Mod spcific accssibility and car ownrship 1 Max Bohnt, 2 Carstn Grtz TU Hamburg-Harburg, Institut für Vrkhrsplanung und Logistik 1 2 bstract This papr analyss th links btwn mobility tool ownrship (cars, bicycls public transport passs) and mod spcific accssibility, using data from th Hanovr Rgion. n utility basd indicator is proposd, which capturs accssibility by diffrnt mods diffrntiating prsons with and without car availability. Incorporating accssibility indicators in discrt choic car ownrship modls hlps to bttr xplain spatial influncs on car ownrship. Th analysis rvald that, whn controlling for socio-conomic variabls, car ownrship lvls doubl in aras with a poor accssibility to dstinations for prsons without cars. 1. Introduction In th last dcads motorisation lvls havily incrasd in most countris. t th sam tim, car-orintd land us pattrns and spatial rlationships mrgd not only in North mrican citis that ar hardly accssibl without privat cars. Prsons which do not (anymor) hav accss to a car find thmslvs in an accssibility trap. This is an issu of major importanc looking at a rapidly aging suburban population. cars pro 1.000 capita tlanta 700 Houston Prth 600 San Francisco 500 Munich Nw York 400 Brn Barclona 300 mstrdam Tokyo Copnhagn 200 Lisbon 100 Singapor Hong Kong 10000 20000 30000 40000 50000 GDP pr capita (US$) Figur 1: Car-ownrship and GDP pr capita in 100 mtropolitan aras (Sourc: Data from th Millnnium Citis Databas [1]) Howvr, som citis and rgions, whr car ownrship lvls hav rmaind rlativly low dspit high incom lvls, could prsrv a transportation systm and urban structur whr walking, cycling and public transport provid a rlativly comptitiv accssibility compard to accssibility by car. In ths placs, othr mobility tools lik bicycls and public transport passs ar oftn attractiv altrnativs to owning a (scond) car. Car ownrship is a major dcision for housholds not only bcaus of th conomic impact on th houshold budgt. It is an important link btwn long and short trm mobility dcisions (Figur 2). On th on hand, car availability dtrmins short trm mod and dstination dcisions for daily activitis. It influncs which dstinations (.g. shops) a prson visits and by which mod th trip is carrid out. On th othr hand long trm job and rsidntial location dcisions dpnd on th availability of mobility tools. Housholds only tak locations into account that thy prciv accssibl by thir availabl mods.
Figur 2: Th land us-transport fdback cycl [2] Th fdback mchanism of accssibility on car ownrship dcisions has rarly bn studid. Th purpos of this papr thrfor is twofold: Sction 2 discusss mod spcific accssibility and proposs utility basd mod spcific accssibility indicators to assss th comptitivnss of walking, cycling and public transport compard to car us. In Sction 3 ths indicators ar usd to stimat discrt choic modls of car ownrship. Th impact of accssibility on car ownrship dcisions of housholds and prsons ar analysd, controlling for socio-conomic variabls. Finally th rsults ar discussd and som indications on furthr rsarch nds givn. 2. ccssibility in th Hanovr Rgion Th Hanovr Rgion mtropolitan ara has 1.1 Mio. inhabitants. Half of thm ar living in th city of Hanovr, th othr half in th 22 suburban municipalitis. Nustadt Wdmark Burgwdl Wunstorf Garbsn Langnhagn Isrnhagn Burgdorf Utz Slz Hannovr Lhrt Barsinghausn Ghrdn Wnnigsn Spring Ronnnbrg Hmmingn Pattnsn Laatzn Shnd cars pr 1.000 capita irport 250-300 301-350 351-400 401-450 451-500 501-550 551-600 601-612 Rail Ntwork Light-Rail Ntwork Figur 3 and 4: Mobility tool ownrship in th Hanovr Rgion. motorisation (lft, Data: [3]) and public transport passs (right, Data: [4]) Hanovr traditionally has rlativly low car ownrship lvls [5]. Th motorisation is highst in suburban municipalitis lacking (good) rail accss (Figur 3), whil in cntral locations in Hanovr only on out of 4 prsons owns a privat car. Highst rats of public transport pass ownrs ar found in a 500 m-radius around th (light-) rail stops in Hanovr and som suburban municipalitis (Figur 4). Pass ownrship drops by 25% in a distanc of 500-1.000 m and by 50% in a distanc of mor than 1.000 m to a rail stop.
Data sourcs s socio-conomic charactristics play an important rol in mobility dcisions, thy hav to b controlld for whn analysing th link btwn land us and mobility. This study analysd th data of th Grman Mobility Survy (MiD) 2002 [6]. Th survy includs information on 25.800 housholds with 61.700 prsons throughout Grmany. For th Hanovr Rgion, dtaild spatial information on rsidntial location and on trip origins and dstinations ar availabl for a 4.581-houshold sampl. transportation ntwork modl [7] and a GISdatabas of jobs and shopping facilitis wr usd to calculat travl tims and accssibility indicators. Th transportation modl of th Hanovr Rgion (863 zons) and a GIS-databas of jobs and shopping facilitis wr usd to calculat travl tims and accssibility indicators. Dspit th growing significanc of car sharing (2.800 usrs in th Hanovr Rgion) no data was availabl in th MiD data, so th influnc of car sharing as an altrnativ to privat car ownrship could not b studid. Prcivd accssibility in th Hanovr Rgion In th MiD 2002 th rspondnts wr askd to assss th accssibility to usual dstinations by car and public transport. Th prcivd accssibility by car is good or vry good for most rspondnts. Only in Hanovr s innr nighbourhoods a significant shar (13%) asssss th accssibility as modrat to vry poor (Figur 5). This might rflct scar parking facilitis in th innr nighbourhoods. Th accssibility by public transport is ratd vry good by 62% in th innr nighbourhoods of Hanovr. This shar drops to 22% in rural aras (Figur 6). Compard to othr Grman rgions, accssibility by public transport is gnrally ratd bttr in th Hanovr Rgion. This might b du to th xtnsiv light and commutr rail systm in Hanovr City, that covrs 71% of th Citis population within 500 m and 94% within 1.000 m of a light or commutr rail station. In th suburban municipalitis, 19% liv within 500 m and 41% within 1.000 m of a light or commutr rail station. 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Hanovr innr nighbourhood Hanovr outr nighbourhood vry good good modrat poor vry poor first suburban ring mid-lvl cntr basic cntr rural ara Hanovr Rgion Grmany: monocntric conurbations Grmany: citis ovr 500.000 inh. Grmany Hanovr innr nighbourhood Hanovr outr nighbourhood vry good good modrat poor vry poor first suburban ring mid-lvl cntr basic cntr rural ara Hanovr Rgion Grmany: monocntric conurbations Grmany: citis ovr 500.000 inh. Grmany Figur 5: prcivd accssibility by car (N=7.403, lft) and by public transport (N=7.677, right) in th Hanovr Rgion and in Grmany. Data: [6] n utility basd mod spcific accssibility indicator Thr ar many ways to masur accssibility, ranging form simpl, infrastructur basd to mor complx utility basd indicators [8]. prson without car availability might accss som dstinations by foot, som by bicycl and othrs by public transport. To assss an intrmodal accssibility, th qustion ariss, how to combin th accssibilitis by various mods. This objctiv was to dvlop an utility basd indicator basd upon an simultanous dstination and mod choic modl. Th indicator is abl to captur th availability of mobility tools and intgrats th accssibility of dstinations by diffrnt availabl mods of
transport. s th paramtrs ar stimatd from MiD survy travl data, it rflcts th rspondnts prcption of distanc dcay. Choic Dstination 1 (...) Dstination j (...) Dstination J Car Public Transport Cycl Walk Figur 6: Nstd logit modl of mod and dstination choic nstd logit modl for mod and dstination choic (dstination choic at nst lvl, mod choic at th bottom lvl) was stimatd (Figur 6), taking up th modl proposd in [9: p. 18 ff.]. modl for J zons has 4*J altrnativs. Th utility formulation is writtn in (1): U ijm mav ( j fm( gcmij ) + i = C + ln ε (1) Th utility U ijm of mod m and dstination j for prson i incrass with th attractivity j of dstination j, masurd by numbr of jobs or rtail floorspac. It dcrass with rising gnralisd costs gc ij (travl tim, costs, transfrs ) btwn th rsidntial location of prson i and dstination j. It is possibl to rlat th travl costs to th incom of prson i. This has not bn don within this study. ε i rprsnts a stochastic trm, which is assumd to b gumbl distributd and rflcts th unobsrvd prfrncs of obsrvation i. 100% 90% 80% 70% 60% Walking Cycling Public Transport Car 50% 40% 30% 20% 10% 0% 0 10 20 30 40 50 60 70 80 90 gnralisd costs Figur 7: Distanc dcay functions for diffrnt mods Th shap of th mod spcific distanc dcay function f m (gc mij ) dscribs, how fast th utility of dstinations dcrass with incrasing gnralisd costs (Figur 7). Diffrnt distanc dcay functions (xponntial function, Box Turky transformation, loglogistic, EV) wr tstd. Bst fits wr obtaind with th EV function (2). f ( w) = 1/ ( 1+ w) F ( ( 1 w / G 1+ ) ) (2). C mav is a constant trm for mod m. It rflcts th gnral prfrncs for mod m and its availability for prson i (mod car: always, tmporal or nvr availabl). For a prson 1 with tmporal car availability, C mav is lowr than for a prson 2 with a car always availabl. This ffcts, that th accssibility by car of a dstination contributs lss to th ovrall accssibility of prson 1 than for prson 2, bcaus prson 1 cannot accss this dstination
all th tim by car. Th modl could also captur th availability of othr mobility tools (bicycl, transit pass) by incorporating corrsponding constants. Th probability for prson i on th givn rlation to dstination j (within nst j) to choos mod m from th st of availabl mods M is writtn in (3). λ is th nst-paramtr capturing th covarianc of th rror trms of th altrnativs within th nst. P ij m M = U m M ijm U / λ ijm / λ (3) I ij = m M U ijm ln (4) / λ Th log of th dnominator in (3) is writtn as I ij in (4). It rflcts th utility that dstination j contributs to th ovrall utility of th altrnativs. It is usd to stimat th probability to choos dstination j out of th st of all dstinations J (5). P i j J = λi ij λi ij (5) ac i = λi ij ln (6) j J j J Th log of th dnominator in (5) can b intrprtd as th accssibility ac i for prson i to all dstinations J with all availabl mods M (6). s th utility formulation of th modl (1) includs th trm C mav which dpnds on th availability of mod m of prson i, th accssibility indicator ac i rflcts mod availability. Th modl was stimatd with th MiD-Survy data (1.889 journy to work trips). To ach chosn altrnativ (dstination j, mod m) 47 non-chosn altrnativs wr sampld. In ordr to rduc stimation tim, 11 non-chosn dstinations hav bn slctd as a stratifid random sampl: for ach k [1..11], on dstination j k within th subst S k. S k is th choic st of all dstinations in a distanc class k from th rsidntial zon of prson i. If S k is mpty, an additional dstination from distanc S k+1 is chosn. Th information on th gnralisd costs of all 48 altrnativs and th attractivnss jk, wightd with th invrs probability of zon j k to b chosn from th choic st S k wr addd to th datast in ordr to stimat th modl paramtrs. Th modl paramtrs wr stimatd using biogm [10]. ccssibility ll mods 9,1-9,5 9,6-10,0 Railway Frway ccssibility without car 9,1-9,5 9,6-10,0 Railway Frway 5,4-6,0 10,1-10,5 Fdral Highway 5,4-6,0 10,1-10,5 Fdral Highway 6,1-7,0 10,6-11,0 6,1-7,0 10,6-11,0 7,1-8,0 11,1-11,5 7,1-8,0 11,1-11,5 8,1-9,0 11,6-12,5 8,1-9,0 11,6-12,5 Figur 8: ccssibility with all mods (lft) without car availability (right)
Figurs 8 show accssibility masurd by this accssibility indicator. To compar and intrprt utility basd accssibility indicators, thy hav to b scald with a marginal utility of incom paramtr λ [8: p.63]. Using paramtrs drivd from Tabl 1 blow (λ=xp(β EVni / β pinkln ) = xp(2.961)=1.40), an incras of on indxpoint of this accssibility indicator rprsnts a 40% bttr accssibility. Th lft map in Figur 8 shows th job accssibility indicator for a prson with all mods availabl. It rangs from indx valus of 8.3 in priphral aras up to 12.1 in cntral Hanovr and along som motorway xits. This rprsnts an 3.5 tims bttr accssibility in cntral locations for car ownrs. Th right map in Figur 8 shows th job accssibility without car availability. It is lowr than th accssibility by car, ranging from 5.4 at th priphry to 10.5 in cntral Hanovr. Howvr it can b sn that prsons in cntral Hanovr hav a highr job accssibility than prsons in priphral aras at th fring of th rgion, vn if thy don t hav accss to a car. In Figur 8 th light-rail tracks and th aras around th railway stations can b clarly idntifid as aras with substantial highr accssibility than in th surrounding. High job accssibility for prsons without cars is also found in som mid-rang cntrs lik Laatzn, Lhrt and Wunstorf with a fast rail accss to Hanovr and a good local job supply accssibl by walking and cycling. Th accssibility gap of prsons with no car availability In aras, whr accssibility by car is much highr than accssibility by othr mods ar particularly car dpndnt. Prsons without car availability hav clar disadvantags compard to car ownrs, thy fac a big accssibility gap. ccssibilty gap indicators ar proposd by [11] and [12]. This study proposs th diffrnc btwn th prsntd accssibility indicator for prsons with and without a car availability as accssibility gap indicator. Th advantag is, that it covrs th accssibility by diffrnt mods (walking, cycling, public transport) as altrnativs to car us. low accssibility gap indicats aras whr walking, cycling and public transport ar rlativly comptitiv to cars (Figur 9). ccssibility Gap without car 1,3-1,5 1,6-1,8 1,9-2,0 2,1-2,3 2,4-2,5 2,6-2,8 2,9-3,0 3,1-3,3 3,4-4,0 4,1-5,0 Railway Frway Fdral Highway Figur 9: Job accssibility gap without car compard to prsons with car availabl For commut trips, th accssibility gap is highst (5 indxpoints) in priphral aras but also in many aras clos to Hanovr with a fast road accss and a poor public transport supply. It is low in cntral Hanovr (1.25 indxpoints) and along th rail stations and in som scondary cntrs.
Th indicator also could b calculatd to assss th accssibility gains providd by othr mobility tools, such as bicycls or public transport passs as acm = acm ac0, whras ac m is th accssibility indicator with m, a combination of mobility tools (.g. car availabl and transit pass ownrship) and ac 0 th accssibility with rfrnc altrnativ 0 (.g. car nvr availabl, no transit pass). Th study focusd to job accssibility. For an ovrall pictur, not only job accssibility also accssibility to othr activitis lik shopping, schools or srvics should b considrd. 3. Car ownrship and accssibility without a car In this sction th utility basd accssibility indicator is usd to analys th rlation btwn accssibility and houshold car ownrship lvls. Car ownrship is widly dtrmind by socio-conomic charactristics lik houshold siz, incom, or ag [13]. To control for ths variabls, a discrt choic modl of houshold car ownrship was stimatd. Houshold car ownrship modl Th dpndnt variabl is a st of discrt altrnativs with an ordinal structur: 0, 1, 2, 3 or mor cars pr houshold. For that rason an ordinal logit modl has bn chosn (7): P n ln = α n X + εi P β (7) 1 n It stimats th probability P n for a houshold to own n cars or lss. X is a vctor of th indpndnt variabls, (charactristics of houshold i at its rsidntial location) and β is a vctor of th paramtrs of X. α n is a constant (thrshold valu) for th probability to own n cars or lss. ε i rflcts th unobsrvd prfrncs of houshold i. It is a stochastic trm, which is assumd to b gumbl distributd (an ordinal probit modl assuming normal distribution of ε I dlivrs narly th sam rsults). Th modl has bn stimatd using th ologit (ordinal logit) modl of Stata10 [14]. First, a basic modl only including socio-conomic variabls was stimatd. s thr ar svral non-linar rlationships, th following spcification of th significant variabls was found to fit th modl bst: th catgorical variabl nt houshold incom (8 classs) has bn transformd into a continuous variabl log of nt prsonal quivalnt incom, taking th numbr and th ag of houshold mmbrs into account [15]. numbr of adults (1, 2, 3, 4 or mor as dummys) numbr of childrn (1, 2, 3 or mor as dummys) th sum of th (ag - 65) of all mn oldr than 65 yars and th sum of th (ag - 65) of all womn oldr than 65 yars Th rsults ar displayd in th column basic modl of Tabl 1. s xpctd, th probability to own mor cars incrass with incom and numbr of adults in a houshold, and lss strongly with th numbr of childrn. It dcrass with th ag of snior prsons in th houshold, particularly among womn.
COEFFICIENT LBELS basic modl including including accssibility mobility tools FRU_65 womn sum of -0.110*** -0.118*** -0.127*** yars ovr 65 (-0.012) (-0.012) (-0.013) MNN_65 mn sum of -0.0427*** -0.0424*** -0.0567*** yars ovr 65 (-0.014) (-0.014) (-0.014) pink_ln log(quivalnt 1.762*** 1.836*** 1.805*** incom) (-0.098) (-0.1) (-0.1) _InzErw_2 2 adults 2.420*** 2.322*** 2.319*** (-0.11) (-0.11) (-0.13) _InzErw_3 3 adults 4.057*** 3.961*** 4.198*** (-0.18) (-0.18) (-0.21) _InzErw_4 4 and 5.747*** 5.559*** 6.011*** mor adults (-0.3) (-0.29) (-0.33) _InzKindr_1 1 child 0.772*** 0.707*** 0.732*** (-0.12) (-0.12) (-0.13) _InzKindr_2 2 childrn 0.943*** 0.781*** 0.767*** (-0.13) (-0.13) (-0.13) _InzKindr_3 3 and 1.000*** 0.921*** 0.851*** mor childrn (-0.26) (-0.27) (-0.27) ZPP750LN log(m² rtail -0.0276** -0.0327** floorspac in 750 m) (-0.013) (-0.013) EVni accssibility wihout -0.620*** -0.509*** car availabl (-0.048) (-0.049) fahrrad numbr of 0.300*** bicycls (-0.071) _Izitkart_1 1 public -1.354*** transport pass (-0.12) _Izitkart_2 2 public -1.832*** transport passs (-0.24) Constant 1 car 12.41*** 7.013*** 7.611*** (-0.7) (-0.8) (-0.81) Constant 2 cars 16.10*** 10.96*** 11.81*** (-0.74) (-0.83) (-0.85) Constant 3 cars 18.91*** 13.97*** 14.94*** (-0.76) (-0.84) (-0.86) Obsrvations 3872 3872 3862 Psudo R-squard 0.2602 0.2998 0.3344 Robust standard rrors in parnthss *** p<0.01, ** p<0.05, * p<0.1 Tabl 1: Estimatd modl paramtrs for houshold car ownrship To this basic modl, two accssibility indicators ar addd: EVni : Th job accssibility without car availability drivd from th joint dstination and mod choic modl, using an EV-distanc dcay function, as dscribd in sction 2. ZPP750LN: th log of rtail floorspac (for priodic shopping purposs) within in 750 m walking distanc. Including ths variabls, th modl can b improvd significantly. Ths variabls hav a strong ngativ influnc on car ownrship. Th modl can b usd to stimat th
probabilitis for a houshold to own 0, 1, 2 or mor cars, dpnding on th accssibility without a car at th rsidntial location. Figur 10 displays th car ownrship probability for a houshold with 2 adults undr 65 yars without childrn with an avrag incom (1.850 nt monthly houshold incom, which corrsponds to 1.230 prsonal quivalnt nt incom), dpnding on th accssibility without cars at th rsidntial location. Th probability to own no car incrass from 1 % at th last accssibl location to 17 % at th most accssibl location of th Hanovr rgion. Manwhil th probability to own 2 or 3 cars dcrass from 70% to 10%. 100 90 80 70 Thr Cars Two Cars 60 % 50 40 30 20 10 0 worst accssibility On Car No car 5 6 7 8 9 10 ccssibility Indx without car availability bst accssibiltiy Figur 10: Probability of car ownrship of a 2-adult houshold with a nt houshold incom of 1.850 by accssibility without car at rsidntial location This mthod can b applid for a micro-simulation of car ownrship of a synthtic population, as dscribd in [16]. In Figur 11 th modlld avrag numbr of cars for such houshold of 2 adults undr 65 yars without childrn with a nt monthly houshold incom of 1.850 ar displayd. It shows that th motorisation of this houshold typ is narly twic as high in priphral locations and that within suburban aras it is clarly lowr around railway stations and clos to th scondary cntrs. So an avrag 2-adult houshold spnds yarly 4.000 on fixd car costs in poorly accssibl aras instad of 2.000 wll accssibl locations (calculatd with cost rats from [17]). In addition, in poorly accssibl aras thy driv much mor kilomtrs and bar svral tims highr variabl vhicl costs [18]. Th last column of Tabl 1 displays th rsults of anothr modl spcification which includs th ownrship of othr mobility tools. strong ngativ corrlation btwn car ownrship and public transport pass ownrship is found, which indicats that public transport passs might srv as an altrnativ to a (2 nd ) car for many housholds. For bicycls, a positiv corrlation was found btwn housholds which own mor cars and own mor bicycls. Modlling th mobility tool dcisions in a joint modl could hlp to bttr undrstand ths intractions [19].
Railway Cars pr houshold 0,94-1,00 1,01-1,10 1,11-1,20 1,21-1,30 1,31-1,40 1,41-1,50 1,51-1,60 1,61-1,70 1,71-1,85 Outlook Figur 11: Expctd avrag numbr of cars pr houshold for a 2-adult houshold with a nt monthly houshold incom of 1.850 in th Hanovr Rgion Th rsults show, that thr ar significantly lowr car ownrship lvls at locations wll accssibl without car, vn whn controlling for socio-conomic houshold charactristics. Th study rvals, that housholds living at cntral locations with an attractiv public transport ar using othr mobility tools lik public transport passs as a substitut to a (scond) car and sav a high amount of mobility costs. Howvr, th qustion of slf-slction rmains opn: to what xtnt housholds with prfrncs for an automobil lifstyl choos locations wll accssibl by car and housholds with prfrncs for intrmodal mobility mov to locations wll accssibl by foot, bicycl and public transport? To what xtnt do housholds adjust thir motorisation whn thir accssibility changs du to land us or infrastructur changs, or aftr rlocation? To analys ths intractions btwn accssibility, car ownrship and location choic, longitudinal panl data analysis can provid valuabl insights [20]. Furthr rsarch along ths lins is ndd for a bttr undrstanding of th land us - transportation intraction. cknowldgmnts I would lik to thank th Hanovr Rgion (particularly Tanja Göblr and Klaus Gschwindr) and th üstra (Harald Paul) for providing th data and Carstn Grtz, Matthias Winklr and ndra Broaddus for thir hlpful commnts on this papr. Rfrncs [1] Knworthy J, Laub F: "Millnium Citis Databas", In: UITP, (d.). Bruxlls, (2000). [2] Wgnr M, Fürst F: "Land-Us Transport Intraction: Stat of th rt", Bricht aus dm Institut für Raumplanung: Univrsität Dortmund, (1998). [3] Landshauptstadt Hannovr: "Strukturdatn dr Stadttil und Stadtbzirk 2007". Hannovr, (2007).
[4] üstra Hannovrsch Vrkhrsbtrib G: "bonnntndatn", (2007). [5] Dckrt P: "Di voraussichtlich Entwicklung ds Pkw-Bstands in Hannovr bis zum Jahr 2000", In: Landshauptstadt Hannovr, (d.) Schriftn zur Stadtntwicklung Hft 39. Hannovr, (1988). [6] infas, DIW: "Mobilität in Dutschland 2002 - Kontinuirlich Erhbung zum Vrkhrsvrhaltn - Endbricht", In: GmbH i-ifas, (DIW) DIfW, ditors. Bonn / Brlin: Bundsministrium für Vrkhr; Bau- und Wohnungswsn (BMVBW), (2003). [7] WVI: "Vrkhr in dr Rgion Hannovr - Zustandsanalys und Handlungsansätz", Intgrirt Vrkhrsntwicklungsplanung Rgion Hannovr. Braunschwig: Prof. Dr. Wrmuth Infrastrukturplanung GmbH, (2004). [8] Gurs K, Ritsma van Eck J: "ccssibility Masurs - rviws and application", In: Environmnt NIoPHat, (d.) RIVM Rport. Bilthovn/Utrcht, (2001). [9] Lohs D, Schillr C, Tichrt H, Vrtic M, Fröhlich P, Schüsslr N, xhausn KW: "Ein zwisitig gkopplts Modll zur simultann Brchnung dr Vrkhrsrzugung, Vrkhrsvrtilung und Vrkhrsauftilung: thortischr Hintrgrund und praktisch nwndung für in nationals Modll dr Schwiz", Vrkhrsforschung Onlin (2006);1(3). [10] Birlair M: "BIOGEME: fr packag for th stimation of discrt choic modls", 3rd Swiss Transportation Rsarch Confrnc,(2003), scona, Switzrland. [11] Kwok RCW, Yh GO: "Th us of modal accssibility gap as an indicator for sustainabl transport dvlopmnt", Environmnt and Planning (2004);36(5): pp. 921-936. [12] Stin : "Gordntr Raum und öffntlichr Vrkhr - Di Roll ds öffntlichn Vrkhrs bi dr Sichrung nachhaltigr Strukturn am Bispil ds Raums Drsdn", Erknr: Lit Vrlag Brlin, (2006). [13] d Jong G, Fox J, Daly, Pitrs M, Vonk L: "udit of Car ownrship Modls", In: Transport Rsarch Cntr (VV) of th Nthrlands Ministry of Transport PWaWM, (d.): RND Europ, (2002). pp. 379-408. [14] StataCorp.: "Stata Statistical Softwar: Rlas 10", In: LP S, (d.). Collg Station, TX, (2007). [15] Dckl S: "Indikatorn dr Einkommnsvrtilung in Dutschland 2003", Statistischs Bundsamt Wirtschaft und Statistik (2006);2006(11): pp. 1178-1186. [16] Bohnt M: "Influnc of Land us and Transportation Systm on Car Ownrship", CUPUM 07-10th Intrnational Confrnc on Computrs in Urban Planning and Urban Managmnt,(2007), Iguacu, Brazil (School of Enginring of São Carlos), CD ROM. [17] Zumkllr D, Chlond B, Ottmann P, Kuhnimhof T, Kagrbaur M: "Panlauswrtung 2005 - Datnaufbritung, Plausibilisirung, rst uswr-tungn zu dn Erhbungn zur lltagsmobilität 2004/06 sowi zu Fahrlistungn und Tribstoff-vrbräuchn 2005/07 für das Mobilitätspanl - Zwischnbricht", In: Univrsität Karlsruh IfV, (d.). Karlsruh, (2007). [18] Bohnt M, Gutsch J-M, Mnz : "Standortabhängigkit dr Vrkhrswirkung von Projktn ds Wohnungsbaus und ds Einzlhandls. Möglichkitn ds Einsatzs von Vrkhrsmodlln bi dr Standort-bwrtung für Einzlvorhabn", In: Logistics ECfTa, (d.) ECTL Working Papr. Hamburg: Tchnisch Univrsität Hamburg- Harburg, (2006). [19] Ciari F, Balmr M, xhausn KW: "Mobility Tool Ownrship and Mod Choic Dcision Procsss in Multi-gnt Transportation Simulation", 7th Swiss Transport Rsarch Confrnc. Mont Vrità, scona, (2007). [20] Prillwitz J, Harms S, Lanzndorf M: "Impact of Lif Cours Evnts on Car Ownrship", Transportation Rsarch Rcord (2006);1985: pp. 71-77.