1 of 20

Orcasound hydrophone network & the

Acartia data cooperative

Scott Veirs on behalf of the growing Orcasound community & Acartia team

sveirs@ gmail.com | orcasound.net/talks

2 of 20

Our orcas: the southern salmon seekers

Southern Resident Killer Whales (SRKWs) are:

  • Southern = ranging from northern California to SE Alaska
  • Resident = historically re-occurring within the Salish Sea (inland waters of WA and BC)
  • Killer = apex predators, salmon specialists
  • Whales = cultural icons, both historically & as modern “charismatic megafauna”

How can we collaborate across State and International boundaries to conserve these orcas?

3 of 20

How? cooperate as a hydrophone network

2002: Orcasound Lab (Haro Strait)�2008-12 NOAA funding (expansion: 2 to 5 nodes)�2013-15 Philanthropy only (decline to 2 nodes)�2016-now Cooperative network (20 NGOs in 2024)

Resilient funding & cooperative administration:2017 crowd-funding ($20k Kickstarter)�2018 open-sourcing, crowd-sourcing, open data...�2020 hackathons, Amazon/Microsoft cloud credits�2021 Google Summer of Code, Experiment.com�2022+ hackathons, open source collective�2023+ industry and State partnerships, WDFW/Ecology grant (5-8 nodes in 2024)

4 of 20

Thank you to the A-teams! (A=Acoustic)

North San Juan Channel: hosted by Orca Behavior Institute: Monika Wieland Shields, Jason Shields, David Howitt (& family!)

Orcasound Lab: hosted by Beam Reach): Val Veirs, Leslie Veirs, David Howitt, & many others!

Port Townsend: hosted by Port Townsend Marine Science Center: Eric McRae, Darryl Hrenko, Betsy Carlson, Diane Quinn

Volunteer with an A-Team via admin@orcasound.net !

Bush Point: hosted by Orca Network: Trevor Snow, Howard Garrett, Lon Brocklehurst, Florian Graner, Scott Veirs

Sunset Bay: hosted by Orca Conservancy: David Bain, Tamara Kelley, Joel Thomas, Jeffrey Thomas

Point Robinson: hosted by Sound Action: Amy Carey, David Bain, volunteer divers

MaST Center: Rus Higley, Aeriel Wauhob, Scott Veirs, Cora Reese

General: advice & hydrophones from Joe Olson

5 of 20

How? collaborate openly in soft-/hard-ware

<2002-17 �-- humans listening via Shoutcast mp3 streams�-- Val building custom software, alone�-- Scott using static websites, Google sheets, manual Twitter/email notifications...

2017-18: software + hardware (Kickstarter for v1 web app), live.orcasound.net launched Nov ‘18

2019-20: v2 UI beta-tested in Nov ‘19, launched in May, 2020; Microsoft hackathon ML MVPs

2021-24: v3 UI + related projects (48 public Orcasound Github repositories in 2024)

6 of 20

  • cabled nearshore sites streaming 24/7
  • Streaming hardware cost: ~U$300 per node
  • Hydrophone $300-1500
  • ~$1k per listening node
  • PiSound HAT = 2 channel, �24 bit, 48/96/192 kHz
  • Free open source code!

Community scientists detect orca & novel sounds in real-time via a web app -- live.orcasound.net

Easy to listen live. Scales inexpensively. Lets users tag data.

7 of 20

Artificial Intelligence for orcas

AI for orcas (#ai4orcas) -- ai4orcas.net -- OrcaHello | Pod.Cast | OrcaAL

towards (more) open (marine) bioacoustic data science… and autoencoders!

8 of 20

ML pipeline with humans in loop

9 of 20

AI+human detection is optimal

For many end-users, expert validation of acoustic &/or visual detections is still important.

5-year results (2018-23):

  • 253 bioacoustic events
  • 186 SRKW bouts
  • 32 Bigg’s bouts
  • 16 humpback bouts
  • Of 41 known SRKW transits in 2021: AI detected 54%; humans 81%; and 97% combined

10 of 20

Demo: OrcaHello AI & humans listening live

  • Moderator emails
    • AI detection
    • Human detection
  • OrcaHello moderator portal
  • Orcasound reports page

In progress (2024):

    • OrcaHello post to Orcasound API
    • Definition of detection confidence
    • Real time posts to Acartia.io API
    • Automated & customized subscriptions
  • End-user notifications
    • 4400 Orcasound subscribers (mid-2024)
    • locate@orcasound.net listserv (140 members)
  • Event log (Google sheet)
  • Orcasound blog posts

Moderator documentation (Github wiki)�Administrator guide (Github wiki)

11 of 20

Sighting & listening together

Great synergies lie in integrating acoustic and visual real-time observations!

Acartia goals (2020-24):

  • Data cooperative for sharing:
    • vetted sightings
    • verified human detections
    • moderated OrcaHello detections
  • Public API & Creative Commons license
  • Tests with Facebook sightings:�Orca Network | San Juans (OBI)

12 of 20

Acartia data cooperative

  • a distributed data cooperative & example of a web3 distributed app…
  • …for sharing any marine animal locations (including in real time)...
  • …across the Salish Sea & the range of the SRKWs (California to Alaska).
  • Initiated in 2019; TypeHuman completed v1 in 2021; tested in 2022-23 with…
  • Phase 1 data sources: Orcasound, Orca Network (220k Facebook followers), & Conserve.io mobile apps Whale Alert & Ocean Alert
  • Phase 1 progress presentation at 2022 Salish Sea Ecosystem Conference

13 of 20

Thank you to the Acartia creators

  • Phase 1 led by TypeHuman
    • Beam Reach, Orca Network, Resolve Conservation, Conserve.io
  • Phase 2 led by Ali (2023+)
    • 2023 fork to Salish Sea organization on Github
    • Swinburne U. student team (2024) building user-centered designs from a volunteer UX team from Orcasound community

14 of 20

Acartia end-users: extant & planned

15 of 20

Acartia data flow: phase 1-2

Data input manual now (via Whale Alert or Cascadia web app)

Any new sightings of key species confirmed by Orca Network are posted to WRAS API every 5 minutes

WRAS integration notes (Acartia repo project card)

16 of 20

Acartia.io phase 1 prototype & public API

17 of 20

Questions?

18 of 20

2024 challenge: lower hardware costs

Calibrated 48 kHz solutions for <$100?�� $15 Raspberry Pi Zero 2W (Oct 2021)�+$15 USB ADC (16bit, 48kHz)�+ $70 low-cost DIY 10m hydrophone? →�

19 of 20

Workshop challenges

  • What do individual hydrophones need to achieve capacity to make updates/sustain hydrophone and data transmission?
    • Hydrophone arrays or DAS to localize
    • Calibrated hydrophones to measure absolute levels, not just relative
    • Faster data connections for remote nodes
  • What does Orcasound collectively need to detect SRKW in real-time/send detections to WRAS?
    • Use of Acartia API by Ocean Wise
    • More models/algorithms, including deep learning models
  • What are the unknowns? What questions need answers?
    • When will we have a low cost (<$1000) buoyed solution?

20 of 20

Workshop challenges

● What needs to happen to move forward?

Empowerment of community scientists (attribution by default, rather than anonymization by default)

More symmetrical transboundary data sharing

Open performance comparisons of AI models and other detection algorithms (& sharing by default if US Federal or State funding supports development)

  • How accurate can/should SRKW detections be to be sent on to a mariner?
    • Ideally, all should be validated by experts so 100% confident
    • But >80% confidence for OrcaHello or any trained Orcasound listener should suffice
  • What are next steps for: hydrophones, Orcasound, Ocean Wise, Quiet Sound, others?
    • WDFW funding to maintain key nodes and add one in Rosario
    • Puget Sound DAS experiment informed by at least Bush Point and Sunset Bay
    • Haro Strait DAS experiment at Orcasound Lab