01697666

Component

Incidents occur regularly on urban metro systems, usually causing delays in the operation, which hence reduces the service level and can also bring safety problems of crowded passengers. Therefore, it is necessary for metro operators to understand the occurrence of disruptions and analyse their impact to prepare better recovery plans. This paper presents a propensity score matching (PSM) framework to estimate the disruption impact on London Underground. Using smart card data and other supplementary datasets the authors demonstrate that disruption occurrence is not random. Based on this, the impact is measured at multiple aggregation levels. From the system perspective, disruptions will increase 37.4% of the exit ridership and 1.0% of the average journey time, meanwhile reducing 5.0% of the entry ridership and 2.4% of the average travel speed. At station level, the disruption impact varies from station to station. Spatially, the stations with the most affected ridership are gathered in central London areas, while stations that locate in outer London areas experience larger changes on average journey time and speed. Additionally, the authors build second-stage regression models to identify determinants of the station-level impact. Factors such as daily ridership, number of surrounding bus stops and area of non-domestic buildings around metro stations are found to be significant for the disruption impact on travel demand.

This paper was sponsored by TRB committee AP000 Public Transportation Group.

19-03762

English

5p

2019

Transportation Research Board 98th Annual Meeting

Digital/other

References

Rail transit stations; Rapid transit; Ridership; Service disruption; Smart cards; Travel time; Urban transit

London Underground

Propensity scores

London (England)

Planning and Forecasting; Public Transportation; Railroads; Security and Emergencies; Terminals and Facilities

Transportation Research Board Annual Meeting 2019 Paper #19-03762

TRIS, TRB, ATRI

Dec 7 2018 9:34AM