The encounter that delays an inbound vessel doesn't always happen in the harbor mouth. Most of the time, it happened 40 miles offshore — before any vessel traffic service had line-of-sight, before any approach buoy logged the ping.
The Setup
Industrial and logistics-adjacent CRE near major container ports prices throughput risk. Rents in port-adjacent submarkets respond to dwell time, vessel frequency, and congestion signals. The dominant data source for those signals is approach-channel AIS density: what vessels are visible entering and exiting the harbor.
That is a 50,066-encounter view of the world. The other 246,754 Rule 17 events — open-sea give-way inaction encounters in the last 90 days — happened before any involved vessel reached the approach buoy.
Los Angeles/Long Beach is the most data-rich maritime submarket in the index: 2,042 H3 cells, average development pipeline score of 23.7, average amenity demand of 21.0. The port-adjacent industrial corridor is tracked, scored, and mapped. What is not in those scores is upstream throughput risk.
The Chain
The mechanism is indirect but traceable. A give-way inaction event in open water forces a course alteration on the stand-on vessel. Depending on severity, that produces a speed reduction, a deviation from the filed routing, or a repositioning delay before re-entering the traffic lane. The vessel arrives later. If it carries time-sensitive cargo or is connecting to a scheduled berth window, that delay compounds at the terminal.
The 246,754 open-sea give-way encounters in the last 90 days had a median DCPA of 0.114nm — tight enough to require a real deviation in most cases. 51.1% ended in a documented Rule 17 deviation. Even if a fraction of those produce a meaningful ETA impact, the volume means port-adjacent demand signals are receiving a degraded upstream input without a correction factor.
The approach channel giveway dataset (50,066 encounters) is what VTS-derived throughput feeds can see. The open-sea dataset (246,754 encounters) is a 4.9× multiplier that approach-only models miss entirely.
The Implication
The mismatch has a directional consequence for LA port-adjacent industrial pricing. Throughput risk models that undercount upstream Rule 17 events by 4.9× produce a systematic demand signal that is lower than actual maritime traffic stress. Lease pricing and vacancy forecasts for the San Pedro and Wilmington submarkets may be basing throughput assumptions on a one-fifth view of the real encounter queue.
This is not about catastrophic miss rates. It is about signal attenuation. LA port-adjacent cells average a development pipeline score of 23.7 — meaning new supply relative to demand is already thin. In a supply-constrained submarket, a consistently attenuated demand signal means less new supply reaches the market than conditions justify.
What to Watch
- Whether LA port-adjacent pipeline scores show movement as upstream encounter data becomes better connected to throughput models
- Which Pacific and Gulf coastal submarkets have the highest concentration of port-adjacent industrial cells with low pipeline scores — those carry the most unpriced open-sea exposure
- Port dwell time at LA/LB over the same 90-day encounter window — if open-sea Rule 17 events are producing real ETA drift, average dwell should be trending longer than the approach-only signal would predict
Limitations
The connection between open-sea Rule 17 encounters and port throughput delay is causal in direction but unmeasured in magnitude. The Locus cell scores for port-adjacent cells are built from permit velocity, planning records, and demographic data — they do not currently incorporate maritime encounter data. The 246,754 encounter figure includes vessels that may not be bound for LA-area ports; disaggregation by vessel destination is not yet available in the pairwise_encounter table. The throughput mechanism described here is a structural inference, not a regression result.
Data as of July 14, 2026. Sources: Axiom Overwatch pairwise_encounter (open-sea giveway_inaction, April 29 – July 14, 2026); Axiom Locus cell_scores (LA metro, n=2,042 cells).