California Stewardship Fund • AB 747 / Gov. Code §65302.15 • v3.4 ΔT Standard • March 2026
For City Councils, City Managers & City Attorneys
JOSH is a first-principles calculator built entirely from established, credentialed expert standards and publicly verified data. Every number it produces can be reproduced by any licensed engineer with a spreadsheet. The software offers one thing: speed — turning a calculation that once took weeks into an instant result with a documented, hand-auditable audit trail.
The methodology is not novel. Every input is published. Every threshold is derived from arithmetic. The city does not adopt a JOSH opinion — it adopts the standards JOSH mechanically applies.
That distinction is the foundation of appellate survivability.
One rule governs fire marshal occupancy limits in every occupied building in California. JOSH applies the same rule to the roads those buildings depend on.
"You cannot put more people behind an exit than the exit can handle in an emergency. Measure the exits. Count the people. Do the math."
— The same principle a fire marshal applies when posting a Maximum Occupancy sign, applied at city scale to evacuation roads.
A fire marshal does not ask who is already in the nightclub before deciding whether 50 more people can safely enter. The question is whether those 50 people can get through the doors fast enough in an emergency. The answer depends on the doors and the 50 people. It does not depend on who else is in the room.
A wildfire evacuation road is a door. A housing project is a group of people asking to drive through it. JOSH measures whether those people can get through in time. The road either has the capacity or it does not. That is a physical fact — measured in minutes, from published data, using arithmetic.
Why this principle is legally durable:
Fire marshal occupancy limits are the most litigation-tested life safety standards in American law. Courts have upheld them uniformly because they rest on a physical reality: exits have a fixed throughput; people have a fixed size; time is time.
JOSH applies the same logic at city scale. The inputs change — roads instead of corridors, vehicles instead of people on foot, fire hazard zones instead of kitchen fires — but the structure of the argument is identical, and it has been accepted by courts for over a century.
After the 2018 Camp Fire killed 85 people in Paradise, California, the National Institute of Standards and Technology documented, minute by minute, exactly how long residents had before the fire front arrived. That timeline is now the empirical foundation for JOSH's escape-window thresholds. There is no opinion in that number — it is a federal agency's post-disaster forensic finding.
The California Legislature recognized this when it passed AB 747 in 2019, placing evacuation capacity analysis in the Safety Element — not CEQA, not the Circulation Element. It is a life safety question. JOSH implements what the Legislature directed.
JOSH was designed for appellate courts before it was designed for planning departments. Every architectural decision traces back to one question: can we defend this in front of a judge who has never seen a traffic study?
Every standard JOSH applies existed before JOSH was written. No input was invented by the city, by California Stewardship Fund, or by any consultant. The methodology is an assembly of published, credentialed expert sources:
Every number JOSH uses is available to the public, free of charge, from the agency that published it. A developer's engineer, a plaintiff's attorney, or a judge's clerk can pull the same source document and verify every input.
No proprietary model. No black-box weighting. No expert opinion that can be cross-examined. The data speaks for itself because it already has — in every state DOT, in every building department, in every fire marshal's office in the country.
Every determination JOSH produces comes with a plain-text audit trail. The full calculation — every input, every intermediate result, every threshold comparison — fits on one printed page.
A city attorney preparing for litigation can read the audit trail and explain every line to a judge without calling an expert witness. The calculation is four arithmetic operations. There is nothing to challenge that cannot be checked with a calculator in under ten minutes.
One number: how many minutes does this project add to evacuation clearance time? Every step uses only published national standards or federal data. No step involves anyone's opinion.
Multiply the project's housing units by vehicles per household (Census ACS), then apply the 90% mobilization rate from NFPA 101 — the same design basis used to size stairwells in every building in the country. NFPA 101 assumes 100% occupant evacuation; the 10% reduction accounts for zero-vehicle households measured directly by Census.
Sources: U.S. Census ACS B25044 • NFPA 101 Life Safety Code
Examine every road between the project and safety. Each road's capacity is taken from the Highway Capacity Manual (HCM 2022) — the national standard used by every state DOT. If a road passes through a Cal Fire Fire Hazard Severity Zone, its capacity is reduced by the HCM's own published factors for smoke and lane blockage. The road with the lowest effective capacity is the bottleneck.
Sources: HCM 2022 Exhibits 10-15, 10-17 • Cal Fire FHSZ (Gov. Code §51175)
A family on the sixth floor must walk down six flights of stairs before reaching a car. NFPA 101 and the International Building Code provide published methods to calculate this delay: 1.5 minutes per story for buildings over four stories, capped at 12 minutes. Single-family and low-rise multifamily projects receive no penalty.
Sources: NFPA 101 • International Building Code
Divide project vehicles by bottleneck capacity. Convert to minutes. Add any egress penalty. Compare the result to the zone-specific threshold: 5% of the escape window documented by NIST for the project's fire hazard zone. That percentage is the one policy value the city adopts — everything else is arithmetic from published sources.
Sources: NIST TN 2135 (escape windows) • City resolution (5% share)
Step 1 — Project vehicles: 45 units × 2.5 veh/HH × 90% = 101 vehicles Step 2 — Bottleneck capacity: 1,350 vph (two-lane, 30 mph) × 35% (VHFHSZ) = 472 vph Step 3 — Egress penalty: 3-story building → 0 min (below 4-story threshold) Step 4 — ΔT: (101 ÷ 472) × 60 + 0 = 12.8 minutes Threshold (VHFHSZ): 45 min × 5% = 2.25 minutes
No input was chosen by the city, by California Stewardship Fund, or by any consultant. Every value is traceable to a specific exhibit, table, or dataset published by a credentialed national authority.
| Standard / Dataset | Publisher | What JOSH Uses It For | JOSH Parameter |
|---|---|---|---|
|
Highway Capacity Manual, 7th Ed. (2022)
Exhibits 10-15, 10-17, 12-6, 12-7, Ch. 15
|
Transportation Research Board National Academies of Sciences |
Road capacity (pc/h/lane) by road type and speed; capacity reduction factors for smoke and incidents | Road capacity & degradation |
|
NFPA 101 Life Safety Code
Current edition
|
National Fire Protection Association | 90% mobilization rate (design basis: full building evacuation); building egress time calculation for structures ≥ 4 stories | Mobilization & egress penalty |
|
NIST Technical Note 2135 (2021)
Camp Fire Minute-by-Minute Timeline
|
National Institute of Standards and Technology U.S. Department of Commerce |
Safe egress windows by hazard zone: VHFHSZ = 45 min, High = 90 min, Moderate/Non = 120 min | Escape-window thresholds |
|
Fire Hazard Severity Zone (FHSZ) Maps
Gov. Code §51175–51189
|
Cal Fire (State of California) Office of the State Fire Marshal |
Determines which road segments receive capacity degradation; sets escape-window zone for the project site | Road degradation zone |
|
American Community Survey 5-Year Estimates
Tables B25001, B25044
|
U.S. Census Bureau U.S. Department of Commerce |
Vehicles per household (2.5 California average); zero-vehicle household rate (basis for 10% non-mobilizing share) | Vehicle generation rate |
| International Building Code (IBC) | International Code Council | Egress capacity and stairwell requirements for buildings ≥ 4 stories; minimum exit width standards | Egress penalty (tall buildings) |
| OpenStreetMap Road Network | OpenStreetMap Foundation (public domain) | Road geometry, classification, and connectivity for bottleneck routing within 0.5-mile project radius | Route identification |
JOSH does not add methodology. It adds speed, consistency, and an interface that makes near-instant insight available to planning staff, city attorneys, and council members alike.
The Berkeley demo shows JOSH applied to six representative projects — from a 10-unit infill to an 80-unit hillside development in a Very High Fire Hazard Zone. Each project receives a determination in seconds, with a full brief and hand-auditable audit trail.
Cities can adopt JOSH at three levels of formality. Each tier strengthens legal defensibility. Tier 1 requires no action — a city can use JOSH to support a safety finding in a staff report today under its existing General Plan.
A planner runs JOSH and cites the ΔT result in a staff report as evidentiary support for a safety finding under the Housing Accountability Act (Gov. Code §65589.5(j)(1)). The calculation is substantive evidence of a quantifiable, direct, unavoidable safety impact grounded in objective written standards.
Best for: immediate use while council adoption is pending. Defensibility is strong but not as strong as Tier 2.
Council adopts the ΔT standard by resolution as an objective development standard. This converts the methodology from ad hoc evidentiary support into a pre-adopted, legislatively enacted standard — the exact instrument courts give the highest deference under the HAA. Model resolution language and recitals are provided in the Legal Defensibility Memo.
Recommended first step. Protects prior determinations. Establishes uniform application going forward.
The ΔT standard is incorporated into the city's Safety Element as required by AB 747 (Gov. Code §65302.15). This is the highest level of adoption — a General Plan policy carries the full weight of legislative enactment and cannot be challenged as ad hoc or pretextual. Typically done as part of a scheduled General Plan update.
Strongest legal protection. Required by AB 747 for the next Safety Element update cycle.
Every document below is written for a specific audience. City attorneys reviewing for the first time should start with the plain-language explanation. Staff preparing council materials should start with the Legal Defensibility Memo.
How the evacuation capacity standard works, written for judicial review. Uses the fire marshal nightclub analogy. Walks through all four steps with a concrete example. Explains why this is a life safety question, not a traffic question. Includes the full source documentation table.
Complete legal framework for adoption. Documents how the ΔT standard satisfies all five HAA §65589.5(j) requirements: significant, quantifiable, direct, unavoidable, and objective. Includes three-tier adoption pathway, model resolution language, and responses to anticipated developer objections.
Full technical justification for the ΔT system. Every parameter is traced to its source publication and exhibit number. Covers HCM capacity tables, hazard degradation compositing, NIST escape window methodology, bottleneck selection algorithm, and egress penalty derivation. Written for peer review.
Complete technical specification for the JOSH v3.4 ΔT Standard implementation. Covers all five standards (size gate, routing, hazard zone classification, ΔT calculation, SB 79 transit flag), data flows, scenario architecture, configuration parameters, and audit trail format. Source of truth for all code.
JOSH is a Python command-line tool, not an AI system. Every determination it produces is the output of four arithmetic operations on published data. The code is open source under AGPL-3.0 so any city, engineer, or developer can inspect, verify, and run the calculation independently.
JOSH downloads data automatically from four public sources — Cal Fire, OpenStreetMap, the U.S. Census Bureau, and the city boundary from Census TIGER — and applies the HCM 2022 capacity table, NFPA 101 mobilization design basis, and NIST Camp Fire timeline as a pipeline of deterministic calculations. There is no machine learning, no model training, and no probabilistic inference. The software is an automation layer over arithmetic.
The pipeline has three stages: data acquisition (Stage 1 downloads and caches all inputs), capacity analysis (Stage 2 builds the road network, applies hazard degradation, and identifies evacuation route bottlenecks), and the objective standards engine (Stage 3 runs the ΔT test and generates the determination and audit trail). Every stage reads from config files that expose every parameter, making the full calculation transparent to any reviewer.
Because the methodology does not belong to JOSH — it belongs to HCM, NFPA, NIST, and Cal Fire — a city that stops using the software retains every legal position it established. The determination letters cite the standards, not the software.
data/{city}/