The Climate Action Plan subcommittee meets on Monday the 6th for their bi-monthly session. The change in transportation analysis remains interesting.
 |
| March agenda |
 |
| In yesterday's paper |
Transportation Modeling
In the packet is final report on the new emissions inventory on 2021 data, and as we noted in January on the video presentation, it has some important analytical changes. One of them is on transportation.
 |
| Revised pie chart on transportation (yellow highlighting added) |
About the change, which purports to use empirical data rather than computer modeling, they say:
The 2021 data used in this report were based on...Google's proprietary location history data....The source for local VMT data in 2016 was the Salem-Keizer Metropolitan Area Travel Demand Model. This model provides a computer-generated estimation of vehicle trips and does not use on-the-ground measured data.
Here's the 2023 work plan for the MPO on modeling, first the summary of previous work, then the projects for 2023. If the Google data is so much more accurate, then the MPO really needs to discuss that explicitly and to fold it into their own modeling.
 |
| On previous work |
 |
| On current work |
This may seem like a minor point, but so much of our current planning depends on the SKATS travel demand model, and if there's a real disconnect with reality and actual travel, we are making planning decisions on bad data.
Here for example is a clip from the recent traffic study for the Kuebler Village project just east of the new Costco. It depends on modeled output for 2017, 2037, and 2043.
 |
| Dependence on SKATS travel demand model September 2022 study for Kuebler Village (dark red text and yellow highlighting added) |
If the travel demand model varies so far from actual travel behavior, then we need to look into that more explicitly, asking the MPO to refine the model, and also asking about ways the City and other entities use data from that model. We are looking at something of a garbage in/garbage out problem! Maybe it's minor, within reasonable tolerance for any model; but maybe it's not, and requires a more thorough revision of our approach to modeling.
As our 350.org chapter pointed out, the other significant change in the 2021 inventory is a revised assessment in methane emissions, with a very substantial increase from home heating and cooking with gas.
Not Engaged on Actions with High Reduction Potential
Next on the agenda, and in the packet, is a "complete list of strategies with high GHG reduction potential." It's not very long! There are eleven total. Just over half of them, six, are "not yet reviewed"; four are "on the 2023 CAP Work Plan," at the moment more theoretical than actual; and only one is "underway."
 |
| "Complete list" on most effective actions (yellow added) |
In a holistic sense, we are not very engaged yet with these "high" potential actions.
Committee Framed as Passive and for Compliance Mainly
And in the work plan are four items on transportation, which are mostly compliance with new DLCD rules on climate and transportation. The City's own plan is not originating much policy action. Of those four items, on reduction potential one only is rated "high," two rate "medium," and one "low."
 |
| 2023 work plan on transportation |
The whole structure of the meeting remains pretty passive. The committee is not asked to make any decision or to undertake any action. They are merely to hear an update from others, who are presumed to be doing enough. There's not a lot of urgency.
2 comments:
I share your frustration, but again, I think you're overly harsh in your judgment. One of biggest challenges is that our process to change the built environment is-assess whether to study something, study, decide, plan, fund, design, permit, and build. Similarly, the process for policy is: assess whether to study something, study, vote, fund, implement. The tempo of this process is at a completely different scale than climate change now.
I see that the City has ten (one is a duplicate) high impact strategies. Two are underway, three are on the 2023 workplan, and five are under review. So, there is or will be action in 2023 on half. If you look at the list of low and medium priorities, they city is currently: updating the TSP plan (and transportation as we all know is the biggest driver of GHG), eliminating parking minimums, hire a climate action manager, increasing energy efficiency in municipal buildings, create a network of renewable energy microgrids (Salem received $1 million to pilot the city's first microgrid), promote energy Trust incentives for all electric building construction, create incentives to reduce solid food waste, and update the natural hazards mitigation plan.
I think what you dismiss as "talk" is all the thinking, studying, assessing, coordinating, and planning that leads to action.
There are two issues here: The difference in estimating VMT and the methods used to estimate emissions.
When establishing a base year, the SKATS travel demand model is validated to traffic counts from roads throughout SKATS, travel times from INRIX are used to check the speeds in the model for reasonableness, and ridership from Cherriots is used to check the transit values. The underlying equations for the model were estimated from the 2009/10 household activity survey. While the model has all the roads, the focus is on those roads with the higher functional classifications (e.g., not local roads). And the way demand is loaded onto the network for assignment means not all roads will have demand and the distance from the home to the major road may not be included in the VMT estimate. And this distance will depend on the size of the transportation analysis zone. (This would be clearer with a picture)
Google is collecting cellphone-based data from travel on all roads in the area, and then using their algorithms to assign travel to a mode. Part of their process seems to include inflating the sample collected to represent all travel (means unspecified). [For some details, see: https://insights.sustainability.google/methodology#transportation]. Theoretically we could use this data for validating our model, and other vendors do offer similar datasets (as far as I know Google does not). But it would be more useful for checking flows of demand between areas (origin-destination) than for VMT estimates.
For the climate work there are advantages to Salem in using the Google data: It offers a one-stop for emission estimates for transportation and buildings, the data is free, and it is available yearly. They also allow for different emission factors to be used in the calculation.
The disadvantages are primarily in the lack of any future forecasting and some opaqueness in how they go from data collected to result(s). It is also unknown how the results may change in the future as the machine learning tools keep ‘learning’.
Regarding the emission estimates: There are likely differences in the methodologies used by Google and the prior consultant. Google is using emission factors for each mode (car, motorcycle, etc.) based on estimates of distance traveled and average fuel use. I don’t know what the prior consultant did, but likely just applied a factor to the estimated VMT.
It would be interesting to find out how the actual fleet of vehicles and appropriate emission factors change the results. And that is another option with Google’s product – you can change the average vehicle efficiency and emissions factors assuming you have better data on the fleet. As a side note, DEQ is working on producing emission factors for each county by mode (e.g., motorcycle, automobile, etc.) and it might be possible to get from either DEQ or ODOT fleet info by county.
Ray
MWVCOG/SKATS
Post a Comment