Temperature is important to desert reptiles as it affects many aspects of their lives, including their ability to digest food, avoid overheating, and conserve water. Climate change creates increased urgency to understand the thermal ecology of desert reptiles and to develop tools for advancing such efforts. For decades, biologists have measured reptile internal body temperatures (Tb) using fine-gauge thermocouples inserted into the cloaca. Recent studies demonstrate that new tools such as infrared pyrometry and thermography may offer non-invasive means to estimate internal Tb in small lizards. We investigated the efficacy of thermography to estimate internal Tb in hatchling desert tortoises (n=39) in a controlled laboratory setting as part of a larger headstart program. We exposed hatchlings to one of two experimental thermal gradients (~22–40 or 22–60 °C) for 3 hours and compared paired surface (thermal camera) and cloacal (thermocouple) temperatures under warming (0.5 h), elevated (2.5 h), and cooling (2.5 h) temperature conditions. We then compared surface and internal Tb of hatchlings during normal activity within their home enclosures.  We present findings from these analyses and additional observations that highlight limitations and strengths of thermography in juvenile tortoise research and conservation.

Changing fire regimes are rapidly and extensively reshaping dry forest ecosystems across the Western US. Understanding the implications of contemporary fire regimes for biodiversity is necessary both for the direct conservation of species and to make informed decisions about the costs and benefits of potential forest management actions. However, the hypothesis that atypically large, severe fires in the Sierra Nevada are influencing overall biodiversity – beyond just a few priority species – has proven difficult to test. The confluence of three technologies have created new possibilities: durable, low-cost recording hardware enable landscape-scale monitoring, the machine learning algorithm BirdNET enables the extraction of bird diversity data from the resulting audio, and the Bayesian statistics program Stan enables the implementation of population models capable of accommodating complex avian community data. We conducted a Before-After, Control-Impact study of the effects of the 129,000 ha North Complex Fire on 67 species of diurnal birds. This fire caused site extinction probabilities to increase significantly for 26 species and caused site colonization probabilities to increase significantly for eight species. Thus, this large, high-severity fire has caused both a change in community composition and a net loss of avian biodiversity.

The California spotted owl (CSO) is a subspecies of conservation concern that resides at the center of forest planning efforts in California. CSO populations have been monitored in local areas with mark-recapture based demographic studies for decades, but the distribution, trends, and status of CSO outside of these areas is less well understood. Therefore, we developed and, in 2021, implemented a bioacoustic monitoring program that spanned most suitable habitat in the Sierra Nevada to estimate trends and understand patterns in site occupancy. We deployed autonomous recording units (ARUs) at over 1,700 sites for approximately 5 weeks each, yielding approximately 1 million hours of passively recorded avian vocalization data. Audio data was scanned using a novel machine-learning algorithm (BirdNET) trained to detect five different CSO vocalizations. Here, we present estimates of CSO site occupancy for the Sierra Nevada bioregion in 2021 and, by scaling occupancy estimates to densities within demographic study areas, provide the first estimate of population size for CSO in this part of their range. Our study demonstrates the feasibility of monitoring rare, vocally active species over regional scales with bioacoustics – providing not just an understanding of population trends but also locational information central to forest planning efforts.

Recent advances in acoustic recorder technology and automated species identification hold great promise for avian monitoring efforts. Comparing these innovations to traditional monitoring techniques is vital to understanding their utility to researchers and managers. We compared bird detection among four acoustic recorder models and concurrent point counts and assessed the ability of the artificial neural network BirdNET to correctly identify bird species from AudioMoth recordings. AudioMoths performed comparably to higher-cost recorders, and three of the five recorder models detected more species than the point counts. A combination of long AudioMoth recordings, BirdNET, and human verification detected higher species richness than point counts conducted in similar habitats. These methods enabled us to survey avian community composition with low misidentification rates and limited need for human verification. Subsequently, we have expanded this methodology to 170 survey locations across diverse ecosystems in central and northern California in 2021 and 2022. This approach holds great promise for improving large-scale, multi-species avian monitoring to inform conservation and management of California’s bird species.

Quantifying population parameters provides the opportunity to understand the effects of landscape structure and change on species of conservation interest. Estimating these parameters can be difficult for rare or cryptic species, particularly in rugged and remote landscapes, but noninvasive sampling methods can mediate these challenges. Herein, we developed an integrated spatial capture-recapture model to estimate the abundance, density, and distribution of a large carnivore of conservation interest, the cougar (Puma concolor), in Yosemite National Park, California, USA. We incorporated genotyped scats and remote camera detections collected in 2019 and 2020 to estimate cougar density and detection probability in Yosemite, and estimated the effects of vegetation, topography, anthropogenic and natural linear features, and survey effort on these parameters. We estimated an average of 31 (± 3.96, 24 – 39) cougars in Yosemite, with cougars occurring at higher densities in productive, vegetated areas. Detection probability was higher with increased survey effort, in the second year of surveys, and in areas closer to trails and farther from streams and heavily trafficked roads. Noninvasive sampling and integrated modeling approaches provide a framework that allows ecologists to leverage empirical data to elucidate the status of species of interest and provide inferences that can inform conservation objectives.

Long-term monitoring is crucial to understanding how ecosystems change over time at local, regional, and state-wide scales, which informs management strategies and actions aimed at conserving California’s biodiversity. The California Department of Fish and Wildlife (CDFW) is working with several partners to establish a Climate-Biodiversity Sentinel Site Network to monitor ecosystems and wildlife on public lands and inform land management in the face of climate change and other stressors. As part of this ongoing and growing effort, CDFW sentinel sites are being established on select Wildlife Areas and Ecological Reserves across the state, reflecting a range of climate conditions and ecosystem types. Each sentinel site will host a series of weather and soil sensors for climate monitoring, suites of wildlife cameras and acoustic sensors at four survey points and permanent vegetation plots for biodiversity monitoring, and a Motus tower to monitor animal movement. Resulting data will be processed with multiple automated and machine learning tools and will allow scientists to evaluate links between climate change and effects on local species and ecosystems. Understanding the effects of climate change at multiple spatial, temporal, and taxonomic scales can support adaptive land-management decisions and inform long-term goals and strategies for conserving California’s biodiversity.