| Test | What Changed | Geography |
|---|---|---|
| Base 2019 | No change — 2019 base land use | Small districts |
| HH ×2 | Households doubled in test districts | Small districts |
| HH ×4 | Households quadrupled in test districts | Small districts |
| HH ×12 | Households ×12 in test districts | Small districts |
| EMP ×2 | Employment doubled in test districts | Small districts |
| EMP ×4 | Employment quadrupled in test districts | Small districts |
| EMP ×12 | Employment ×12 in test districts | Small districts |
| HH+EMP ×2 | Both HH and EMP doubled | Small districts |
| HH+EMP ×4 | Both HH and EMP quadrupled | Small districts |
| HH+EMP ×12 | Both HH and EMP ×12 | Small districts |
| Future Today (2050) | 2050 forecast land use in test districts | Small districts |
| Future Today — Centerized | 2050 forecast, growth centered on transit corridors | Small districts |
| Future Today | 2050 forecast applied region-wide | Region-wide |
| Future Today — Centerized | 2050 centerized forecast applied region-wide | Region-wide |
Non-Motorized Travel Sensitivity Analysis
Wasatch Front Regional Travel Demand Model — Non-Motorized Enhancement
Prepared by Bill Hereth · last updated 2026-07-02
Purpose
This document summarizes the results of a structured sensitivity analysis conducted to evaluate how the Wasatch Front Regional Travel Demand Model (WF-TDM) responds to changes in land use density under its enhanced non-motorized travel component. Specifically, the tests examine how non-motorized mode share — the share of daily trips made on foot or by bicycle — changes as household and employment densities are increased in a localized set of small districts.
The intended audience is planners who use or interpret travel demand model outputs. The central question is simple: when land use becomes denser, does the model produce a credible and proportional increase in walking and biking? The answer has direct relevance for evaluating transit-oriented development proposals, infill housing, and employment center growth.
Key Findings
The model responds correctly to density. Non-motorized mode share increases consistently and monotonically as household and employment density increases — behavior that aligns with observed travel patterns and theoretical expectations.
Employment density is a more powerful driver of non-motorized travel than household density. At equal multipliers, employment-driven scenarios produce roughly twice the non-motorized mode share gain as household-driven scenarios. At ×4, the gap is nearly 12 percentage points (30% vs. 18%).
The response is non-linear with diminishing returns. Each successive density increase produces a smaller marginal gain in mode share. On a linear density scale, the slope is steep at low multipliers and flattens at higher ones — a concave curve consistent with how accessibility gains taper as destinations become increasingly reachable on foot.
Walking, not biking, drives the employment density response. In employment-density scenarios, virtually all non-motorized gains are in walk trips. Bike share remains relatively stable regardless of employment density. Both walk and bike increase with household density, but walk still dominates.
Home-Based-Other and Non-Home-Based trips are most sensitive. These purposes — which include midday errands, lunch trips, and non-work travel — respond most strongly to employment density, consistent with research on walkable urban environments.
The 2050 future scenarios represent a moderate, credible response. The 2050 land use forecasts produce a 3.8 percentage point gain in the test districts — a reasonable and interpretable result that sits in the lower-to-middle range of the sensitivity test spectrum.
Planning Implications
For evaluating transit-oriented development and mixed-use proposals: The model is well-suited to distinguish between the non-motorized impacts of residential vs. employment density. Proposals that concentrate employment — particularly office, retail, and mixed-use — will produce substantially larger walk trip forecasts than residential-only projects of similar scale.
For interpreting mode share forecasts: A 2–5 percentage point increase in non-motorized mode share at the district level represents a realistic and meaningful change consistent with the 2050 land use forecast. Scenarios projecting larger gains would require density levels far beyond what the regional forecast envisions — in the range of 4× to 12× current land use — and should be treated with appropriate scrutiny.
For understanding regional vs. local impacts: Non-motorized mode share gains from land use intensification are inherently local in nature. Even large changes in specific districts produce modest region-wide shifts because the majority of the region’s trips originate in lower-density areas. Planners should evaluate non-motorized outcomes at the sub-regional or district level rather than relying on regional aggregates.
For model confidence: The non-motorized component of the WF-TDM produces directionally correct, proportional, and purpose-differentiated responses to land use change. The sensitivity tests do not reveal any unexpected discontinuities or implausible outcomes, supporting confidence in the model for evaluating non-motorized travel in planning applications.
How the Tests Were Designed
Thirteen test scenarios were constructed, each modifying the 2019 base-year socioeconomic (SE) data — the land use inputs to the model — in a controlled way. A fixed set of five small districts (278 TAZs) in varying locations across the region was selected as the test geography. These districts represent a cross-section of urban contexts.
The tests fall into three groups:
Multiplier tests (Tests 1–9): Household counts, employment counts, or both are multiplied by 2×, 4×, or 12× within the test districts. These artificial but extreme values are intentional — they stress-test the model’s behavior across a wide range of densities, well beyond what any real scenario would produce.
Future scenario tests (Tests 10–13): Instead of applying multipliers, the 2050 fiscally-constrained land use forecast is substituted into the test districts (Tests 10–11) or applied region-wide (Tests 12–13). Test 11 and 13 use a “centerized” variant that concentrates 2050 growth along transit corridors.
All results are measured as daily trip productions at the non-motorized mode share level — the percentage of all trips that are walked or biked — compared to the unmodified 2019 base.
Results
Sensitivity by Trip Purpose
Not all trip types respond equally to density. The chart below shows how each purpose responds across the three scenario types.
Notable patterns by purpose:
- Home-Based-Other (HBO) and Non-Home-Based (NHB) trips show the strongest response to employment density. These trip types include lunch outings, short errands, and midday travel — exactly the kinds of trips that become walkable in high-employment environments.
- Home–Work (HBW) trips respond more modestly and more evenly across both household and employment increases. Commute trips have longer distances on average and are more constrained to auto travel even in dense areas.
Walk vs. Bike: Who Drives the Gains?
Non-motorized travel is not a single mode. The chart below separates the change in mode share into its walking and biking components.
The key finding here is stark: virtually all of the non-motorized gains in employment-density scenarios are driven by walking, not biking. At 12× employment density, the walk share rises by roughly 36 percentage points while bike share barely moves (+0.3 percentage points). This reflects how employment density operates — it creates short, walkable trips to nearby destinations.
Household density scenarios tell a different story. At 12× households, bike share increases by about 2.5 percentage points, while walk share increases by approximately 10 percentage points. Denser residential areas generate more walking and more biking, but walking still dominates.
This has a practical implication: if a planning scenario involves predominantly residential infill, the model will show meaningful but modest non-motorized gains, with some biking component. If it involves employment intensification — office development, mixed-use density — the model will show large walk trip increases. Both outcomes are directionally consistent with what travel behavior research shows for those land use types.
How Do Future 2050 Scenarios Compare?
The final chart places the multiplier sensitivity range in context by comparing it against the 2050 forecasted land use scenarios — both at the local (test district) and region-wide scale.
Before looking at the mode-share result, it helps to see how much the underlying land use actually changed between 2019 and the 2050 Future Today forecast, at each scale:
| Scenario | Local HH Growth | Local EMP Growth | Region HH Growth | Region EMP Growth |
|---|---|---|---|---|
| Base 2019 | +0% | +0% | +0% | +0% |
| Future Today | +108% | +61% | +76% | +46% |
| Future Today - Centerized | +118% | +61% | +76% | +46% |
The test districts see much more growth than the region as a whole: households roughly double locally (+108%) versus +76% region-wide, and employment grows +61% locally versus +46% region-wide — expected, since these districts were chosen because they see concentrated future growth. Centerization adds a further +10 percentage points of local household growth (+118% vs. +108%) by redistributing where regional growth lands, without changing the region-wide totals at all — which is why the two Future Today variants come out identical at the region-wide scale but differ slightly at the local scale.
For context against the multiplier sensitivity tests above: local household growth here (~2.1×) lands close to the ×2 multiplier tested, while local employment growth (~1.6×) falls between the ×1 and ×2 tests. Readers can use this to judge whether the non-motorized share gains below (+3.3 points locally, +0.8 points region-wide) look proportional to what the sensitivity range would predict for land-use change of this magnitude.
At the test district scale (local): the 2050 future land use scenario produces a non-motorized mode share of approximately 19.4% in the test districts, up from 16.1% in the 2019 base — a gain of roughly 3.3 percentage points.
Comparing this to the multiplier sensitivity range, the 2050 local forecast falls in a range roughly comparable to a 2× household or employment increase — a moderate, realistic result.
At the region-wide scale: the 2050 scenario shows a more modest increase — from 11.0% to 11.8%, a gain of roughly 0.8 percentage points.
Geographic Effect of Centerization
The centerized variant does not change total regional growth — it redistributes households from outlying suburbs toward urban cores while keeping region-wide totals nearly identical. At the medium district level, this redistribution produces a clear and consistent relationship: districts gaining density gain Non-Motorized share; districts losing density lose it.
Ogden Core — the district gaining the most households (+13,500) — sees a +3.0 percentage point increase in non-motorized share, rising from 12.1% to 15.1%. West Valley/Taylorsville (+9,400 HH) gains +0.4 percentage points; the Clinton/West Point/Clearfield area (+9,100 HH) gains +0.8 percentage points. On the other end, North Ogden (−6,600 HH) drops 0.5 percentage points and Layton (−2,900 HH) drops 0.3 percentage points.
This confirms the model is behaving consistently: the same density-to-Non-Motorized-share relationship observed in the multiplier tests carries through to spatial redistribution in the 2050 scenarios. Where growth concentrates, walking and biking increase — whether that growth comes from a direct density multiplier or a shift in where 2050 households locate.