Current Practices in Evaluating Freeway HOV Facilities

TRANSPORTATION RESEARCH RECORD 1299 63 Current Practices in Evaluating Freeway HOV Facilities KATHERINE F. TURNBULL, ROBERT W. STOKES, AND RUSSELL H. HENK Evaluating the impact f high-ccupancy vehicle (HOV) facilities has been a tpic f cnsiderable interest and discussin amng transprtatin prfessinals in recent years. Prject bjectives, ptential evaluatin criteria, apprpriate effectiveness measures, and data cllectin methdlgies have been a majr fcus f sessins at natinal HOV cnferences, as well as f numerus reprts. A review f the majr befre-and-after evaluatin studies that have been cnducted n HOV facilities in the United States since the pening f the Shirley Highway (Nrthern Virginia) bus lane in 1969 is presented in this paper. The review includes a summary f the appraches used with the different evaluatins and the identificatin f cmmn elements. The results f this analysis advances understanding f the majr cmpnents that shuld be cnsidered in the design f HOV evaluatin studies. The paper shuld prve beneficial t agencies in the prcess f designing befre-and-after evaluatin prgrams fr new HOV facilities. In additin, it shuld be f help t transprtatin prfessinals in areas that have perating HOV facilities and may be interested in imprving their current evaluatin methds. High-ccupancy vehicle (HOV) facilities are being used in many metrplitan areas t address grwing traffic cngestin prblems. Many areas are faced with significant increases in traffic cngestin and prjected travel demands beynd what can reasnably be served at current vehicle ccupancy rates. Attempting t address these cngestin and mbility prblems in a time f limited financial resurces and right-f-way availability has led many transprtatin prfessinals t cnsider a range f ptential slutins. The use f HOV facilities, which fcuses n increasing the persn-mvement capacity f the radway facility, represents a viable alternative in many areas f the cuntry. As the use f HOV facilities increases, interest in the prper appraches and techniques t evaluate their effectiveness is cntinuing. Discussins f prject bjectives, ptential evaluatin criteria, apprpriate effectiveness measures, and data cllectin methdlgies have been a majr fcus f recent cnferences, reprts, and articles. T advance the discussin f these tpics and assist in the establishment f a mre unifrm apprach, this paper presents a review f the majr befre-and-after evaluatin studies that have been cnducted n HOV facilities during the past 20 years. The review includes a brief discussin f the develpment f HOV facilities and evaluatin prgrams, a summary f the appraches used in the different evaluatins, and the identificatin f cmmn elements amng the studies. The results f this analysis shuld be f benefit in the design f befre-and-after studies f new HOV facilities and t trans- Texas Transprtatin Institute, The Texas A & M University System, Cllege Statin, Tex. 77843. prtatin prfessinals in areas with perating HOV facilities wh may be interested in imprving current evaluatin prcedures. Thus, this paper advances the state-f-the-art understanding f the majr elements that shuld be addressed in evaluating HOV facilities. DEVELOPMENT OF HOV FACILITIES AND EVALUATION PROGRAMS The first HOV lanes in the United States were implemented in the late 1960s and early 1970s. The initial sectin f the Shirley Highway exclusive bus lanes pened in 1969 in Nrthern Virginia. This was fllwed in 1971 by the exclusive bus lane apprach t the Lincln Tunnel n I-495 in New Jersey and the first phase f the Ls Angeles-San Bernardin Freeway Busway in 1973. Since the late 1970s, the use f HOV lanes has expanded greatly. As f April 1990, there were apprximately 40 HOV facilities in peratin n either freeways r in separate rights-f-way in 20 metrplitan areas in Nrth America (1). Althugh these facilities ften differ in design and peratin, all have similar purpses. In general, HOV facilities are intended t help maximize the persn-carrying capacity f the radway. This is dne by altering the facility's design, peratin, r bth and prvides pririty treatment fr HOVs, which are defined as buses, vanpls, and carpls. A primary cncept behind these pririty facilities is t prvide HOVs with bth travel time savings and mre reliable travel times. These tw benefits serve as incentives fr individuals t chse a higher ccupancy mde. This, in turn, can increase the persn-mvement efficiency f the radway by carrying mre peple in fewer vehicles. In sme areas, additinal incentives such as reduced parking charges r preferential parking fr carpls and vanpls have been used t further encurage individuals t change their cmmuting habits. Thus, the intent is t prvide a cst-effective travel alternative that a significant number f cmmuters will find attractive. The reasns fr implementing HOV facilities and the appraches used vary greatly. Many HOV applicatins have been attempts t develp relatively lw-cst methds fr increasing the capacity f cngested freeways. This is especially true with sme f the early facilities, many f which were implemented as part f transprtatin system management (TSM) prgrams. Appraches used in these cases include restriping t add an HOV lane t an existing facility, using the shulder fr an HOV lane, and ther relatively lw-cst treatments. Other HOV facilities represent permanent lng-

64 term imprvements. These include the cnstructin f highcst prjects, such as bus-nly facilities n separate rights-fway and exclusive HOV facilities n freeways. Evaluatin methds and appraches have als varied greatly amng prjects. Because f limited resurces, adequate befre-and-after data have nt always been cllected. In sme cases, an initial evaluatin may have been cmpleted shrtly after implementatin, with little r n nging data cllectin r evaluatin activities. Thus, evaluating the effectiveness f many facilities has been difficult. In additin, the purpse r gal f the HOV facility has nt always been well defined. Furthermre, the evaluatin measures, alng with the data cllectin techniques and evaluatin criteria, have differed amng areas. Thus, n standard apprach r set f evaluatin measures, criteria, and data cllectin techniques exist fr evaluating HOV prjects. It shuld be nted that the lack f cmmn evaluatin measures and extensive befre-and-after studies is nt unique t HOV prjects. Many transprtatin imprvements, bth radway and transit, have been implemented withut a clear statement f the gals fr the prject and a definitin f the evaluatin measures that will be used t determine their effectiveness. Gd befre-and-after studies are nt all that cmmn with highway r transit prjects. Hwever, these shrtcmings d nt dismiss the need fr accurate evaluatins f the effectiveness f HOV facilities. Because they require spending public funds, HOV lanes need t be evaluated t ensure that these investments are prviding the desired benefits. T identify past and current practices used t evaluate HOV facilities, an extensive literature review was cnducted. Several reprts and articles were examined frm many surces. These included dcuments frm federal agencies, such as the Federal Highway Administratin and the Urban Mass Transprtatin Administratin, state departments f transprtatin, lcal transit agencies, metrplitan planning rganizatins, cnsulting firms, university transprtatin research rganizatins, and ther grups. These reprts cver a wide range f tpics including general guidelines fr HOV prject planning and develpment, prject specific case studies, and verviews f HOV prjects arund the cuntry. Reprts dealing with befre-and-after prject evaluatins were examined in detail. In additin, telephne cnversatins were held with representatives frm agencies respnsible fr different aspects f the evaluatins t btain further infrmatin. The results f this review are summarized in the next tw sectins. An verview f selected evaluatin studies is then prvided. This is fllwed by a summary f the cmmn elements amng the different studies and general cnclusins resulting frm the review (2). REVIEW OF SELECTED HOV EVALUATION STUDIES This sectin examines sme extensive befre-and-after evaluatin studies that have been cnducted n freeway HOV facilities and the appraches that have been used in the different metrplitan areas where the facilities are lcated. Evaluatin studies frm the fllwing areas are included: Washingtn, V.C.-Nrthern Virginia, Ls Angeles, Hustn,,, TRANSPORTATION RESEARCH RECORD 1299 Seattle, and Minneaplis and Orange and Santa Clara cunties in Califrnia. Althugh this review des nt include all the evaluatin studies cnducted f HOV facilities, it des prvide a sample f the types f studies, level f detail, and appraches that have been used with different freeway HOV facilities. The analysis is n available evaluatin reprts and additinal telephne cnversatins with representatives frm agencies in the different areas. A brief descriptin is prvided f the prject, evaluatin prcess, prject bjectives, evaluatin measures and criteria, data cllectin techniques, and any unique features f the study. Shirley Highway HOV Lanes, Nrthern Virginia As nted previusly, the Shirley Highway HOV lanes were the first majr HOV facility in Nrth America. A 5-mi busnly lane pened in 1969 and additinal segments f the facility were pened in 1970 and 1971, with the 11-mi, 2-lane, reversible, barrier-separated HOV facility cmpleted in 1975. Buses were the nly vehicles allwed t use the facility during the initial stages; vanpls and carpls were permitted later. Several evaluatin studies have been cnducted n the Shirley Highway HOV lanes during the past 20 years. The first f these was cnducted n the Express-Bus-On-Freeway Demnstratin Prject frm 1971t1975. This demnstratin prject, which was the largest bus and highway prject ever spnsred by the U.S. Department f Transprtatin, had three majr cmpnents. These were the 11 mi f HOV lanes, the use f new feature buses in express service, and the use f new park-and-ride lts crdinated with the express bus service. The primary and secndary gals f the prject, related bjectives, evaluatin measures, and data fr the prject are shwn in Figure 1 (3). Specific threshlds r perfrmance standards were nt set fr each f the bjectives r measures upn which the prject's success wuld be evaluated. Rather, such general terms as "imprve," "increase," and "reduce" were used t describe the desired results. Extensive data cllectin and analysis were cnducted t evaluate the demnstratin prgram. These included traffic cunts, surveys, and the use f analytical prcedures t estimate sme impacts. The majr data cllectin and analysis effrts and their use are described in the fllwing. Vehicular vlumes and persn-trip cunts. Peridic cunts were made f peak-perid vehicular vlumes and persn-trips (bus and autmbile) crssing an 8-statin screenline that intercepted the main radial arterials in the crridr. This infrmatin helped determine verall changes in travel within the crridr, and specifically n the Shirley Highway facility. It was used t analyze changes in the persn-mvement efficiency f the crridr, as measured by changes in ttal persn-trips, bus persn-trips, bus market share, and autmbile ccupancy rates. Mnitring bus schedule adherence. Surveys f bus schedule adherence, as measured by cmparing the actual arrival time f buses at the first dwntwn stp with the time listed in the printed schedule, were cnducted befre the pening f the entire busway and at seven times thrughut the demnstratin.

Turnbull et al. 65 Primary Gal: Demnstrate that express bus-n-freeway peratins can imprve the quality f bus service and lead t an increase in the peple mving capability f peak perid transprtatin facilities fr the entire urban crridr. Objectives: Increase reliability f bus service Reduce travel time fr transit and aut cmmuters Increase cverage by bus rutes Increase bus passenger cnvenience and cmfrt Increase bus's share f crridr cmmuters Measures: Operating speed Dr-t-dr travel times Reliability f service Cverage area f bus service Passenger cmfrt and cnvenience features (seat availability, fewer transfers, etc.) Increase in bus patrnage and market share Increase in carpling and reductin in single ccupant autmbiles Grwth in persn vlumes (bus and aut) per lane n the Shirley Highway and resultant changes in the quality f service encuntered by bth bus and aut cmmuters. Secndary Gal: Demnstrate that the technlgy can have a favrable impact n the transprtatin-related envirnmental and scial cnditins within a crridr and n the ecnmic cnditin f the transit peratr. Objectives: Reduce peak perid aut pllutant emissins Reduce peak perid gasline cnsumptin Increase mbility f the transprtatin disadvantaged Increase prductivity f the bus peratr Measures: T determine changes in the scial and ecnmic bjectives, the fllwing changes were measured: Ecnmic Impact perating csts and capital expenditure, and savings frm increased prductivity fr the bus peratr Envirnmental Impact gasline cnsumptin and autmbile pllutant emissins Scial Impact use f bus service by transit dependent husehlds FIGURE I Shirley Highway Express-Bus-n-Freeway Demnstratin: gals, bjectives, and measures. Mnthly bus data. The bus peratr prvided mnthly infrmatfn n passenger levels, aggregate system csts, revenues, and perating statistics. These were used t evaluate the impact n the transit peratr. Bus and autmbile travel times. These were measured directly by travel time and speed urveys, and indirectly by questins n the mail-ut and passenger surveys. General and specific cmmuter surveys. Several surveys were used t identify changes in cmmuter behavir, the reasns fr the changes, and cmmuter ' general perceptins f the significance f these change. In-depth, mail-back surveys f autmbile and bus cmmuters in the crridr were cnducted in the initial and final stages f the demnstratin. Surveys f park-and-ride lt users and bus users were cnducted. Analytical prcedures. Specific analytical prcedures were develped t estimate bus market share, cmmuter travel time savings resulting frm HOV lanes, bus perating csts, and reductins in autmbile vlumes, gasline cnsumptin, and air pllu tant emissins. As the first evaluatin f an HOV facility, the Shirley Highway prvides a gd example f an in-depth evaluatin. The evaluatin prcess fr the Shirley Highway Express-Bus-On Freeway Demnstratin did nt include any surveys fcused n nnusers f the HOV lane. This was a nted weakness f the evaluatin in the final reprt, and it was suggested that such a survey be part f future evaluatins. In additin, the reprt recmmended the develpm nt f a b tter prcedure t identify and analyze imprvements in vehicle and driver prductivity attributable t the higher speeds n the HOV facility. San Bernardin Freeway Busway, Ls Angeles The 11-mi San Bernardin Freeway Bu way was pened in 1973. The tw-lane, tw-directin faci lity was initially re-

66 stricted t buses. Carpls and vanpls were allwed t use the facility starting in 1976. An initial evaluatin f the busway was cnducted ver the first S years f peratin, frm 1973 t 1978. Accrding t the final reprt (4), the gals and measures f effectiveness identified in Table 1 were used in the cst-effectiveness analysis. The relative imprtance f these different gals was identified by the Bu. way valuatin mmitte and the crrespnding values were used in the analysis. The data cllectin and analysis were cnducted ver the 5 years and included the fllwing majr elements: Travel time studies. Caltrans cnducted befre-and-after tachgraphy (time and speed) runs n the busway, freeway, and n several majr parallel radways. Vehicular vlume and ccupancy cunts. Caltrans als cnducted befre-and-after vehicle and ccupancy cunts. This included electrnic cunts at fur lcatins alng the San Bernardin Freeway and ne lcatin n the Santa Ana Freeway. Mechanical vlumes cunts were taken at several lcatins alng the busway, at ramps alng the Santa Ana Freeway, and n five parallel surface streets. Occupancy cunts were taken at ne lcatin n the freeway and tw n the busway. Vilatin rates. Caltrans mnitred the number f vilatrs (vehicles using the lane nt meeting the ccupancy requirements) during the 5 years and the Califrnia Highway Patrl prvided recrds f enfrcement activities. TRANSPORTATION RESEARCH RECORD 1299 Bus ridership and travel times. Suthern Califrnia Reginal Transit District prvided weekly ridership cunts until 1977, after which they were dne mnthly. Safety. The Califrnia Highway Patrl cllected cntinuus accident data fr the busway. User and nnuser surveys. Several surveys evaluated user and nnuser perceptins f the facility and specific reasns fr mde selectin and mde shift. These included an unbard survey f bus riders, a handut-mail-back survey f carplers, and a fllw-up telephne survey, a handutmail-back survey f freeway mainlane users, and an interview survey f bus drivers. It is interesting t nte that the San Bernardin Freeway Busway evaluatin cntained three elements nt included in the Shirley Highway HOV facility evaluatin. These are the inclusin f safety and accident data, vilatin and enfrcement data, and surveys f nnusers f the busway facility. Hustn Transitways The first HOV facility in Hustn was the 1-45 Nrth Cntraflw Lane (CFL) demnstratin prject, which was implemented in 1979. The successful implementatin f this prject led t develping additinal HOV facilities, called transitways, n ther freeway crridrs. Currently 47 mi f a planned 96-mi system f transitways are in peratin n fur TABLE 1 SAN BERNARDINO FREEWAY BUSWAY EVALUATION PROCESS Gal Effectiveness Measures Relative Imprtance Prvide added crridr Increased carrying capacity f 20% capacity the crridr (persns per peakhur r perid) Reduce envirnmental impacts Reduced emissins f air 10% f crridr travel pllutants (tns per year) energy savings (BTU- 10% equivalent gallns f gasline Imprve the level f service Travel time savings (minutes 20% per persn trip) and the value f such savings, in dllars Reduce the cst f persnal User cst savings, (cents per 20% travel persn trip, including parking cst savings) Imprve the safety f crridr Number f accidents avided 15% travel and the assciated dllar savings t sciety Prvide fr future Adaptability f the busway fr 5% cntingencies (e.g., a rail line, such situatins, plus their future grwth, etc.) likelihd and timing

Turnbull el al. freeways in Hustn. These are all ne-lane, reversible, barrier-separated facilities lcated in the freeway median. The evaluatin prcedures, measures f effectiveness, and data cllectin t evaluate the effectiveness f the Hustn transitways have evlved ver the past 11 years. The prcess used with the initial evaluatin f the 1-45 Nrth CFL demnstratin prject is reviewed first, fllwed by a review f current practices. Listed in Figure 2 are the bjectives and evaluatin measures used in this prject (5). Varius data were cllected and analyzed t supprt the evaluatin, including traffic and CFL vehicle-vlume data. Several surveys f cmmuters in the crridr were cnducted; bus passengers, van drivers, van passengers, peak directin autmbile drivers and passengers, and ff-peak directin drivers were surveyed. The use f surveys t measure the attitudes and reactins f bth users and nnusers was the mst extensive use f surveys t date in evaluating HOV facilities. As ther transitway facilities were planned and implemented, a standardized evaluatin prgram and crrespnding mnitring and data cllectin prgram began t emerge. The Texas State Department f Highways and Public Transprtatin and the Metrplitan Transit Authrity f Harris Cunty have spnsred this effrt, which has been cnducted by the Texas Transprtatin Institute (TTI). The majr elements f this prcess fcus n data cllectin effrts needed t evaluate the fllwing bjectives (6). Effective persn-mvement capacity f the freeway shuld be increased; Transitway implementatin shuld nt unduly impact freeway mainlane peratin; Transitways shuld be cst-effective; Develpment f the transitways shuld have public supprt; and Transitways shuld have favrable impacts n air quality and energy cnsumptin. T evaluate the transitways n the basis f these general bjectives varius infrmatin is regularly cllected by TTL The fllwing list prvides a summary f the majr elements f the mnitring prcess. Vehicle and ccupancy cunts. Vehicle and ccupancy cunts are taken n the transitways and the general-purpse freeway lanes. The same cunts are taken n tw freeways that d nt currently have transitways. These facilities act as a cntrl grup. In additin, vehicle and ccupancy cunts are taken n eight arterial streets that serve as alternative rutes t the freeway facilities. These cunts are taken quarterly. Park-and-ride lt cunts. Vehicle cunts are cnducted at the park-and-ride lts assciated with the transitways and the tw cntrl freeway crridrs. Travel Time Runs. Travel time runs are cnducted quarterly n the fur transitways and the tw cntrl freeways. User and nnuser surveys. Surveys f bus users, carplers, and vanplers using the transitways and singleccupant vehicles in the general-purpse lanes are cnducted annually. These surveys gather infrmatin n user and nnuser perceptins f transitway use, reasns fr mde chice selectin, and general attitudes tward the impact f the transitways. Accident data. Accident data are cllected by the Hustn Plice Department fr the freeway mainlanes and by Metr fr the transitways. Vilatin rates. Metr Transit Plice mnitr the vilatin rates n the transitways. The vehicle and ccupancy cunts als prvide a check n vilatin rates. The Hustn transitway evaluatin prgram is the mst extensive and cmprehensive mnitring and evaluatin prgram currently being cnducted n HOV facilities. As such, it represents ne f the better mdels. 1-5 HOV Lanes, Seattle Six miles f cncurrent flw HOV lanes were pened n I-5 in Seattle in 1983. Three-mnth and 20-mnth evaluatins f 67 Prject Objectives: Decrease (r slw the grwth f) crridr vehicle miles f travel (VMT) and assciated fuel cnsumptin and vehicle emissins Increase vehicle ccupancy in the crridr Reduce cngestin and, thus, decrease travel time Encurage acceptance and usage f public transprtatin Evaluatin Measures: Persn and vehicle utilizatin Characteristics f bth cn.traflw lane users and nn-pririty travellers Impact n nn-pririty users f the freeway Influence in prmting bus and vanpl use relative t ther crridr imprvements Assciated safety and enfrcement issues Public acceptance Impacts n crridr VMT, fuel cnsumptin and vehicle emissins Assciated csts FIGURE 2 1-45 Nrth Freeway Cntraflw Lane Demnstratin: bjectives and evaluatin measures.

68 the facility were cnducted by the Washingtn State Department f Transprtatin (7,8). Six general measures f effectiveness were used t evaluate the impact f the HOV Janes: 1. Number f vehicles traveling in the lanes; 2. Number f peple served by the lanes; 3. Extent t which peple are beying the Jaws gverning the HOV lanes; 4. Time savings fr freeway cmmuters; 5. Effect n accident rates; and 6. Public reactin. N specific threshlds were identified fr these measures. It appears that nly a limited amunt f data was cllected fr the befre-and-after analysis fl-5. Vehicle and ccupancy cunts were taken in the HOV Jane after 2 weeks, 3 mnths, and 20 mnths f peratin. Freeway mainlane vlumes were 111easu1ed by the average daily traffic fr the same time perid. Vilatin studies were cnducted during the secnd and third mnth f peratin, and after the implementatin f the HERO prgram (a telephne htline prgram n which HOV lane ccupancy vilatrs can b reprted), which wa initiated in February 1984. Travel time saving were mea ured a part f the annual metrplitan travel time study. Accident infrmatin was nt available fr the 3-mnth study and is nt mentined in the 20-mnth study. Public reactin was measured by the number f letters and telephne calls received by the department. The I-5 HOV prject and ther HOV facilities in Seattle are mnitred cntinuusly as part f the FLOW System. FLOW is the name given t TSM techniques in the Central Puget Sund area. Evaluatins f the system cmpnents, including ramp metering, HOV lanes, freeway mainlanes, bus service, and accidents, are being cnducted by the Washingtn State Department f Transprtatin n a regular basis (9). TRANSPORTATION RESEARCH RECORD 1299 general t all three evaluatin perids, the evaluatin prgram nted that the degree t which they are achieved will vary fr each perid (10). The bjectives were as fllws: Increase the peak hur carpl-vanpl mdal split fr the 1-394 crridr; Increase the peak hur transit mdal split fr the I-394 crridr; Imprve the level f service fr carpls and vanpls n I-394; Maintain r imprve the existing level f service fr mixed traffic n I-394; Maintain r imprve the accident rate alng I-394; Achieve and maintain a lw vilatin rate f the HOV lanes n I-394; and Cnstruct a cst-effective HOV facility n I-394. Fr each f these bjectives, specific perfrmance measures were identified and a crrespnding perfrmance threshld was established fr each perid. The threshlds were established by an analysis f the existing cnditins and the frecasted use fr the different perids. The perfrmance measures and threshlds identified fr the bjective relating t increasing the peak hur mde split are shwn in Table 2. This prvides an example f the apprach and level f detail invlved in the 1-394 befre-and-after evaluatin. The I-394 evaluatin prgram is supprted by cntinuus data cllectin. This prgram includes many elements similar t thse described fr ther studies. These include regular vehicle and ccupancy cunts n the HOV lane, mainlanes, and parallel facilities; travel time runs; accident data; vilatin rates; surveys f users and nnusers; and evaluatin f the different marketing and public infrmatin prgrams. Like the Hustn prgram, the I-394 evaluatin prgram is ne f the mre extensive and cmprehensive evaluatin prgrams currently being cnducted. 1-394 Minneaplis An interim HOV lane is currently in peratin in the Highway 12, I-394 crridr. The interim facility includes 3 mi f a reversible, barrier-separated HOV lane lcated in the median f the highway and additinal segments f cncurrent flw lanes. The final design f I-394, which is scheduled t pen in 1993, includes 3 mi f 2-lane, reversible, barrier-separated HOV lanes and 8 mi f cncurrent flw lanes. The interim HOV lane was pened in Nvember 1985. An extensive befre-and-after study fthe interim and final HOV Janes was initiated befre the pening f the interim facility. Under the guidance f the Minnesta Department f Tran prtatin a set f prject gals and bjective wa identified by ihe 1-394 prject management team fr bth the interim and cmpleted facilities. These frmed tbe ba i t develp the evaluatin prgram (10). The evaluatin prgram was develped befre the interim facility was pened in.1985. Three different perids were identified fr the prject evaluatin. These were the cn tructin perid, when the ijllerim facility wuld be in peratin; the start-up perid fr the cmpleted facility; and the table perating perid. The Ellwing bjectives were identified fr the HOV facility. Althugh these bj ctive were t apply in Rute 55 Cmmuter Lane, Orange Cunty, Califrnia In 1985, 11 mi f cncurrent flw HOV lanes were pened n Rute 55 in Orange Cunty. Called cmmuter lanes, they were the first exclusive carpl facilities in Orange Cunty. The lanes were initially pened as part f a 90-day demnstratin prgram spnsred by the Orange Cunty Transprtatin Cmmissin and Caltrans. An evaluatin prgram fr this demn tratin was develped by the Rute 55 Advisry Cmmittee. The key evaluatin issues used in the 90- day evaluatin are utlined in Table 3 (11). The data cllectin activities net:ued t supprt the evaluatin were als identified, as were the initial threshlds the prject shuld meet t be cnsidered successful. Data that were t be examined included vehicle and ccupancy cunts, travel time and speed runs, accident reprts, buffer and vehicle ccupancy vilatins, and mnitring the letters and telephne calls received. Tw a peels f the data cllectin activities are f intere I. Fir t videtaping vehicles at selected lcatins was used as ne methd f mnitring vehicle and ccupancy cunts. Secnd, n surveys were cnducted f users and nnusers during the initial demnstratin r the first few years f peratin. Hwever, in 1987 an extensive Rute 55

Turnbull et al. 69 TABLE 2 I-394 EVALUATION OBJECTIVE AND PERFORMANCE MEASURES: EXAMPLE Objective Perfnnance Existing Cnstructin Start-Up Stable Measure Cnditins Perid Perid Operatin 1984 1985-1990 1991-1992 1993-2000 Increase the Aut Number f 625 Peak Hur Carpls/Vanpls Mdal Split Carpl/Vanpl/ 1,380 Occupants Carpls as % f 19% Auts Carpl Occupants 34% as% f Aut Occupants Aut Occupancy Rate 1.23 700 1,075 1,585 1,540 2,515 4,900 21% 25% 29% 36% 42% 56% 1.25 1.3 1.6 travel behavir study was cnducted that included extensive surveys f bth users and nnusers f the HOV facility (12). Using the same general apprach, evaluatin studies were cnducted n Rute 55 after 9 mnths, 12 mnths, and 18 mnths f peratin. In additin t the travel behavir inventry, additinal special studies were cnducted n safety issues (13). These studies fcused specifically n changes in traffic safety resulting frm the HOV lanes. Tw elements f the Rute 55 evaluatin prcess are wrth stressing. First, the Rute 55 prcess appears t be ne f the first t use vide cameras extensively t btain vehicle, ccupancy, and vilatin infrmatin. Secnd, and prbably mre imprtant, the cncern ver and emphasis placed n safety and accident infrmatin prvides an element fr cnsideratin in ther evaluatins. Santa Clara Cunty Cmmuter Lanes HOV lanes, called cmmuter lanes, are currently in peratin n three expressways and ne freeway in Santa Clara Cunty. These facilities were pened between 1982 and 1988 and represent the first phases ( a larger system f HOV lanes that had been planned. In 1988, the Santa Clara Cunty Transprtatin Agency spnsred a detailed assessment f the cunty's existing cmmuter lane netwrk (14). The purpse f the study was t cmpile and summarize available data n cmmuter lane use, travel time savings, safety, and enfrcement. This infrmatin was then used t evaluate the effectiveness f the fur existing prjects. The final reprt n the evaluatin des nt identify any specific gals r bjectives the cmmuter lanes were designed t meet. The evaluatin relied primarily n histrical data available frm Santa Clara Cunty, Caltrans, and ther public agencies. The reprt presents a summary f available data n each facility and identified data deficiencies. The fllwing infrmatin was examined in the evaluatin: Vehicle vlume cunts fr bth the cmmuter lanes and the mixed-traffic lanes; Occupancy cunts fr bth the cmmuter lanes and the general-purpse lanes, which were used t cmpute the vilatin rates and rideshare percentages; Travel time data dcumenting speed and travel time fr vehicles in the cmmuter lanes and general-purpse lanes; Enfrcement statistics n the number f citatins issued fr illegal use f the HOV lanes; Accident data dcumenting the number f accidents in the HOV and general-purpse lanes; and Surveys f drivers n rutes with cmmuter lanes. This infrmatin was used in the evaluatin fr a general summary f the perating trends n the facilities. The results were nt measured against any benchmark; the evaluatin is descriptive and prvides an verall summary f the histrical trends. Summary f Cmmn Elements The preceding studies represent sme f the better examples f befre-and-after evaluatins f HOV facilities. Nte that the facilities cvered in these evaluatins cnstitute a relatively small percentage f the apprximately 40 HOV prjects currently in peratin. Frmal evaluatins have nt been cnducted n many HOV facilities and in sme areas the data needed fr such evaluatins has nt been, and is nt nw being, cllected. General cnclusins that appear t be apprpriate cncerning the current state-f-the-art in HOV prject evaluatins fllw. Frmal evaluatins f majr HOV facilities and thse with significant federal funding have been mre extensive and cmprehensive than lwer-cst TSM-type HOV facilities. Mst f the HOV prjects reviewed in this paper invlved significant investment in majr facilities. Many f these als included federal funding nt nly fr the facility, but als fr the evaluatins and data cllectin activities. The limited number f evaluatins n ther facilities appears t have partially resulted frm the nature f these facilities, many f which were implemented as TSM activities that limited funding fr evaluatin. Althugh frmal evaluatins were ften cnducted during the initial demnstratin stages f sme prjects, such as the

TABLE 3 SUGGESTED OBJECTIVES, MEASURES OF EFFECTIVENESS, AND THRESHOLD RANGES Objective Mea1ure1 r Effectiveneu 0 The HOV racility ehuld imrrve the c1pability r 1 cngelled rreeway crridr l mve 0 Aculll and percent incrs ae in the penin mvement efficiency mre peple by inc;rea1ing the number f penm per vehicle 0 Actual and percent increase in average vehicle ccupancy rate 0 Actual and percent increaae in carpls and vanpl 0 Actual and percent lncreaae in bu1 riden 0 The HOV facility ehuld increhe the persti111 efficiency r bu1 1ervice in the rreeway 0 Imprvement in vehicle prductivity (perating COil per vehicle-mile, perating cst per crridr paasenger, perating cst per phsenger mile) 0 Imprved bu1 schedule adherance (n-time perfnnance) 0 Imprved bua ufety (accidenll ratel) 0 The HOV facility shuld prv_ide travel time uvinga and a mre reliable trip time l 0 The peat-perid, peak~irectin travel time in the HOV lane( ) shuld be le11 than the travel time HOV1 utilizing the HOV facility in adjacent freeway lanes 0 lncreaae in travel time reliability fr vehiclee using HOV lane( ) 0 The HOV facility llhuld have favrable impacll n air quality end energy cnsumptin. 0 Reductin in emis1in1 0 Reductin in tlll fuel cnsumptin 0 The HOV facility shuld increaee the per lane efficiency r the ttal freeway facility. 0 Reductin in the grwth f vehicle mile1 f travel (VMT) and vehicle hun f travel 0 Imprvement in the peak-hur per lane efficiency r the lll! facility 0 The HOV facility lhuld nt unduly impact the peratin f the freew1y mainlane1 0 The level f 1ervice in the freew1y mainlanea lhuld nt decline 0 The HOV facility lhuld be ufe and lhuld nt unduly impact the ufety r the freeway general-l'urpae mainl1ne1 0 Number and 1everity f accidents fr HOV and freew1y l1ne1 0 Accident rate per millin vehicle mile1 f travel 0 Accident rate per millin passenger milea f invel 0 The HOV facility llhuld hlve public 1upprt. 0 Supprt fr the facility 1mng usen, nn-usen, general public, and plicy mak.en 0 Vilatin rate1 (percent f vehicle nt meeting the ccupancy requirement) 0 The HOV facility llhuld be a cst-effective transprtatin imprvement. 0 Benefit-cst rati

Turnbull et al. Shirley Highway HOV lanes and the San Bernardin Freeway Busway, nging evaluatins have nt always been cnducted. The data cllectin and evaluatin prcess used n the Hustn transitways appears t represent the mst extensive and cmprehensive effrt currently being cnducted. Many HOV facilities have been implemented withut clearly defined gals n the purpse and bjective f the prject. This lack f a clear understanding f the purpse and gals f the prject make evaluating the effectiveness difficult. There is n way f knwing if the gal has been reached when the gal has nt been defined. Cmpunding this prblem in sme cases is the use f bjectives that either cannt be measured r are inapprpriate fr the desired gal. Many evaluatins have been cnducted using very general evaluatin criteria. These measures may be as simple as a statement that the HOV lane shuld reduce travel times fr bus and autmbile cmmuters, withut identifying the level f time savings that shuld ccur. Thus, n benchmark r specific threshld is identified against which the prject will be measured. If the HOV facility leads t any imprvement in the general evaluatin measure, the prject is cnsidered successful. Fr example, the Shirley Highway Busway evaluatin did nt set specific imprvement levels, but rather used terms like imprve and increase. The I-394 and Rute 55 evaluatin prgrams represent tw f the better examples f the use f specific threshlds. There des nt appear t be a cnsensus amng transprtatin prfessinals n the apprpriate criteria r measures t evaluate HOV facilities. Varius measures have been used with different facilities. Althugh cmmn elements exist, many different appraches are currently being used. Further, a cnsensus des nt appear t exist regarding what levels f imprvement r change are f sufficient magnitude t cnclude that prject has been effective. These appear t be greatly influenced by lcal cnditins and perceptins. Sme evaluatin studies have fcused just n the HOV lane, withut cnsidering the full range f impacts, such as the effect n nnusers in the general purpse lanes and the peratin f the ttal system. Mst f the evaluatin studies examined in this paper accunted fr sme r mst aspects f the nn-hov lanes and the peratin f the ttal facility. This appears t be an issue that needs t be evaluated, but because f csts is nt always examined as extensively as might be desired. Many evaluatins are perfrmed n data that may preclude statistical analysis f the significance f any changes. In many cases, "befre" data is scarce r nnexistent. This, cmbined with limited samples f "after" data and little nging data cllectin, has ften lead t the inability t make statistically meaningful cmparisns. The evaluatin methdlgy, definitin f terms, and data cllectin methds ften differ amng prjects making cmparisns between them difficult. A clse examinatin f the data cllectin methds and definitin f terms used in the preceding evaluatins identified many such differences. Fr example, the definitin f the length f the peak perid ften varies, as des the exact time f the peak hur. There des nt appear t be cnsensus amng researchers n the apprpriate way t deal with utside changes that may affect the results f the HOV prject. In additin, sme lcalities use a cntrl facility t mnitr verall changes, while thers d nt. The Hustn evaluatin prcess, which mnitrs nt nly the fur transitways, but als tw cntrl freeways, prvides ne f the better appraches fr identifying ptential utside influences. The frging examinatin f the evaluatin prcesses in the seven metrplitan areas and a review f ther relevant literature identified a number f cmmn gals r bjectives and crrespnding measures f effectiveness that appear t be mst apprpriately used with HOV facilities. These are presented in Figure 3. An indicatin f the types f data needed t supprt the evaluatin are prvided in Table 4. In additin, the fllwing steps are imprtant elements in the develpment and implementatin f a cmprehensive nging evaluatin prgram fr HOV facilities. These steps are nt new-they reflect the majr activities that shuld be undertaken as part f any evaluatin prgram. Clear articulatin f prject gals and bjectives. The gals and bjectives the HOV prject are intended t accmplish need t be clearly defined. This is a critical step, because the remainder f the prgram will be designed t btain and 71 What is the effect f the Cmmuter Lane n verall freeway peratins? Is the Cmmuter Lane being utilized effectively? What is the effect f the demnstratin prject n verall freeway utilizatin? What effect has the Cmmuter Lane had n travel times and cngestin levels in the general use lanes? Hw d the Cmmuter Lanes affect verall freeway safety and vilatins? Are the Cmmuter Lanes safe t use? T what extent are freeway users vilating the painted buffer and entering and exiting the lanes? T what extent are the lanes being used by nn-carplers? Has the Cmmuter Lane had an effect n traffic vlumes n the parallel arterials and/r freeways? What is the public's attitude tward the prject? Are there any special factrs r circumstances that shuld be reprted and cnsidered? FIGURE 3 Rute SS cmmuter lane evaluatin elements.

TABLE 4 SUGGESTED OBJECTIVES, DATA COLLECTION EFFORTS, AND MEASURES OF EFFECTIVENESS FOR EVALUATING HOV FACILITIES Data CUectia Effrts Vehicle and Ocupancy Cunt Travel Time Runs Survey ' Objecti"e Freeway' HOV Lane Freeway' HOV Lane Freeway HOV Lane Olher CrnspUI& M~ r Effectiveness (MOE's)' Increase vehicle.... Actual ne! percent incre11e in peak-hur, peak-directin pel"90n ccupancy vlume; increae in 1verage vehicle ccup1ncy; and mdal hifl Bus perating Imprved vehicle prductivity; imprved bu1 chedule adherance; nd efficiency imprved bu111fety Travel time 11ving1... Amunt f 1nvel time 11ving by HOV uaen; reliability f trip time fr HOVuaen Energy and air..., Reductin in cm.iuina; reductin in energy cnaumptin Pe r lane efficiency.. lncreae in peak-hu.r per lane efficiency f ttal freeway facility. Freeway Maintain r imprve level f 1ervice n freeway m1inlane1 pcralin1 Safety Number and acverity f accidcnta; accident rale per millin vehicle m.ilea f travel and per millin puaenger mile f invel Public supprt Percent f uacn, n111h1aen, and general public wh apprve f HOV facility; vilatin rate Cst effective Benefit..:at rati Indicates the tp pririty data cllectin effrt nce<>ed t evaluate the bjectivea. lndicatca data cllectin effrll which huld ideally be cnducted, but arc nt abelutely neceaaary t evaluate the bjectivea. Invlves peridic uac f survey f HOV uaera (bu1 ndera, cuplcn, ind vanplcn), nn-hov Ultra in the 1eneral traffic lanes, ind in eme c11t1, the 1eKral public. h i111rnj!ly1ugge1ted that thi1 d1ta be cllected fr bth the freeway lanc11djaccnt t the HOV facility and the cntrl freeway. Sme, hut nt ncccaunly all, f the..uggcatcd MOE.I auciatcd with gauging the altainmcnt f the bjective are hwn. Vehicle and ccup1ncy cunta n alternate arterial rute t identify any ch1nge1 in thrughput fr the crridr, cunta al park-and-ride Ita, ind vehicle and ccupancy cunll n a "cntrl" freeway. Befre-and-after bu1 1trvice levela, vehicle prductivity, chedule adherence, number ind 1tverity f bu1 accidenll, vehicle perating cats, and change in l1br, fuel, and ther Clla. Mnitring bua n time perfrmance ind chedule adherence befre-and-a Iler implementatin f the HOV l1ne(1). Mnitring air quality levcl1 alng the crridr ind use r 1imulatin mdeb t estimate impact. Mnitring freew1y accident ralea ind type befre-and-after implementalin f the HOV l11ie(1), 11 well 11 btaining accident ratea n the HOV facility. Identifying vilatin rte1 fr the HOV lane (i.e., thse vehicle t meeting the minimum ccupancy rc.iuirement). Mnitr cmplaint, media, and plicy actiu.

Turnbull et al. evaluate infrmatin that will be used t determine if these bjectives have been met. The develpment f measurable bjectives is nt an easy task, but time spent n this effrt will help ensure a fcused evaluatin. Identificatin f measures f effectiveness and the threshld level f change. Fr each bjective, the apprpriate measure r measures f effectiveness shuld be identified, alng with the desired level f change t be used t determine if the facility has met the bjective. It is imprtant that this activity fcus n identifying the key measures that mst accurately relate t the bjectives, and that realistic threshld levels be established. Identificatin f the infrmatin needed t cmplete the evaluatin. This step identifies the infrmatin needs f the evaluatin prcess. Fr each measure f effectiveness, the data needed t determine if the bjective has been realized must be identified. The apprpriate methds t btain and evaluate the infrmatin shuld als be identified. Develpment f the study design. The previus three activities shuld be brught tgether.in the develpment f the cmprehensive study design. This shuld include the listing f bjectives, measures f effectiveness, and data cllectin and analysis needs and prcedures. It is imprtant that the scpe f thi effrt crrespnds t the funding and staffing limitatins f the agencie invlved. Cnduct "befre" data cllectin. In this step, data are cllected befre the implementatin f the HOV prject. This step is critical. If n "befre" data are cllected, it is virtually imps ible t determine the impact f the HOV facility. Cnduct "after" data cllectin and evaluatin. In this step, the "after" data are cllected. Usually a number f different evaluatin pints are identified, such as after 1 year, 2 years, and n an cntinuing ba is. The befre-and-after data are then evaluated, and the prject effectiveness is examined. Onging mnitring and evaluatin. After the initial evaluatin, an nging mnitring and evaluatin prcess shuld be maintained. Because different areas will have different resurces available fr this, the prgram shuld be designed t ensure that the key infrmatin is cllected and analyzed within the resurces available. CONCLUSION A review f current practices used t cnduct evaluatin studies f perating HOV facilities has been presented in this paper. It has included a review f examples f seven evaluatin prgrams cnducted n HOV prjects and a discussin f the cmmn elements f these studies. The results f this analysis shuld be beneficial fr areas in the prcess f designing befre-and-after evaluatin prgrams fr new HOV prjects and fr areas with perating HOV facilities that may wish t imprve current evaluatin methds. It is imprtant t nte that differences will prbably cntinue t exist in the evaluatin prcesses used by different metrplitan areas. This is apprpriate because f the need t fcus evaluatins n the issues that are f lcal cncern. Hwever, in the lng run, all areas culd benefit frm a mre standardized and fcused apprach t evaluating HOV facilities. ACKNOWLEDGMENT The wrk described in this paper was undertaken by TTI, a part f The Texas A & M University System. The examinatin f current befre-and-after evaluatin studies f freeway HOV facilities was ne element cnducted as part f a larger study, The Assessment f Freeway High-Occupancy Vehicle Lane Prjects. This assessment is funded by the Urban Mass Transprtatin Administratin thrugh the Texas State Department f Highways and Public Transprtatin. REFERENCES 1. K. F. Turnbull and J. W. Hanks, Jr. A Descriptin f Hig'1-0cc11pa11cy Vehicle Facilities i11 Nrth America. Texas Tran prta1in In titute, Cllege tatin, Tex., 1990. 2. K. F. Turnbull, R. H. Henk, and D. L. Chri tiansen.. 11gge ted Prcedures fr Eva/11ali11g the Effectiveness f Freeway l IOV Fa cilities. Draft. Texa Trnn prtatin In 1i1u1e. llege 'talin Tex., 1990. 3. J. T. McQueen, D. M. Levinshn, R. Waksman, and G. K. Miller. Evaluatin f the Shirley Highway Express-Bus-On-Freeway Demnstrlllin Prject-Final Rep/'/. U.S. De partment f Cmmerce, 1975. 4. Crain and Assciat. a11 8emardi110 Freeway Busway Evalu min f Mixed-Mde Operatins. uthern Califrnia Assciatin f Gvernments, Ls Angeles, 1978. 5. T. J. Athertn and E.. Eder. Nust()11 Nr1/1 Freeway Cntmfliv Lane Dem11s1ra1i11. ambridge Sy ternatic, Inc.; U.S. Depanment f Tran prtatin, Austin, Tex.. 1982. 6. D. L. Christiansen and D. E. Mrris. The Sta1us f the Effec liveness f the Hustn Transitway System, 1989. Tcxa Transprtatin In titute Cllege Srntin, Tex.; Tcxa State Department f Highway and Public Tran prtatin, Austin, Tex., 1990. 7. S. M. Betts, L. N. Jacbsn, and T. D. Rickman. J-5 HOV Lanes: Three Mnth Reprt. Washingtn State Department f Trans prtatin, Seattle, Wash., 1983. 8. 1-5 HOV Lanes: A Twenty Mnth Reprt. Washingtn State De partment f Transprtatin, Seattle, Wash., 1985. 9. K. C. Henry and M. Mehyar. Six-Year FLOW Evaluatin. Washingtn State Department f Transprtatin, Seattle, Wash., 1989. 10. trgar-rsce, Inc. 1-394 Trai1spnmi11 y.vtem Mm111g1 me111 Plan Objectives. Minnesta Department r Transprtatin, St. Paul, Mi nn., 1984. 11. Rwe 55 Cmmtt1er Lmw Demnstratin Prjec1: 90 Day Eval- 1101i11 Reprt. Orange Cunty Trnn prtatin Cmmi in; Cal ifrnia Department f Tran prtatin, 1986. 12. S. Green and J. apellc. Rute 55 Travel Behavir tudy, Orange Cunty, Califrnia. Presen1ed at the 681h J\nm111I Meeting f the Transprtatin Research Bard, Washingtn, D.C., Jan. 1989. 13. An Analysis f Traffic Safety Relative t the Cmmuter Lane Prjects 011 R-55 and SR-91 in Orange and Ls Angeles C1111ty. Technical Memranda I and 2. Institute f Transprtatin llldie-, Univcr ity f 'alifrnia. Irvine, 1987. 14. Cmmuter Lane Perfrmance Evaluatin. SYSTAN, Inc.; Santa Clara Cunty Transprtatin Agency, San Jse, 1989. The pinins expressed in this paper are thse f the authrs, nt UMTA r SDHPT. Publicatin f this paper spnsred by Cmmittee n High-Occupancy Vehicle Systems. 73