GIS, Telemetry, Remote Sensing I

Contributed Oral

High Sensitivity and Reliable Automated VHF Tag Detection
Michael Shafer, Paul Flikkema

Very high frequency radio tags are used extensively in wildlife tracking applications due to their low mass, low cost, reliability, and extensive existing reception infrastructure. For these reasons, these devices will likely remain a tool for tracking into the future. The low transmission power of many VHF tags, coupled with challenging topography for both the transmitter and receiver, makes signal detection a challenge. Historically, detection and localization of these devices has required human hearing to isolate audio pulses within a noisy signal. For marginal signal-to-noise ratio conditions that often occur when transmitter and receiver distances are large, detection of single pulses amongst the noise becomes challenging and eventually impossible. In order to improve detection for these traditional tracking scenarios, as well as enable reliable automated detection through signal processing, our group has worked to develop a novel method of pulse detection that relies upon incoherent pulse integration to improve signal-to-noise ratios. This presentation will include a brief review of the signal processing theory that enables this novel detection method and some results of noisy signals processed for tag detection. We will show how this method is better able to detect signals than previous amplitude thresholding methods, and how the algorithm is able to detect signals at and below the levels detectable by the human ear. This presentation will also highlight software made available by our team which allows users to process their own radio data for tag detection.

Small Unmanned Aircraft Systems Acoustic Analysis for Noninvasive Marine Mammal Response: An Exploratory Field Study
Genni Brookshire

Small unmanned aircraft systems (sUAS) have the potential to benefit pinniped (Pinnipedia; e.g., Phocidae [seals], Otariidae [sea lions], and Odobenidae [walruses]) response efforts. The employment of sUAS could give responders a close-up look at animals in distress in order to determine their condition as well as develop a response strategy. However, pinnipeds may respond to the distinctive sound generated by small, multirotor sUAS. This reaction may include retreating into the water en masse, which could put the target individual out of reach of the response team and cause stress to the animals. To potentially prevent this outcome, this exploratory field study established sUAS acoustic profiles through quantitative and qualitative measures for multiple aircraft across a range of distance and altitude. These data were collected in both a secluded rural environment and a coastal environment. To reflect the type of sUAS that would likely be available to small, non-profit marine mammal response groups working in remote locations, the aircraft studied were limited to compact models $3,000 or less. The results indicate that sUAS sound pressure levels at least 20 dBA (re 20 μPa) below the ambient noise floor are required to completely mask the distinctive sound of the aircraft. The results were used to create aircraft operational envelopes to mitigate disturbance to wildlife while optimizing visual information. Although our study was completed with pinniped response and monitoring in mind, the flight envelopes could be applied to many wildlife species, both marine and terrestrial.

Quantifying the Vegetation Structure Around Wildlife Crossing Structures in Cameron County, Texas
Thomas Yamashita, Humberto Perotto-Baldivieso, Michael Tewes, John Young, Jason “Jay” Lombardi

Wildlife crossings are often constructed for specific target species in locations where previous road mortalities have occurred or where the vegetation directly surrounding the road is favorable. This placement method neglects considerations of whether the habitat surrounding the road is appropriate to support the target species. In South Texas, wildlife crossings are being constructed along highways to aid in the conservation of the endangered ocelot (Leopardus pardalis). Ocelots in South Texas have a strong preference for woody cover, especially Tamaulipan thornshrub, which is characterized by dense vertical and horizontal canopy cover. Many species also prefer this habitat, including bobcats (Lynx rufus), a sympatric felid with similar ecological needs to ocelots. We examined landscape-level and internal vegetation structure around wildlife crossings built for ocelots and other wildlife along three roads in Cameron County, Texas using aerial imagery and LiDAR. We determined landscape structure of woody vegetation in a one-km area around 14 wildlife crossing locations by calculating percentage of landscape, patch density, edge density, mean patch area, distance to nearest patch, patch size, and landscape shape index from classified 1-m resolution NAIP imagery. Within each area, we also calculated canopy height (m) basal area (m2/ha), and leaf area density (m2/m3) from LiDAR to assess internal vegetation structure. We compared these metrics to those in 14 randomly determined locations away from major roads to assess how the vegetation structure around wildlife crossings compares to non-disturbed areas. We also compared how vegetation structure was related to bobcat detections from camera traps at wildlife crossings. We expect that there will be larger patches of woody cover with higher leaf area density in areas away from roads compared to wildlife crossing locations and that bobcats will preferentially use wildlife crossings with more dense woody cover.

Satellite Examination of a Walrus Haulout: A Step Towards Range-Wide Monitoring
Anthony Fischbach, David Douglas

Pacific walruses (Odobenus rosmarus divergens) haulout on land to rest when sea ice is unavailable over continental-shelf waters.  Sea ice has retreated from the Bering Sea during summer for millennia, but only recently has it begun retreating north of the continental shelf in the Chukchi Sea where it is too deep for walruses to feed.  When hauled out on land walruses clump together in large numbers and are much easier to survey than when in offshore habitats where they occur in small aggregations spread across vast regions.  Because aerial surveys at remote haulout locations are logistically problematic, we sought to identify walrus haulout occurrence and size through visual interpretation of satellite imagery (both multi-spectral and synthetic aperture radar) which may be collected range-wide.   We evaluated our satellite imagery interpretations with UAS aerial imagery surveys collected in 2018 and 2019 at a large coastal haulout near Point Lay, Alaska.  We found that publicly and commercially available satellite imagery, ranging in resolution from sub-meter to 40 m, can be used to detect the presence and size of walrus aggregations when hauled out.  Not only were different observers able to consistently identify when walruses were present, they also were able to delineate the geographic extent of the aggregations with considerable agreement and cross-validation with the aerial surveys.  These findings suggest that satellite reconnaissance could provide cost-effective information toward monitoring walrus haulout occupancy and dynamics, as may be required by managers to mitigate disturbance issues at this haulout. These methods may also have range-wide utility if they are ultimately found to be effective at other haulouts with more complex local topography with bluff or cliffs that obscure satellite views or cast shadows.

Monitoring Early-Seral Wildlife Habitat Through Landsat Time Series Disturbance Mapping: A Case Study for Golden-Winged Warbler Habitat in Minnesota
Lisa Elliott, Jody Vogeler, Alexis Grinde, Mark Nelson, Patrick Fekety, Joseph Knight

Changes in natural disturbance regimes and land management practices have led to declines in early-seral habitats and corresponding conservation concerns for species that rely on this habitat. Remote sensing offers new and more efficient tools to monitor and maintain a mosaic of forest seral stages across the landscape to provide suitable habitat for species that specialize in edges and interiors. Our study utilized a new dataset on forest disturbance and seral stages to identify and monitor early seral forest at scales that are relevant to wildlife management and conservation. We used Landsat time series data collected from 1984 to 2019, to train an algorithm to identify patches of abrupt change in forested areas and subsequently classify agents of change. We paired this with point count data on a neotropical migrant of conservation interest, the golden-winged warbler (Vermivora chrysoptera; GWWA), which primarily breeds and nests in early seral upland forests. Point count surveys were conducted in 2017 in the Chippewa National Forest in Minnesota. We used our annual disturbance products to calculate predictors relevant to the species such as seral stage (time since disturbance), magnitude of change, agent of change, combinations of all those, canopy cover, and NLCD classes. Results of our boosted regression tree analysis suggest that GWWAs are negatively associated with tree canopy cover within 100m of point count locations and are positively associated with forest area disturbed 2-10 years ago with low-magnitude change within 1000m of the point count locations. Occurrence peaks at low (but non-zero) values of forest area disturbed 11-20 years ago with moderate-magnitude change within 1000m. Thus, both seral stage and magnitude of change are useful metrics for this species and we are confident these data products will have other potential applications for monitoring and management of wildlife habitat.

Effects of Helicopter Net-Gunning On Space Use and Movement Behavior of Nilgai Antelope
Jeremy Baumgardt, Aaron Foley, Randy W. DeYoung, Kathryn Sliwa, J. Alfonso Ortega-S., David G. Hewitt, John Goolsby, Kimberly Lohmeyer

Movement and space use studies typically involve capturing individuals and releasing them with some tracking device. The act of capturing and handling, as well as acclimation to the tracking device may impact the individual’s behavior and thus, temporarily bias movement data. To avoid these issues, researchers typically left-censor data for some period of time. Understanding the duration and magnitude of an animal’s response to capture is important for avoiding biased data while maximizing sample size. However, it is difficult to accurately assess these impacts since pre-capture data are typically unavailable, requiring researchers to make assumptions and extrapolate from observed behaviors following a capture. We recaptured 21 GPS radio-collared, free-ranging nilgai antelope (Boselaphus tragocamelus) with a net-gun fired from a helicopter in South Texas, USA during 2019–2020, resulting in hourly location data before and after the capture. We calculated mean daily movement rates (DMR) over each 24 hr period for 60 days beginning 30 days prior to a capture. We also calculated the net squared displacement (NSD) for each individual by measuring the distance to each hourly location for the 30-day period following a capture from the centroid of all the animal’s locations for the 30 days prior to capture. Nilgai had a mean DMR of 134 m/hr over the 30 days prior to capture, which increased 65% on the day of and day following capture, then dropped to 60% of the pre-capture rate before returning to normal around 6 days after capture. Our NSD analysis revealed high diversity in response to capture, with most nilgai returning to pre-capture space use patterns within 1-2 weeks. Our results suggest that for nilgai captured via helicopter net-gunning, censoring a minimum of 5 days following capture for analyses related to movement rates and inspecting NSD on an individual basis for space use analyses.

Contributed Oral
Location: Virtual Date: November 2, 2021 Time: 3:00 pm - 4:00 pm