Time Interval Photography Monitoring of Cinder Butte Snake Hibernaculum (2014)
Return to Research 2014
Investigators and Affiliations
- Charles R. Peterson, Ph.D., Department of Biological Sciences, Idaho State University, Pocatello, ID
- Jeremy P. Shive, Environmental Surveillance, Education, and Research Program, Gonzales-Stoller Surveillance, LLC., Idaho Falls, ID
- David Bush, Department of Biological Sciences, Idaho State University, Pocatello, ID
- Idaho State University Department of Biological Sciences
- Idaho National Environmental Research Park
The T-17 wildland fire burned approximately 17,807 ha (44,000 acres) in 2011, including the area around Cinder Butte (Figure 2). The basalt outcropping near Cinder Butte supports multiple snake hibernacula, including the primary North den, which has been monitored by the ISU Herpetology Laboratory for over 15 years. Anecdotal field observations following the T-17 fire found there was a lot of soil and sand movement in the areas devoid of vegetation. The wind-blown sand was beginning to fill in the interspaces of the basalt rock and there was some concern whether access to the den would be restricted, and the individuals returning for winter hibernation would be stranded with no alternative refuge.
We used a Reconyx PC900 Hyperfire Professional IR camera positioned to image the main den opening and the surrounding vicinity of ledges and rock overhangs where snakes are commonly observed (Figure 3). We also deployed temperature data loggers using snake physical models to record operative temperatures rather than just air temperature. The camera was configured to collect images on a one-minute fixed time interval from sunrise to sunset.
The primary goal of monitoring the Cinder Butte snake hibernaculum is to document the continued use of the den site, and to identify which species of snakes remain present following the T-17 wildland fire. Additional objectives include comparing seasonal activity patterns with an established seasonal baseline to better understand if populations are increasing or decreasing, and to assess rates of detectability using various time-intervals to maximize accuracy and minimize sampling effort.
Accomplishments through 2014:
In 2014, we collected over 82,000 images during the spring from 3/24 – 6/25, and over 32,000 images in the fall from 9/15 – 10/27. All images were initially reviewed once and each observation event was recorded. An observation event is defined as one snake observed for one or more consecutive images. If an individual moved out of view or retreated back into the den, it concluded the observation event even if an individual was seen back at the same spot minutes later. Because we cannot be sure it was the same individual, we treated each instance as a new observation event.
An Access database was designed to more efficiently store and manage the observation events recorded since 2013. The database provides a location to archive the monitoring data, and facilitate future data analyses of multiple seasons of image data.
Time-interval photography continues to be an effective method for monitoring snake species at the Cinder Butte hibernaculum. All four species (Great Basin Rattlesnake, Crotalus oreganus lutosus; Gopher Snake, Pituophis catenifer; Striped Whipsnake, Coluber taeniatus; and Terrestrial Garter Snake, Thamnophis elegans) previously documented at the Cinder Butte hibernaculum by ISU Herpetology Laboratory were successfully detected and present during both spring and fall of 2014. There was only a single observation of a Terrestrial Garter Snake each season in 2014, and this species remains present despite seemingly limited number of individuals.
The camera was deployed earlier this spring compared to 2013, and we likely captured first emergence (i.e., the first snake to emerge from the hibernaculum following winter hibernation) on April 20. Even though air temperature was warm and there were many sunny days, no snakes were observed for the first three weeks of imaging. The last spring snake observation was made on June 5, when a Great Basin Rattlesnake head has visible near the ledge system above and right of the main den opening. Further analyses are planned to quantify the observations, however, there tended to be more Gopher Snakes observed early in the season with more Great Basin Rattlesnakes observed later in the season.
Fall imaging began on September 15, and we already had four observation events of Great Basin Rattlesnakes the first day. This suggests we missed the first return to the hibernaculum (i.e., the first individual to be observed returning to the hibernaculum before winter hibernation) and numerous individuals had already migrated back the hibernaculum before the imaging start date. Generally, there were more observation events for Great Basin Rattlesnakes in the fall compared to the spring.
Plans for Continuation:
Once multiple seasons of data have been collected, a baseline of observation events can be established. Additional time interval photographic monitoring can be conducted in the future and compared to the established baseline to understand changes in population status of each species.
We would like to experiment with different camera settings to optimize the accuracy of snake detections while minimizing overall sample effort. Specifically, we want to test the omission error rate of the current time-interval setup. We plan to alter the time interval and capture an image every second (rather than every minute) from sunrise to sunset and essentially record a continuous sampling of the hibernaculum. Then that dataset would be artificially subsampled at different time intervals (e.g. 30-second, 1-minute, 2-minute, etc.), and reviewed for observation events. This analysis will provide insight about how the time interval affects detection rates. If fewer images need to be collected, and detection rates do not vary considerably, we could minimize the required processing time
We plan to model operative temps using physical snake models attached to a temperature datalogger. By comparing the internal camera system thermometer with the physical models, it will allow us to understand the relationship between camera measurements and the temperatures the snakes are more realistically experiencing at the den.
We also developed a laboratory exercise to incorporate undergraduate students from Idaho State University to assist with the image review process. Manually reviewing all of the images collected during a single season is the most time consuming step of the monitoring process. Comparisons between different observers have shown the omission error rate to be low, but there are observation events that were missed by both observers. If students prove to be effective conducting independent review of seasonal images, our observer error rate can be minimized while providing observation data for further analysis.
Publications, Theses, Reports:
Additional data analysis and statistical modeling are planned for 2015 but have not been initiated.