Remote Monitoring

Network-connected cameras and acoustic recorders that read band numbers and pick up nēnē calls. Coverage in places we cannot walk every week.

What this is

Most nēnē monitoring happens on foot. Someone with binoculars reads bands and writes down sightings. That works in accessible areas. It breaks down in remote breeding habitat, at odd hours, and over big landscapes. Remote monitoring fills those gaps with cameras that read bands automatically and recorders that capture nēnē calls.

Three network-connected cameras (a mix of custom-built and commercial units) are currently deployed alongside acoustic monitoring units. Custom software handles band number extraction from images and call type classification from audio. The point is to make individual-level data collection cheaper and more continuous, so high-elevation breeding areas on Mauna Loa and Haleakalā are not only visited when staffing allows.

What the system captures

Imagery from fixed cameras

Network-connected cameras at known nēnē use areas, streaming back continuously.

Band number reads

Software pulls band combinations out of camera frames so individuals get identified without anyone needing to be on site.

Image processing

Filters out empty frames, sorts hits by individual, and stages confirmed reads for review.

Acoustic recording

Wildlife Acoustics units record continuously. A trained model picks out nēnē calls and classifies call types.

Equipment and method

Cameras

Three network-connected units in the field, a mix of custom-built rigs and commercial trail cameras.

Acoustic recorders

Wildlife Acoustics units logging continuous audio at fixed sites.

Placement

Sites selected with LOHE Lab at UH Hilo and DLNR wildlife biologists. Forest reserve and known use areas.

Processing pipeline

Custom code handles band number recognition from images and call detection from audio. Outputs go to managers for review.

Where this is going

The current three-camera footprint is a proof of concept. Next steps are more breeding sites, additional translocation receiving sites on Hawaiʻi and Maui, and acoustic transects in seasonal use areas.

Two practical reasons remote monitoring matters for nēnē specifically. The population has re-established altitudinal migration (Hess et al. 2012 documented roughly 974 m of seasonal movement on Hawaiʻi Island). And the Maui population has declined sharply since 2017 for reasons that are not yet well understood. Both questions are easier to answer with continuous coverage than with periodic field visits.

What this is useful for

Coverage in places we cannot reach often

High-elevation breeding sites, fenced exclosures, and seasonal use areas get continuous monitoring instead of periodic checks.

Better individual-level data

Automated band reads mean more sightings logged per bird, which tightens up survival and movement estimates.

Shared with state and federal partners

Outputs go to DOFAW, USFWS, NPS, and other partners working the same birds, instead of living in a project silo.

Project status

Development phaseIn progress
Initial deploymentComplete
Data analysisOngoing

Partners

  • LOHE Lab at UH Hilo
  • DLNR Wildlife Division
  • Mālama Kioea
  • Birds of Hawaiʻi Past and Present
  • Hui Aloha Kiholo