Track, classify & protect marine life
Spot a whale, dolphin or porpoise — anywhere in the world? Snap a photo or record its call and our AI identifies the species instantly. Every sighting you share builds a global picture of cetacean life.
Across US waters — from Alaska to the Caribbean — we combine those sightings with AIS vessel traffic, satellite ocean data and climate projections to model where whales and ships are most likely to collide, so the right areas get protected.
0.0M+
H3 Cells Modelled
0.0B+
AIS Pings Processed
0.0M+
Cetacean Sightings
Community Reports
0
Species Tracked
0
Climate Decades Projected
The Problem
Whales are being killed by ships
Every year, large whales are struck and killed by commercial and recreational vessels. Most strikes go undetected — the crew often doesn't know it happened, and the whale sinks before it can be found. Scientists estimate that reported strikes represent only a fraction of the true number.
20,000+
Estimated whale deaths from ship strikes per year globally
A 2024 study estimated that global ship strikes kill over 20,000 whales annually — far more than previous estimates suggested. The vast majority are never recorded.
60,000+
Large vessels in these waters each year
Shipping lanes overlap directly with whale feeding, breeding, and migration routes. Whales surface to breathe and rest — putting them in the path of vessels that are often too large to stop or steer in time.
80–90%
Reduction in lethal strikes when ships slow to 10 knots
Speed reductions are the single most effective mitigation. Below 10 knots, the probability of a strike being fatal drops dramatically. Seasonal speed zones and route adjustments save lives — but only where risk is known.
That's why we built Whale Watch. By mapping collision risk in real time — combining vessel traffic, whale sightings, ocean conditions, and community reports — we help mariners, regulators, and researchers know exactly where whales are most at risk, and what to do about it.
Focal Species
8 priority species, 77 taxa tracked
Our platform tracks 77 cetacean taxa across species, genus, family, and higher ranks — unifying OBIS sightings, Nisi ISDM predictions, and NMFS strike records through a single crosswalk. These eight species are the focal targets of our collision risk models because they are the most frequently struck in US waters.
North Atlantic Right Whale
The most endangered large whale on Earth. Slow-moving and coastal, they overlap heavily with shipping lanes along the US East Coast. Ship strikes and entanglement account for the majority of known deaths.
Humpback Whale
Despite overall recovery, several distinct population segments remain endangered. Their coastal feeding habits and surface behaviours — lunging, breaching, resting — put them in direct conflict with vessels.
Fin Whale
The second-largest animal ever to live. Fast swimmers, but they rest at the surface and are frequently struck by large commercial vessels — the single largest source of human-caused fin whale mortality.
Blue Whale
The largest animal on Earth, still recovering from whaling. They feed in productive coastal upwelling zones that overlap with major shipping routes, particularly off California.
Sei Whale
One of the fastest baleen whales, but unpredictable surface feeding makes them vulnerable. Poorly studied and often misidentified, making accurate monitoring critical.
Sperm Whale
Deep divers that rest motionless at the surface between dives — making them nearly invisible to approaching ships. Strikes often go undetected in deep water.
Minke Whale
The smallest and most abundant baleen whale in the study area, but their small size makes them hard to spot. Frequently struck by recreational boats and ferries in coastal areas.
Killer Whale
Southern Resident killer whales (only ~75 individuals) are critically endangered. Vessel disturbance, noise pollution, and occasional strikes compound prey depletion threats in the Salish Sea.
Risk Modelling Area
US Atlantic, Pacific & Caribbean waters
You can report sightings from anywhere, but our collision-risk models focus on US waters — the continental US, Alaska & the Aleutians, Hawaii, Puerto Rico, and the US Virgin Islands. The modelled area spans 2°S to 52°N latitude and 180°W to 59°W longitude at H3 resolution 7 (~1.22 km cells).
Platform Features
Map, classify, report & collaborate
Interactive Risk Map
Explore whale–vessel collision risk across the study area with 7 expert-weighted sub-scores per H3 cell. Toggle between macro heatmap overview and high-resolution hex detail.
Open Map→
Photo Classification
Upload a whale photograph and our EfficientNet-B4 model identifies the species from 8 target classes. GPS-tagged photos get automatic H3 risk context.
Classify Photo→
Audio Classification
Submit underwater audio recordings and our XGBoost/CNN pipeline segments them into 4-second windows, extracting 64 acoustic features to identify whale species.
Classify Audio→
Interaction Reports
Report whale interactions with optional photo and audio evidence. Our AI classifies the species, assesses local collision risk, and generates real-time advisories.
Report Interaction→
Vessel Violations
Flag vessels speeding through active slow zones, entering marine protected areas, or suspected of disabling AIS transponders. Community-reviewed reports improve collision risk data.
Report Violation→
Climate Forecasting
Explore how whale habitat and collision risk shift under CMIP6 climate scenarios (SSP2-4.5 & SSP5-8.5) from the 2030s through the 2080s. Compare projected SST, ocean conditions, and species distributions against today.
View Projections→
Risk Scoring Model
7 expert-weighted sub-scores
Each H3 hex cell receives a composite collision risk score from 7 sub-scores. All are percentile-ranked (0–100%) relative to all US coastal cells, then combined with expert-elicited weights from published literature (Vanderlaan & Taggart 2007, Rockwood et al. 2021, Nisi et al. 2024).
Traffic Intensity
25%Vessel speed, volume, lethality potential, draft risk, night operations
Cetacean Presence
25%OBIS sighting records, baleen whale concentration, recent observations
Proximity Blend
15%Distance-decay from whales, strikes, and unprotected areas
Strike History
10%Historical NOAA ship strike records at this location
Habitat Suitability
10%Bathymetry (shelf edge) and primary productivity
Protection Gap
10%Distance from no-take zones, MPAs, and seasonal management areas
Reference Risk
5%Nisi et al. 2024 global collision risk baseline
Data Sources
Built on authoritative marine datasets
AIS Vessel Traffic
MarineCadastre
Cetacean Sightings
OBIS (~1M records)
Ship Strike Records
NOAA (261 incidents)
Bathymetry
GEBCO 2023
Ocean Covariates
Copernicus (SST, MLD, SLA, PP)
Marine Protected Areas
NOAA MPA Inventory
Speed Zones
50 CFR § 224.105 (SMAs)
CMIP6 Projections
SSP2-4.5 & SSP5-8.5, 2030s–2080s
Community Reports
Verified user submissions
Community & Insights
Citizen science meets expert analysis
Everyone from recreational boaters to marine researchers can contribute sightings, verify reports, and access role-specific dashboards — turning collective observations into actionable protection.
Community Sightings
Submit whale sightings with photo and audio evidence, pinpointed on an interactive map. Community members vote to verify reports, building a trusted database. Earn reputation as a contributor and climb the tier system from Observer to Expert.
Report
Photo + audio + GPS
Verify
Agree / dispute votes
Earn
Reputation & tiers
Browse Community Feed→
Stakeholder Insights
Five tailored dashboards translate complex risk data into actionable guidance for each audience. Route-specific advisories for captains, regulatory analysis for policy makers, SDM outputs for researchers, threat mapping for conservationists, and traffic analytics for ports.
Captains
Policy
Research
Conservation
Ports
Explore Insights→
Machine Learning
12 trained models, climate-projected habitat
Species Distribution
ISDM+SDM ensemble for 6 whale species: expert-trained ISDM and OBIS observation-trained SDMs fused into habitat probability maps. Scored on CMIP6-projected covariates for 2030s–2080s under two emission scenarios.
Photo Classifier
EfficientNet-B4 fine-tuned on Happywhale data. 8 species classes including an "other cetacean" rejection class. 380×380 input, differential learning rates, cosine annealing.
Audio Classifier
XGBoost on 64 acoustic features (MFCCs, spectral shape) or CNN on mel spectrograms. 8 species, 4-second segments, 97.9% (XGB) / 99.3% (CNN) accuracy. Three-stage class balancing.
Climate Forecasting
See how risk shifts through 2080
We project whale habitat and collision risk into the future using CMIP6 climate model outputs under two emission scenarios. Our ISDM+SDM ensemble predicts how ocean warming, changing productivity, and shifting currents will reshape where whales go — and where new collision hotspots will emerge.
Ocean Projections
CMIP6 ensemble-mean SST, MLD, SLA, and primary productivity projected at ~0.25° resolution across the full study area.
Species Redistribution
6-species ISDM+SDM ensemble predicts how blue, fin, humpback, sperm, right, and minke whale habitat shifts under warming.
Four Decades
Risk projections for the 2030s, 2040s, 2060s, and 2080s — far enough to inform long-term infrastructure and policy.
Two Scenarios
SSP2-4.5 (moderate emissions) and SSP5-8.5 (high emissions) bracket the plausible future, showing best- and worst-case risk.
58 million projected risk cells
1.8M H3 cells × 4 seasons × 2 scenarios × 4 decades. Every cell carries 6 sub-scores, per-species whale probabilities, and projected ocean conditions — all explorable on the interactive map.
Ready to explore?
Open the interactive risk map, explore climate projections through 2080, upload whale media for AI classification, report a sighting, or dive into stakeholder-tailored insights.