Hawk Happenings: Final Report

November 27, 2020—Posted by Rachael Mady on behalf of Bird Cams Lab

Key Takeaways

  • Red-tailed Hawk cam viewers teamed up with scientists to co-create an investigation and learn how the hawks’ behaviors changed in relation to time and weather.
  • During the 2020 breeding season, 323 cam viewers collected 12,585 observations of six behaviors in real time.
  • In exploring the data online and in a live webinar with scientists, the Bird Cams Lab community documented how behaviors varied with time, both within and across days, as well as with temperature.
  • Notably, the total number of events involving vocalizations, feeding, prey delivery, flapping, walking, and mantling (defending prey) increased with temperature. Additionally, the likelihood of a prey delivery or a nestling mantling (defending prey) increased in the early morning and late afternoon.
  • Exploratory statistical modeling revealed that the number of prey deliveries per hour increased as the season progressed and the number of feedings per hour increased as the season progressed and as temperature increased.
  • The new discoveries from this investigation form the groundwork for future investigations using cams to study the behavior of Red-tailed Hawks and other raptors at their nests during the breeding season.

Background

Each year, the Cornell Lab of Ornithology’s Red-tailed Hawk cam captures the interest of viewers from around the world as two Red-tailed Hawks raise their nestlings on a light tower overlooking Cornell University’s athletic fields. This year The Bird Cams Lab community of viewers and scientists co-created an investigation called Hawk Happenings, the first study with this cam to use data collected live. (In the previous investigation, Hawk Talk, participants collected data from recorded video clips.)

A screenshot of Big Red looking down and standing next to the three nestlings in the nest. One of the nestlings is looking right at the camera.

As the Red-tailed Hawks Big Red and Arthur were busy incubating their eggs in spring 2020, the community joined an online forum called the “Wonder Board.” There participants shared their observations while watching the cam and discussed questions that they were interested in investigating together. They further refined these questions in another forum titled the “Question Design Board,” then they voted on which question to pursue.

After 340 people cast their votes, two questions rose to the top about (1) incubation and (2) how hawk behavior relates to weather (Figure 2). We set aside the question about incubation to investigate later since incubation had just ended and we could no longer collect real-time data.

The results of the vote by the participants on which questions to investigate, numbered 1 to 5.Figure 2. Rankings of the five community-proposed questions after 340 votes. Community members ranked the questions using four criteria: interesting, specific, measurable, and feasible. The two questions in green are the questions we considered investigating.

So we set out to answer questions about the following behaviors: brooding (when a parent sits on top of the young to keep them warm), prey delivery, feeding, vocalizations, and nestling flapping, walking, and mantling (spreading wings and hunching over prey). We asked, “What are the frequencies of these behaviors, and how do they vary with the weather?”

Although Red-tailed Hawks are one of the most common hawk species and are found across North America, there are not many detailed accounts of their behavior at the nest, including how behavior might relate to weather. Traditionally, researchers studying Red-tailed Hawks have been able to watch only from afar or to check nests up close quickly and infrequently to avoid disturbance. The Cornell Hawks cam gives us the opportunity to watch right at the nest at all hours, thanks to hundreds of observers who collectively watch for far longer than any single researcher could.

Methods

Between May 21 and June 14, more than 320 participants watched the Red-tailed Hawk cam and documented more than 12,500 observations using an online tool to tag events as they occurred in real time (watch a one-minute data collection tutorial video). Every time participants observed a behavior, they clicked a button to log it in the database. Because we collected data in the latter half of the breeding season, the adults were brooding the nestlings less and less. On May 27, we switched out the brooding button with the nestling activity button so participants could document the nestlings flapping, walking, and mantling.

A screenshot of the data collection tool with the video footage in the middle-left, buttons to click on the right, and data recorded appears in the bottom.Figure 3. A screenshot of the live data tagging feature that participants used to document their observations of behaviors at the Red-tailed Hawk nest. When participants clicked the buttons pictured on the right side, their observations were recorded under “THIS SESSION YOU REPORTED.” They could also access a Field Guide and a Tutorial, shown in the bottom-left corner.

Once we finished data collection, the staff downloaded and turned the data into interactive visualizations that we could explore as a community. In preparing the data, we realized that we did not observe enough brooding to assess the relationship with time or weather. This was because we started collecting data when the nestlings were becoming large enough to maintain their own body temperature (thermoregulate) and their parents spent less time brooding them. We did, however, have large enough sample sizes to look at vocalizing, feeding, prey delivery, and nestling activities.

Big Red feeding young nestlings that look like small white fluffballs.Figure 4. Big Red, the adult female, feeding her nestlings pieces of prey during a feeding event. Other prey can be seen around the nest.

For each behavior, we “binned” (combined) observations that were within one minute of each other, using the first recorded time to indicate when the behavior occurred. Sometimes observers collected data at the same time, and recorded multiple observations of the same event. With these “binned” datasets, we determined the total number of times each behavior occurred by day or hour. Additionally, we calculated the percentage chance that each behavior would occur each hour. Read about these methods in more detail here.

To match the behavioral data to weather data, we reached out to the closest weather station, located at the Cornell Apple Orchards, for the temperature recorded each hour. We then created interactive visualizations that explored the relationship between the number of times a behavior occurred and temperature. At the request of Valerie Curtis, a Bird Cams Lab participant, we also looked at wind speed and precipitation.

In addition to visualizing the raw data, we ran an exploratory statistical analysis to determine the variable that best predicts the number of prey deliveries and feedings per hour: date, hour of day, or weather (temperature and wind speed). We could not look at the effect of precipitation because there was very little rain during the data collection period. We were able to account for variation in sampling effort (i.e. number of minutes watched by at least one person each hour) in the model. We ran multiple models with all possible combinations and chose the model that fit the best (using a measure called Akaike Information Criterion).

Preliminary Results and Discussion From Data Visualizations

We discussed the data visualizations in online forums and a live webinar, “Hawk Happenings: A Look Into the Cornell Hawks’ Nest.” More than 100 people attended the live session, and an estimated 630 viewers watched the recording afterwards. During the live event participants joined in an engaging discussion about the investigation, the scientific process, and interesting patterns seen in the visualizations.

Research Question #1: What is the frequency of certain hawk behaviors?

During May 21–June 14, the total number of vocalizations and prey deliveries increased, most dramatically after June 4 (Figure 5, Figure 6). There was not a clear relationship between day and feeding, but there was a noticeable increase in the total number of flapping, walking, and mantling events after June 4. As discussed in the live webinar, the increase after June 4 could reflect actual increases in these behaviors, or could have been due to an increase in the sampling effort.

More Vocalizations Were Recorded As The Season Progressed

A clustered bar chart with green, orange, and blue bars showing that vocalizations increase in number after June 4th.Figure 5. A stacked bar chart in which the height of each bar represents the total number of vocalizations recorded each day during daylight hours. Each color refers to one of three time periods: morning is 5:00–10:00 A.M., midday is 10:00 A.M.–4:00 P.M., and evening is 4:00–9:00 P.M.

More Prey Deliveries Were Recorded As The Season Progressed

A clustered bar chart with green, orange, and blue bars that show that the total number of prey deliveries increased as the season progressed.Figure 6. A stacked bar chart in which the height of each bar represents the total number of prey deliveries recorded each day during daylight hours. Each color refers to one of three time periods: morning is 5:00–10:00 A.M., midday is 10:00 A.M.–4:00 P.M., and evening is 4:00–9:00 P.M.

Within a day, the percentage chance of vocalizations was lowest in the morning and highest just before dark (Figure 7). June, one of the participants, noted their surprise to see this pattern because they would have predicted nestlings to be loudest in the morning when hungry and active. Another participant, miksohca, proposed that the hawks might be more quiet in the morning to avoid being detected by predators while the adults were off hunting.

Percent Chance Of Vocalizing Increased Throughout The Day

Figure 7. A bar chart in which the height of each bar represents the percentage chance that there is a vocalization during each daylight hour. Each day was split into 1-hour intervals from 5:00 A.M. to 8:00 P.M.

For feeding, flapping, or walking, there was no discernable relationship with the hour of day. In contrast, there were two peaks for the percentage chance of a mantling event or prey delivery, one in the morning and one in the afternoon (Figure 8, Figure 9).

Percent Chance Of Prey Delivery Peaked In The Morning And Late Afternoon

A bar chart with blue bars that has two peaks in percent chance of prey delivery.Figure 8. A bar chart in which the height of each bar represents the percentage chance that there is a prey delivery during each daylight hour. Each day was split into 1-hour intervals from 5:00 A.M. to 8:00 P.M.

Percent Chance Of Mantling Peaked In The Morning And Late Afternoon

A bar chart with blue bars that has two peaks in percent chance of mantling.Figure 9. A bar chart in which the height of each bar represents the percentage chance of a nestling mantling prey during each daylight hour. Each day was split into 1-hour intervals from 5:00 A.M. to 8:00 P.M.

Participants proposed, based on their own observations and experience, that the two peaks of activity might reflect when the prey are most active. The Red-tailed Hawks on the cam have been observed bringing small mammals, songbirds, and snakes to feed the nestlings. One participant, Chirpiesmom, noted that small mammals and songbirds are most active in the morning and in the evening at their own feeder. We wondered if mammals, birds, and other prey are more readily available during those hours, in which case, the hawks might be more successful in hunting at those times. An increase in prey deliveries would then lead to an increase in nestling mantling because nestlings can’t mantle prey unless it’s brought to the nest.

Research Question #2: Does the frequency of certain hawk behaviors vary with weather?

Regardless of the time of day, the total number of events with vocalizations, feeding, prey delivery, flapping, walking, and mantling increased with temperature up to 70–79 degrees Fahrenheit before slightly decreasing at 80–89 degrees Fahrenheit. Participants again wondered if prey activity could explain the relationship between prey deliveries and temperature. As temperature increases, the activity of prey could increase, creating more favorable conditions for hunting.

It’s also possible that as the temperature increases, the nestlings can afford to flap, walk, and mantle prey more because they need to spend less energy maintaining their internal body temperature. There might be a threshold, however, as the temperature approaches the upper range of 80–89 degrees Fahrenheit. During the live webinar, Kate Niedner and Mikal Deese suggested that the nestlings might decrease their activity to avoid overheating when it becomes very hot.

Preliminary Results and Discussion From Exploratory Analysis

We found that the number of prey deliveries per hour was not related to temperature, wind speed, or the hour of the day. However, the number of prey deliveries per hour increased as the season progressed (Figure 10), which reflects what many cam viewers said they had observed at the nest in previous years. Interestingly, this contrasted with findings in the scientific literature suggesting that prey deliveries do not increase across the Red-tailed Hawk breeding season (Preston and Beane 2020).

Higher Rate Of Prey Deliveries As The Season Progressed

A line graph with a gray 95% confidence interval around it, increasing from left to right.Figure 10. The model-predicted mean number of prey deliveries as the season progressed. The black line is the model-predicted mean and the gray band around the black line is the 95% confidence interval, which shows the range of values that we can be 95% certain contain the true mean.

Similar to the prey delivery rate, the number of feedings per hour was not related to wind speed or hour of day. In contrast, the number of feedings per hour increased as the season progressed and as temperature increased (Figure 11, 12).

Higher Rate Of Feedings As The Temperature Increased

A line graph with a gray 95% confidence interval around it, increasing from left to right.Figure 11. The model-predicted mean number of feedings per hour as the temperature increases. The black line is the model-predicted mean and the gray band around the black line is the 95% confidence interval, which shows the range of values that we can be 95% certain contain the true mean.

Higher Rate Of Feedings As The Season Progressed

Figure 12. The model-predicted mean number of feedings per hour as the season progressed. The black line is the model-predicted mean and the gray band around the black line is the 95% confidence interval, which shows the range of values that we can be 95% certain contain the true mean.

Conclusion

Our investigation into the behavior of Cornell’s Red-tailed Hawk family during the 2020 breeding season showed how behaviors varied with time, both within and across days, as well as with weather, in particular temperature. An additional exploratory statistical analysis revealed that the prey delivery rate increased as the season progressed and that the feeding rate increased with both season and temperature.

As many participants noted, we have only scratched the surface studying Red-tailed Hawk behavior. We studied one nest over one breeding season. However, we did uncover patterns that have not been previously mentioned in scientific literature about Red-tailed Hawks, and collected preliminary data to inform future investigations which could look at more nests over a longer time. Increasing the sample size and sampling period would allow us to see if the patterns we discovered reflect broader trends.

Hawk Happenings is the second Bird Cams Lab study to collect data in real time (the first was Panama Live), demonstrating the potential of co-created investigations to reveal new insights about bird behavior observed with live streaming cams. Additionally, Hawk Happenings builds on the findings in Panama Live that there are important considerations when collecting data in real time: variation in sampling effort, accuracy of observations, and combining observations about the same event.

Join us in the next Bird Cams Lab investigation in which we’ll collaborate with Cornell Lab of Ornithology scientist Dr. Eliot Miller to learn more about the aggressive interactions between tropical bird species on the Panama Fruit Feeder Cam. Sign up by taking a survey here.

Acknowledgements

Thank you to everyone who participated in any or all parts of Hawk Happenings: watching the cam, asking questions, collecting data, exploring the data,  and sharing the findings. We also thank the 21 volunteer reviewers who provided input on the final report. Hawk Happenings was a co-created investigation and would not have been possible without our many participants every step of the way.

Thank you as well to the Cornell Lab’s Web Communications team for designing the live data tagging tool and online collaboration spaces needed for this project.

This project was funded by the National Science Foundation grant 1713225. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.

References

Preston, C. R. and R. D. Beane (2020). Red-tailed Hawk (Buteo jamaicensis), version 1.0. In Birds of the World (A. F. Poole, Editor). Cornell Lab of Ornithology, Ithaca, NY, USA. https://doi-org.proxy.library.cornell.edu/10.2173/bow.rethaw.01