Our understanding of naturally flying bats is limited, and a better knowledge of the natural history of their aerial ecology could play a vital role in their conservation. The expansion of wind energy infrastructure has been detrimental to bat populations, primarily through collision mortality. We used advanced 3-Dimensional (3D) Geo-Tracking technology to provide a deeper understanding of bat flight behaviors in the wild. This Thermal Tracker 3D system was developed by the Pacific Northwest National Laboratory and commercialized by Sightir, We deployed the Thermal Track 3D (TT3D) to passively track bats with thermal stereoscopic imaging. The TT3D system relays real-time positional data of tracked bats by recording multiple parameters of bat flight, including altitude, speed, wing beats per second, and trajectory. We used a Wildlife Acoustics SM4 bat detector to help identify bats to species as we recorded them with the TT3D. Sequences of the recorded 30 thermal images per second were then analyzed using algorithms to interpret the biophysics of a bat’s flight. In addition to advancing our knowledge of the aerial ecology of bats, this ongoing research may also provide a toolset to reduce wind energy impacts to bats.

Wildlife monitoring on highway crossing structures informs conservation efforts. Traditional methods with motion-activated cameras and manual analysis are time-consuming, expensive and prone to human error. We focused on monitoring wildlife crossings that are part of the I-90 Snoqualmie Pass East Project, WA. The wildlife monitoring program for this project involves 15 networked thermal cameras, generating over two million images and videos annually. We created a computer vision model capable of classifying thermal animal footage in real-time. The model was trained on simulated thermal data, created through a custom image morphing algorithm, as well as real thermal data. The purpose of the model was to remove false positive detections from the motion detecting camera, and classify footage by species (deer, elk, otters, pumas, etc.). The model was trained on a dataset of 26,000 simulated thermal images across 10 animal classes as well as 720 images of deer collected from WSDOT cameras. In one test, the model achieved a precision of 100% and a recall of 98.80% on 270 videos of 813 deer from an overcrossing on the I-90 corridor. The model is currently deployed in a WSDOT data center for real-time classification of footage from cameras on wildlife crossings.

Effective monitoring of ecosystem change over time is essential for natural resource management, guiding decisions from permitting and restoration efforts to conservation prioritization. A key limitation is often the lack of resources to effectively monitor change, particularly over periods that extend beyond traditional project windows (i.e., 2-5 years). Furthermore, comparisons across seasonal events or landscape disturbances like wildfire, drought, or floods, or planned events such as land use changes or restoration actions are often opportunistic. One method that can be used is a trail camera for timelapse photography. Timelapse data can be used to document landscape change, phenology, and seasonality by taking photos at set intervals of time from a stationary location. Using photos to estimate vegetative change provides a robust integrator of the effects of year-to-year climate variability and longer-term ecosystem change. R scripts are used to process imagery, allowing users to draw a polygon around an area or vegetation type of interest. This polygon is used to extract data from all images and calculate indices of greenness and senescence. We describe pilot efforts to monitor seasonal change at a series of watersheds in California and highlight the potential implications of using timelapse photography to track restoration efficacy.

The Mariana monitor (Varanus tsukamotoi, ne indicus; Chamorro name: hilitai) is a large lizard, growing up to 4 ft in total length, and is an opportunistic predator. They were once considered an introduced species in the Mariana islands, however recent evidence suggests they are endemic. Wildlife managers are nonetheless concerned with potential predation on the endangered Guam rail (Hypotaenidia owstoni, Chamorro name: ko’ko’). The two species co-occur on an islet called Islan Dåno (Cocos Island) off the coast of Guam where the ko'ko' was reintroduced in 2010 after going extinct in the wild in the 1980s. While a control program for the hilitai was implemented in 2009, little is known about their density and distribution on Islan Dano. We used hand captures, camera traps, and GPS telemetry to inform a spatial mark-recapture (SCR) model for estimating the density and abundance of hilitai, evaluating the effect of lethal removal and informing future management actions. Our estimate of 17.5 individuals per hectare is more than twice as high as previous estimates using line-transect methodology. The use of downwards facing trail cameras paired with drift fences detected other rare species on Islan Dåno as well, including the ko’ko’.