Can the Number of eBird Check Lists Predict the Number of Species at a Site?

By Waterbird Intern Alicia Manfroy

Ulistac Natural Area is not as quiet as one would expect on a Saturday morning. The sound of rushing cars is replaced by the twitter of birds as I move farther towards the park’s interior. I made this trip to the meandering park in Santa Clara because I wanted to see one of the eBird birding hotspots mentioned in a recent data analysis by SFBBO volunteer Chris Johnson. The results of his analysis showed that the number of eBird checklists a site has can explain much of the variation in species number between birding locations. 

​Johnson used data from eBird to find the top birding locations ...

around Santa Clara County.  eBird is a very popular bird tracking app created in 2002 by the Cornell Lab of Ornithology. A birder can download the app onto his or her phone (see our article for directions) or access the eBird website. Birders can then select a  location and add the birds they observe to a checklist. New checklists can be created for each birding trip. eBird creates an online platform for birding knowledge, including species observed, hotspots, and sites sorted by number of species. This data is shared among citizen and scientific communities. But this begs the question Chris wanted to answer: Is the number of eBird checklists a good predictor of the number of species found at a location (species richness) and hence a predictor of a “good” birding site?

Chris was soft spoken, but passionate about this project when I spoke with him on the phone recently.. Chris found birding at a young age and has been performing data science for resource management purposes since college. Chris decided he wanted to combine his two passions, birding and data analysis, to examine the relationship between eBird checklists and species richness at 38 creekside Santa Clara County sites.

First, Chris downloaded the checklists from eBird and then used a spreadsheet to create graphs to analyze the relationship between the number of eBird checklists at a site and the number of species at the same site. He found the relationship to be logarithmic, meaning there is a fast increase in species observed at first, but that increase slows over time. On average, a birder has a 3% chance of seeing a new species on any give outing once a certain threshold (the linear or flat portion of the log curve in Figure 1) has been hit. Chris found that the number of eBird checklists was a powerful predictor for the number of species a birder can see on a given trip, explaining almost 80% of variation between sites.

​At this point, Chris narrowed his analysis from the original 38 sites spanning four Santa Clara County creeks and rivers to the top 20 “hotspots,” defined as locations with over 150 species recorded. The number of checklists varied widely over all 20 locations. This variation could depend on convenience and accessibility to birders. To combat this variation, Chris extrapolated the data to make all the checklists “even”, essentially eliminating the number of checklists as a factor for the number species found at a site. While some sites showed surprising changes in the number of species, other locations held their rank. The top sites before the extrapolation were the Palo Alto Baylands Preserve, Don Edwards National Wildlife Refuge, and Shoreline Park. The Top sites after the extrapolation were Don Edwards National Wildlife Refuge, Coyote Creek, and the Sunnyvale Baylands. Below is a graph comparing the number of species before and after the extrapolation.  

​While this analysis of eBird data is a great way to find new birding locations based on species richness, Chris does warn that this analysis may not tell the full story. Other factors that might influence the number of species are habitat type, habitat quality, and the way eBird participants enter data.  

Despite the other variables that may affect species richness, eBird checklists are a great way for the average birder to find new birding locations. If you want to get involved with eBird, join the Cornell Lab on May 4, 2019, during their Global Big Day. Have a great time birding and thank you, Chris, for the interesting data science analyzing ebird checklists as a predictor of good birding sites.

Alicia Manfroy is a waterbird intern at SFBBO. She graduated from Santa Clara University with a degree in Environmental Science. A South San Jose native, Alicia grew up birding at nearby wetlands, percolation ponds, and reservoirs. She is excited to pursue her passion for waterbirds with SFBBO in the upcoming months.