Great Western Coffee Shop

This talk of catchments set me wondering, because it doesn't seem right to assume that buses and train/tram stations have the same catchment size. A study for Bournemouth (https://www.bournemouth.gov.uk/councildemocratic/CouncilMeetings/CommitteeMeetings/environmentandeconomyoverviewandscrutinypanel/2018/09/25/agenda/appendix-high-level-study-on-light-rail-and-other-rapid-transit-systems-for-bournemouth.docx) supports this view, suggesting that whilst heavy rail or metro systems might have a catchment radius of 800m to 1.2km, for tram systems it is nearer 400m to 600m. Grahame's 250m radius may well be right for buses. In essence, the catchment increases in proportion to the quality (in all its aspects) of the service.

Comment is made about some local use made of Lelant Saltings, offering Lelant or the bus as alternatives. Perhaps this traffic which is within the catchment of both stations has gravitated to the Saltings because of the much better service and will revert to Lelant in the future?  I am noting the extra calls at Lelant are lunchtime and school-end time.

The Euclidean method of measuring a rail station’s catchment, a circle on a map, is widely discredited in most literature; the failure of it to allow for natural and manmade obstacles means a more detailed method is required for accurate results [14]. Landex and Hansen [14] recommend the Network Analyst approach of mapping the real walking routes along the street network to the station. They map the catchments for rail stations in Copenhagen using both methods to show how the Euclidean method’s inability to take rivers into consideration leads it to include areas with no actual access to the stations whereas the Network Analyst approach identifies these obstacles and makes the relevant changes to the catchment area.

The Euclidean method can often have a detour factor built into it e.g. even if people are willing to walk 1km the circle will have a 800m radius to take account of obstacles. Once again however Landex and Hansen [14] point out that a one size fits all detour factor is not useful because as the urban grain of areas differ so would the detour times. Ker and Ginn [3] point out that even in areas of high pedestrian permeability, e.g. inner city neighbourhoods, circular walking catchments still only reach 60% efficiency which drops as low as 10% in poorly connected, often outer suburban, areas.

Actual walking routes are rarely as direct as the Euclidean method assumes meaning that the actual catchment area is smaller than expected; if the actual walking route is a third longer than the Euclidean method uses then the actual catchment radius is 75% of the original assumption and the actual population could be 56% of that projected. Barton et al. [13] support the use of the Network Analyst approach to map the real routes travelled and as such better plan the location of public facilities including rail infrastructure.

NHS used to use Euclidean to estimate nearest GP surgery, did imply a lot of patients swimming the Thames to see the doctor.

So - how should we define the catchment area around a public transport access point?  How do we estimate abstraction of traffic from one access point by a proposed new one?  How do we correlate market penetration of an access point based on the frequency and desired destiantion of the services that do or will call there?

Euclid or rivers are the least of your problems, as the catchment size is very dependent upon the type of service provided. In fact, if you have a mix of destinations from a station, you will have different catchment sizes for each destination! Chippenham is likely to be a fair example of this people will drive quite a long way to catch a train to London, but for trips to Bath most travellers are likely to start within walking distance of the station.

Looking at Bridport, we have stations at Axminster, Crewkerne, Dorchester and Maiden Newton, all are around 25 minutes away.
Axminster has a half hourly bus service and Dorchester an hourly one, there is also an hourly service to Weymouth. Crewkerne can be reached by bus but, due to recent funding issues, you normally have to change at Beaminster( and again at Crewkerne for Yeovil). There are no late evening services.
The train service on the West of England line is hourly while that from Dorchester is half hourly.
Car parking is limited at Crewkerne by 10 am, while Axminster has a larger car park.
I would presume modelling of passenger flows from Bridport would need to take these factors into account.