The Zone Barely Changes
We ran RAM PE for each month using the corresponding WOA2023 sound speed profile at the IHA’s assumed 6 dB mitigation level. The 160 dB shutdown zone varies from 4.9 km (September) to 5.4 km (February) — a seasonal spread of just 0.5 km. This is consistent with Investigation 02’s finding that sound speed profiles contribute less than 1 km of zone uncertainty.
If you only looked at the acoustics, season doesn’t matter. The shutdown zone is roughly 5 km year-round. You could pile in any month and the acoustic footprint would be essentially the same.
This is the wrong conclusion. The shutdown zone defines where the sound is dangerous. Whether it matters depends on whether a whale is inside that zone. The acoustic question and the biological question give very different seasonal answers.
But Density Varies 21×
Roberts et al. (2024) provides monthly NARW density estimates for the southern New England shelf. The variation is enormous. February peak density is 21 times higher than August minimum. The whales are migrating through the lease areas in winter and spring, and largely absent in late summer.
| Month | Density (animals/100 km²) | Relative to Min | IHA Status |
|---|---|---|---|
| Jan | 1.079 | 19× | Restricted |
| Feb | 1.200 | 21× | Restricted |
| Mar | 1.032 | 18× | Restricted |
| Apr | 0.891 | 16× | Restricted |
| May | 0.712 | 13× | Restricted |
| Jun | 0.177 | 3× | Allowed |
| Jul | 0.123 | 2× | Allowed |
| Aug | 0.056 | 1× (min) | Allowed |
| Sep | 0.091 | 1.6× | Allowed |
| Oct | 0.109 | 2× | Allowed |
| Nov | 0.234 | 4× | Allowed |
| Dec | 0.717 | 13× | Allowed |
December is a policy gap. December density (0.717) is nearly identical to May (0.712), which IS restricted. The IHA restriction boundary appears to follow calendar convention (January–May) rather than density data. This is the kind of insight that emerges from combining acoustic modeling with biological data.
Caveat: The IHA restriction may incorporate behavioral factors beyond density — calving, nursing, acoustic sensitivity — that justify the January–May boundary even where December density is comparable. Density alone may not capture the full biological risk.
Density from Roberts et al. (2024), Duke Marine Geospatial Ecology Lab. Monthly resolution, 10 km grid. Values for the Vineyard Wind lease area centroid.
The IHA Restriction Works
The Incidental Harassment Authorization for Vineyard Wind restricts pile driving from January through May. Our analysis validates this restriction: the five restricted months contain 77% of the annual NARW density-weighted risk. The restriction is well-calibrated to the biology.
The risk is not uniformly distributed. January alone accounts for 18% of annual risk. The five restricted months (Jan–May) account for 77%. The seven allowed months (Jun–Dec) share the remaining 23%, with July through October contributing just 6% combined.
Optimal Construction Windows
If a developer can choose when to pile, the risk reduction from timing alone is dramatic. No additional technology, no additional cost — just scheduling.
| Window | Months | % of Annual Risk | Risk Reduction |
|---|---|---|---|
| 4-month optimal | Jul – Oct | 6% | 94% avoided |
| 6-month optimal | Jun – Nov | 12% | 88% avoided |
| IHA allowed (7 months) | Jun – Dec | 23% | 77% avoided |
| Year-round | Jan – Dec | 100% | 0% |
The 4-month window (Jul–Oct) captures only 6% of annual risk while still providing enough construction time for a typical foundation campaign. Extending to 6 months (Jun–Nov) doubles the risk to 12% but adds critical schedule flexibility for weather delays.
Per-Event Probability
The density numbers translate to concrete encounter probabilities. For a single pile driving event (one foundation, one day) during the allowed months, the probability that at least one NARW is within the shutdown zone ranges from 3% to 7%, depending on the month.
Per-event risk is low. Per-season risk is not. A typical Vineyard Wind foundation campaign drives ~60 piles over 4–6 months. At 3–7% per event, the probability that at least one encounter occurs during the entire season is approximately 90% under the conservative (6 dB bias) model. This is why Investigation 04 builds the full Monte Carlo take estimate.
Per-event probability: pevent = density × zone area / normalization. Per-season: 1 − (1 − pevent)Npiles. At 5% per event and 60 piles: 1 − 0.9560 = 95%. The ~90% figure uses the month-weighted average across Jun–Oct densities with 60 piles under the conservative (6 dB) zone.
The 90% per-season encounter probability sounds alarming, but context matters. An “encounter” means a whale is within the 160 dB zone — it does not mean injury. With an active shutdown protocol (Protected Species Observers + passive acoustic monitoring), piling stops when a whale is detected. The per-event risk is the probability that mitigation needs to activate, not the probability of harm.
Density: Roberts et al. 2024. Acoustics: RAM PE with monthly WOA2023 SSP. Risk = density × zone area. Vineyard Wind lease area. Encounter probability assumes Poisson process with conservative (6 dB bias) zone estimate.