Not Just Pictures: Utility of Camera Trapping in the Context of African Swine Fever and Wild Boar Management

Abstract

African swine fever (ASF) is a highly contagious disease affecting all suids and wild boar (Sus scrofa). Since 2007, ASF has spread to more than 30 countries in Europe and Asian regions, and the most recent outbreak has been in mainland Italy (reported on January 2022). When the genotype II of the ASF virus infects a population, a mortality rate close to 90% is usually reported. This drop in wild boar abundance produces a cascade effect in the entire ecosystem. In this context, effective monitoring tools for deriving management parameters are a priority aspect, and the utility of camera trapping could have been overlooked. Here, sampling the infected area in north Italy, we showed the utility of camera traps in the context of ASF infection. Specifically, we used 43 camera traps randomly distributed to (i) estimate movement parameters and population density of wild boar, roe deer (Capreolus capreolus), and wolf (Canis lupus); (ii) quantify wild boar recruitment; and (iii) assess whether the human restriction rules are being met. On the first spring after the outbreak detection, our results for wild boar indicated a density of 0.27 ind·km−2 ± 0.11 (standard error, SE), a daily activity level of 0.49 ± 0.07 (i.e., 11.76 h·day−1), a daily distance travelled of 9.07 ± 1.80 km·day−1, a litter size of 1.72 piglets·group−1, and a 72% of pregnant females. Despite human outdoor activities being restricted in the infected zone, we recorded human presence in 19 camera traps. The wide range of parameters estimated from the camera trap data, together with some intrinsic and practical advantages of this tool, allows us to conclude that camera traps are well positioned to be a reference approach to monitor populations affected by ASF. The population-specific parameters are of prime importance for optimizing ASF control efforts.

Epidemiological analysis of African swine fever in the European Union during 2022

“…In addition, hunting data are not easy to interpret, as hunting numbers are highly influenced by many other factors not related to the abundance of the populations but with management decisions. Recently, the use of camera traps has demonstrated its utility to monitor wild boar populations in ASFaffected areas, using the newly infected regions of the north of Italy as an example (Palencia et al, 2023). It appears that different numerical responses of wild boar populations to ASF introduction are related to the surface area of the country affected by the disease.…”

Inter‐population variability in movement parameters: practical implications for population density estimation

“…The high inter‐population variability reported in day range and activity, together with our lack of being able to statistically explain this variability, implies that researchers should estimate movement parameters for the target population when estimating population density with cameras. While the estimation of the day range from camera data usually was time‐consuming, recently developed photogrammetry methods allow for deriving movement data by clicking on the location of the animals in the picture, which reduces the time needed to process the images (Wearn et al 2022, Palencia et al 2023). In addition, locating the animals within the field of view is the starting point of most unmarked methods (Table 1) to estimate movement parameters (e.g., speed) and to estimate other necessary parameters such as the probability of detection.…”

“…Additionally, the REM accounts for activity level and day range (Table 1), which allows us to explore the effect of both parameters on the density estimate, and discuss the effect of not estimating population‐specific activity level parameters on those methods in which only activity is required. Thus, we reanalyzed a recently published and independent dataset (Barroso et al 2023, Palencia et al 2023) of 25 red fox, red deer, roe deer, and wild boar populations in Spain and Italy to test how such practices might affect density estimates. We compared the published reference density, which was estimated using specifically measured activity and day range parameters for the target population and applying REM (i.e., observed density), with the density that would be derived when using predicted day range and activity from the mixed models in this paper (i.e., predicted density).…”

How abundant is a species at the limit of its distribution range? Crested porcupine Hystrix cristata and its northern population

“…Animal‐to‐camera detection distances are required to apply CTDS (Howe et al., 2017). We applied a photogrammetry approach to estimate the location of the animals in the field of detection and lastly estimate animal‐to‐camera distances (Palencia et al., 2023; Wearn et al., 2022). Briefly, the photogrammetry approach describes (i) the relationship between the size of the calibration object in the image (in pixels) and its actual size and distance from the camera; and (ii) the relationship between the X‐Y pixel position in the image and the angular distance from the camera's principal axis.…”