From the Field: Paul Millhouser and tracking migrating kestrels with low cost, light based geolocators

Satellite tracking is the state of the art of technology for mapping wildlife movement. But what if your budget is not state of the art?

In our second From the Field interview, we're talking with WILDLABS member and GIS consultant Paul Millhouser. Sometimes the challenge of using less advanced technology can have unexpected benefits, as he discovered while analyzing the migration of lesser kestrels that nest in Mongolia using data from light-based geolocators.

Date published: 2017/03/29

Q: What challenge has this technology helped you overcome?

Lesser kestrels are one of the smaller raptors, and weigh no more than about 8 ounces (225 grams to everyone outside the US), so very small tracking gear is required. A satellite tracker small enough for this application would cost more than $3000, and the budget would have covered only one tracker. Instead, the field team used the same budget to fit ten birds with light-sensing geolocators that were pioneered by the British Antarctic Survey.

Kestrel Tag

Q: How did you first get the idea to use this technology for your work?

A clever researcher (not me) from the Denver Zoo, which spearheaded this project, realized that, even assuming the tracker stayed on the bird and did not malfunction, the project would be betting a lot on that one bird surviving the migration cycle and on it taking a route representative of its group. The light-based geolocators are rugged devices that consist of simply a light sensor, clock, and memory embedded in clear epoxy—no transmitter. The geolocator merely records the average light intensity for the preceding two minutes for each minute. Software with tables of sunrise/sunset times at different locations is used to calculate daily longitude based on the time of sunrise and sunset, and latitude based on the length of day.

Q: Do you use specific criteria to select the technology or model you use?

Because I usually work with conservationists and researchers with limited budgets, I try to find the most cost-effective technology or data that will get the job done. There is a saying that “the perfect is the enemy of the good”; if we insist on using only the best technology, a lot of valuable work won’t get done. Often, just good is good enough.

For example, the geolocators used on this project seem inferior to a satellite tracker in every way: they must be individually calibrated to determine what light levels correspond to sunrise and sunset at a particular zenith angle, they are very imprecise (positions +- 185 km), they provide no data on cloudy days or close to equinoxes, they require recapture of the bird to retrieve data. But because of their low price, the effect of losing a bird or tracker during migration is reduced, and, if multiple birds are recaptured, there is confirmation that one bird’s path was not anomalous. For this preliminary study, simply knowing whether the birds wintered in Africa or India would be a sufficient outcome, so a geolocator’s inaccuracy would not be a problem.

Q: What were some of the biggest challenges you faced using this technology in your location?

Retrieving the data! The light-based geolocators store all of their data onboard and have no transmitter, so the birds have to be recaptured. Despite the birds returning to the same nesting grounds each summer, that is a big task for the field team. Lesser kestrels won’t survive long in a trap in the Mongolian summer, so traps need to be checked every few hours during the day. As a result, we only have data on three birds, although it is possible that more will be recovered. Still, that’s better than the maximum of one bird that would have been possible using a satellite tracker.

Q: What are some of the shortcomings of the technology you’re using for your work that you’d like to see addressed?

The biggest problem is that for a period of about two weeks before and after an equinox, the latitude data is unreliable—the variance of length of day at different latitudes is so slight that making a distinction is difficult. As those are times when the kestrels are in mid-migration, data about the exact route they take is spotty. If the light sensors were more precise, and also more consistent from unit to unit, it might be possible to at least partially overcome that limitation. The software for converting the light/time data into positions is probably due for a redesign; if the amount of user intervention required could be decreased, a lot of time could be saved.

Q: Have there been any unexpected positives for using this technology? What are the most surprising findings that the technology has helped you to discover?

Because three birds were tracked rather than just one, we have a better sense of the variation in migration patterns than if a single satellite tracker had been used. While all three birds went to roughly the same latitudes in Africa, there seems to be a good bit of variation in their longitude, even allowing for the imprecision of the geolocators. The ornithologists had expected kestrels’ wintering grounds to be much more compact, and the mix of grasslands and forest in their winter habitat was a surprise as well.

Kestrel Map

Map showing the migration patterns of the three recaptured birds.

Q: What advice would you give other groups such as yours that might be thinking about using this technology in their work?

If budget is an issue and they can live with the limitations I described, this technology can provide a surprising amount of information at a reasonable cost. Remember to calibrate the sensors before deployment, as every geolocator is a bit different. To maximize the quality of data, you will need to spend a good amount of time in the software to assess the validity of each day/night transition, so be sure to account for that in your planning.

Are you tracking wildlife in your work? Interested in designing new tags? Our Wildlife Tracking group is where WILDLABS tracking experts are discussing designing new tags, sharing their experiences using current open source models, and sourcing recommendations for tags suitable for specific animals (including turtles, polar bears and small animals). Become part of the WILDLABS community and join the conversation. 


About the Author

Paul Millhouser is an independent GIS consultant who advises conservation and research-oriented nonprofits. His work has included analyzing satellite data to assess human impact on protected areas, processing radio collar data to determine animal home ranges and migration patterns, and determining the impact of oil and gas drilling on endangered species. His current focus is on using satellite imagery to evaluate the success of conservation efforts on changing landscapes. Using Python and other programming languages to automate GIS data entry and analysis is another long-term interest. 

Get in touch with Paul through his profile or visit his website