Hudson River AIS Intelligence Brief

This intelligence brief documents normal and anomalous AIS traffic patterns on the Hudson River, providing baseline operational context for maritime surveillance and anomaly detection research.

Normal Speed Ranges and Flagging Thresholds

Vessel Type	Normal Speed	Flag Threshold
Tanker (80-89)	6–10 kt	>12 kt
Cargo (70-79)	8–12 kt	>14 kt
Tug with tow (31-32)	4–6 kt	>8 kt
Tug running light (52)	8–12 kt	>15 kt
ATB (Articulated Tug-Barge)	10–12 kt	>14 kt
NYC Ferry (60-69)	20–25 kt	Not flagworthy
NY Waterway	28–30 kt	Not flagworthy
Seastreak	29–38 kt	Not flagworthy
Staten Island Ferry	16–17 kt	Not flagworthy
Pleasure craft	5–25 kt	>30 kt
Law enforcement	0–30+ kt	Flag by name

Hudson River Traffic Context

Approximately 13 million tons of cargo annually transit between NYC and Albany
20–50 commercial vessels on the river at any time in the NYC–Albany corridor
Tanker traffic is routine (petroleum and chemical delivery to upriver terminals)
Tug and barge operations form the backbone of Hudson River commerce
Main channel has no blanket speed limit; 5 mph applies within 100 ft of shore
Hudson River Protection Act (December 2025): Prohibits new anchorages between Yonkers and Kingston

Major Operators (Routine, Non-Flagworthy)

Towing and Barge Operators

Reinauer Transportation (tugs marked "X REINAUER")
Vane Brothers, Kirby Corporation, Moran Towing, McAllister Towing

Passenger Operators

NYC Ferry/Hornblower: 38 vessels
NY Waterway: 32 vessels
Seastreak: 10 catamarans
Staten Island Ferry, Circle Line, Spirit Cruises

Law Enforcement Agencies on the Hudson

NYPD Harbor Unit: 27 boats across 3 bases, 150 officers. Typically 6–10 active vessels at any time
USCG Sector NY: Routine presence; operates icebreakers PENOBSCOT BAY and STURGEON BAY
NJ State Police Marine Services Bureau
Westchester County PD and Rockland County Sheriff Marine Unit
CBP Air and Marine Operations (presence not publicly documented on Hudson)
Hudson River Pilots Association (AIS type code 50)

Sensitive Sites and Maritime Security Zones

Indian Point Nuclear Complex: Decommissioned but maintains active 300-yard security zone per 33 CFR 165.169 (coordinates: 41°16'N, 73°57'W)
West Point Military Academy: No documented maritime security zone, but military installation with restricted airspace
USCG Station Bayonne: Primary station with jurisdiction extending upriver to Albany

AIS Data Quality and Limitations

Known Issues

Greater than 50% of US towing vessels transmit incorrect AIS data (USCG 2018 finding)
Type code 0 = "Not available" — typically small/recreational boats, not inherently suspicious
Speed readings >50 kt on Hudson almost certainly represent GPS error
Static data (name/type) broadcasts every 6 minutes vs. position updates every 2–10 seconds — type may lag reality
MMSI errors: 30–50% of vessels have incorrect MMSI or store it incorrectly in name field

Military AIS Behavior

Warships are not required to broadcast AIS (exempt from regulations)
Post-2017 collision mandate requires AIS in congested waters, but commanding officer discretion remains
When military vessels do broadcast: typically generic identifiers like "US GOV VSL" rather than actual ship names
USCG vessels generally do broadcast with identifiable names and callsigns

AIS Type Code Reference

Critical codes for Hudson River surveillance:

0: Not available (most common, usually small boats)
31–32: Towing vessel — routine
35: Military Operations — always flag
50: Pilot vessel — routine
51: Search and Rescue — notable (indicates SAR activation)
52: Tug — routine
55: Law Enforcement — always flag
58: Medical transport — notable
60–69: Passenger vessels (ferries) — routine
70–79: Cargo vessel
80–89: Tanker

What's Actually Suspicious on the Hudson

True anomalies warrant investigation:

AIS transmission dark near sensitive sites (Indian Point, West Point, major bridge crossings)
Loitering in non-anchorage areas (sustained slow speed with frequent course changes)
Ship-to-ship rendezvous in mid-river outside established terminals or anchorages
Vessel with no name broadcast after 30+ minutes of continuous observation
Invalid MMSI patterns (000000, 999999, or all identical digits)
Tanker or cargo vessel north of Cuomo Bridge not heading toward Albany or Kingston terminals
Anchoring north of Cuomo Bridge outside designated areas (now illegal under 2025 act)
Night-only operational pattern
Navigation status mismatch (status=underway but speed <1 kt and not anchored)

OSINT Tools and Data Sources

Free/Public Resources

MarineCadastre/NOAA AccessAIS: Historical US AIS data back to 2009 (free)
Equasis: Vessel ownership and inspection database (free)
IMO GISIS: Official vessel registry
Global Fishing Watch: Free tool for dark vessel research
AisAbnormal (GitHub, Danish Maritime Authority): Grid-based statistical anomaly detection
GeoTrackNet (GitHub): Neural network anomaly detection for maritime vessels
TREAD Methodology: NATO framework for Traffic Route Extraction and Anomaly Detection

Cold Start and Baseline Methodology

When establishing baseline traffic patterns:

Bootstrap from MarineCadastre historical data (years of established patterns, free)
Begin with rules-based approach (speed thresholds, geofences, type-based filters)
Transition gradually toward statistical and machine learning methods
Apply progressive thresholds: Higher thresholds initially (2+ weeks), lower over time as confidence builds
Maintain known-good vessel whitelist to reduce noise
Use compound scoring over single-indicator alerting to reduce false positives

Spoofing Detection Heuristics

Methods to identify spoofed or manipulated AIS tracks:

Position jump detection: Reported movement exceeding theoretical maximum (speed × time delta)
Repeated track segments: Identical course/speed patterns cycling unnatural
Anchorage stability: Real anchored vessels exhibit drift; stationary coordinates over hours may indicate spoof
Reception geometry anomalies: AIS signals received by shore station beyond ~50 nm range
Origin geometry: Circle patterns consistently starting from northernmost point (artificial behavior)