Introduction
Camera trapping has become a powerful research tool for collecting data on wildlife because it can be carried out at a relatively low cost compared to some other survey or monitoring methods that would require extended human presence in the study area. Moreover, its non-invasive nature in data collection enables the monitoring of elusive species, also in remote locations (Burton et al. 2015, Caravaggi et al. 2017). This approach has traditionally been used to gather information on various aspects of large terrestrial mammals such as occurrence (Salvatori et al. 2023), abundance (Taylor et al. 2022, Santini et al. 2022) and behaviour patterns (Li et al. 2020, Gracanin and Mikac 2022), but nowadays it is used increasingly also for birds (Caravaggi et al. 2017). It is especially useful in the study of ground dwelling birds, where game cameras have been recently used to describe, for example, activity patterns (Nykänen et al. 2021), foraging (Sperry et al. 2021), habitat use (Firth et al. 2020, Puffer et al. 2021), abundance (Kanka et al. 2023) and predation pressure (Laux et al. 2022).
Despite the great potential of game cameras as an effective research tool in a range of applications (Wearn and Glover-Kapfer 2019), they may also have several limitations associated to them, such as variability in camera performance or challenges in sampling design (Rovero et al. 2013, Jacobs and Ausband 2018, Palencia et al. 2022, Santini et al. 2022). One key aspect in camera performance is the trigger mode: in motion sensitive triggering a passive infrared (PIR) sensor triggers the camera to capture an image, whereas in time lapse triggering the camera is programmed to take images at a predefined time interval. Problems may occur, if the camera produces false triggers leading to vast amounts of blank or empty images and therefore drains batteries and fills memory card space. On the other hand, detections of target species can be missed, if the camera is not triggered appropriately. This all causes extra work for researchers and may bias the results of studies.
Here, we compare the performance of motion sensitive and time lapse camera settings in gathering occurrence and relative abundance data of the bean goose (Anser fabalis ) in its breeding areas in Finland. The bean goose is breeding sporadically in remote and inaccessible habitats in the arctic and boreal zones from Fennoscandia to Western and Eastern Siberia (Scott and Rose 1996, Kear 2005). The Western Palearctic population of the species, consisting mainly of individuals belonging to the subspecies taiga bean goose (A. f. fabalis ), has declined in recent decades (Fox et al. 2010, CAFF 2018), the conservation status of the subspecies being considered Vulnerable in Finland (Lehikoinen et al. 2019). While efforts have been put into increasing the accuracy of methods used to estimate taiga been goose numbers in the nonbreeding season (Piironen et al., 2023), monitoring its numbers in the breeding season remains a challenging obstacle to efficient conservation of the subspecies. During the breeding season taiga bean goose is highly elusive, and any survey method involving disturbance caused by the presence of human observers will further reduce the detectability of the species (Pirkola and Kalinainen 1984).
Hence, the main objective of this study was to gain a better understanding of game camera trap performance for reaching more reliable results with the most cost-effective data collection procedure for ground dwelling avian studies, bean goose serving as a research species. A specific aim of the study was to find out which trigger type, motion sensor or time lapse, captures greater goose numbers or is associated with higher daily capture probability, the latter being critical in providing occurrence data.