The “3-1-3” service was designed so there’s 3 hourly roundtrip trains in the morning, one midday roundtrip, and 3 hourly roundtrips in the evening between western Durham and Clayton. That was requested by Johnston County to look at cheaper ways to serve Clayton as opposed to the default 8-2-8-2 service (trains every 30min in four-hour periods in the morning and afternoon, and then two roundtrips 2hrs apart midday and in the evening).
Because many of us have asked for an all-day regional rail service instead of something commuter-focused. So naturally, GoTriangle’s calling our bluff and seeing whether we really prefer it over the alternative.
Obviously, the best thing we could have is a train that runs frequently and all day. But we only have a limited budget, so we have to make an informed decision on whether to lean one way or the other. This presentation is one step towards us (well, GoTriangle’s appointed leaders) making that decision by having more information.
This could certainly explain the increase being small, and could theoretically (though not plausibly) explain zero increase. Extending some of the trains to Clayton, resulting in 3-1-3-0 service at the end of the line, is pretty poor service- most people would probably choose to use the park and ride.
But in my opinion this cannot possibly explain a negative change as shown on that chart. It doesn’t make sense at all.
True, a ridership decrease from a service extension’s weird. …if there’s a meaningful decrease there in the first place.
The simulations suggest a drop of about 200 passengers (about 1.8% of 12,033) when the suggested margin of error is about 5x of that (1000 passengers, or 8.3%). Without using more proper techniques of error analysis that I highly doubt GoTriangle leaders would think of – let alone understand or request – I think it’s fair to assume that decrease is nothing more than a statistical fluke.
Yeah, this was surprising to me. However, the second slide deck, titled Demographic Analysis of Trip Production Zones, notes the following:
This analysis is not a forecast of the demographics of future Commuter Rail riders
This is a description of the current demographics of the home locations of forecast trips on the Commuter Rail
The data is stratified by vehicle ownership and geographic area
Commuter Rail riders from a particular geography are assumed to have consistent demographics with that area
The demographics of a geographic area are assumed to remain constant into the future, consistent with the 2050 MTP
The analysis assumes that transit/rail use among different demographic groups is correlated with household vehicle ownership
So, if I’m understanding this correctly, these ridership forecasts assume zero changes in demographics, vehicle ownership, or (and this is key) development along the corridor. In other words, we could theoretically blow these projections out of the water with a) a shift away from car-dependency and b) transit-oriented development.
We’re already (hopefully) in the early stages of a shift away from car-dependency here in the Triangle. In the last two or three years, we’ve seen a strong focus denser development and significant improvements in bicycle and pedestrian infrastructure. Our first bus rapid transit corridor should break ground this year, and, in theory, we should have at least three routes in service by the time commuter rail is finished.
As for transit-oriented development, the City of Raleigh already has the tools for it because of the bus rapid transit projects. It should not be very difficult to apply those same tools to commuter rail. The real question, of course, is if the other municipalities involved would follow suit. Most of them are promising to, but we won’t know for sure until we start seeing more construction sites.
All that said… I don’t know if either of those things have a big impact on a commuter-focused service. I’ve said it before and will say it again: COVID killed the 9-5 commute, probably forever. But consistent all-day service, even if it’s just hourly for parts of the day, could really be appealing to people who live in TOD that’s walking distance from a station. Would I rather see 20min service all day? Absolutely. But I think I’d take consistent all-day service over 15min headways during rush hour. I think that scales better in the long run.
My understanding is that all-day service would be only a marginal increase in cost, since both the 8282 and 30/60 scenarios require the same number of trainsets and operators (since they both have 30-min peak headways). Trains cost the same to buy whether they’re used for four hours or 12 hours a day; operators work 8-hour shifts, and you need two shifts to cover the full day, so why not have the trains and operators doing work midday?
Because waiting for one hour beats waiting for two hours, which is the alternative.
If we’re talking about statistical samplings of real world phenomena, an error bar like that certainly makes sense. Talking about confidence intervals for how well a ridership model actually predicts ridership - ok! I’m on board.
But an error bar when comparing one modeled scenario relative to another? WTF. Are these sort of ridership models non-deterministic? How could you even test your model if the same run with the same initial parameters doesn’t reliably yield the same result?
Admittedly I do not know much about this discipline, but it certainly gives credence to Jarrett Walker’s assertions that ridership modeling is a dark art, and furthermore, bullshit.
No - the alternative is to say eff it and drive. commuters are not going to put up with that. At some point it stops becoming commuter rail, if commuters can’t count on it to get to their jobs on time. Same reason I don’t currently take the bus. It would take me 1.5 hrs for a 30 minutes car commute.
TL/DR: I think they are (but not in the way you're thinking), and I think it's ...fine I guess... (click for more!)
I don’t work in transportation modeling (though I’m making new ways to model and measure the biomechanics of human tissues for my PhD research, so tomāto tomàto?) so someone like @atl_transplant may have a better answer, but I don’t think it’s deterministic. This study’s ridership model is based on the federally-standardized STOPS software, and you’re supposed to input local transportation models like the Triangle Regional Model (TRM) into it. Since you’re supposed to run the TRM iteratively until it converges on some solution, that makes me think it’s stochastic -which would make the entire ridership model a stochastic system.
It’s okay to have slightly different numbers every time you run the model as long as your model can behave in “predictable” ways, in two senses of the word:
You have noisy result but the noise is reproducible. You could do this by having a record of the random initial conditions where you start running your models, or by having control over how noise is introduced into your results.
Your results don’t wildly change given small changes in your input. As a dumb example, if you added 50 more Garner-based workers into the TRM’s census data input, you should not get a 7000% increase off-peak train passengers every night.
You can still compare two noisy modeled scenarios (predictions) against each other. It could be harder to tell how much you’re wrong since the error of each scenario would propagate, but you can still look at correlations or do hypothesis tests that way.
I read your concern to mainly be about the first point (random model outputs) and my understanding of the TRM’s documentations is that that’s not a problem. The second point, though, is what raises yellow flags in my mind -and GoTriangle acknowledges this in their memo, as well:
I think the problem here is that the consultants, GoTriangle, Norfolk Southern etc. are all doing awful jobs at presenting their data. Not that they’re uniquely incompetent, though; it’s really hard to talk about what your data means without butchering your conclusion -so hard that even researchers at the National Cancer Institute struggled with and felt the need to publish a workbook for grownups on.
This is why, in statistics and instrumentation, we like to say that all models are wrong, but some are useful. Your results can be mired in bullshit, but there can be gold hidden inside if you know what you’re doing.
Yeah, that’s how I read that too. I think it’s smart, then, that they’re sticking to present-day demographics since that helps us get a more conservative and grounded estimate for our return on investment.
And if 12,000 is a conservative estimate, I think we’re in pretty good shape: at that number, we’d be ranked 15th overall in daily commuter rail ridership (assuming Q4 2019 numbers), beating out services like the South Shore Line, SunRail, and the Keystone Service. And, looking at this list, I think that’d make us the smallest population in the top fifteen by a rather significant margin.
I’m going to be really frustrated if this gets whittled down to crappy service or killed altogether. The fact that this corridor is bookended by the second and fourth largest cities in NC while also serving the seventh largest makes it such a no-brainer for rail transit. There are very few unserved corridors in the US with this much potential. This is such an easy win if they do it right.
Amen !! This project can be such a game changer for the region. The dense development from Raleigh to Cary could be continuous. This would give Morrisville a purpose lol. It literally could be TRANSFORMATIVE
Edit: I love Durham, and I’d love for Durham to have a new station and better rail alignment. The infrastructure challenges do seem deep for Durham tho .
In so many respects, Durham is the both the outlier and the most critical piece. This project doesn’t make sense without Durham, but it’s also so difficult for them both logistically and politically.
What I’m hoping officials in Durham see is what you’re alluding to in your first paragraph: this transforms the entire region. Everyone wins if this is paired with proper TOD. You’re building little walkable hubs along a regional spine that attract jobs and housing alike. That slows gentrification creep into neighborhoods like East Durham and concentrates more development in specific areas.
This is less about getting cars off of I-40 and more about what we want the future of the Triangle to look like. We saw with I-540 that development follows transportation investment. Nationally, we see similar trends with transit investment, but, instead of inducing sprawl, good transit induces dense, walkable, sustainable development. This idea is politically challenging because the average person doesn’t think about development nearly as much as we do here, and Durham has every right to be cynical (especially after DOLRT), but they have to start thinking about the future of this region as a whole.
I know I’m preaching to the choir (mostly) at this point, but I keep beating this drum because it’s so critical for long-standing residents to understand: people are moving to the Triangle region regardless of what we do. Our job, then, is to figure out where to put all of them. Do we let suburban sprawl creep further into neighboring towns and counties while allowing Raleigh and Durham to slowly become exclusive playgrounds for the wealthy, or do we find ways to fit more people into a smaller amount of space? Commuter rail, paired with TOD, is one of the best tools we have to ensure the latter.
The commuter rail section of the draft’s video summary didn’t mention any percentages for how much funding Durham will pour into that project, though. It seems like the idea is to finish up the ongoing feasibility study (whose byproducts include more realistic cost estimates) and then open it up to further discussions.
It’s not clear what that 27% number means, though, since we only know from the website that the plan would make use of “over $1 billion” in sales tax revenue over 20 years. This plan could make Durham County better prepared to keep up with their share of the bills -but that’s also according to last year’s numbers. So it’s probably for the best that we don’t have a solid interpretation of what that number means, and that we wait to make the finances work until after we have better data.
Would you prefer they pull another DOLRT, where they just shove a project down everyone’s throats because they feel like it’s worth bulding -only for it to go up in flames for reasons that could’ve been prevented with better evidence? Would you prefer that GoTriangle lost all credibility to do regional transit by being bullish on this project, too, and be left with transit agencies that are divided for good?
I know you don’t mean that, but I’m certain other people on here see the endless studying thing more maliciously. Sweating these details matter a lot for a taxpayer-funded agency with a history of big, reckless promises.
