Small mammals are keystones in ecosystems at northern latitudes (Ims & Fuglei 2005), and understanding their population dynamics are therefore important. Obtaining reliable empirical data on population abundances is a general challenge in ecology (Forsyth et al. 2022), and in particular for small mammals as most are nocturnal and elusive. Hence it is difficult to observe them directly. Historically, estimating abundances of small mammals has required either live trapping or snap trapping. These methods only provide estimates of abundance within a few day period each season, making it difficult to separate different processes that can affect the number of trapped individuals, like activity, habitat use and abundance. Further, these methods are invasive and often with animal welfare concerns. The use of camera trapping techniques now provides a non-invasive alternative that allows to continuously monitor abundances and activity of mammals (Burton et al. 2015), now with established protocols also for small mammals (Soininen et al. 2015; M?lle et al. 2022; Kleiven et al. 2023; Linds? et al. 2025). Yet, analysis of multiannual camera trapping data is scarce from forested areas of southern Scandinavia, and processing large data sets from camera trapping manually is time consuming.
Based on data from 2018-2020, we have previously shown that activity peaks from camera traps occured later in fall for wood mice (Apodemus sylvaticus) compared to bank voles (Clethrionomys glareolus) in Akershus, Norway (Linds? et al. 2025). The hypothesis is that this is related to harvesting of agricultural grain in fall, and reflects habitat use more than timing of reproduction and abundance. However, data from longer time periods is required to validate this. There is now established a semi-automatic workflow for processing small mammal camera trap images (B?hner et al. 2023). Currently, this procedure can successfully separate lemmings, shrews and voles common in alpine habitat and northern ecosystems, in addition to non-target groups like least weasel (Mustela nivalis), stoats (Mustela erminea) or birds. Yet, the tool has not been trained including data on wood mice or pygmy shrew (Sorex minutus), or non-target groups like snakes occurring in our dataset being at lower elevations and in forested habitat.
The aim of this thesis is to use a longer time series of camera trapping data (2018-2025) to estimate seasonal population abundances and activity of a community of small mammals. Questions to be answered:
-
To what extent can artificial intelligence algorithms be trained to separate wood mice from bank voles, and to separate the common shrew from the pygmy shrew?
-
To what extent is the seasonal peak of activity of different species consistent across years?
-
To what extent is prevailing weather influencing activity patterns of mice, voles and shrews during the different seasons?
Methods
Design, data capture and availability
There is ongoing trapping of small mammals in spring and fall (2014-now), while camera trapping with 12 cameras in boxes is continuous since 2018 in sites established near Son in Vestby municipality, Akershus county, Norway. Data from 2018-2020 is published and available (Linds? et al. 2025), while data from 2021 onwards will be used for the current thesis. The Master candidate will participate in monitoring sites by aid of camera trapping.
Work related to the thesis
Field work will involve taking part in the checking of the camera traps (a day). The master candidate will either manually go through pictures from 2021-2025, or train the AI based on data with known species from 2018-2020. This will form the basis for subsequent analysis of seasonal activity patterns.
Supervisors
- Main supervisor: Professor Atle Mysterud, CEES, UiO
- Co-supervisor: Dr. Yngvild Vindenes, CEES, UiO
References
- Burton AC, Neilson E, Moreira D, Ladle A, Steenweg R, Fisher JT, Bayne E, Boutin S. 2015. Review: Wildlife camera trapping: a review and recommendations for linking surveys to ecological processes. – J Appl Ecol 52: 675–685.
- B?hner H, Kleiven EF, Ims RA, Soininen EM. 2023. A semi-automatic workflow to process images from small mammal camera traps. – Ecological Informatics 76: 102150.
- Forsyth DM, Comte S, Davis NE, Bengsen AJ, C?té SD, Hewitt DG, Morellet N, Mysterud A. 2022. Methodology matters when estimating deer abundance: a global systematic review and recommendations for improvements. – J Wildl Manage 86: e22207.
- Ims RA, Fuglei E. 2005. Trophic interaction cycles in tundra ecosystems and the impact of climate change. – Bioscience 55: 311–322.
- Kleiven EF, Nicolau PG, S?rbye SH, Aars J, Yoccoz NG, Ims RA. 2023. Using camera traps to monitor cyclic vole populations. – Remote Sensing in Ecology and Conservation 9: 390–403.
- Linds? LK, Rivrud IM, K?nig EA, Herland A, Mysterud A. 2025. Diel niche of sympatric small mammals revealed by year-round camera trapping. – Ecology and Evolution 15: e71590.
- M?lle JP, Kleiven EF, Ims RA, Soininen EM. 2022. Using subnivean camera traps to study Arctic small mammal community dynamics during winter. – Arctic Science 8: 183–199.
- Soininen EM, Jensvoll I, Killengreen ST, Ims RA. 2015. Under the snow: a new camera trap opens the white box of subnivean ecology. – Remote Sensing in Ecology and Conservation 1: 29–38.