Animal Species Recognition with Deep Convolutional Neural Networks from Ecological Camera Trap Images

Accurate identification of animal species is necessary to understand biodiversity richness, monitor endangered species, and study the impact of climate change on species distribution within a specific region. Camera traps represent a passive monitoring technique that generates millions of ecological images. The vast numbers of images drive automated ecological analysis as essential, given that manual assessment of large datasets is laborious, time-consuming, and expensive. Deep learning networks have been advanced in the last few years to solve object and species identification tasks in the computer vision domain, providing state-of-the-art results. In our work, we trained and tested machine learning models to classify three animal groups (snakes, lizards, and toads) from camera trap images. We experimented with two pretrained models, VGG16 and ResNet50, and a self-trained convolutional neural network (CNN-1) with varying CNN layers and augmentation parameters. For multiclassification, CNN-1 achieved 72% accuracy, whereas VGG16 reached 87%, and ResNet50 attained 86% accuracy. These results demonstrate that the transfer learning approach outperforms the self-trained model performance. The models showed promising results in identifying species, especially those with challenging body sizes and vegetation.

Biogeographic consequences of shifting climate for the western massasauga ( Sistrurus tergeminus )

The western massasauga (Sistrurus tergeminus) is a small pit viper with an extensive geographic range, yet observations of this species are relatively rare. They persist in patchy and isolated populations, threatened by habitat destruction and fragmentation, mortality from vehicle collisions, and deliberate extermination. Changing climates may pose an additional stressor on the survival of isolated populations. Here, we evaluate historic, modern, and future geographic projections of suitable climate for S. tergeminus to outline shifts in their potential geographic distribution and inform current and future management. We used maximum entropy modeling to build multiple models of the potential geographic distribution of S. tergeminus. We evaluated the influence of five key decisions made during the modeling process on the resulting geographic projections of the potential distribution, allowing us to identify areas of model robustness and uncertainty. We evaluated models with the area under the receiver operating curve and true skill statistic. We retained 16 models to project both in the past and future multiple general circulation models. At the last glacial maximum, the potential geographic distribution associated with S. tergeminus occurrences had a stronghold in the southern part of its current range and extended further south into Mexico, but by the mid‐Holocene, its modeled potential distribution was similar to its present‐day potential distribution. Under future model projections, the potential distribution of S. tergeminus moves north, with the strongest northward trends predicted under a climate scenario increase of 8.5 W/m². Some southern populations of S. tergeminus have likely already been extirpated and will continue to be threatened by shifting availability of suitable climate, as they are already under threat from desertification of grasslands. Land use and habitat loss at the northern edge of the species range are likely to make it challenging for this species to track suitable climates northward over time.

Searching for rare and secretive snakes: are camera-trap and box-trap methods interchangeable?

Abstract ContextAdvancements in camera-trap technology have provided wildlife researchers with a new technique to better understand their study species. This improved method may be especially useful for many conservation-reliant snake species that can be difficult to detect because of rarity and life histories with secretive behaviours. AimsHere, we report the results of a 6-month camera-trapping study using time lapse-triggered camera traps to detect snakes, in particular the federally listed Louisiana pinesnake (Pituophis ruthveni) in eastern Texas upland forests in the USA. MethodsSo as to evaluate the efficacy of this method of snake detection, we compared camera-trap data with traditional box-trapping data collected over the same time period across a similar habitat type, and with the same goal of detecting P. ruthveni. Key resultsNo differences in focal snake species richness were detected across the trap methods, although the snake-detection rate was nearly three times higher with camera traps than with the box traps. Detection rates of individual snake species varied with the trapping method for all but two species, but temporal trends in detection rates were similar across the trap methods for all but two species. Neither trap method detected P. ruthveni in the present study, but the species has been detected with both trap methods at other sites. ConclusionsThe higher snake-detection rate of the camera-trap method suggests that pairing this method with traditional box traps could increase the detection of P. ruthveni where it occurs. For future monitoring and research on P. ruthveni, and other similarly rare and secretive species of conservation concern, we believe these methods could be used interchangeably by saturating potentially occupied habitats with camera traps initially and then replacing cameras with box traps when the target species is detected. ImplicationsThere are financial and logistical limits to monitoring and researching rare and secretive species with box traps, and those limits are far less restrictive with camera traps. The ability to use camera-trap technologies interchangeably with box-trap methods to collect similar data more efficiently and effectively will have a significant impact on snake conservation.

Why didn't the lizard cross the road? Dunes sagebrush lizards exhibit road-avoidance behaviour

Context Research has shown many negative effects of roads and traffic on wildlife and other biodiversity. The direct and indirect mechanisms through which roads and traffic harm animal populations vary across taxa, making mitigation of road effects a great challenge for conservation. As such, a large toolkit of species-specific management techniques may be needed to mitigate the negative effects of roads for wildlife and other biodiversity. The dunes sagebrush lizard, Sceloporus arenicolus, is a psammophilic (sand-loving) habitat specialist endemic to the Mescalero–Monahans Sandhills ecosystem of New Mexico and Texas. Within this ecosystem, roads fragment shinnery oak sand-dune landforms occupied by the species. Aims In the present study, we conducted behaviour trials in experimental enclosures to test whether the smallest roads restrict movements of the dunes sagebrush lizard. In addition, we also conducted trials to evaluate whether a sand-filled wildlife-crossing feature could facilitate road crossing. Methods We conducted behavioural trials on 24 dunes sagebrush lizards in our control enclosure and 22 lizards in our road and sand-filled wildlife-crossing enclosure. Movements were recorded for 15min. The final locations at the end of each trial were analysed using circular statistics to determine whether movements in the road or the sand-filled wildlife-crossing enclosures were different from the control. Key results Our results supported the hypotheses that dunes sagebrush lizards avoid roads and do so according to a surface-avoidance mechanism. We also found that the wildlife crossing-feature design tested here had no effect on the movements or road-crossing frequency of dunes sagebrush lizard. Conclusions Surface-avoidance behaviour indicated that roads will persistently affect the movements of dunes sagebrush lizard, even when traffic is not present. Also, more research into an effective wildlife crossing is needed to increase connectivity of fragmented populations. Implications These findings help evaluate the impact of roads in creating isolated populations that experience increased demographic stochasticity and, in some instances, localised extirpation in this species. Our study can guide conservation plans for the dunes sagebrush lizard, and contribute to our understanding of road effects on biodiversity in general.