdoor lightsources such as HPS lamps in terms ofphotopic luminous efficacy. However, theresults of the present field demonstrationshow that the spectral characteristics ofLED lighting systems can provide additionalleverage over HPS systems by providinggreater scene brightness and perceptions ofsafety and security at lower power densities.From a practical perspective, then, there aresubstantial opportunities to use less energy,produce lower peak power demand andreduce lighting system costs by designingoutdoor lighting that supports perceptionsof safety and security when LED sources areconsidered in comparison to HPS sources.More generally, the present study offers a‘brightness illuminance’ design methodologyby which any light source can be utilised tomaintain perceived safety and security foroutdoor lighting.

Acknowledgements

The authors acknowledge the contribution ofDr. Levin Nock representing the BonnevillePower Authority who served as the projectmanager. Also acknowledged are the staff at

10 MS Rea et al.




the Seattle Lighting Design Laboratory fortheir invaluable assistance in coordinating sitelogistics and photometric measurements. Inparticular the authors would also like tothank Kurt Nielsen, Jeffrey Robbins, EricStrandberg and Edward Smalley, who wereinstrumental in making the project possible.From the Lighting Research Center, Dr.Mariana Figueiro is acknowledged for herassistance in collecting subjective evaluations,Andrew Bierman and Timothy Plummer fortechnical support, and Dennis Guyon foreditorial and graphic support. Also acknowl-edged are the University of Washington andSound Transit for enabling us to use theirparking lots for this research project. Finally,the subjects who volunteered their eveningduring the busy holiday season are gratefullyacknowledged.

Declaration of Conflicting Interests

The author(s) declared no potential conflictsof interest with respect to the research,authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the follow-ing financial support for the research, author-ship, and/or publication of this article: Thisproject was sponsored by the BonnevillePower Authority (TIP 329).

References

1 Bullough JD, Donnell ET, Rea MS. To illu-minate or not to illuminate: roadway lighting asit affects traffic safety at intersections. AccidentAnalysis and Prevention 2013; 53: 65–77.

2 Illuminating Engineering Society on NorthAmerica. Recommended Practice for ParkingLot Lighting, RP-20-14. New York, NY:IESNA, 2014.

3 Rea MS (ed) IESNA Lighting Handbook:Reference and Application. 9th ed. New York:Illuminating Engineering Society of NorthAmerica, 2000.

4 Boyce PR. Human Factors in Lighting. 3rd ed,Boca Raton, FL: CRC Press, 2014.

5 Boyce PR, Eklund NH, Hamilton BJ, BrunoLD. Perceptions of safety at night in differentlighting conditions. Lighting Research andTechnology 2000; 32: 79–91.

6 Leslie RP, Rodgers PA. The Outdoor LightingPattern Book. New York: McGraw-Hill, 1996.

7 Fotios SA, Cheal C. Lighting for subsidiarystreets: investigation of lamps of differentSPD, Part 2: brightness. Lighting Research andTechnology 2007; 39: 233–249.

8 Fotios S, Cheal C. Predicting lamp spectrumeffects at mesopic levels. Part 1: spatialbrightness. Lighting Research and Technology2011; 43: 143–157.

9 Rea MS, Radetsky LC, Bullough JD. Towarda model of outdoor lighting scene brightness.Lighting Research and Technology 2011; 43:7–30.

10 Rea MS, Mou X, Bullough JD. Scene bright-ness of illuminated interiors. Lighting Researchand Technology. First published 21 April 2015.DOI: 10.1177/1477153515581412.

11 Bullough JD, Radetsky LC, Rea MS. Testing amodel of scene brightness with and withoutobjects of different colours. Lighting Researchand Technology 2011; 43: 173–184.

12 Bullough JD, Radetsky LC, Besenecker UC,Rea MS. Influence of spectral power distribu-tion on scene brightness at different lightlevels. Leukos 2014; 10: 3–9.

13 Rea MS. Value Metrics for Better Lighting.Bellingham, WA: SPIE Press, 2013.

14 American Society of Heating, Refrigeratingand Air-Conditioning Engineers andIlluminating Engineering Society. EnergyStandard for Buildings Except Low-RiseResidential Buildings, 90.1-2013. Atlanta, GA:ASHRAE, 2013.

15 RS Means. Electrical Cost Data. Norwell, MA:RS Means, 2015.

16 Rea MS, Bullough JD, Akashi Y. Severalviews of metal halide and high pressure sodiumlighting for outdoor applications. LightingResearch and Technology 2009; 41: 297–320.





17 Ferguson H, Stevens W. Relative brightness ofcoloured light sources. Transactions of theIlluminating Engineering Society (London)1956; 21: 227–255.

18 Bullough JD. Spectral sensitivity modeling andnighttime scene brightness perception. Leukos2015; 11: 11–17.

19 Rea MS, Bullough JD, Brons JA. Spectralconsiderations for outdoor lighting: designing

for perceived scene brightness. LightingResearch and Technology. First published 23October 2014. DOI: 10.1177/1477153514556126.

20 US Department of Energy. LED ApplicationSeries: Outdoor Area Lighting. Washington,DC: USDOE, 2008.

Appendix

The Likert-type scale questionnaire given to experiment participants.

12 MS Rea et al.




parkinglotlightingbasedupon ......to the photopic luminous efficiency function, v( ), also shown in...

Documents