Developing Trip Production and Mode Choice Model for Home Based Trips: The Case of Hawelti Sub-City Residents, Mekelle - Ethiopia

Abstract

Hawelti sub City considered one of the most rapid increases in population and crowded subcity in Mekelle city. Hawelti sub-city does not have a transportation model; hence, developing a model is considered very important for predicting the average daily trips and use of each mode. The aim of this study is to develop trip production and mode choice model for home-based trips in Hawelti sub-city residents. The trip production includes a trip for work, education, shopping, social and recreation whereas the mode choice focuses on work and education trips. A household interview survey was used as a data source in addition to secondary data obtained from the municipality of Hawelti Sub-City. Trip production was analyzed using multiple linear regressions and multinomial logistic regression was utilized in the analyses of mode choice with statistical data processing software SPSS. The modes of transport considered in the model are Walking, Bajaj, Private car, Minibus Taxi, and Service (Company or school provided transport). Minibus taxi is found to dominate with 44% share followed by the company provided service with a 22.8% share for all trips. In the developed model the factors that affect the trip production are the number of students, employed persons, monthly income, the number of females and the age above 45 in the household. Moreover, the factors that significantly affect the choice of transport modes for employees were total travel time, work type and family size. The mode choice model developed for the case of students the significant factors were found to be travel time and gender. The developed models were significant at 95% confidence level and exhibits a good fit for the data with the coefficient of determination (R2 ) of 0.889 for the general trip production and the adjusted goodness of fit measure statistics rho square (ρ2) of 0.823 and 0.824 for the work and student models respectively for the whole study area. The models are validated to be capable of predicting 90.97% and 87.2% of the preference of trip makers for the work and education mode choice models respectively.