Factors influencing urban on-street parking search time using a multilevel modelling approach

Vehicles searching for on-street parking create environmental and economic externalities through increasing network traffic flow and congestion, heightening pollutant emission levels, creating additional noise, giving rise to time delays for through vehicles, and leading to potential safety hazards caused by vehicles manoeuvring into or out of on-street spaces. Despite extensive negative impacts on individual drivers and on society, parking search is an under-researched area, particularly in more recent years and within the UK. Furthermore, current statistical modelling techniques applied to parking search time have not utilised a more comprehensive analysis in which hierarchically structured data on multiple levels could be addressed. The aim of this thesis, therefore, is to investigate and compare the factors that influence drivers urban on-street parking search time and its policy implications. A mixed methods approach was applied that comprised qualitative interviews conducted with local government authority Council Officers and a quantitative revealed preference on-street parking survey (sample size, 1,002 observations) undertaken in four cities in the East Midlands region of the UK in order to obtain individual driver-level socio-economic and other parking related factors that may influence parking search time. Statistically significant variables for each of the cities were identified by employing separate linear regression models. A multilevel mixed-effects model in which drivers (Level 1) are nested within streets (Level 2) was then applied to the pooled dataset. Significant factors in the multilevel (street level) model were identified as: time of arrival at a parking place (for which every time period after the 07:00-07:59 reference case indicated increased search time); parking habit; parking tariff; the number of parking places previously visited (on the same trip); trip time from origin to parking place; area type; trip purpose; weather; vehicle type; and walking time from a parking place to a destination. Comparison of the factors that influence parking search time revealed important differences in statistically significant variables and coefficient values between the single-level and multilevel regression modelling approaches. Policy recommendations based upon the findings of the parking survey, modelling analysis, and further interviews conducted with local authority Council Officers, focus around time of arrival at a parking place, area type, parking charges and the potential technological advances that, if implemented, could have a considerable effect on parking search times within urban areas. Robust data collection and subsequent monitoring of parking search activity within each city should be undertaken in order to provide an evidence base which would support the introduction of future policy measures to reduce parking search activity.