Our Monday night Citizen Science presentation this week featured a webinar talk from David Bonter, PhD, who spoke about overcoming the challenges associated with operating a large citizen science project like FeederWatch. Dr. Bonter is the Arthur A. Allen Director of Citizen Science at the Cornell Lab of Ornithology and oversees the scientific and educational portions of FeederWatch. His work will prove invaluable to future citizen science projects and has helped FeederWatch become the proven monitoring tool it is today. Thanks for the presentation Dr. Bonter!
Project FeederWatch began in the late 1980s after Canada's Long Point Bird Observatory wanted to expand their survey program and partnered with the Cornell Lab of Ornithology to do so. The Lab of Ornithology's access to both sophisticated computer systems and thousands of bird enthusiasts nationwide allowed the Long Point Bird Observatory to apply their experience at managing feeder surveys to a much larger area. Today, FeederWatch is still a joint effort between Bird Studies Canada (formerly Long Point Bird Observatory) and the Cornell Lab of Ornithology, and has over 20,000 participants.
These 20,000 volunteers for FeederWatch have generated millions of survey checklists at over 50,000 count sites, leading to two books and 26 published scientific papers. As gathering accurate data is the ultimate goal, volunteers are provided with booklets and posters to try to prevent initial errors in identification. Some birds are especially difficult to identify, so the FeederWatch website also has a Tricky Bird IDs page to assist participants. The data FeederWatch requests from volunteers is the largest number of a given species seen in the two count days, and a Tally Sheet is used to keep track of counts throughout. Most data is then submitted through the FeederWatch website, which has a number of built in safeguards to maintain data accuracy. Data entered shows up on a sidebar, so participants can see if they entered any obvious errors. In addition, the final page reports all data back to the citizen scientist so they can do a final quality check.
Once a volunteer is satisfied with their entry and clicks submit, a smart filter immediately reviews the entry to see if it matches with geographic and temporally specific parameters set using historic bird data. If the entry is within the normal range for that location, it is stored in the database. If the entry is flagged as being possibly invalid, the volunteer will receive an automatic email asking them if they are sure what they typed in was correct. If so, the entry goes under expert review, and the volunteer is usually asked to provide photo evidence of the unlikely entry. By using this expert review system combined with photo evidence, many birds have been discovered that are far from their normal ranges and are featured on the FeederWatch Rare Bird Report webpage. One of the current problems with this system is that the smart filter only catches obvious errors based on temporal or geographic inconsistencies. Some less obvious errors, like misidentifying a bird as another local species likely seen during that time, may go unnoticed.
Through a combination of preemptive participant support and behind-the-scenes data analysis, FeederWatch has developed a unique process to facilitate public data collection. From all the points seen on the map, data streams to FeederWatch and is analyzed and used for research. Although the system isn't perfect, continual development will help improve this already successful citizen science project in the years to come. Some of the amazing results of this data collection can be seen in the Winter Bird Highlights published each year by FeederWatch. Thank you for contributions citizen scientists!
If you're interested in learning more about FeederWatch, check out their website! http://feederwatch.org
Want to learn more about citizen science? Come to our next discussion on Monday February 6th with Claire Runge from NCEAS about using e-bird data, talks are at 6pm in Harder South, Rm 1013. See you there!