MAGNOLIA CROSSING RESILIENCE - NATIVE SUMO MESOSCOPIC EDITION

Start with Magnolia-SUMO-brief.pdf and METHOD.md. The CSV contains all 143
closure/demand comparisons. Counts refer to one-way vehicle trips, not cars owned.
The proposed-home and car-use ranges are assumptions, not a buildout forecast.

These are native SUMO mesoscopic road queues. They include finite receiving
storage, signals, priorities, rerouting, origin waiting and unfinished trips.
They do not model a car physically blocking the interior of a junction. The
microscopic congestion stress tests remain rejected, not silently certified.
Reference traffic and generated fixed signals are uncalibrated.

The 143 cases are parameter scenarios using one main seed, not 143 independent
replications or a probability interval. Use modeled differences to request a
crossing-specific resilience study and better data. They do not prove a real
catastrophe, an exponential law, or an error in the city's work.

The source/ directory includes the pinned OSM input, dependencies, experiment,
engine adapters, tests and site/PDF builders. Follow source/README.md to rerun.
Full native trip and route outputs are large and are not duplicated in this
packet; SUMO-native-audits.json records their hashes. Rebuilding reproduces the
simulation inputs and numerical outputs; native headers contain build times
and paths, so file-byte hashes can differ across fresh builds.

Interactive simulator: https://yourfriendlynephew.com/magnolia-crossing-simulator/
Street data: OpenStreetMap contributors, ODbL. See SOURCES.md.
