Merlin Bird ID & eBird by the Cornell Lab of Ornithology

Reviewed by: Erin Livingston

Review date: March 19, 2025

Site Link:

• Merlin ID Website/App Download
  
• eBird Website/App Download (eBird includes a browser friendly platform)

Archive Link: https://archive.ph/2g1pX (Merlin Sound ID webpage)

Keywords: Digital Ecologies & Communities, Sustainability, Citizen Science

Data Sources:

• Users submit data on their bird sightings in the eBird app or web platform:
  
  • Images of sighted birds
    
  • Bird call recordings
    
  • Location and time data from sighting
    
  • Bird quantities on provided location-based sighting checklists

Processes:

• Standardizing bird sound data into a structured machine-readable format for AI model training:
  • Spectrogram analysis which included collaboration with human annotators to verify bird vocalizations and provide corrections when AI has misidentification
    
  • Ongoing refinement of these classification models as new recordings are collected and new species can be modeled (150 minimum recordings required to incorporate sound ID)
    
• Organizing and categorizing user-submitted recordings into species-specific datasets within Cornell's Macaulay Library

Presentation:

Merlin Bird ID and eBird are two apps created by the Cornell Lab of Ornithology at Cornell University. Users submit sound recordings or images on the Merlin app to identify birds in the field. The Merlin Bird ID app lets users save bird recordings and explore information about the birds in their area. eBird is an app that records how, where, and when users are birding and provides expert bird checklists based on region. Users fill out the type and quantity of birds seen using these checklists on field adventures and the data contributes to conservation and research efforts for tracking bird populations globally. The data submitted through users on eBird also supports the training of computer algorithms for sound and image recognition on the Merlin Bird ID app.

Digital Tools Used:

• Both apps are integrated with Google cloud-based storage
  
• R & Python based data analysis tools for population modeling (note: this is available on the eBird browser platform, which allows users to sign in and interact with population maps)
  
• Machine-learning algorithms and computer-vision technology for sound and image recognition on Merlin Bird ID.
  • Photo ID algorithms created through collaboration of Caltech and Cornell Tech computer vision labs
    
  • Sound ID algorithms created from recordings provided by the Macaulay Library. Recordings converted to spectrograms with manual annotation support for sound ID
    
• Geographic mapping and location services

Languages:

• Merlin Bird ID: The app supports 24 languages, making bird identification accessible to a global audience.
  
• eBird: The website is available in 18 languages, while the eBird app supports 36 languages. Additionally, eBird accommodates common bird names in 99 languages and regional dialects.

Note: This project was reviewed in English

Review

The digital age allows memory to extend beyond human history to capture ecological change. Merlin Bird ID and eBird are two digital citizen science tools created by the Cornell Lab of Ornithology. They function as living archives of bird populations, migrations, and biodiversity and are utilized for global conservation and research. eBird was launched in 2002 as a global database with a goal of aggregating birder community-based knowledge and developing tools to make birding more rewarding. Merlin Bird ID, introduced in 2014, uses AI and machine learning to help users identify birds through photos and sound recordings. These projects were designed to bridge gaps in scientific bird surveys by making data collection more accessible and participatory. Both were developed with significant institutional backing from Cornell University and are funded through grants, sponsors, donations, and research partnerships. These tools represent a new form of ecological memory. Rather than digitizations of analog archives, these are ongoing, digital-first records of bird populations and behaviors. These tools shape our understanding of nature, conservation, and technology’s role in memory. They are not just birding tools, but dynamic memory infrastructures.

Merlin Bird ID and eBird both live primarily in app formats and work in tandem to engage different users, each designed for varied levels of users’ birding enthusiasm. Both tools have web pages that detail their institutional presence, mission, and partnerships, though it’s the mobile app that functions as the primary memory tool. eBird serves as a data repository where users log how, when, and where they are birding. It provides region-specific bird checklists that users fill out along their journey. eBird also offers a web-based browser login for users to explore population maps, bird hot spots, charts, and reports, or to upload external media, with a more advanced configuration than mobile. Merlin Bird ID serves as an identification tool, allowing users to discover birds through sound or image ID algorithms, which are AI based models trained on eBird data. Merlin also includes reference photographs and recordings from over 5,000 expert birders who have uploaded their media to Cornell's Macaulay Library, a key partner in the development of these tools.

The Sound ID tool within the Merlin Bird ID app is an especially powerful component within this assemblage of birding technology, and it is one of the newest developments, with its first public release in June 2021. It was developed in-house at the Cornell Lab of Ornithology, and trained using recordings from eBird and the Macaulay Library. This feature transforms bird vocalizations into dynamic, machine-readable visual data. Using spectrogram analysis, the app visualizes sound waves in real time, allowing users to watch bird songs as they hear them, reinforcing the connection between sound, species, and place. This powerful approach aids identification and deepens user engagement by providing a data visualization of the memory itself–turning fleeting bird calls into a lasting digital record.

Merlin Bird ID and eBird are passion-driven projects that recognize the borderless qualities of both birds and birders. They democratize memory creation, allowing birders to collectively document the natural world. This aligns with decentralized memory practices, where knowledge production is community-driven and ongoing. Further, these tools function as a record of loss. In documenting species decline and extinction, digital ecological memory is a site for both mourning and activism, with data mobilized toward conservation efforts. These aren’t just birding tools–they’re evolving digital memory infrastructures that document environmental change, loss, and conservation.

How are the collaborative aspects reflected in the project and are there elements that work particularly well?

Merlin Bird ID and eBird rely on citizen science, which at its core is a collaborative approach. Birders worldwide contribute data, and the Cornell Lab of Ornithology partners with local audubon societies to access community based knowledge. In addition, community participation fuels the accuracy of Merlin’s Sound ID, as user-submitted recordings help refine AI models and increase the number of bird species able to be identified through recording. The Macaulay Library, eBird, and Cornell’s researchers form an interdisciplinary collaboration, blending ornithology, data science, and AI. The website notes a team of over 5,000 contributors toward the lab's efforts in developing and maintaining these tools. The open-access data sharing through eBird allows scientists, conservationists, and the public to benefit from and build upon a vast dataset, encouraging temporal analysis and understanding of ecology. This project demonstrates how crowdsourced ecological knowledge can scale globally while remaining interactive and locally meaningful.

Do you see an opportunity for collaboration that would be helpful to the project?

It’s not clear from the Cornell Lab of Ornithology home pages if there’s already involvement or value with Indigenous knowledge keepers. This is an essential but missing component of the project’s current presence. Many Indigenous communities have deep, place-based understandings of bird vocalizations, behaviors, and seasonal patterns that are not always reflected in Western scientific datasets. By creating partnerships that allow for respectful knowledge sharing, Merlin Bird could expand its database with culturally significant bird calls and regional dialects of Indigenous communities, incorporating an element of Indigenous stewardship connected to conservation. Collaboration of this nature could also support language revitalization efforts, as many Indigenous languages contain bird names and meanings that hold historical and ecological significance.