Walking with Butterflies

Citizen Science Research of the Future

Citizen Science, Conservation, and Butterflies

Two-tailed swallowtail (Papilio multicaudata). Photo courtesy of KRD Photography ©

Why Study Butterflies?

  • Charismatic species
  • Easily identifiable*
  • Indicator species
  • Common prey species

Why Citizen Science?

  • Large amounts of data
  • Outreach/Education
  • Demystify science

My Research Project

iNaturalist as a Supplement to Pollard Walk Data: Analysis of Butterfly Diversity and Distribution in the Colorado Front Range

Pollard Walks focus on trail-based measurements of butterfly diversity and have strong records of butterfly populations over time but lack widespread distribution. By comparing these butterfly records with those from iNaturalist, it is possible to evaluate the reliability of both methods and find ways to improve upon their biases. Note: All data is from 2019 only.

Pollard Base vs. iNaturalist Survey Methods

An example of a Pollard Walk. Pollard Walks involve following a set trail approximately every week and counting butterflies within a certain radius of the observer (typically 5-6m). Monitors can submit photos but are not required to. Butterfly counts and identifications (ID's) are submitted in a survey which also includes weather and time data. Butterflies can be identified as broadly as family level or as specific as species level. This data is submitted to a database called Pollard Base where it can be analyzed.

An example of an iNaturalist observation. iNaturalist observations are ID'ed based on photographic evidence. They are more similar to convenience sampling and do not have a systematic method of sampling. Only wild butterflies can achieve research grade and ID's must be agreed upon (See 3).

  1. Title line that includes the common name, scientific name, and research quality designation.
  2. Time/date and geographic data including a map showing the nearby area. Red markers indicate nearby locations where the same species was observed.
  3. This is the ID rating. This shows whether ID's agree. The bar is red if ID's disagree. An observation reaches research status by having 2/3 agreement on ID's with a minimum of two ID's required. If ID's disagree, the ID is listed back to the most specific taxa agreed upon. For example, if I ID'ed this as a moth (incorrectly) and someone ID'ed it as a pipevine swallowtail (correctly) it would be listed with the status Needs ID and the taxa Lepidoptera (order of butterflies and moths).
  4. An activity feed that shows who has ID'ed the photo, whether ID's agree, and the ID's users gave.
  5. A place to add additional info such as life stage, sex, and alive or dead.

For my project I created a sampling area to test differences in species distributions and diversity in the iNaturalist and Pollard Base databases which is shown below. All Pollard Base observations are on the trails designated by the markers. Trails were selected if they had >=5 surveys, reasonable accuracy (no significant formatting issues or incorrectly ID'ed photos), and at 50% or more of butterflies identified to at least the family level.

Slight differences in shape due to zoom. Left image is trail locations of Pollard Base monitors. Right image is Pollard Base trails and distribution of iNaturalist observations illustrated by red markers. To view an interactive version of the iNaturalist observations click here:

Sampling area vs. the whole state of Colorado. Most Pollard Base data falls within this yellow sampling area. Note: trails appear to be outside of the sampling area due to the level of zoom and viewing angle.

A comparison of 2019 species distributions. Species that are the same are shown in the same color. The "other" category represents all other species on record. Note how different the top 5 species are despite being in the same sampling area. iNaturalist is likely skewed toward more colorful butterflies while Pollard Base seems to better match population records. To see some examples of what I mean by "colorful species" check out the top species here:

  • iNaturalist has a similar number of species despite having roughly 11,500 fewer observations.
  • Both databases have exclusive species.
  • Pollard Base has a much larger proportion of unidentified butterflies.
  • Pollard Base species accuracy is likely more related to the skill of the observer.

iNaturalist distribution of butterflies in 2019. Similar to Pollard Base, most observations are concentrated along the front range. Despite this, iNaturalist still has much better distribution records.

  • iNaturalist had nearly twice as many species despite having about 14,000 fewer observations.
  • Top species varied wildly even at state level.
  • iNaturalist has fewer unidentified records.

Conclusion

  • Pollard Base lacks distribution data, especially in alpine and plains environments.
  • Pollard Base likely has more accurate population data (less skew toward photogenic species, count every butterfly on walks).
  • iNaturalist has better state-level species data which will improve as its userbase increases.
  • Both databases have exclusive species and unique data for conservation.
  • Pollard Base data is more likely to be influenced by the skill of the observer (iNaturalist observations have verifiable photos).
  • Citizen science is a major source of conservation data, especially with invertebrates.

What I Learned From Citizen Science

Variegated fritillary (Euptoieta claudia) spotted during one of my surveys

  • Getting the right data can be difficult
  • ID's are not as easy as they seem*
  • Limitations are in all research
  • Citizen science is new and changing
  • Citizen science is the future of conservation

Interested in monitoring butterflies? Find a program near you:

Acknowledgements

Special thanks to Kathryn Hokamp, Chris García-Helmuth, Sarah Triplett, and the Butterfly Pavilion for supporting my research.

I would also like to thank Amber Hecko, Dr. Jim Benedix, Dr. Joe Heithaus, and the Environmental Fellows Program for their continued support throughout this process.

References

Hochmair HH, Scheffrahn RH, Basille M, Boone M (2020) Evaluating the data quality of iNaturalist termite records. PLoS ONE 15(5): e0226534.

Pellet J, Bried JT, Parietti D, Gander A, Heer PO, et al. (2012) Monitoring Butterfly Abundance: Beyond Pollard Walks. PLoS ONE 7(7): e41396. doi:10.1371/ journal.pone.0041396

Prudic, L. Kathleen, Oliver, C. Jeffrey, Brown, V. Brian, Long, C. Elizabeth (2018) Comparisons of Citizen Science Data-Gathering Approaches to Evaluate Urban Butterfly Diversity. Insects. doi:10.3390/insects9040186

Riva, F., G. Gentile, S. Bonelli, J. H. Acorn, F. V. Denes, A. D. Crosby, and S. E. Nielsen. 2020. Of detectability and camouflage: evaluating Pollard Walk rules using a common, cryptic butterfly. Ecosphere 11(4):e03101. 10.1002/ecs2.3101

Two-tailed swallowtail (Papilio multicaudata). Photo courtesy of KRD Photography ©

Slight differences in shape due to zoom. Left image is trail locations of Pollard Base monitors. Right image is Pollard Base trails and distribution of iNaturalist observations illustrated by red markers. To view an interactive version of the iNaturalist observations click here:

Sampling area vs. the whole state of Colorado. Most Pollard Base data falls within this yellow sampling area. Note: trails appear to be outside of the sampling area due to the level of zoom and viewing angle.

Variegated fritillary (Euptoieta claudia) spotted during one of my surveys