Background

The Mariana Crow (Corvus kubaryi), locally known as Aga, is a critically endangered forest crow found only on Guam and Rota. With fewer than 200 individuals remaining, urgent conservation efforts are underway, including a captive rear-and-release program that began in 2016. This program hand-raises chicks from wild-collected eggs but lacks natural parental vocal interaction—an element known to affect long-term social behavior and reproductive success in corvids.

Project Objectives

1. Characterize nesting-related vocalizations of wild Aga.

2. Identify stage-specific parent-offspring vocal exchanges.

3. Compare wild and captive-reared juvenile vocal behavior.

Methodology

Data Collection

• Wild Aga Nests:

  • 20 nests monitored with SM3/SM4 ARUs and SMM-A2 extended microphones.

  • Over 4,160 hours of audio collected from nests (7 to 67 days each).

  • Recording times adjusted to capture early morning calls and improve frequency range (44.1kHz–48kHz).

  • Vocalizations were detected using Kaleidoscope Pro software.

• Captive-reared Chicks:

  • 13 chicks monitored from hatching through 2 months, with follow-up at 5–8 months.

  • 390 hours of recordings collected during indoor feeding and outdoor aviary sessions.

  • Post-20-day analysis excluded due to mixed vocal environments.

Analysis

• Created and refined a Kaleidoscope classifier using non-breeding season data to identify Aga calls.

• Selected acoustic features (e.g., peak frequency, duration) analyzed using Raven Pro and R statistical software.

• Applied ANOVA and Bonferroni corrections to test effects of age, sex, and weight.

Key Findings

Vocal Patterns in Wild Nests

• Most vocal activity occurred within the first hour after sunrise, with a secondary peak around 4 PM.
• Preliminary evidence suggests new vocalizations may emerge in response to nest failure (e.g., post-egg collection).

Development of Captive-reared Vocalizations

• • A general decline in frequency (Hz) of calls as chicks aged, likely tied to body growth.
  • Sex-based differences observed:
    • Females showed most change from day 1–7.
    • Males changed most significantly from day 7–13.
  • Captive chicks exhibited a wider variety of calls by 5–8 months, possibly due to earlier independence and "vocal play."

  Challenges in Comparison

  • Differences in recording environments limited direct acoustic comparison between wild and captive chicks.
  • Kaleidoscope proved ineffective for detecting faint chick calls or separating call types.

  Conservation Applications

  • Recordings of wild nests will be used to supplement auditory environments for captive chicks, improving exposure to natural vocalizations.
  • The data provides a foundation for refining captive-rearing protocols, particularly for feeding schedules and socialization.
  • Vocal patterns may help assess individual readiness for release and social integration in the wild.

• Future Directions

  • Repertoire Characterization: Use unsupervised clustering to identify and categorize the full Aga call repertoire, addressing the graded nature of their vocalizations.
  • Juvenile Comparison: Assess repertoire diversity and development in captive vs. wild-reared juveniles (>35 days old)
  • Field Integration: Incorporate findings into broader recovery programs, including training future rearing cohorts with real nest audio.
  • Community Science: Continue involving local conservation partners to monitor potential new nesting sites and track juvenile dispersal.

Conclusion

• • This project has laid the groundwork for understanding the vocal development of the critically endangered Aga and its implications for successful captive rearing and release. By combining field recordings, advanced bioacoustic analysis, and strategic conservation practices, it offers a promising path toward preserving one of the world’s rarest corvids.