[bouldercouncilhotline] Hotline: TMP questions for Aug. 5 meeting

[cmosupport at bouldercolorado.gov]

Sender: Appelbaum, Matt

Colleagues – Here are some comments/questions/concerns regarding the TMP (transportation Master Plan) that will come before us at our Aug. 5 meeting.  Sorry I didn’t get these in earlier.
 
My issues pretty much all revolve around data – but not because I’m hung up on the details or we need greater precision necessarily, but because I think – as I’ll try to explain – that these data might have considerable significance to
 our assumptions and goals.  And in one case, regarding LOS (level of service), I think we’re completely on the wrong track.
 
Since packet pages change, I’m going to use the TMP’s own pagination.
 
Page 1-5:
I remain as uncertain as ever regarding the VMT chart.  Is this all VMT within the city limits, regardless of whether they are caused by residents or non-residents?  I assume that’s the case, but it might just be residents.  For folks just
 passing through, I assume that’s included too (lots of traffic on, for example, Foothills Pkwy, neither begins nor ends in Boulder, I would think)?  Since much of our TMP goals depends on reducing VMT, this seems important to me.  And for those trips that
 don’t begin/end in Boulder, we have much less ability to affect the mode choice, so I would think it would be helpful to have those measurements.
 
I would also note that this table does not match up with the data on P. 3-2, which also shows Annual VMT but yields different results.
 
Also on this page is the SOV Mode Share chart.  This clearly states that it shows data only for Boulder resident trips.  Given the very significant VMT from non-Boulder residents, do we have a similar chart for them?  Would it show improvement? 
 There is a note that three new measures will be created, including VMT per capita for residents and non-residents, which I think will be very helpful – although I’m not quite sure what the “capita” is for non-residents (trips that end in Boulder?), and it
 would be most helpful – if no doubt difficult – to hone in on the purpose of the trips that create that VMT.
 
Page 2-2:
Here the Work Mode Share is the issue.  Do students (CU, but others as well) going to/from school count as going to “work?”  If so, what % of the various totals does that compose?  For me, this is not just an interesting piece of data but
 a critical one.  Let’s say that half of the bike trips in the work mode share are from students.  Since students are essentially transit/bike/walk  dependent and perhaps maxed out already, then most if not all of the gains in those categories must come from
 non-students, which would require a much greater shift in those populations than suggested.
 
Another perennial issue of mine is how we seem to measure the work mode share.  I believe this is done via surveys, which are always taken during the very best time of year.  But we need year-round data, or else our goals of increasing
 the mode share of bike/walk/bus won’t yield the desired overall results even if the survey shows significant increases.  In other words, the shift in mode share might be very different during different times of the year, so that we can’t conclude that, say,
 a 10% increase in non-auto share reflected in the survey actually yields a 10% reduction in auto travel year-round, or perhaps anything close to that.
 
The same issues of course arise in the Boulder Transit Use chart on P.2 – 4 for exactly the same reasons.  And, finally, this entire question obviously then relates to the implications of the Proposed Mode Share Targets on P. 3-7.
 
Page 3-5:
This is my concern with LOS.  Now, it’s important to note that I’m not asking this question, as some have, in order to suggest that we should, or shouldn’t, aspire to make it easier or harder to get around town via car by increasing or
 decreasing congestion.  In fact, I happen to think it would be quite a mistake to increase congestion, at least until a very significant % of travelers have “good” alternative modes of transport, which most certainly isn’t the case now and won’t be for the
 foreseeable future (unless you live in the Broadway bubble).
 
It seems to me that LOS can be useful when examining a single intersection, and watching as it, say, progresses from LOS A to LOS E or F over time.  But using LOS on a citywide basis just makes no sense to me – although I realize that most
 cities do this, no doubt because it greatly understates the actual situation.
 
Here’s my problem – and I might very well be off-base here so some staff assistance would be very helpful.  Simply calculating LOS at each intersection and then treating them all equally to determine that, say, no more than 20% of them
 are at LOS F provides absolutely no indication of the relative traffic and delay at each such intersection.  So a number of lightly used intersections might be LOS A or B, while the intersections that carry the vast majority of traffic are at LOS F during
 peak hours.  And yet we seem to just average them together
to get what I must conclude is a meaningless result.
 
To perhaps better demonstrate the problem, let’s say that over time (as we’ve been doing), we add more signalized intersections on the periphery, where traffic is relatively light and LOS will likely be A or B most of the time.  Just doing
 that will improve our average citywide, perhaps allowing us to claim that, indeed, no more than 20% of roadways are at LOS F.  But since we obviously didn’t improve any of the congested roads via this technique, the outcome is absurd.
 
Certainly, as indicated in the report, obtaining information regarding delay for all users of all modes would be helpful – but only if weighted appropriately (and this effort might, if done right, also resolve the problem with LOS F indicating
 a potentially huge range of delay).  I think we actually already have one metric that is even more useful: trip time along various corridors.  Again, if (and only if) this is done correctly – at various times of the day and year – measuring actual travel time
 on key travel corridors via various modes could, I think, be quite helpful.
 
So
I’d ditch the 20% LOS F metric.  Again, not because I want more congestion, but because I simply have no idea what it’s measuring or how it could possibly be useful.  We’d return to this issue later when we have better metric(s) that
 provide more useful and accurate data.
 
--Matt