Influence Of Neighborhood Forms On Household Energy Consumption/CO₂Emission And Travel Behavior Model

Transportation energy consumption and emission control is essential in urban airquality improvement. Household residential energy consumptions and CO₂emissions ofdifferent neighborhood forms are investigated in this study. A model is established toestimate the effects of urban form on the travel mode choice, which explores the basicmechanism among urban form, travel mode choice and transportation energy/emission.This study provides additional empirical evidence on the relationships between land useand travel behavior.Research data are collected by households’ questionnaire surveys in2009and2010in Jinan. By contrasting the different levels of household energy use and CO₂emissionin different neighborhood forms, the influence of socioeconomic factors and formcharacteristicson household operational and transportation emissions is analyzed.Compared with traditional, grid, enclave neighborhoods, the superblocks has the highesthousehold energy use and emission rate, and its ratio of transportation consumption isalso the highest among the four differentneighborhoodcategories. The household energydecreases with lower household income, smallerhousehold size, more compact, land usemixed and BRT providing.Travel mode choice determines the residents’ commute transportation energyconsumption. Based on the discrete choice model theory, Multinomial Logit Model（MNL） is established to fit the commuter’s mode choice behavior in Jinan23neighborhoods. The model specific utility function is established firstly. Then thepropermodel variables are selected byanalyzing bivariate correlation between modelparameters. To improve the explanatory power of the model, the selected variables aretransformed in different ways. A base model including travel parameters andindividual/household socioeconomic characteristics and a expand model adding withneighborhood form variables are both presented in this study. According to the basemodel results, the travelers with characteristics including middle-aged, male, more thantwo household members are employed and more vehicles per person have a higherprobability to drive to their destinations. The goodness-of-fit of expand model improvedcompared with the base model. Those who live with higher percentage of frontages to be lined with building street wall are less likely to drive to work. Residents in higherdensity neighborhoods are more likely to take bus. A downtown-bound worker is morelikely to drive than a suburb-bound worker. Amount of bus stopsnearneighborhoods isnot a significant factor in the commuter’s decision to take mid-carbon modes to work.Finally, this study tests the outcomes of the model fitting and makes a leave-one-outcross-validation. Accuracy rate of the model fitting results and cross-validation is0.74,0.52respectively. Besides, the transportation energy consumption per capita wascalculated by the model results, which matches the actual consumption levels.