USF Libraries
USF Digital Collections

Roadway level-of-service determination


Material Information

Roadway level-of-service determination
Physical Description:
1 online resource (vi, 82 p.) : ill. ;
Florida Institute of Government
University of South Florida -- Center for Urban Transportation Research
University of South Florida, College of Engineering, Center for Urban Transportation Research
Place of Publication:
Tampa, Fla.
Publication Date:


Subjects / Keywords:
Highway capacity -- Florida   ( lcsh )
Traffic flow -- Mathematical models -- Florida   ( lcsh )
Roads -- Evaluation -- Florida   ( lcsh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Includes bibliographical references (p. 57-62).
Statement of Responsibility:
prepared for the Florida Institute of Government by the Center for Urban Transportation Research.
General Note:
"May 1991."

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 029120306
oclc - 752959493
usfldc doi - C01-00010
usfldc handle - c1.10
System ID:

This item is only available as the following downloads:

Full Text


Center for Urban Tran s port a ti o n Re s e a rc h Roadway Level-of-Service Determination I?" 1.\May 1 991 C UTR rJ8r:i' U niv ers ity of South F l o r i da Colle g e of E n gineer i ng T ampa, F l o r ida "\' ./.; ./.


Roadway Level-of-Service Determination .. ------g..,....., Prepared for the Florida Institute of Government by the Center for Urban Transportation Research May1991


TABLE OF CONTENTS Table of ContentS ....... ...... ........ ... ................ ...... .................... . .... . . ............. . i Preface . ..... ... .. .. . . ..... .. ... .. . . ..... ... .... .. ....... ... . ...... . ...... ...... .. .... .. .. .... . .... ii E S .. xecuuve ummary . ........ ................ ....... .... ......... .......... . ... .... . ....... . . .... u1 Introduction .. . .... . ..... ...... ... ..... . ...... .. ...... . .. .. .. . .... .. . ..... ... ... ... .. .. .. ... .. . I Scope of Study ... .... ... .......... ..... ......................... .... ..... . ..... .... ...... ... . ... I Sources of lnfonnation .......... .... . ............... ... ...... .......... .... .... ...... ... :... 2 Roadway Levels of Service ... . . ....... ...... ......... ...... ... .... . . ....... . .............. 5 Method s of Analysis . ....... ..... ..... .... .... ... ... ... . . . . . . .. .. .... . . ....... .... . . . . . .. 8 State Policies . ....... . .. .. ....... ... .. .... . . . . . . . ... .. . . . . . . .... . ......... ...... ..... 9 Local Options .. .. .. .. .. ... .... . .. . . . .... . .. .. .. .. . .. . . . .. . . .. . .. . ..... . ..... .... I 0 Areawide Levels of Service ............. ... ... ... .... .... . .... .... ...... ..... ..... . .............. 27 State Policies .. .... ....... ........ .. ... ... ..... .. ... .. ..... ... .... .. ... . . ...... . . . . . .... ... 28 Local Options .... .. . .. ...... ........ . ........ ... . .. ... . ..... ... ...... ..... . . .... . . . . 29 Multimodal Levels of Service . ..................... .............................. ................. 40 State Policies .... .... ...... .... .. ... . ..... ..... .... ........ .... ........ ........ .. ...... .... .. ... 40 Local Options ...... .......... ............. . ...... .. .. . . .. . .. .. .. ...... .. ....... .. .. .. .. .... 41 Choice of Analysis Period .... . .... ... ...... ...... ...... ............. .... ......... ......... . ... . . 47 State Policies . .... . . .... . . .... . ........ . .... . .... . .... ... ...... ............. ....... ....... .. 48 Local Options .. ..... .. . .. ... ........ . .... ........ ........ .. ... .... .. .. .... . .... ..... .. ....... 49 Conclusion . .... ...................... ...... ........ ..... .... ......................... . ...... .... ....... 56 Api)Clldix ....... ..... . .... ... ...... ............. . ........... ...... ...... .......... ... ..... ... . . .... .... 63


11 PREFACE Tilis teChnical StUdy, "Roadway Level-of-Service Determination," was prepared by the Center for Urban Transponation Research, College of Engineering, University of South Aorida. The srudy was proposed by the Florida Association of Counties (FA C) and coordinated by Stephen Hogge ofFAC's staff Funding was provided by the Florida Instirute of Govenunent under STAR Gran t 1190. The aulhor is Reid Ewing, who wol1ted under the general supervision of Edward Mierze j ewski. Research assisLance was provided by Edward Hannan. We are gratefu l 10 FAC's Technical Advisory Committee for valuable comments on a draft of this repon. Committee members were Ben Folce Hillsborough County; Pavid Harris, Florida Deparnnent of Community Affairs; Stephen Hogge, Aorida Association of Counties; Herb Kahlen, Palm Beach County; Bill Lecher, Nassau County; Douglas McLeod, Florida Depanmeru ofTransponation; Diane Salz, Aorida League of Cities ; and Clarl< Turner, City of Miami. We are also grateful to the following individuals for clarifying methods of roadway level-of service determination used by various jurisdictions. and for providing insights imo the strengths and wealcnesses of alternative methods: Bob Kamm (Brevard County); Ossama Alaschlr.ow (Broward Coumy); Marl< Woerner (Dade County); Laura Fine! (consu ltant 10 Flagler County); Mark Gillis (consuiLant to Lee County) ; Tun Allen (Tallahassee); Clark Thmer (Miami); Danny Pleasant (Orlando); Ali Atefi (Pasco County); Dick Thomas (Seminole Coumy); and Mahdi Mansour (Tampa). Marl< Bevis of Seminole County, John Zegeer of Banon-Aschman AssociateS. and Glalling Lopez Kercher Anglin. Inc., supplied the traffic counts and travel time study results used in the quantitative sectiOIIS of this repon. Gary L. Brosch Director May 1991


EXECUTIVE SUMMARY Under Florida s 1 985 GroWih Management ACt, adequate infrastructure must be available concurrent wilh the impactS of development. Adequacy is defined by level-of-service standards, adop!ed by local governments as part of their local comprehensive plans. No development order or permit may be issued if levels of service wiU be degraded below the adopted standards. The focus in groWih management is on level-of-service standards. However, the methods used to detennine roadway levels of service affect conclusions about road adequacy as much as do the standands to which they are compared In this repon it is shown that the specific technique used to analyze roadway level-of-service whether HCS. ART-Pl.AN, ART-TAB, or travel lime studies can make at least a two letter grade diff erence in the OUtcome. Likewise, the choice of analysis period or peak hour whether 30th highest hour, 100!11 highest hour average peak hour of the peak season or highest two consecutive hours on the average weekday can make a difference of two or more letter-grades. While harder 10 quantify, the effect of combining levels of service across facilities or modes could be of comparable magnitude. Given the imponance of methodology this repon seeks 10 provide guidance 10 public officials wrestling with the foUowing four issues: (1) What methods should be used to assess roadway levels of service? Appropriate methods of analysis may differ from application to application since the consequences of error are more serious in some cases than others. Less precise methods are probably adequate for local comprehensive planning, while more precise methods of analysis are required for development review and approval To assess the precision of altemalive methods, level-of-service estimates were compared to travel time study results. Highway Capacity Software (HCS) and ARTPl.AN (formerly ART-Al.L) proved better predictors of arterial levels of service than ART TAB (formerly LOS). Since ART-PLAN is much easier to use than HCS. ART-PLAN appears to be the preferred method of arterial analysis. All three methods are necessarily superior to generalized maximum volume tables General iud tables assume roadway, traffic. and sigrtal characteristics averaged across the entire Stale rather than specific 10 a particular facility or locale iii


iv All methods tested tend to unclereljmate travel speeds Travel time studies can be used to better calibrate HCS. ART-Pl-AN, and other programs Altematively, travel time study results can be correlated directly with traffic volumes and other variables in statistically derived models. To illustrate the laner approach, regression analyses were performed relating average peak-hour travel speeds on Seminole County roads to peak -hour traffic volumes and numbers of signalized intersections per mile. The exp lanatory power of the resulting model is probably inadequate for use in forecasting future travel speeds and levels of service. Nonetheless. with 55% of the variation in average travel speed explained by only two variables. a good predictive model could dOubtless be developed with a richer data base (including such independent variables as the green ratio. arrival type, and % turns from exclusive lanes). (2) Should levels of service be averaged or aggregated across road facilities? l-evels of service on roadway segments are routinely combined into one overall level of service for a facility end-to-end. The procedure for doing so is straightforward. and no difficult methodological issues arise. However, we find ourselves in unchaned waters when we begin to combine levels of service for different facitities intO one overall level of service for a roadway network. Three approaches suggest themselves: summing volumes and capacities, averaging levels of service, and adopting performance summaries. All three approaches have precedentS in local comprehensive plans already approved by the Depamnent of Community Affairs (DCA). All three enable local governments to fmance the most cost-effective system improvements rather than isolated roadway improvements dictated by minimum operating standards. Of the lhree, the ayerame method may be preferred for growth management purposes. Travel speeds fall precipitously as traffic volumes approach capacities. With areawide averaging. local governments, concerned about maintaining average travel speed, will have ample incentive to fix locatized traffic problems. l-ess incentive is provided by the Olher approaches. Whichever approach is chosen, roadways must be assigned we ights that reflect their contributions to areawide levels of service. The preferred weighting factor is yemcJetraveled on roadways. since it accounts for the volume of traffic exposed to different traffic conditions.


Also, meaningful analysis areas must be delineated if levels of service are to be estimated on an areawide basis. Transportation concurrency management areas (as DCA refers to them) should be drawn so as to encompass aiJemate routes ayajlable for common p<:ak bour trips. How this general guideline is operationalized is best left to Joeal planne rs. The resulting areas could weU coincide with pre-existing tranSponatlon impaet fee benefit districts. To ensure that localized traffic problems continue to be addressed, indiv idual facilities could remain subject 10 level-of-service standardS. albeit standaros set at the absolute minimum acceptable level. At the same time. areawide average standards could set at a higher desjrable level to encourage cost-effective system improvements. (3) Should levels of service be averaged or aggregated across urban transportation modes (public and private)? There are many problems with combining the levels of service for different modes into one overall level of service. The measures used to define levels of service are different for transit and private vehicles. Transit is a viable option only for auto users with trip endS in the transit corridor; commercial vehicles and other auto users cannot avoid roadway congestion by switching modes. Combined stand3lds could have unintended consequences. Even if buses ran empty and roadS were gridlocked, local government could approve new developments as long as they were "served" by transit Rather than combining levels of service across modes. transit availability may be reflected in re]axe

vi (4) For what time period or peak hour sl10uld levels of service be analyzed? The choice of peak hour is ultimate! y a political ralher !han a technical decision. It involves ba!ancioi tbe public's desire ro hold down taxes (which means fewer road improvement s and more traffic congestion) against their desire to avoid traffic congestion (which requires higher taxe s for road improvements). The same peak hour should be used for bolh design of new roads and management of existing roads but that need not be the 30th highest hour or the year (that is. the hour with the 30th highest traffic volume of the year). While time-honored in highway design, the 30th highest hour has no special economic significance An analysis of hourly volumes from FOOT's penn anent count stations shows that the 30th highest hour ordinarily does not correspond ro the point of diminishing returns in roadway design. FDOT is proposing a shift ro the I OOth highest hour. If it i s necessary to settle on one peak hour for roads statewide. as FDOT has declared. the IOOth highest hour may be the best choice. An analysis of hourly volumes from FOOT's pennanent count stations suggests that the tOOth highest hour is roughly equivalent 10 the average peak hOur of weekdays during the peak season in urt>an areas and weekends in recreational areas and other places with weekend peaking. Thus. the JOOth highest hour reflects daily, weekly. and seasonal peaking of traffic; and given its rough equivalence to the average peak hour of peak days during the peak season the IOOth highest hourly traffic volume would be relatively easy ro estimate and project. Daily peaks tend to spread out as urt>an areas grow and traffic congestion causes motorists 10 adjust !heir travel hours Indeed, the largest cities do not have a "peak hour'' pu se but ra!her a two10 tluee-hour period in the morning and afternoon when commuting is heaviest. A shift ftom peak hour to peak period analysis might be justified by local policies that encourage commuters 10 adjuSt their travel hours. With flex-time in place, government may be relieved of responsibility for accommodating every trip maker's choioe of travel hour.


I. INTRODUCTION Roadway levels of service play a cenual role in Florida's effons to manage growth. The 1985 Growth Management Act and implementing Rule 9J embraced a "pay as you grow" policy, commonly referred to as concuaency Adequate infrasuucrure must be available concurrent wilh the impactS of development. Adequacy is defined by level-of-service standards, adopted by local governments as pan of their local comprehensive plans. No development order or permit may be issued if levels of service wiU be degraded below the adopted standards Local concurrency management systems must be implemented to ensure that level-of-service standards are maintained in the face of development pressures. As one writer noted: "'Fi ve of the infrastrucrure elements have posed few problems for local governments. But the sixth category, roads. is proving to be a nightmare."' The state's 1987 comprehensive plan repon projected a $53 billion infrastructure shonfall by the year 2000 unless growth slowed or infrastructure investment increased. Of that amount. nearly half was for tnnsponation. Roads are the jnfrastrucrure element most Ukely to triner public dissatisfaction. &rowth morawria. and le&al challen&es under lhe Growth Manaeement As;!. Thus. it is crucial that roadway level-of-servi ce determinations be accurate and results be interpreted sensibly. 11tis repon explores alternative approaches to ensure that they will be. A. SCOPE OF STUDY Given the central role of roadway levels of service in growth management, one might expect to find volumes written about level-of-service determination and standard setting. Yet. a litcrarure search uncovered only one published article that approaches the subject from a policy perspective.' Other literarure focuses on technical issues. How do progression adjusDilent factors for signalized intersections vary with intersection spacing? How do geometries affect saturation flow rates? 11tis is the sruff of the level-of-service literature. The present srudy addresses four policy issues: (I) What melhods should be used to assess roadway levels of service? (2) Should levels of service be combined across roadway facilities? (3) Should levels of service be combined across urban transponation modes (public and private)? I


2 (4) For what time period or peal< hour should levels of service be analyzed? A host of other issues arise in the sening of levelofservice standards, the monitoring of levels of service, the approval of projects that degrade levels of service, and the treatment of roads that are backlogged to begin with. While imponant, these issues are beyond the scope of this study B. SOURCES OF INFORMATION Policies and procedures for roadway level-of-service determination are in a state of nux. The Florida Department of Community Affairs (DCA) is revising its rules for local compreben sive plans (LCPs), the Florida Department of'Transponation (FDOT) is revising its lev el-of service guidelines, and many localities are switching 10 more sophisticated methods of roadway level-of-service determination. Adding to the confusion, DCA in its regulatory role often differs with FOOT in its role as advisor to DCA. DCA negotiates "creative" compro mises with localities that are having difficulty meeting roadway level-of-service standards while FOOT adheres to ''professional standards and practices that preclude such compro mises. To get a fix on current policies and practices, this srudy included: (I) A review of pertinent rules of the DepaJtment of Community Affairs (DCA) : Cbapter9J -5, F.A.C., "Minimum Criteria for Review of Local Govenunent Comprehensive Plans and Determination of Compliance.' A draft rule change to Olapter 9J-5, F.A.C. Olapter 9J-2, F. A. C .. "Rules of Procedure and Practice Pettaining 10 Developments of Regional Impact.'' DCA's "Development of Regional Impact: Application for Development Approval" (referred to below as the DRI Applicant Handboolc).


(2) A review of guidelines of the Florida Depamnen1 of Transportation (FOOT): FOOT's Florida's Leyel of Sendee Standards and Gujdeljnes Manual for Plarmjng (referred 10 below as the Level of Service Manual). A draft revision of the Level of Service Manual. FOOT's Guidelines for the Revjew of Deyeloomems of Reona! Impact (referred to below as DR! Guidelines). (3) A review ofTraffic Circulation Elements in 23 LCPs. and where applicable. S!lpulated Settlement Agreements between DCA and local governments aimed at bringing LCPs iniO compliance with state law and regulations. Figure I contains a list of the 23 LCPs. plus summary information about their Traffic Circulation Elements (4) Interviews with transportation professionals responsible for roadway level-of-service determinations in selected localities (see the PJcface for a listing) The localities were chosen for their diversity of approach from among the 23 whose LCPs were reviewed previously The opinions of those interviewed are given much weight in the discussions that follow. 3


Figure 1 M..-ot fedlllles ... __ ......,... 8tc'VI rd CO\.IJ\.'Y vjc ltnioa. (196S HCM) hrlllel RO&dways Higbest Hour 8oc:a !Lacon FDOTTablcs None 30th H ighut Hour Btowald Coutlt'( FOOT Tol>"" Non< 30th KIShbt Hour Char lene CO\tMy FDOT'Tlb)c-' (19791JTPS ver ) Nooe ,\Vg. PCI.k Hr. o( Pnlt Sca$on Cocoo FOOT Tables None H}8I'IC$t Hour COllier COUMY FOOTTobl"' ,., .. AVS Peale Hr. of Peak Seuon COtal Springs klcaJ (Bucd on LOS Patti Nooe Not SptdrJed Ot.4e Cotuncy FDOTTabks (l979l1J'P$ 'Yr:r.) None ,wg. 2 Highcll COnk<: Wlcdy Hrt Dwal County FOOT Table$ Nooe Higl'lc$t Hour Fbpe-r Counry FDOTTab le& atld I O&d.Spedftc None 30th H ighe"$t HOI.It Tablu (B.ued 04'1. LOS Prog;ra-m) Ft. Lauckrda l e FOOT T abkl af'MS Nooe Not Specillecl Analy&l$ With HCM Softwaw ............... Cty. fDOTT ..... NoM 30th Hlgbest Hour Jefrer:son fDOT,...... None ,om KIShest H01.1r "'Y"'"' FOOT HGM None HlSbeat Hour of U.S 1 LeIy PDOTT&b l u (l9?9 t11?S w:r.,) Jtoedf ilhln pea)( Hour of ........ """"" PakSaaoo Loon"""'"'' fDOTT ..... None 30th Hi$hest Hour T-........ and 1Dc:al None H.ighat Ko'ur Cu>cy Tablel (Bued on LOS """" v/c IWlOt (196S HOI) Joedl a 1nNJt Fac. of 2 HiJbat: WIUlln Gon"idors Conlec:ut ive Weekday Hours Odondo FOOT T&blu and Arteria l lcled$ wilhln Tralftc JOOch Highest Hour with .uT.W. Performatlot DiA.tricU Pal:m Beach Cty. FOOT ....... ..... 50th Hl&bCilll How Puc;o Gounry FDOT Ta.b&et l.ftod l.ocal """" 50th Hiibest Hour Tables (Bued on LOS PfOJft.n'l) T._ 1.oc:al Muimum Volume T&bks Joe4s within Di&a:ricU Hitbmt HoW Oluod on I 96S HC>I) and Citywide Voluloia County FDOT'hl>b None NOt Specified 4


C. ROADWAY LEVELS OF SERVICE Roadway levels of service are reponed using the same grading scale as a school report card. with "A" being the highest gracle a road can receive and "F" the lowes t (see Figure 2). The 1985 Hi&bway C;macily Manual (HCM) defines lev e ls of service using numerical measures. The measures vary by type of facility (see Figure 3). Level-of service measures in the 1 985 HCM emphasize the experience of motoristS. 'This represents a departure from the 1965 HCM, which focused on the condition of facilities. In panicular, before 1985,levels of s ervice on arterials were defined by the volume/capacity ratios. They are now defined by the average trave l speeds of mo10ris!S (see Figure 4). 'This study focuses on urban and suburban anerials because so many arterials are already oongested. The precise method of level-of service detennination may weD determine whether these arterials are designated adequate or inadequate under adop!ed level-of-service standards, whether developmentS impacting them are approved or disapproved, and whether LCPs are found to be in compliance with state law and regulations. Figure 2 Roadw3y Levels of Service Beoome Newsworthy grades forSt. st. Pet:ersbwg earned reladvely good marks on Wednesday' r epon card. The city earned fiVe A's and had a average. SL Petersburg Times, article entitled roads earn improved grades", Februrary 14, 1991 5


6 Measures of Effectiveness for Level of Service Definition TYPE OF FACILI'rY Freeways Basic freeway segments ...... Weaving areas .............. Ramp junctions ............ Mullilane Highways . ....... .. Two-Lane Highways .......... Signalized Intersections .... ... Unsignalized Intersections ...... Arterials ..................... Transit ....... ......... .. .. .. Pedestrians ........... ...... MEASURE OF EFFECTIVENESS Density (pc/mi/ln) Average travel speed (mph) Flow rates (pcph) Density (pc/mi/ln) P ercent time delay (%) Average travel speed (mph) Average individual stopped delay (sec/veb) Reserve capacity (pcph) Average travel speed (mph) Load factor (pers/seat) Space (sq fl/ peel) Source: Transportation Research Board, Hiehway Capacity Manya l (198S) p. 1 The 1985 HCM provides the following descriptions of levels of service for arterials: Level-of-service A (LOS A) describes primarily free flow-operations at average travel speeds about 90 percent of the free now speed of lhe arterial class. Ve hicles are completely unimpeded in their ability to maneuver within lhe traffic stream. Stopped delay at signalized intersections is minimal. Level-of-service B (LOS B) represents reasonably unimpeded operations at average travel speeds about 70 perceru of the free flow speed of lhe arterial class. The ability to maneuver within the traffic stream is only slightly resoicted and stopped delays are not bothersome. Drivers are not subjected to appreciable !ension. Level-ofservlce C (LOS C) represents stable conditions. However, ability to maneuver and change lanes in midblocl: locations may be more resuic!ed lhao in LOS B. and longer queues and/or adverse signal coordination may conoibute 10 lower average travel speeds of about 50 percent of the free flow speed of the arterial class. Motorists experience an appreciable tension while driving.


Level-of-serviceD (LOS D) borders on the range in which small increases in flow may cause substantial increases in approach delay and. hence. decreases in anerial speed. 1bis may be due to adverse signal progression. inappropriate signal timing. high volumes, or some combination of these. Average travel speeds are about40 percent of free flow speed. Leve l -of-service E (LOS E) is characterized by significant approach delays and average travel speeds of one-third the free flow speed or lower. Such operations are caused by some combination of adverse progression. high signal density, ex tensive queuing at critical interseetions. and inappropriate signal timing. Leve l -of-service F (LOS F) characterizes anerial flow at extremely low speeds below one-third to one-quarter of the free flow speed Intersection congestion is likely at critical signalized locations, with high approach delays resulting Adverse progression is frequently a contributor to this condition. ARTEIUAL CLASS Rmtg e of Free Flow Speeds (mph) Typical Free Flow Speed (mfh) LEVEL 01' SERVICE A B c D E F figure 4 Anerial Levels of Service I II 4S to 3S 3S to 30 40mph 33 m p h m 3S to 2S 27 mph A VEAAO TllA VEl. SPEED 3S 2S 28 24 19 22 18 13 17 14 9 13 10 7 < 13 < 10 < 7 Sourc:e: Transportation Research Board, Hjabway Capacity Maoyal (1985), p. 11-4. 7


8 D. METHODS OF ANALYSIS Methods of detennining roadway levels of service may be arrayed in tenns of data and analytical requirements. and corresponding precision of estimates. It is usually assumed !hat the simplest methods are the least precise. the most complex methods the most precise (as shown in Figure 5). nus is believed 10 be the first study 10 systematically test !his assump tion. From least to most complex, methods used in Florida include : FDOT gene!3lized maximum volume tables based on .. gene!3lized"traffic. roadway, and signal parameters for the State of Florida. Levels of service are detennined by comparing traffic flows 10 the maximum volumes at different levels of service for roads of a given type. Figure 5 Alternative Methods of Arterial Analysis PREOSION e Local Maximum Volume Tables e Generalized Maximum Volume Tables e Travel Time and Dela y Srudies e Network Simulation Programs e Arterial Analysis Programs COMPLEXITY


Local ma>

10 FDOT's DR! Gujde!jney These guidelines require a levelof-service delenninalion for e very intersection in the immediate area of a PRJ, plus an anerial analysis for each anerial in the larger project study area. Intersection analyses are to be done with intersection capacity analysis procedures based on the 1985 HCM. Anerial analyses may be based on one of several methods, where the preferred method depends on the planning horizon' From I to 5 years into the furure, planning and traffic operational methods are suggested. From I 0 to 20 years, planning methods or generalized tables may be used.' DCA's DR! Aw>licam Handbook The handbook provides for roadway level-of-service detenninations based on "the most recent procedures of the Transportation Research Board (1RB) and FOOT."' This is a reference to 1985 HCM procedures (developed by TRB) and the most recent maximum volume tables (developed by FOOT). Like FOOT, DCA is receptive to other methods if agreed upon in advance. In summacy, current FDOT and DCA poticies allow generalized maximum volume tables to be used in LCI's and ORis. More precise traffic operational methods or travel time srudies are favored in certain apptications, but are not required except in DR! intersection analyses. The only methods specifically precluded by FOOT and DCA policies are those based on the 1965 Capacity Manual orTRB's Circular 212. which have been superseded by the 1985 HCM. B. LOCAL OPTIONS Different Methods for Different AgplicaJions Under the Growth Management Act and implementing Rule 9J, commwlities must analyze roadway levels of service in their local comprebensjye plans. They must artalyze levels of service before issuing development orders or permits as pan of concurrency manaeement. And to assess progress in implementing their comprehensive plans, loc.aliJies must analyze levels of service on a "continuous" basis as pan of a monjtorine and eyaJuation process. Methods of analysis may vary from application to application. Indeed, they probably should vary since the consequences of error are more serious for some applications than others. If level-of-service projections in the comprehensive plan are erroneous, a jurisdiction will probably catch the error later during roadway monitoring or detailed project review.


However, if errors are made in the development review process. a project may be approved lll:itbout adequate mitigation or may be disapproved without justification. leaving a jurisdiction's decision open [I) legal challenge. ntis suggests tbe following guideline: les s precise methods of analysis are adequate for local comprehensive plarming, while more precise methods of analysis are required for development review and approval. Roadway monitoring demands an intennediate level of precision. From interviews IIIith transponation professionals, we find Utat many jurisdictions follow this guideline. In their LCPs they rely on FDOT's generalized maximum volume tables or local maximum volume tables generated with FOOT's LOS prognm (see Figure 1). In development review and approval, these same jurisdictions typically allow or encourage developers [I) use more precise methods to assess levels of service on impacted roadways. The trick is to achieve a measure of consistency between roadway levels of service in LCPs and traffic impact studies conducted by different developers. When different analysts use different methods and make different traffic assumptions, consistency is sacrificed. Jurisdictions have responded [I) the need f or consistency in various ways: TallahaSsee supplies developers with maximum volume tables for individual roadways generated with ART-ALL These are the same maximum volume tables used in Tallahassee's LCP. Developers are required to use these tables in their uaffic impact studies. Broward County supplies developers with projections of traffic volumes. including traffic generated by proposed developments assigned [I) the road netWork. using the county's own uavel model. Developers analyze roadway levels of service using county-prescribed methods of analysis and local traffic. roadway. and signal data. Roadway capacities derived by developers are then adOpted by the ooumy as replacements for the generalized capacities assumed in Broward County's LCP. Brevard County is integrating an anerial analysis procedure from the 1985 HCM in[l) a regional travel model. The model will be run for each proposed development, yielding consistent estimaies of uaffic volumes and levels of service on major roads in the area. The Oty of Tampa plans to begin charging subswllial fees for development review. The fees will allow the city [I) carefully check the assumptions and results of developer-prepared traffic impact studies. 11


12 At the v ery least. the need for consistency demands that localities promulgate assumptions about traffic, roadway, and signal characteristics for developers' use in traffic impact studies. Developers should be allowed to make a case for alternative assumptions. If the case is compelling, lhe developers' assumptions should be adopted by the locality for use in LCP updates and furure traffic impact studies Consistency also demands that the methods used 10 analyze levels of service produce consistent results. It has been shown that even assuming lhe same traffic. roadway. and signal characteristics, FDOT's LOS program (the source of generalized and local maximum volume tables) may yield roadway level-of-service estimates that differ significantly from those obtained with software based directly on the 1985 HCM (HCS and ART-ALL software).1 Figures 6a-6c illustrate the magnitude oftbe differences. FDOT is in lhe process of updating both LOS and ART-ALL. FDOT wants the resulting software packages. renamed ARTTAB and ART-PLAN. respectively. to be c onsistent with one another and with HCS. Mjxioe and Matchioe of Me!hods In the unpublished draft revision of the Level of Service Manual, fDOT raises a cautionary nete: fDOT does liOl regard mixing and matching of different evaluation teChniqueS as an acceptable practice. For example, if !be ART-PLAN computer model is being used in the preparation of a local government comprehensive plan. it should generally be used for all arterials and the generalized tables should not be used for some anerials ... .Some areas of the state are using the tables as a planning tOOl to identify pOtential problem sections of roadways and then using more refined techniques to analyze them 10 see if lhe problem is real, (in) a tiered approach. Cenainly, this approach of using the tables and then using lbeir deriving models bas merit; however, there is a danger in the approach when applying it to cities with many road sections.11 FDOT's concem is well-founded Selective application ofmelbods of analysis tends 10 minimize apparen1 traffic problems. One jurisdictiort. for example, applies FDOT's generatized maximum volume tables 10 all roadways; if levels of service fall below adopted standards, the county then develops maximum volume tables for specific roadways that eliminate many of the apparent problems. Another jurisdiction allows developers 10 first analyze roadway levels of service based on generalized tables; for roads wilh apparent problems. developers may then apply more detailed melbods of analysis; and for any roads that remain problematic, developers may conduct travel time and delay studies. If roads pass this final test, developments may proceed


Figure 6a S.R. 50 LEVEL OF SERVICE COMPARISON uutotl MMUIJil OJII\'IWA'I' IOif:JitiiDOM ........ tL. .. . .. I ---l,1'l0 t,IOl I S.tiS Lt" c ARTI .. IAL ANIJ.VIII I ... e .. OVERALL LOS D" SPUD tJ.f MPH \.01 IO"WAIJII IPilC.,I C I MAX. Vot.UIIII LOI ... .. .. ,.DOT QINIRALI%1D I MAX. VOLUWII ., . .. .,. Figure6b S.R. 535 LEVEL OF SERVICE COMPARISON .... u AIIIKWA'f MOUL l .tl WI:IT I IAII "'AU ILYO. IWI" IUMPI .. ,. AfiiTI.I'IIAL NtALYJ:II LOI tonWMI 141'l UICIPIC IIAX., PDOT-..WIIO MAl. YO&. ...... 1 ... 1 ... I Lot \ttl LU.a LtU ... ... ... O'f'lltALL LOS IPEID M.l IIIIPH ... ... .. ... SouJ'tt: Reid Ewing, Anto ine Khoul)', Glatting Lopez Kercher Angtin, Inc. Paper delivered at the 1990 ASCE Florida Section A nnual Meeting. I 13


14 "'' tt:t .... AUifUAL AHAl.Ys.,01 lOS IOn'WME S1Tl snetnc WAl, VOLUMES "'' FOOT OlNifiiAUZED MAJ VOLUYU ... SOIIJ'ce$: Figure 6c U S 192 LEVEL OF SERVICE COMPARISON c .. tiC WAif' kYO, ,. . . .. HOUMY ....... f'Ot..,.. .,. .... OVERA&.I. &.OS SP.0 JO, t "'" .. ; ,,, ... ... .......... ...... 2.).)1 . .. Reid Ewing, Antoine Khoury Glatting Lopez Kercher Anglin Inc., Paper delivered at the 1990 ASC Florida Section Annual Meeting. .... '" The selective application of methods creates a ratchet effect. Roadway levels o f service can only improve as estimaies are refined, since successive me1hods are only applied to roadways with apparent problems. The solution is 00110 reject a "tiered" approach 10 level-i!f service determination. Rather it is to apply more precise methods 10 all roadways identified as possible problems whether they appear 10 operate above or below adOpted level-of-serviu S1alldanls. In this manner, localities and developers can benefit from the efficiencies of a tiered approach without biasing results. Brevard Coumy has designed a tiered approach along these lines. When road traffic reaches 85% of !he generalized maximwn acceptable volume, a road is designated ''distressed." It is then subjected 10 a more delailed analysis using ttaffic, roadway. and signal data specific 10 that roadway. Minimal bias is introduced because the 85% threshold recognizes a margin or error in preliminary Jevel-ilf service estimates on both sides of the "uue condition."


Qeneraliud Maximum VOlume Tables Wbile simple methods of analysis have !heir uses, lhe ultimate in simplicity generalized maximum volume tablesmay be too" general" 10 be useful. If estimaied and actual levels of service are 100 far apan. even a tiered approach will not catch traffic problems. To illustraie lhe limitations of generauzed tables, we need only consider the experiences of three Jocatities. Lee County switched from generalized maximum volume tables 10 local maxi mum volume tables derived with FDOT's LOS program The result was a 6 10 7 percent increase in maximum volumes at level-of-service E Aagler County developed maximum volume tables for state roads. after finding lhatthe driving experienc e on those roads did not correspond 10 levels of service in FOOT's generalized maximum volume tables. Orlando began using ART-ALL 10 analyze state roads after plaMers found lhat level-of-service estimateS with ART-ALL differed by one or two l etter grades from those obtained with FDOT's generalized maximum volume tables. Generatized maximum volume tables are based on "generatized" assumptions about traffic. roadway. and signal characteristics. It would be pure coincidence if the amount of green time devoted 10 arterial-through movements, the percentage of vehicles using exclusive tum lanes, the arrival panem of vehicles at intersections, and other characreristics of a given roadway approximated the input values assumed in the generalized tables. Of course. no method of analysis is as easy to use as generatized maximum volume tables. The tables require no aaffic. engineering experieooe. no time-consuming analyses. and no data other than daily aaffic volumes, signal spacing, and number of through lanes. Nonetheless. the additional effon involved in estimaiing levels of service with FOOT soft ware (ART-TAB or ART PLAN) may be justified by the increase in precision The programs are user-friendly and the data requirements are modest. FOOT will provide traffic counts and estimated K-factors and D-factors for an state roads. City and county traffic engineers can provide information on traffic signal timing, spacing, and free flow speeds. 1Urning move ment coums can be done at low cost for morning and afternoon peak periods. In a matter of minutes, data can be eruered into a microcomputer, and the level of service determined for a given roadway. 15


16 Past Comparisons or Melhods 1lle choice among methodS shOuld be based on the ease or usc and the accuracy of their l evel-of -service estimates. To our knowledge, there have been no published srudies compar ing actual roadway levels of service to levels of service estimated by different methodS. ReportS and anicles comparing traffic models and methodS of analysis have stopped shon of field resting.'' 1llere have been some unpublished attempts to val idate methods of analysis. A transporta tion consulting firm in Tallahassee compared average travel speeds on Apalachee Parkway 10 level-of-service estimates based on maximum volume tables and on .ART-ALL. ART-ALL estimates came closer to acrual levels of service than did maximum volume tables, but all estimated levels ofseNice were lower than acwallevels derived from travel time srudies." The Dade County Planning Department compared travel time srudy results for eight county roads to levels of service estimated with HCS, ART-ALL, and FDOT's LOS program. According to preliminary analyses, none of the methods duplicates field data. Indeed, when methodS predict LOSE or F. roadways are often at LOS C." An Qrland().based consulting firm compared peak-hour travel speeds on 12 anerial sections to level-of-service estimates generated with HCS, ART-ALL. and FDOT's generalized maximum volume tables. The generalized tables produced the lowest levels of service; the travel time srudies almOSt always produced the highest levels of service." These results have profound implications. If common methodS of analysis underestimate average travel speeds, past LCPs and DRJs may have reached enoneous conclusions about roadway levels of service. Furure LCPs and ORis could be compelled 10 apply new methods or revamp existing methodS. A More Rieorous Eyaluation 10 assess common methods of analysis, traffic and speed data for three anerials were acquired from consulting finns 1\vo of the arterials. Kirlcman Road and Tulkey Lake Road. are in Orlando." The fonner has high traffic volumes and low signal density, tbe laaer relatively low traffic volumes and higher signal density." Intersections were first analyzed witb HCS. and results were fed into the HCS arterial analysis program to obtain estimateS of overall average travel speed for each arterial." Assumptions from HCS rurts were then carried over to ART-PLAN and ART-TAB rurts. This meant tbat the three methods of estimating levels of service could be compared to travel time rurts with some assurance that all were measuring the same conditions.


Estimated and actual average travel speeds and levels of service are compared in Figures 7 and 8. Given two anerials, two peak periods, and two directions, eight comparisons ean be made It appears that actual travel speeds are significantly higher !han estimated speeds in nearly all cases. They areS 10 10 mph bigher in most cases than H CS-de rived travel speeds. AM PM Figure 7 Avaase Travel Speeds and Levels of Service Kirkman Road ;.: 50 30 20 10 Northbound Southbound Northbound Southbound AM PM HCS E ART-PLAN CJ Speed Study Levels of Service HCS ART-PLAN ART-TAB Speed Study Northbound B B B A Southbound c c c c Northbound c B c c Southbound B B c B 17


AM PM 18 FigureS Aver33e Travel Speeds and Levels of Service TUrkey Lake Road Averaoe Trayel Speed (mph) Northbound Southbound Northbound Southbound AM PM HCS Study Levels of Service HCS ART-PLAN ART-TAB Speed Study Northbound B B B A Southbound B B c B Northbound B B B A Southbound c B c A


ART-PLAN predicts much higher speeds on !Goonan Road than does the more sophisticated tiCS software. and slightly higher speeds on Turkey Lake Road. Thus. ART-PLAN is more in line with field data. ART-TAB does not estimate average travel speeds but only maximum volumes at different levels of service. Still, comparing levels of service in Figures 7 and 8, ART-TAB seems to be the least precise of the three methods The othe r anerial analyzed was Bmward Boulevard in Fon Lauderdale/Plantation.20 Compared to Kirlcman Road, Broward Boulevard has higher traffic volumes on side streets and thus can claim a smaller ponion of the signal cycle to accommodate its heavy traffic volumes.21 Estimated travel speeds on Broward Boulevard are a fraction of acrual speeds (see Figure 9).12 If results for !Goonan and Turlc.ey Lake Roads suggest that methods of analysis under estimate travel speeds, results for Broward Boulevard indicate that methods break down entirely when demands are too heavy relative to intersection capacity Figure 9 Estimated 'VS. Actual Travel Speeds Browanl Boule9anl 30r 20f---Eaetbound Westbound Eastboun d Westbound AM PM ARTPLAN -Speed Study 19


20 Why Estimates Differ fipm Acrual Ttayel Sgeells To bclp explain why actual travel speedS are higher than estimates, results fOr Kirkman and Turkey Lake RoadS were analyzed by link and by component of total traveltime, the compo nents being delay at intersections and ruMing time between intersections. Typical results (fornonhbound AM movements) are presented in Figures 10 and II. SlOPped delays in the travel time runs are usually shoner than estimated with HCS and ARTPLAN. Differences are exaggerated for intersections with longer delays Figure 10 Aver.age Slopped Delay (in Seconds) Northbound AM MovemeniS HCS ART PLAN Kirkman Road Carrier 10 Intenlallonal 10. 0 11.4 Intema!lonal 10 Major 7.3 7.3 Major 10 VinP"'ad 9.8 11.1 Vineland 10 Conroy 29.3 28.8 Average 14.1 14.7 Turkey lake Road Sand lake 10 Wallace 3.4 3.4 Wallace 10 Paalher 2.6 2.6 Panlher 10 Paw 1.0 1.0 Paw 10 Hollywood 1.9 1.8 Hollywood 10 Producdoa 1.2 1.2 Producdoa to VIneland 0.9 0 8 Average 1.8 1.8 SPEDSTUDY 12.15 6 0 11. 2 6.4 9.0 12.3 2.6 0 0 5.7 0 3.5


Figure 11 Average Running Speed (In MPH} Northbound AM Movements --AliT PLAN Kirkman Rood earner to lnkmallonal 32.1 }1.4 l ntemat.lonal to Major 36 8 <16.7 Major 10 Vlndand 33. 2 33. 1 Vlndand 10 Conroy 35. 2 3SJ AftnJO 35.2 39. 6 1\u-keyl.aRIIood Sand tak< to Walla<:< 38.9 38. 9 wau-10 Pantber 38.2 382 PMt.ber to Paw 32. 1 32.4 Pa,. 10 llollytvood :!9. 8 29.9 llollywood 10 Pn>ductlon 30. 5 30. 6 Ptoclualon to Vlodand 36.1 36. 1 Awrase 36.6 36.7 -SPEW STUDY 29.7 51. 5 39. 1 6 49.6 46 2 48.4 30. 8 29. 0 38. 9 41. 7 The degree of chance inherent in a small sample of uavel time nms c.auses acrual delays to vary greatly from run to run and inlenection to intersection. While the models predjct some delay at all signalized intersections, no delay was acrually experienced at certain intersections with high green ratios and good progression. 'This has a pronounced effect on average uavel speeds, not directly but through its effect on running speeds (see below). Rupnjna speed$ in the uavel time nms are significantly highe r than estimated with either model, and as such, account for most of the difference between actual and estimated average travel speeds. FoUowing convention. the posted speed limit was talcen as the free flow speed in model nms Yet. roads are often designed for safe speeds in excess of the posted speed limit, and as casual obseJvation sugg ests. drivers have a tendency to drive at design speeds on long. Wlintenupted segments wilh moderate traffic volumes." 21


22 There is another reason why running speeds in travel time runs are higher than estimated with HCS and ART-PLAN. The models assume some stOpped delay at all intersections. and hence some acceleration and deceleration which depress average running speed significanlly on shon segmerus."' However, in lhe real world, vehicles do not slow down unless they have 10 Stop at a uaffic signal, and the free flow speed may be maintained over the entire length of a shon segment or even a series of shon segments. It is easy 10 see why actual travel speeds would differ from estimated IJ'avel speeds. But why would HCS and ART-PLAN estimates diffe!1 After all. they are supposedly based on the same equations from the 1985 HCM and lhe same assump1ions (or at least consistent assumptions) about traffic, roadway. and signal characteristics. Sll'aight running times between intersections are the same for HCS and ART-PLAN. with one exception. HCS (Release 1.5, the latest ) incorrectly extrapolates running times for segmerus of more than one mile when the free flow speed is 45 mph. Kirlanan Road contains one such segmen1. which depresses lhe average travel speed on Kirlanan Road in the HCS runs." This error went undetected until results were analyzed by link and by component of total uavel time. Adjusted Stopped delays differ slightly between HCS and ART-PLAN. The reason is not obvious. since bolh are based on lhe intersection delay equation in the 1 985 HCM (Equation 9-18). ART-PLAN's delay estimates are often lower than HCS', and therefore ART-PLAN's estimates of running speed and average travel speed are higber.Z' This brings them slightly closer to travel time study resultS. It appears, then, !hat ART-PLAN js at leaS[ as good a predictor of anerial leyels of service as js HCS. And since it is much easier 10 use. ART-PLAN would appear to be the preferred method of arterial analysis. Local Maximum Yo!umc Tables As previously noted. l evel-of-service estimates based on ART-TAB are less precise than those based on HCS and ART PLAN (at least for !he two arterials tested). ART-TAB estimates also tend to be lower than level-of-service estimates based on the other methods, two leuer grades lower in one case than observed in the field (see Figures 7 and 8). In one sense, ART-TAB is the miiTOr image of ART-PLAN. ART-PLAN predictS levels of service, given u-affic volumes and characteristics of traffic. roadways, and signals. ART TAB predicts maximum traffic volumes, given l evels of service and characteristics of traffic,


roadways, and signals. ART-PLAN workS lhrough lhe intersection delay equation (Equation 9-18 in lhe 1985 HCM) from left to right; ART-TAB works lhrough lhe same delay equation from right to left. However, !here is one fundamental difference between ART-PLAN and ART-TAB. Willi ARTPLAN, uaffic volumes. %turns from exclusive lanes. signal cycle lengths, green ratios. and intersection arrival types are defined segment-by-segment. The overall level of service for a roadway is determined by summing lhe contributions of individual segments In conuast. wilh ARTTAB, traffic volumes. % turns from exclusive lanes, signal cycle lenglhs. etc have 10 be averaged over an entire section of roadway." Average traffic volumes are !hen compared 10 maximum traffic volumes derived wilh ART TAB, which also represent a kind of average, to determine lhe overall level of service Valuable information is lost in lhe process of averaging. and lhe estimated leve l-of-serv ice need not be lhe same for lhe two melhods since lhe relationsh.ip between intersection delay and lhe characteristics averaged is not always linear. This accounts for ART-TAB's lack of precision. But why are ART-TAB's level-of-service estimates apparently biased downward? The answer is lhat FDOT's Level of Service Manual requires lhe use of a wei&)lted average green ratio (g/C ratio) when developing maximum volume tables, ralher than a sjmple average. Green ratios of all intersections along an arterial are first averaged, and then this simple average is combined wilh the lowest green ratio of any intersection along lhe anerialto produce a weighted average. In this manner. lhe intersection wilh lhe lowest green ratio is given many times as much weight as any oilier intersection. To determine how much effect this weighting has on levels of service, maximum v olume tables were created for Kirkman and Thrkey Lake Roads first using FDOT's weighted average green ratio, and !hen using lhe simple average Average traffic volumes were compared to maximum volumes for each set of tables 10 determine levels of service. Figure 12 presents the results. In many cases, lhe use of a simple average raises levels of service by one letter grade. Usually. this brings ART-TAB results in line wilh ART-PLAN results However. in one case, it causes ART-TAB to predict beller levels of service !han does AliT-PLAN. It is not clear from this preliminary investigation whelher FDOT's weighted average green ratio, a simple average. or perhaps a different weighted average shoold be used in level-of-service determinations


Figure 12 Levd-of-Scrvk:c Estimates ART TAB ART-TAB AllTPtAN ('tl'd!lbl< {Simple -'"""18< giC) A....-ageg/C) KIFkmanllood -...s 8 B B AM Soutbbc>wwl c B c -"" c B II I'M c B II l.aJu: Rood Nonl>boun4 B B B AM so.nbbouod c B u Nonbboun4 8 8 B ... 8 Soulbbound c u Agplication of Dave! nmc Srudics Only three of II tranSpOrtation professionals inteJViewcd for this study said that their jurisdictions have conducted travel time srudies. or plan 10 conduct them, as a method of determining levels of seJVice. One reason Is the relatively high cost of such studies; this factor may be rendered moot eventually by advances in automatic vehicle location (AVL) teebnology. Tbe more importanl reason is the perception that travel time study results apply only 10 the specific time period when travel time runs are done that they cannot be used 10 projea future levels of seJVice, as required in growth management.


'This perception is incorrect Travel time study results could be used in at least two ways to predict future levels of service. Both warrant funher study in light of the subslanlial disparity between actualttave\ speeds and estimates made wilh s1a11dard methods. Travel time study results could be used to better calibrate HCS. ART-PLAN, and other programs. Default values assumed by these programs may not be applicable to a particular locale'" HCS or ART-PLAN could be run with progre s sively higher saruration flow rates and free flow speeds until estimated intersection delays, running speeds and overall travel speeds better approximate traveltime srudy results. The better-calibrated models could \hen be used to forecast future lev els of service. Alternatively, traveltime study results could be correlated directly with traffic volumes and o\her variables in statistically derived models. In its unique approach to level-of-sel"llice detennination, Seminole County developed local maximum volume aables by averaging daily traffic volumes for all two-lane roadways whose average uavel speeds placed them at a given level of service. For example. the daily traffic volumes on roadway sections operating at LOS C were averaged to obtain the "maJCimum "traffic volume for LOS C. Repeating this process for roadway sections operating at other levels of service, local ma.tmum volume tables were developed that mllect actual travel speeds within the locale. The approach taken by Seminole County could be improved upon with statistical techniques. Rather than averaging traffic volumes at different levels of service. which disregards the variability of traffic volumes within a given level-of-service grade. a technique such as multiple regression analysis could relate traffic volumes to average travel speeds in a continuous manner Multiple regression analysis has the added advantage of allowing additional characteristics of roadways to be factored into equations as independent variables. To illusuate the approach. average peakhour uavel speeds and traffic volumes were acquired for 17 two-lane roadways in Seminole County. With a.m. and p.m. peak hours, and nonhbound and southbound directions. speed and volume data were available for a total of 68 movements. Average pealt-hour travel speed was regiCSsed on pealthour traffic volume and two other variables. number of signalized intersections per mile and free-flow speed. Both linear and nonlinear fonns of the regression equation were tested. The best fit to the data was obtained with a linear equation in two independent variables. pealt-bour traffic volume and signalized intersections per mile (see Figure 13)." Free-flow speed proved statistically insignificant. The other two variables were significant at the O.oi level or beyond. Together they explain SS% of the variation in average peak-hour travel speed among roadways. To use a model of this type in forecasting average uavel speeds and associated levels of service, it would ort1 y be necessary to subStitute projected values of the independent variables 25


26 into the regression equation. Altematively, the model could be used to forecast changes in average travel speeds as peak-hour traffic volumes increase or other variables change The explanatory power of the mOdel t$Dmated for Seminole County is probably inadequate for use in forecasting future travel speeds and levels of service. The standard enor of the estimate. 5.3 mph, could result in a one-or even two-leuer grade difference between estimated and actual levels of service Nonetheless, with 55% of the variation in average travel speed explained by only two variables, one has to believe that a good predictive mOdel could be developed with a richer data base (including such independent variable$ as the green ratio, arrival type, and % rums from exclusive lane$). The regression mOdel's simplicity may be a virtue. The complicated models and multitude of parameters in the 1985 HCM only give the appearance of precision. In light of results for Kir!cman Road, Turkey Lake Road, and Broward Boulevard, the added complexjry may 1101 translare jmo added QRcisjon, Average Travel Speed figure 13 Regression Equation Relating Average Travel Speed 10 Peak Hour Traffk Vohme and Signals Per Mile Seminole County :z LaDe Roads 44.7 R' 0.0087 ]1: (3.1:Z) Standard Erro r Peak Hour Traffic Volume 0.55 5.3 Number of Oblservations Degrees of Freedom 68 65 t statistics shown in parentheses 7.74 ]1: (6.6S) Signak Per Mile


W.AREAWIDE LEVELS OF SERVICE It is routine in uaffic impact studies to detennine leve l s of service for: a lane group at an intersection an entin: intersection, a roadway segment from intersection to intersection and a section of roadway with multiple intersections along its length. AS w e move up the hierarchy from intersections to entire roadway sections we are averaging or aggregating levels of service. The procedure for doing so is straightforward. at least for urban and suburban arterials For signalized intersections. levels of service are measured in terms of average stopped delay Add to this the delay approaching intersections and the running time between intersections, and we obtain lOla! uavel time on a roadway segment. Divide thiS by the length oftbe segment, and we arrive at average uavel speed. which deter mines the level of service. Do the same for several segmentS in a series, and we obtain an estimate of level of service for a roadway section No difficult conceptual issues arise. and no one questions the basic logic of averaging or aggregating levels of service in this context However. we find ourselves in uncharted waters when we begin to combine levels of service across facilities. 'l1lere is no standard. professionally accep!ed method of averaging or aggregating levels of service within: a ttavel corridor, a uaffic district or an entire road ltelwork. 27


A. STATE POLICIES FDOT's Level of Service Manual FDOT's Level of Service Manual sets minimum operating levels of service for state roads and presents generalized maximum volume tables for state and local roads. Local governments and DIU applicants often assume that the standards and maximum volumes apply to individual roadway segments. from intersection to intersection. lYPically, neither FDOT nor DCA challenges segment-by-segment analysis when it appears in LCPs or ORis Yet the manual stales explicitly at one point that maximum volume tables should be applied to "sections of roadways"; the sections should be at least pne mile in downtown areas and two mUes elsewhere. Either average or median traffic volumes for entire sections may be compared to maximum volumes in the tables to determine levels of service. In effect, several segments and intersections on a single facility are being aggregated. FOOT's rationale for combining them is: "Although directly related and important, congestion at individual intersections or links may be acceptable so long as the overall quality of flow is acceptable ... Rule 91-5 This rule requires the establishment of level-of-service standards for all roads in a local jurisdiction as pan of the local comprehensive plan. Standards must conform to Subsection 91-5.005(3), whieh reads in pan: "Eaeh local government shall establish a level of service standard for eaeh public facility .... Such level of service standards shall be set for eaeh individual facility or facility type and not on a systemwide basis.'' DCA has interpreud this rule to preclude LOS measurement and standard selling for collections of road facilities within conidors or distriets. Legislation introduced during the 1990 Aorida Legislative Session would have allowed local governments 10 adopt aye@&e areawide level-of-service standards for "transportation concurrency management areas" (TCMAs). DCA was directed to establish criteria for delineating TCMAs and 10 develop methods of defining, measuring, and monitOring average levels of service within TCMAs. The legislation containing this provision (Senate Bill 1794) was never enacted, but DCA is seeking a rule ehange toward the same end. Under a new section of Rule 91-5, local governments could designate TCMAs in their comprehensive plans and adopt an oye@ll level-of service standard for the tranSportation facilities and services within each TCMA.


Overall level-of-service standards for TCMAs would replace St3IIdards for individual f acilities wilhin TCMAs. As the draft currently reads: "Use of a uansponation concurrency management area system shall allow development orders and pennits to be issued wilhin the uansponation ooncuiTency management area without consideration of the uansponation level-of-service of individual uansponation facilities or services so long as the overall level of-service SWldard adopted for the uansponation concurrency management area is l'rccedents DCA has accepted averaging or aggregation of roadway levels of service as pan of at least two stipulated senlement agreements with localities The agreement with Brevard County allows the averaging of parallel roa

30 networl< exists, an individual may have many routes available for a given trip. Ordinarily, the routes will offer different levels of service since they are made up of segments with varying travel demands upon them from other trip makers. If any route provides an acoept able level of service, government may have met its responsibility to the individual trip maker. Hence, roadway levels of service might reasonably be aggregated within travel eonidors The concept of alternate routes raises an important issue: 1b what degree is government responsible for accommodating individual travel preferences? lllis issue will surface again in the sections on "Multimodal Levels of Service" and "Choice of Analysis Period." Have concurrency requirements been satisfied if a trip can be made a1 an acceptable level of service by some using some mode, at some B during the peak period? Or mUSt acceptable levels of service be provided for every route, mode, and hour of travel that might be chosen by trip makers? In effect, ellisting Rule 9J-5 interprets the Growth Management Act as requiring "adequate facilities" along every route, for every mode and hour of travel. The proposed amendment to Rule 91-S would relax this interprewion. Methods of Aggrcaation How can levels of service on individual facilities be aggregaled into one areawide level of service? While there is no standard approach, three possibilities suggest themselves. All three have precedentS in LO>s already approved by DCA. They are: (l) summing volumes and capacities, (2) averaging levels of service. and (3) adopting performance swnmaries. The first approach is to compare the sum of traffic volumes on roads in an area to the sum of roadway capacities (where capacities are equal to mallimum volumes at adopted levels of service and both traffic volumes and capacities are weighted appropria!ely). If, in the aimite, capacities exceed traffic volumes, the area might be deemed to meet level-of service standards.


Under its senlement with DCA, Lee County may aggregate uaffic volumes and roadway capacities and use any surplus capacity to justify degradation of already 'back logged" roads." Aggregation is accomplished by: (I) summing peak-hour traffic volumes on functionally classified roads in each uaffic district. (2) summing capacities of functionally classified roadS in each district, (3) subuacting aggregate uaffic volume from aggregate capacity to obtain surplus capacity in each district. and (4) comparing surplus capacity to the growth of IJ'affic anticipaled in each district (see Figure 14). A second approach to aggregation is 10 ayera&e levels of service for facilities of a given type in a navel corridor or IJ'affic disuia.. Wbile averaging in this context is novel. averaging navel speeds on anerials has been the norm since the 1985 HCM was released. It is not difficult conceptually or methodologically to go from averaging speeds of vehicles operating on diffemn sections of an aneria!IO averaging speeds of vehicles operating on differept anerials." For anerials, the areawide average speed could be compared to minimum travel speeds at different levels of service (from the 1985 HCM) to determine the areawide level of service. If an area had anerials from more than one anerial class. minimum navel speeds for different classes could be weighted proportionally to produce one set of standards against which average uavel speed could be judged. For freeways and rural multilane highways, density in passenger cars per mile per Jane could be similariy averaged and compared to level-of-service standaJds in the 1985 HCM. The only obvious limitation on averaging is that it cannot be done across types of facilities whose levels of service are measured differenlly Under its settlement agreement with DCA, Brevard County may avcrue levels of service on a main anerial and parallel interconnected colleciOrs. The prescribed procedure involves: (I) dividing IJ'affic volume on each road by the maximum volume at the adopled level of service, and (2) weighting and summing the resulting ratios over all roads to arrive at the average proportion of maximum acceplable volume being utilized by traffic. 31


... .... --......... Figure 14 Tramc Volum<:s vs. C a pacllles (By Dt.ltri

Being among 11le first to prepare an LCP. Brevard County used volume/capacity ratios 10 define levels of service. However, 1lle basic approach would be 11le same if travel times were averaged instead. consistent wi11111le 1985 HCM. A third approach 10 aggregation is 10 adopt a performance summazy for roads in an area, which specifies 11le percentage of roads at or above level-of-service Standards. Unlike 1lle preceding approach, which applies an average standard to an average roadway, litis approach retains individual Standards for individual roadways. Roadway level of service is aggregated only in 1lle sense 11lat certain roadways are allowed to deteriorate as long as o11ler roads are upgraded even more. II is 1lle performance summary. not individual roads. 11lat must show improvement over time. At least 1llree local governments have used performance summaries to meet concurrency requirements (see Figures 15 and 16). Orlando's approach is exacUy as

Source: Figure 15 District Performance Criteria Traffic Percent of Lane Miles Operating Performance at or above Level of Service Standard District 1995 2010 1 73%. 75% 2 69% 73% 3 33% 52% 4 32% 34% s 79% 88% 6 81% 84% 7 88% 88% 8 80% 95% 9 60% 66% 10 SS% SO% 11 59% 62% 12 89% 97% 13 100% 100% 14 85% 91% 15 51% 58% City of Orlando Transportation Planning Oureau "Growth Management Plan Transportation Elements," Presentation to the Municipal Planning Ooard, September 11, 1990.


management. Even if level-of-service sWJdards remain tied 10 individual facilities, areawide averaging will gain all of the legitimacy required for growth management if it is prescribed by DCA rule. Factors Wbicbever melhod is chosen. roadways must be assigned weights that reflect lheir contributions 10 overall levels of service. Lee County weighted traffic volumes and capacities of roadway segments by tbeir respective lengths (i.e., by centerline miles). Brevard County also used segment lengths as a weighting fac10r. Pasco County weighted its performance summary by the number of vehicle-miles traveled on different roads; Orlarulo used the number of lane-miles; and Tampa used centerline miles in one performance summary and vehicle-miles traveled in another Use of yehicle-miles accounts for the volume of traffic exposed 10 different traffic conditions. Since it is the .. average" experience of travelers we wish to caprure in an overall level-of service measure. t\01 the average condition of roadways, vehicle-miles would seem 10 be the Scluttt: Figure 16 System Performance Plan For All DisUicls City ofTampa so 23 1987 1995 2000 2010 YEAR City of Tampa Comprehensive Plan Traffic Circulation Element, p. 3 35


36 preferred weighting factor. Use of other weighting factOrs could encourage improvements to low-volume roads simply to meet concurrency requirements, while higher volume roads go unattended. Delineation o('fCMAs Localities will require some guidance as !hey begin to delineate conidors or distriCts within which levels of service are aggregated This discussion will refer to such areas generically asTCMAs If TCMAs are too large. uaftic problems will be glossed over and development decisions will be subject to challenge. We might expect property owners near !he edges of large TCMAs, for example, to challenge project disapprovals prompted by uaftic congestion at central locations or opposite edges. lfTCMAs are too small, flexibility to respond to systemwide needs will be sacrificed. In !he extreme. TCMAs will cease to reflect motorists experiences on typical trips or !heir choice of alternative routes and simply become surrogateS for individual facilities. Many criteria have been suggested for delineating TCMAs. Several seem intended to ensure !hat TCMAs caprure !he experience of motorists on typjcal trips. Transportation professionals interviewed for this srudy suggeSted !hat TCMAs be sized to reflect average trip lengths, be bounded by narural baniers or freeways, have few enuy and exit points, and correspond to distinct communities within !he metropolitan area. Where naruraJ baniers exist between TCMAs. entry poims are limited, or 1hese other criteria are met, !here is likely to be more trip malcing within TCMAs !han between TCMAs. Thus, as intended. TCMAs will capcure !he experience on typical aips." Other criteria seem intended to ensure !hat TCMAs reflect !he availability of alternate TO!Jtes. In its draft rule, DCA baS limited aggregation 10 roads !hal serve "related purposes." What this apparently means (from examples given) is !hat roads running in !he same general direction may be combined since !hey serve as alternate routes. DCA also baS limited TCMAs 10 areas containing "a complete, integrated networ\: of arterial and collector roads. A complete, integrated networ\: ensures that TCMAs will provide alternate routes for any given trip. 1\vo criteria proposed by DCA in its draft rule deserve special anention. A1 a public workshop on !he draft rule, both criteria elicited strong reactions, pro and con. Tbe draft rule would confine TCMAs to special areas where development will assist in achieving state planning goals." Workshop participants who planned to divide up their entire


jurisdicalions objected to this limitation. The City of Orlando, in particular. anticipateS problems in justifying traffic performance districts throughout iiS jurisdictional area under the draft rule (see FtSUre 17). Figure 17 Traffic Performance Districts City of Orlando I 12 [i]-13 Source: Clty of Orlando Planning & Development Department, Traffic Circulation Elemem: Growth Mana11ement Plan. 37


38 In effect, DCA'S draft rule would render TCMAs equivalent to FOOT s special transportation areas (STAs). S TAs are compact geographic areas in which growth management considerations outweigh roadway level -ofservice objectives. STAs are designated by local governmentS and must be approved by metropOlitan, regional. and state agencie s Within approved STAs, roadway level-i>fservice standards may be relaxed The state plaruting goals cited as justification for TCMAs relate principally to land use. However, the State Comprehensive Plan also contains goals related to transportation system efficiency' These goals might justify the designation of TCMAs throughout a jurisdiction. even if land use goals do not. After aU, the purpose of aggregating levels of service across facilities is to promote cost-effective system improvements instead of. isolated roadway improvementS diCtated by minimum operating standards. Another point of contention at the workshop was the relationship between TCMAs and tranSpOrtation impact fee benefit districts. Some workshop participants argued that TCMAs should coincide with impact fee benefit districts. In Lee County and Tampa. the two coincide for the most pan. Olher participants contended that TCMAs bear no relationship 10 impact fee benefit districts. Orlando's impact fee benefit districts were said to be much tOO l ;uge 10 serve as uaffic performance districts. The draft rule applies the same basic requirement 10 TCMAs as the courts have applied 10 tranSpOrtation impact fee benefit districts ... Under the ''dual rational nexus" teSt, impact fee benefit districts must be small enough to ensure that impact fees spent on uansponation improvements anywhere in a district will be of benefit to feepayers anywhere in the district" This test will be met only if feepayers or their successors are likely 10 make use of uansponation improvements in their routine trip making. that is, only if their typical trips take them all over their respective impact fee benefit districts. By this standanl, impact fee benefit districts are conceprually no different from TCMAs For guidance in delineating TCMAs, the concep!S of typical trips and alternate route s may be combined in the following general guideline : TCMAs should be drawn so a s ro encoD!INS< altenwe rowes ayailable for cgmmon peak-hour trips. How the general guideline is operationalized is best left to local planners Let it suffice 10 say !hat the guideline could be operationalized. For example, regional uavel models could be used 10 generate tables of trip interchanges between traffic zones, and from these, common origin-destination pairs could be identified Because level-of-service standanls apply to peak hours, primary consideration might be given 10 work trip interchanges TCMA's boundaries could be drawn so that traffic zones between which a majoriry of trip interchanges occur are pan of the same TCMAs


Areawjde and Tiere!l S!andards In a recent issue of the Florida EngjDAArin& Soctety Journal. two authors debated the merits of averaging roadway level of service within a corridor, district. or entire urban area. One author contended that averaging could result in a ''glossing over of transpor!ation problems. "41 The other aulhor countered that requiting each roadway link to operate at a minimum level of service could cause 'short-term incremental improvements rather than long-tenn comprehensive improvements."" Bolh aulhors are right. The cballenge is to devise level-of-service S!andards !hat will encourage a long-term comprehensive approach to transportation improvemem programming while still addressing locali1.ed traffic problems. One possibility is to es!ablish areawide S!andards only but measure levels of service in a manner that gives considerable weight to locali2ed traffic problems. As noted previous! y, averaging levels of service across facilities. and weighting facilities by the vehicle miles traveled on them, shoul d have the desired effect Another possibility is to establish tiered standards. Standards for individual facilities could be set at the absolute minimum level, while areawide average s!andards are set at a higher desirable level. For example, the minimum standard for individual facilities could be LOS E or worse (say 2 mph below the minimum travel speed for LOS E). A1 the same time. the minimum areawide average might be LOS B in rural areas, LOS C in urbanized areas. and LOS D in special transpor!alion areas. 39


IV. MULTIMODAL LEVELS OF SERVICE With so muCh of Florida's growth occurring in the age of the automobile, growth management has naturally focused on the adequacy of roadways. Low-density development panems and poor transit service have left mass transit in most Florida urban areas with insignificant maJI

Miami's approach 10 level of service is based on the philosophy that .. 'roadway capacity' and iiS derivative 'level of service are functions not of how many person trips are being made. but ralher of the collective decisions that choose by what manner they are made ... within a reasonable range. traffic congestion exisiS as a maner of collective choioe, not physical law ..... In other words, if commuters wish to subject themselves to congestion by uaveling in single-QCCupant vehicles during the peale hour, it is their decision. no\ local government 's. DCA seemed 10 agree when it found Miami's LCP to be in compliance with state Jaw and regulations. B. LOCAL OPTIONS Role ofTransjl Nearly all of the uansportation professionals interviewed for this study agreed that where bener trariSit service exiSIS, roadway levels of service need not be as high as elsewhere. However, \bey differed on just how good transit service must be before it compensates for roadway deficiencies. They also differed on whether transit availability should be reflected in relaxed roadway level -{If-service standards or in combined level-of-servioe Slalldards for autos and transit Some atgUed for consideration of transit availability only where it provides a level of service competitive wilb \be automobile. Good transit service is not enough; \be servioe bas to be great. ''Great'' means that transit offers 101a1 trip times comparable 10 \be automobile. Considering \be out -{Ifvehicle time lost wilb transit. only syStems wilb exclusive or semi exclusive guideways are lilcely 10 measure up. and then only when congestion on \be street system slows down automobiles. 0\ber.; felt that if transit systems have excess capacity. government has no responsibility 10 provide roadway capacity for every commuter who might choose \be automobile. Miami's LCP. and 10 some extent Dade County's LCP, embrace this philosophy. To be free of \be usual roadway level -{If-service standards. roads need only lie within corridors served by bus lines with W-minute peak-hour headways and excess bus capacity. The two philosophies differ in their reading of where mvemment responsibility ends and jndjyjdual prefqences begin. The capacity-oriented view holds that transit need only have \be capacity 10 serve auto user.;, while the level of service-{)riented view adds \be requirement that transit service be good enough to anract auto users. 41


42 Where the level of service-oriented view may have a conceptual advantage is in its realiza tion that an individual commuter has control only over his own mode choice decision, not over the collective decisions of other commuters Metrorail in Miami offers a high level of service A commuter in a Metrorail corridor can, by his individual choice of transit, avoid silting in traffic. In conuast. a commuter facing the choice of auto or bus travel on congested city streeiS really has no choice. True, if a large number of commuters were to switch to transit, reduced congestion would result in high levels of service for both modes But an individual's choice of transit will have no perceptible effect on congestion levels and leave him with a wall< and a wait in addition to a long travel time. Ahematjye AJWPacbcs As indicated above, Miami defines levels of service for multimodal tranSpOrtation corridors (see Figure 1&). Where a transportation corridor contains a rail line or a bus route with a peak-hour headway of 20 minutes or less. !he ''practical capacity'' of the corridor is esti mated for all modes combines!. The practical capacity of automobiles is taken as 1.6 per sons-per vehicle, higher than current auto occupancy. The practical capacity of transit vehicles is ISO% of seated load for local buses. 125% for express buses, and 130% for rapid rail transit. Both autos and transit have "excess capacity" under these assumptions. Once the practical capacity of a tranSportation corridor is known, it is divided into the person-trip volume within the corridor to obtain a volume/capacity ratio. The level of service is detennined by comparing the volume/capacity ratio for the corridor to the following standards: y/c ratio .01-.60 A .61-.70 B 71-.80 c .81-.90 D .9 1 -1.00 E 1.01 or more F


Figure 18 Transportation Corridors 1988 and :zooo City of Miami I a:a ""Lt-n_._ /-...... I \ ........ _.., ..... ..,.. AfiiUICtlllllf City of Miami Planning Development, Transportation Corridors: Meeting the Challenge of Growth Management in Miami (1989). p. 12. Miami s companion government, Dade County, accounts fo r u-ansit availability by relaxing roadway level-of-service standards in uansit-served areas. Beginning in 1995, Dade County will apply a three-tiered standard 10 roadways in Urban 1nfi1l Areas: Where no transit exists roads may not exceed LOS E traffic volumes. Where tranSit service with beadways of 20 minutes or less is provided within 1/2 mile distance, roads may not exceed 120% of LOS E traffic volumes. 43


44 Where "excraordinary" transit service such as commuter rail or express bus service exists, paralle l roadways within 1/2 mile may not exceed 150% of LOS E traffic volumes. Similar tiered slalldards will apply to roadways in Special Transponation Areas and roadways in rural areas. Both approaches (Miami's and Dade County's) have been accepted by DCA. Both allow roadway levels of service to slip as transit service improves, and in so doing, both presumably encourage transit usage Miami's approach treaiS autoS and mass transit as "equal partners in the trip moving business" (to quOte the planner wbo devised the approach). A seat is a seat, whelherin an automobile or on a bus Equating modes has theoretical appeal. but it must be reconciled with consumer preference for the automobile. Autos and transit are not equal partners in Aorida 's sprawling urban areas. Any system that treats them as equals could have unintended consequences. Transportation improvementS might be funded on the basis of cost-elfectiveness in meeting concunen<;y reguiremems. 1101 cost-elfectiveness in meeting uayel demands. Even if buses ran empty and roads were gridlocked, local government could approve new developments as long as they were "served" by There i s also the practical problem of combining levels of service for different modes. Miami measured levels of service in terms of volum!Vcapacity ratios, as prescribed by the 1965 HiK!JwaY CapaciLY Manua]; it was a simple maner of relating the volume of person trips in a corridor to the sum of the capacities of autos and transiL With the arrival of the 1985 HCM, this methodology has been superceded. Levels of service on urban and suburban arterials are now measured in terms of average travel speeds, while transit levels of service cootinue to be measured in terms of load factors and the like (see Figure 19). Even if localities could develop average trip speed da1a for transit users, including walk, wait, and transfer times, the da1a could not necessarily be combined with average traVel speeds for auto users since the two modes may be subject to very dilferent level-of service slalldards. A ISmph aver.1ge travel speed may be unacceplable to auto users but perfectly acceplable to transit users who can work or read while they ride. The reverse could also be true. Without surveying public anirudes, it is impossible to say.


FljJure 19 Transit Level-of-Service Measures Local Comprehensive Plan Measures Jacksonville Head'9. r ay / load factor/route density Orlando Headway T allahassee Route mileage Tampa Pea k hour load factor Source: Mass Transit Elements, Local Comprehensive Plans. FDOI's Transit Corridor Standards The preceding sections may be summarized as follows: wbere transjt service on ellC)usiye or scmi-exclusiye etrideways is ayaUable. it is reasonable to relax standards for parallc! roa<1ways. As it happens, this policy is already embodied in FOOT's "Swcwide Minimum Acceptable Operating Level of Service Standards for lhe State Highway SySiem'' (see Figure 20). In FOOT's Level of Service Manual (draft revision), roadways paralic! to exclusive transit facilities are subject to levc!-of-service standards generally one letter-grade below lhe oorm. To qualify, roadways must be "parallel to and within one-llalfmile of a physically separated rail or roadway lane reserved for multi-passenger use by rail cars or buses serving large volumes of home/worlt ttips during peak travel hours.' 45


46 Perhaps even lower standards (e.g .. 2 mph below the LOS E SWidard) could be justified when high-quality transit service i s a vailable in the corridor. The only caveat is that some road users have no transit option; this is true of commercial vellicles and auto users with trip ends outside the transit corridor. A minimum roadway standard must be maintained for these users, if not for the auto users who choose gridlock when government has provided them with an attractive transit alternative Flgure20 Statewide Minimum Acceptable Operating Level of Servk:e Standards for the State Highway System Jloadway 'rl-1> ff'HWiyt Principal Arterial t Minor Artl'rials 6 Othen Freeways Principal Arttriala Minor Arterials It Othe n 'l'r a atitioa.ia.c Esittinr OU..r Utboniztd or Utbal:liud Eziltia.c Ineorporalod Rural Areas Cati.S AAu AtUI D c c c D c c c E p D D SI'ECIALCONSIDERA1'10NS P....U.I 1<1 Spodal Esc lui ... T.._ortation Traa..ait CoDJtrain.edi B a c.kloued An81 F a cility Facility Facilitv D D --" trnp,.,. E -Mamtain&lm,..... E E -Mamtain&!m,..... FOOT, Florida s Level of Service and Guidelines Manual for Planning (1969), p 2 -3.


V. CHOICE OF ANALYSIS PERIOD Both Rule 91-5 governing LCPs and Rule 9J -2 governing ORis require that peak-hour volumes be used 10 detennine levels of service. Use of the peak hour is consistent with standard engineering practice in facility design, traffic operations. and traffic control Whenever localities have sought to esublish level-of-service standards based on lliiJ.x ralher !han peakhour rraffic, DCA has objected. It is not enough to prescribe the use of "peak-hour" volumes, though As Bill McShane and Roger Roess note in Traffic .. .if peak-hour volume is to be used as a common focus of design, operations. and control analyses, it is critical to understand which peak hour is being used ."" Among the multitude of choice s are: the single highest hour of the year, the 30th highest hour of the year. the lOOih highest hour of the year, the average peak hour of the peak season. or the average peak hour of the entire year. Which peak hour is selected could have a dramatic effect on estimated levels of service. From Brevard County's LCP, traffic in the 30th highest hour is 10% of annual average daily traffic (K30 = 0. 10), traffic in the JOOih highest houris 9% (KIOO = 0.09), and traffic in the average peak hour is 8%. Thus. the traffic volume during the 30th highest hour is 25% higher than the volume during the average peak hour. !hat extra traffic could make as much as a four lener-grade difference in the estimated level of service. The K-factors of the preceding paragraph are called design hour faCIOrs. They represent the proponlon of the annual average daily traffic occurring during the design hour, that is. during !he hour of the fuwre year which sets the standard for highway design. Customary practice in the United States is to base design on the 301h highest hourly volume in the 20th year of service. This means that facilities are expected to operate at accqxable levels of service all but 29 hours of that year. 47


48 A. STATE POLICIES FpQT' s l&yel of Service Manual FOOT's current manual establishes level-<>f-service standards for the 30th highest hourly traffic volume of the year. The generalized maximum volume tables in the manual are also based on the 30th highest hour (acrually, the highest IS-minute period of tile 30th highest hour). Even the daily maximum volume tables are based on the 30th highest 1\our ... The revised manual will set sundards and compute maximum volumes based on the tOOth highest oour of the year. The tOOth highest hour is aid to approximate the typical peak hour of a during the peak travel season. While FOOT acknowledges that there may be valid reasons for using a time period Olher than lite 30th (or IOOth) hour, the agency inSists that its choice 'is of the FOOT's kvcl-of-service standard$ not to be a(tere(j (emphasis added) .''" Thus, local gov ernments and DRl applicants would appearto have JJnle latitude under FDOT's policy." FDOI's DR! Guidelines These guidelines call for the usc of K30 (presumably to be replaced with K 100 when the Level of Service Manual is revised). This policy has some built-in fle.\ibility, though. The measured peak-to-f-service detenninations based on peak hours other than the 30th higbesL DCA has accepted: The average of the cwo highest consecutive hours of an average wed:day (in Miami and Dade County). The l OOth higheSI hour or the avetage peak hour of days lhroughout the year, provided "the best engineering ptaelices" suggest one of these is "more appropriate"liwllhe 30th highest hour(in Brevard County).


The average peak hour of days in the peak season (in several counties). B. LOCAL OPTIONS lmerplax of Peak Hour and LOS Standard s The choice of peak hour cannot be divorced from the setting of !evei.Pf serv ice standards. The effect will be the same if we appl y lower standards to a higher-volume hour, or higher standa rds to a lower-volume hour. In its LCP. Lee County adopted two standards LOS D for the annual average peak hour and LOS E for the average peak hour of the peak season. The lower standard (LOS E ) applied 10 the peak season may be Jill& restrictive than the higher standard (LOS D ) applied 10 the entire year (as in Figure 21 ) HOURLY TRAme VOU.IME Figure 21 Jzvel of Service Related t o Choice of Peak Hour LOSE Muimum Volume Avenge Peak Hour ofPeakSeuon LOS D Maximum Volume Annual Average Peak Hour mGHESr H OURS OF THE Y.EAJl 49


so Does this mean that there is no prefened peak hour for concunency management purposes? Not hardly, but it does mean that the cbojce of oeak hour must be made on some baSis other tl]an !be desjre to promo!< gr resujct development (which can be accomplished wilh any peak hour by simply lowering or raising level-of-service standards). Beleyance of the Design Hour FDOT presumes that !be appropriate peak hour for cancvrrency maoawnem purposes is the same as the peak hour for desien purposes." When DCA has allowed local jurisdictions to base levels of service on some peak hour other lhan the 30th highest. it has been out of expediency. Traffic conditions were so dire that the use of the 30th highest hour would have made it impossible to meet FDOT level-of-service Standards. despite a "good faith" local effon. But should the same peak hour be used for cancurrency management and design purposes? The tranSportation professionals interviewed for this study had mixed reactions. Some aJgUed that we cannot afford to hold existing facilities to lhe same standard as new facilities. In their view. we design new facilities to desired standards but operate existing facilities at minimum standards because the resources required to upgrade them are limited. Others felt that the same standards must apply to new and existing roads for growth management policies to be internally consistent. In their view, roads identified as deficient by one standard in the Traffic Circulation Element of an LCP should be upgraded to the same sWldard in the Capital Improvement Elemem. The latter view appears more defensible than the former. Once a community bas decided how much congestion it is willing to aceept, and has set a level-of-service standard for a pallicular peak hour to implement its decision, that standard would logically apply 10 .balll new and existing roads. If it is coosidered acceplable for new roads to operate below adopted standards for, say, 29 hours in the design year, it should be acceptal!le for existing roads to operate below adopted sWldards for 29 hours in the currem year, but no more than 29 hours. If jurisdictions cannot afford to maintain eldsting roads at the standard for new roads. the solution is not to establish a lower standard of existing roads but rather 10 re-evaluate the standard for new roads. Standards for new roads are based on theoretical considerations and professional dictates, but funding ayaj!abilily should be a factor. too. Indeed. it might be aJgUed that the appropriate standard for new roads is that which is sustainable for existing roads. given funding availability.


New and existing roads differ in one fundamental respect-traffic volumes are known for existing roads and only estimated for new roads. 11lis difference has been used to argue for higher standards in roadway design chan concurrency managemeru.. where roadways are designed for a "maJgin of error" in traffic forecasts. If the result of underestimating future traffic volumes (congestion) is more serious !han lhe result of overestimating lhem (wasted capacity), it may be advantageous to make liberal assumptions in traffic forecastS. However, !here is no reason for lhe choice ofpealr. hour 10 reflect uncertainty in traffic forecasts, not when lhe forecasts lhemselves can be adjusted to reflect uncenaimy. The 30Ih Hiehest Hour Use of lhe 30Ih highest hourly volume in roadway design dales back 10 lbe 1950 Hieh,way Capacity Manual. Traffic Studies of that era had observed extreme variations in traffic now on facilities from hour 10 hour day 10 day, and season to season. Wben hourly traffic volumes for an entire year were graphed in order of descending magnitude, lhe resulting curves often dropped sharply at first and leveled off quickly (as in Figure 22). The 301h highest hourly volume was found 10 fall on lhe "knee" of lhe curves, where lbe slope changed dr.!matically." Based on chis early work, it has become conventional wisdom that: The 301h highest hourly volume is the point of FlpreZZ Peakbour Volume Characteristics of Typical Highways .. ., [": !11-. -....;; """' "'I\ "" !"" I I .. ., ' Matson, Smilh, and Hur d. Ial1k Enajnccring, (19SS), p. 86 51


52 diminishing returns in roadway design. h is uneconomical to design for volumes to the left of the 30th highest hour, since a great deal of capacity is required to meet demands that occur only a few times a yeaL It is shonsighted to design for volumes to the right of the 30th highest hour, since little additional capacity is required to accommodate demands that occur frequently. 11le conventional wisdom may be wrong in this case. Traffic in many localities does not follow the panem in Figure 22. Cutves tend to flatten out rather than remain static as areas become more developed; the knee of the curve becomes a moving tatgeL Even if the hourly volume cutve has a predictable ruming point. that point has no economic significance; the optimum design of a facility can only be determined by comparing the costS of alternative designs with the benefits to motorists." Plots of hourly traffic volumes for 20 representative FDOT permanent count stations in FY 1989 illustrate the arbitrariness of the 30th highest bour (see Appendix). Several of the curves never level off and/or have no point at which the slope changes dramatically Even where there is a discemable "knee, it does not correspond to the 30th highest hour with any degree of consistency. The cboice of design hour is ultimately a political rather than a technical decision. It involves balancine the public's desire to hold down road user taxes (which means more traffic congestion) against their desire to avoid traffic congestion (which means higher user taxes). Con s ensus Choice? If not the 30th highest hour, what peak bour should be used for level-of-service determinations? As previously noted FDOT is proposing a shift to the I OOth highest hour. Meanwhile, DCA is acoepting and in some cases even requiring the use of the average peak hour of days in the peak season. Are the two agencies at odds? In FDOT's unpublished draft revision of the Level of Service Manual it is stated several times that the IOOth highest hour of the year is "approximately equivalent to the typical peak hour of a day during the peak season.'' Studies in Collier County also suggest the two are equivalenL If true. FDOT and DCA are converging on a common peak hour 10 check for equivalency, we again analyzed hourly traffic counts for FY 1989 from the 20 FDOT permanent count stations Figure 23 shows the 30th, IOOth, and 200th highest hourly counts a1 each station, plus the average count for peak hours of days during the peak season (the peak season being the three highest-volume consecutive months). At all but two


Fltl"re 1'na.ffi< Counts a t Different Design Houl"'l PAit HOllRIPEAK St.MON PEAK Jl(,)lJR/PEA.k SFASON/DA Y5 ITAtroN )0111 KIGMF.ST 100111 HICIIEST 200'111 HIGIIE5J' HOVR lfOta HOUR A.VUAG EQUIYAfJ!NT AVEJIAGF. Qt.IIYAIJ!.'U COOI

54 the average count for peak hours during the peak season is le s s than the lOOth highest hourly count At 10 of 20 stations. it is even less than the 200th highest hour l y count. The 10 stations are mosUy in urban areas with stable commuting patterns and little seasonal peaking At these Stations, a single day (typically a Friday) may contribute several of the top 200 hourly traffic counts Clearly, the IOOth highest hour is not equivalent to the average peak hour during the peak season. However. it may be roughly equivalent to the average peak hour of weekdays during the peak season in urban areas and weel\ends in recreational areas and other places with weekend peaking. Figure 23 shows average peak hour counts on weekdays or weekends (depending on when local peaking occurs) during the peak season at the 20 stations The average count i s le s s than the lOOth highest hourly count at some stations, more at others. and probably about the same on average. If it is necessal)' to settle on one peak hour for roads statewide as FOOT has declared, the IOOth highest hour may be the best choice. The IOOth highest hour reflects daily, weekly, and seasonal peaking of traffic; this is apparent from the rough equivalence of the I OOth highest hour and the average neak hour of peak days during the peak season The IOOth highest hOurly volume would be easy to estimate, assuming thi s rough equivalence is borne out. A locality could take a 24-hour count on a typical weekday or weekend (again, depending on when local peaking occurs) during the peak season and use the highest hourly count for that 24-hour period as its estimate of the IOOth highest hourly volume. This approach would improve on the practice in many localities of applying a generalized K faaor 10 a single 24-hour traffic COUDL Additionally, the lOOth highest hourly volume would be relatively easy 10 project. Widely used regional travel models forecast uaffic volumes for the average weekday during the peak season. Th obtain estimates of the IOOth highest hourly volume. it would only be necessal)' to apply a peak-to-daily ratio to model outputs. At present, modelers must first conven model outputs 10 annual average daily traffic volumes, and then apply a generalized K-factor t o the resUIL The Peak Period Daily peaks tend 10 spread out as urban areas grow and traffic congestion causes motoriSts to adjust their travel hours Indeed, the largest cities do not have a "peak hOUr" per se but rather a two-to tluu-bour period in the morning and afternoon when commuting is heavieSL Roads become capacity-constrained and K-factOrs come to be detennined by supply rather than demand." We can expect even more spreading of the peak as traffic congestion worsens and communities seek 10 better manage t ransponation demands.


With DCA's approval. Dade County and !he City of Miami based levels of service in their LCPs on average traffic volumes for !he two hiehest consecutive bours of !he average weekday. While analysis of a two-hour peak period flies in !he face of time honored design convention (which utilizes a "design hour"), the convention may prove too limiting. It will not be !he spreading of the peak that, in time. justifies a shift from peak bour period analysis. This spreading is already reflected in traffic counts and should be in !he K factors used in traffic projections."' Rather. it will be !he adoption of policies that encourage commuters to adjust !heir travel hours. Let us say a locality adoptS a trip reduction ordinance requiring all employers to instirute flex-time. Employees would !hen have !he option of commuting at Jess congested hours. Such a measure might justify !he averaging of traffic volumes over flexible starting and ending hours As when routing and modal choices are made available, a case could be made that with flex-time in place, government no longer has responsibility for accommodating every trip maker's choice of travel hour. 55


56 VI. CONCLUSION One observer has likened concurrency to "the engine of a car, the vehicle of growth management by comprehensive planning ... ,. Canying the analogy one or two stepS further, he stateS: "The level of service (LOS) determination is analogous to the carburetor which feeds fuel to the engine. As with any carburetor, it must be capable of working well enough to start the engine and operate it at an efficient level. ... the LOS of the carburetor should not operate to choke the motor by setting LOS unrealistically high." 'This observer and many others seem to assume that level-of-service stapdards determine whether the vehicle of growth management will operate efficieruly. In however, the methods used to determine roadway levels of service affect conclusions about road adequacy as much as do the standands themselves Carrying the vehicle analogy to its logical extreme. installing a new carburetor may affect the performance of the engine as much as adjusting the fuel-air mixture on the old carburetor. The "new carburetor" in this analogy is a new method oflevel-of-service determination. Our analyses suggest that the specific technique used 10 analyze roadway level-of-service whether HCS. ART-PLAN, ART-TAB. or travel time studies --can make at least a two-letter grade difference in the outcome illewise, the choice of analysis period or peak hour whether 30th highest hour, I DOth highest hour, average peak hour of the peak season, or highest two consecutive hours on the average weekdaycan make a difference of two or more letter-grades. While hanler 10 quantify, the effect of averaging/aggregating levels of service across facilities or modes could be of comparable magnitude. Thus, even adopting the same level-of-service standards. level-of-service determinations for, say, Miami and Jefferson County have entirely different implications for motorists. Jefferson County has opted for FOOT's "by the book" approach. comparing the 30th highest howiy traftic volumes on individual roads to the maximum volumes at different levels of service from FOOT's generalized maximum volume tables. In contrast. Miami has adopted an innovative but uocooventional approach with DCA's approval. comparing person-lrip volumes for the two highest hours on tbe average weekday to the practical capacities of multimodal uaosporwioo corridors. To some extent. the different approaclles reflect the different circumstances of the jurisdictions and therefore should be viewed in a positive light Yet. there is a nagging concern that Florida's Growth Management Act may be un

and comparability across the state. It is hoped that this repon will provide some guidance to state and local officials searching for meaningful approaches to roadway level-of-service determination. 1 John Koenig. "Down to the Wire in Florida." Planning (October 1990), p. 6. 2 Richard G. Dowling, "Controlling Growth with Level-of-Service Policies." 'Iransponatioo Resealkb Record 1237 (1990), pp. 39-46. FDOT. Florida's Level of Service Stan

58 1be source of lhis infonnation is Mark Woerner of the Dade County Planning Depanment. John J. Moore and Antoine I. Khoury, "Level of Service Analysis: Urban and Suburban Anerials," Communication to the 11E Growth Management Subcommittee, Florida ITE. Data were supplied by Glatting Lopez Kercher Anglin, Inc .. an Orlando-based planning finn. 17 For each anerial, turning movement counts and travel time runs were done during the same peak hours on the same weekdays. Thus. by design. actual uavel speeds (derived from uavel time runs) and estimated travel speeds (dependent on turning movement counts) rela!e 10 the same time periods. In a series of HCS intersection analyses, liberal assumptions were made about: the saruration flow nues at intersections on these arterials (1.850 vehicles per hour after adjusunents) the amount of green time devoted to arterial-through movements (the maximum possible, given the timing plans of these semi-actuated traffic signals). arrival typeS of vehicle p latoons (the best pOSsible progression, given signal spacing and signal timing offsets), and the peak-bour factor (a value of 1.0 was assumed, as if flow rates were abso lutely constant during the peak hour). A peak bour factor of I .0 was assumed to achieve a measure of consistency between HCS estimaleS and travel time runs. If actual peak-hour factorS had been used instead, HCS estimaleS would have applied 10 the peak IS-minute period of the peak hour. while the travel time results were averaged over the entire peak hour. 19 While not yet accepted by FOOT, ART-PLAN and ART-TAB represent the latest generation of such programs and thus are more fittingly evaluated than are their predecessors. ART-AU. and LOS. "' Data were supplied by John Zeeger of Banon-Asclunan Associates, Inc., Fort Laudenlale. 21 Average travel speeds and peak hour volumes for Broward Boulevard were gathered on comparable weekdays of the same month. Portions of the cycle devoted 10 the through movements on Broward Boulevard (so-called green ratios) were observed at


the same time uaffic counts were taken 1lley were subsequently confinned from signal timing plans. Thus, while green ratios for Broward Boul evard appear very low, there is no reason t o doubt the validity of the values supplied by the consulting linn. "' The section of Browaro Boulevard analyzed is between N. W 18th Avenue and State Road 7. HCS runs could not be done for Broward Boulevard because informa ti on about cross sueet uaffic and signal timing is not available. ,., From travel time runs. the stopped delay at intersections and the total running time between intersections are known. From HCS runs, estimates of stopped de lay. total intersection approach delay, and running time between intersections are available. StOpped delay from travel time runs can be compared directly to estimates from HCS runs. However, to compare tOtal nmning time and hence running speed between intersections. an HCS estimate of running time must be inflated by the difference between total approach delay and stopped delay at the downstream intersection. 'This adjusanent capture s the extra nmning time associated with deceleration of vehicles approaching the downstream intersection. "' 'This is not to sanction the practice of speeding but to acknowledge that it occurs. The problem lies with the level of service measure chosen for arterials in the 1985 HCM. Many alternative measures were considered during the update of the HCM. Pemaps a measure such as average delay would have been more suitable than average uavel speed. 15 Delay associated with acceleration is incoiporated intO nmning lime estimates for segments ofless than one mile (see Table 11-4 in the 1985 HCM). Delay associated with deceleration is incorporated intO intersection approach delay estimates. 24 From Table 11-4 of the I 985 HCM, the running time for a one-mile segment at a 45 mph free flow speed is 80 seconds. Properly exuapolated, the running time for a 1.08 mile segment (the segment in question) shOuld be 86 4 seconds. HCS instead a running time of 110. 2 seconds in the northbound direction, 106.9 sec onds in the southbound direction. Extrapolation of nmning times seems to be a problem for HCS only at free flow speeds of 45 mph. Recall that our estim a tes of average running speed incoqx>rate intersection approach delays. which are set equal to 30% of StOpped delays in HCS and ARTPLAN runs 21 Not only were all of these parameters averaged in the ART TAB runs, but free flow speeds were effectively averaged for Kirkman Road by interpolating between maximum volume tables representing different free flow speeds. Kirlcman Road has one segment with 45 and 50 mph speed limits, while the other three segments have speed limits of 40 mph. ART TAB will only recognize free flow speeds of 45, 40, or 35 mph for Oass I arterials, and thus only by interpolation of maximum volumes was it possible to duplicate an average free flow speed or 43 5 mph. 59


60 ,. A planner with the City of M;ami wants to outfit the city's fleet of vehicles with tranSminers thai supply continuously updated travel speed data "' John D Zegeer, "Field Validation of Intersection Capacity Factors TranSpOrtation Research Record 1091 (1987), pp 67-77. The speed-volume relationship is known to become nonlinear as road capacity is ap proached. Apparently, Seminole County roads operate in a flow range that is ade quately approximated by a linear equation. FDOT, florida's Level of Service Standards and Guidelines Manual for Planning p 3.3 -5. Workshop draft of Rule 91-S. F. A. C Minimum Criteria for the Review of Local Government Comprehensive Plans and Determination of Compliance (draft dated 1 / 111')1), FDOT defines backlogged roads as roads operating below leve l -of service Standards and not programmed for improvement in the first three years ofFDOT's adopted work program or in the capital improvement element of any local govemmenL We would first divide vehicle-miles trave l ed (VMT) by average travel speed (ATS) for each road in an area to obtain vehicle-hours traveled (VH1) on each road: then, sum the result over all roads to arrive at total VHT, finally, divide total VHTinto total VMT to arrive at average travel speed for all vehicles on all roads Using mathematical symbols, the calculation is: VHT = VMT./ATS foreachroadi ' VHT I = 1: VHT i. where VHT is summed over all roads j ATS 1 = VMT, NHT1 DCA calls them "TCMAs" (transportation concurrency mattagemeru areas) They are called "traffic performance districtS" in Orlando, "traffic districts" in Lee County, and just "districts" in Tampa.


" Of the criteria used to delineate traffic perfonnance districts in the Orlando LCP. at least three appear aimed at keeping trips internal to districts : Major activity cen1ers are contained within single districts. District boundaries generally follow geographic features, limited access facilities, or lightly uaveled streets but generally dO not follow arlerial or collecu>r roads. Districts generally are aligned along major commuting or traffic circulation panems. The goals specified in the draft rule are: discouraging the proliferation of urban sprawl. encouraging the revitalization of existillg downtowns and designated redevel opment areas. proJecting natural resources. and promoting mass transit. Exampl es of suitable areas cited in the draft rule ate: oommunity redevelopment ateas. areas covered by approved downtown developments of regional impact, regional activity centers designared in a comprehensive regional policy plan. areas designated in a comprehensive regional policy plan as appropriate for increased development of regional impact lhresllolds. and central business districts ,. 1\vo policies of the Stale Comprehensive Plan particularly speak to the issue of system efficiency -- Coordinate uansponation invesunents in major rravel corridors to enhance system efficiency and minimize adverse environmental impacts. Ensure that the uansponation system provides Florida's citizens and visitors wilh timely and efticiem access to services, jobs, mmets, and auractions .. The draft rule states that a TCMA shall "not be latger than the area within which a development .. that pays a development exaction could be reasonably assumed to receive a reasonable benefit fran an improvement funded by the exaction if the improvement were consuueled anywhere wilhin the area.' For a discussion of Florida coun cases. see Florida Advisory Council on lnlergovenunental Relations, Impact fu;s jn Florida (1966). pp. 25 28 and impact fee Use in Florida : An Updaie (1989). pp S2-SS. 41 Richard A. J{all, "Concurrency Managemem for Transponation." Florida Society Jouma) (Sepcember 1990), p. 20. Timothy T. Jackson, "Transponation Concurrency: How Can It Be Achieved?" EJ,orida Eoejneerin& Society Journal (September 1990), p. 24. .. Oty or Miam i Planning Depanment. Ixansoortation Corridors: Meeti!l!! the Clla!1en&e of Growth Manaeemem in Miami (1989). p. 5. William R. McShane and Roger P. Roess, traffic Eniineerine (1990 ). p. 63. 61


62 "' Maximum IS-minute peak directional volumes are divided by the peak-hour factor (PHF) to obtain peak hour directional volumes; peak hour directional volumes are divided by the directional factor (0) to obtain total peak-hour volumes; and finally, total peak-hour volumes are divided by the design hour factor (K30) to obtain daily volumes. Thus, all tables are based ultimately on peak-hour conditions and are inlemally consisteru. "' FOOT, Florida s Level of Service Standards and Gyjdeljnes Manual for Plaonin2 p. 3 5 4. 41 The fact that everyone is required to use K30 (or KlOO) in lev el-o f-service detenninations does not mean that one K-factor fits all K30 (or KIOO) varies from locale to locale, and even from roadway to roadway within a locale. Generally, the K-factor declines as a place becomes more urbanized; traffic congestion causes some spreading of tile peak period. The revised manual recommends the following K tOOs for anerials in differan areas: Ala Rural Undeveloped Developed Urban population 500,000 or more 0.100 0.095 0 093 0.092 0.090 FOOT, Gujde!jnes for !he Review ofDeveiO!)!Dems of Reeional Impact CDR!s), p. 18. JO Ibid. FOOT's Level of Service Manual (current version) and FDOT's Roadway Plans J>reoaratjon Manual. '

Appendix 63


-.., -c Q) 0 .;...) "' 0 l-. ci "' "' :::> d .D ,; -l-. --::::> > c 0 .., (.> ..., I I .., "' ... -64 J I I (") "' -0 ----.LOVV JO 'l U90.1&d Q) 0 0 "' 0 .... -0 "' -0 0 -0 Q) 0 10 0 ... 0 "' ... "' >-" "' '-0 ., ... 0 :r: -., "' .<: .. 0::


"' Q) ,.. ....,> c: 0 ;j -... .. 0 ... en $..., oi a: (lj en .!::l $..., E :::> .. :::0 "' "' -' M N --' ' 0 "' -0 0

"' M c ... 0 66 "' -' ., -' ... -' M "' --.... 0 .x.avv JO 'l uao.1ed "' 0 0 N 0 "' 0 ... 0 N 0 0 0 "' 0 "' 0 ... 0 N ... "' :.-" .c -... c ., ... 0 :r - ., .c .. :r


.., "' c 0 Q) -" .;-) -C/) ;j 0 "' "' :..... "' "' C/) :::> .. .. .. -"' "' ..., C/) .., I "' T ... I (') N I I 0 JO "' 0 0 N 0 "' 0 "' -0 ... 0 N 0 0 0 "' 0 "' 0 ... 0 N ... .. >-" .c -'-0 ... 0 "' -., .. .c .. "' 6 7


-.,., c a> 0 ..,_) ;j "' 0 $.., N c U> ::> ctl .t:J .; $.., ., "' ;:::> 0 .... -0 "' -0 0 -0 0 "' 0 .... 0 N ... .. "' >-"' .<:: -..... 0 .. ... 0 :c -"' "' .c .. :c


Q) 0 "' ..,_) -" 0 0 -... $..,. .. ... rn :>., ..,_) "" ..... N u (tj 0 ... 8 .. "' {/) -u "' -"' ... -M -' N -' 0 -' "' co 0 0 N 0 ... -0 N -0 0 -0 "' 0 "' 0 ... 0 N .... .. .. ... ., .s= ... .... 0 "' .. 0 :I: --.. .. .s= .. :I: 69


v ;:1 0 ..... {.) -ro ..... lJ) 70 r-<:: 0 -.. -en 0 0> UJ ::> .. <:: <:: .. -... .. :::;: ID -' "' -' ... -' M -' "' -' 0 -' 0> 0 0 "' 0 0 ID 0 ... 0 "' 0 0 0 "' 0 ID 0 ... 0 "' ... .. ,.. .c: -..... 0 "' ... 0 :>:: - .c: .. :>::


S1nall City route SR 520. Station 113 16 15 14 -113 Q < < .. 12 0 ., 1: 11.. .. .. Q, 10 k 9 8 -7 lnu"""'i''''''*''''''''"''i''''''"''''''"'''l""""'""""''i''''''".,"''''"'l'''''''"''''''"''l""'''''"''''" l'''"'''""''"''l"''''"""'''"''i'""''"'"''''"'1 20 40 60 80 100 120 140 160 180 200 Highest Hours of the Yar -.) -


;:j S m all City r oute Palatka. US 1 7 S tation 105 1 6 ,-----------------------------------------------------------. 15 -1 4 !-Q IJ < < .. 12 0 I\ c: 11 " ... .. Q, 1 0 -9 -8 -7 lootoo i ohiilhiliiijhllhlliiiflllllh(lliiiiliiliihlliiijlliiiiilhiiiifli;;f'o + ohilliiiO hoo ooloooooou o oont. h .il'iiihllllhiiiiii'jil i ihlilliliOOiiiiJ'"""''""' ''''I"'''''''"' '"''''I 2 0 40 60 80 100 120 140 160 180 200 Hours of the Year


) ' I I I "' "' ... ("' "' 0 -----.LOVV .JO '\ U&O::..JOd I I a> CD 0 0 N 0 CD 0 "' 0 N 73


... "' c 0 -ilJ .. .,...> {/) ;::1 0 -... h {/) :::> -(lj >, h ;::1 c 0 0::: u .. .. 0 u "' "' .... M ---74 _,.J I I I (\1 0 - "' 0 "' 0 .., 0 ... 0 (\1 0 0 0 co 0 .., 0 ... 0 (\1 .. .. .... .c '-0 ., .. 0 :c ., .c .. :c


>. c ::1 0 u 0 -0 .. 0 "' -I I ... (") I "' --I "' I "" 0 0 N 0 "' -0 "' -0 ... -0 "' -0 0 -0 "" 0 <0 0 ... ., :.-'" .c -0 .. ... 0 "' -.. ., .c .. -::t: 75


"' "' c 0 -<1> -U> ;:s .,; 0 "' .... "' ::> Ct$ >. .... -c ;:s 0 u .. -.Q E .. " "' I I I I "' "' ... "' "' -----76 / I I I 0 "' I "' 0 0 "' 0 "' 0 "' 0 ... 0 "' 0 0 0 "' 0 <0 0 ... 0 "' .. " >-., .c .... 0 ., ... 0 0:: -"' ., .c 0::


(1) .._) ;:::s 0 1-. {lj 1-. ;:::s "' ... 0 -.. -"' ...: "' "' "' -" 0 u .!: .. ... .. .... .. _, I "' I ... I .., I "' I -I 0 .l.QVV JO 1UOO.J:d I "' I Q) 0 0 N 0 Q) 0 "' 0 ... 0 N 0 0 0 Q) 0 "' 0 ... 0 "' L .. .. >.. .<: -.... 0 V> .... 0 "' -V> .. .c .. -"' 77


Q) -+-> 0 ;.... ......... ;.... 78 <0 -a 0 -" "' 0 M "' ::> t0 u c: 0 ... .. :::;: <0 I "' I ... I M I N I I 0 JO I "' I to 0 0

<-... c 0 . c ::l 0 0:: u .. , z <0 I ., I ... I M -I N -I -I 0 -I en I "' 0 0 "' 0 "' 0 10 0 ... 0 "' 0 0 0 "' 0 <0 0 ... ... .. "" .c ... 0 ., .. 0 :c -"' .c .. :c 79


M c 0 (l) ..-"' .._) ..-(/) ;:::5 0 -;... .... (/) ...... :::> ;... >. ..-;::l c 0 u 0 "' .. "" I

Recreational route Key Largo. SR 5. Station 164 16 !5 14 ... c 13 <( <( ... 12 0 II u 0. 10 -9 -8 -7 1illliiiilillifiiliijllhhiiiiiiliih$ijiilillillilihiiiiljlliiiilhifihilh ijiiiiiililiiiiiiiiiijliiiliiiiliililiilijiiilhhiiiiiih.iijl11iiilililhhhlijiiiihhl i i i lilhlljiiilfiiiillhiii;al 20 40 60 80 100 120 140 160 180 200 Highest Hours of the Year 00 -


10 -82 I "' I ... I .., I "' I I 0 I "' I "' 0 0 (\1 0 "' 0 10 0 "'


Recreational route Palm Beach. SR A-1-A. Station 87 16 15 -1 14 -1 s 13 -1 ... 12 -l 0 ... 11 0 .. II, 10 -9 -8 -7 l*''"''''"'"""'l"'" "'''''''"" l'""""'"''''''''l"''''' ''''""' '''l'''"'''''' '''"'l"' '''''''""''l''''''"'''''''' "; l ''''; iiilii .Ot. ttll jiii iiiil.oli+olhiljllliiiiiiiimonool 20 40 60 80 100 120 140 160 180 200 Highest Hours of the Year el

xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam 2200349Ia 4500
controlfield tag 001 029120306
005 20110916113840.0
006 m o d
007 cr bn|||||||||
008 110916s1991 flua ob 000 0 eng d
datafield ind1 8 ind2 024
subfield code a C01-00010
0 245
Roadway level-of-service determination
h [electronic resource] /
prepared for the Florida Institute of Government by the Center for Urban Transportation Research.
[Tampa, Fla.] :
b University of South Florida, College of Engineering, Center for Urban Transportation Research,
1 online resource (vi, 82 p.) :
"May 1991."
Includes bibliographical references (p. 57-62).
Description based on print version record.
Highway capacity
z Florida.
Traffic flow
x Mathematical models.
2 710
Florida Institute of Government.
University of South Florida.
Center for Urban Transportation Research.
1 776
i Print version:
t Roadway level-of-service determination.
d [Tampa, Fla.] : University of South Florida, College of Engineering, Center for Urban Transportation Research, [1991]
w (OCoLC)31343064
Center for Urban Transportation Research Publications [USF].
4 856