Discussion
Microhabitat variables describing vegetation within 10 metres of a nestbox and local human disturbance effects influence occupancy by hazel dormice. In particular, we found that nestboxes were more likely to be occupied by dormice in sites with higher abundance of key vegetation resources (hazel, hawthorn and honeysuckle), with tree abundance that we hypothesized offered multiple movement routes but without creating fully closed canopies that would limit understorey growth and solar radiation, and when located further from human disturbances (footpaths and woodland margins). These are features that should be considered when selecting where to place nestboxes in mitigation and conservation actions and also to guide specialised woodland management promoting favourable features to encourage use of nestboxes by hazel dormice.
Hazel dormice can adapt their diet to their surroundings (Eden 2009), but proximity to preferred and suitable resources is likely beneficial and previous research found a positive effect of tree diversity locally on nest box occupancy (Mortensen et al. 2022). In our study area, increased probability of nestbox occupancy was linked to abundance of three plant species (hazel, hawthorn and honeysuckle). Hazel and hawthorn were also associated with higher occupancy in a managed forest in Denmark (Mortensen et al. 2022). High abundance and cover of hazel trees was a key predictor of occupancy, perhaps not too surprising for a species named after the plant. Hazel is a source of high-calorie food between July and October when dormice need to build up fat reserves to overwinter (Tooke & Battey 2010, Bracewell & Downs 2017). Hazel leaves are also a favoured nest material, thus, close range availability should reduce energy costs and predation risk (Prentice & Prentice 1988). Nestboxes located in areas with higher abundance of hawthorn were also more likely to be occupied. Previous research found dormice seek out hawthorns when emerging from hibernation (Juškaitis 2013). This tree flowers in April/May providing valuable resources at a critical time. The third key plant was honeysuckle, one of the few plant with leaves eaten by dormice (Richards & Hurrell 1984). Honeysuckle leaves can represent half of dormice’s diet during May (Richards & Hurrell 1984) and are also a commonly used nesting material (Bracewell & Downs 2017). Honeysuckle bark can be peeled away from the stem in small strips offering a light, easy to transport, material that is readily available as dormice come out of hibernation. Dormice can use leaves of other plants for nesting (eg. beech and oaks) but often these leaves are not available until later in spring (Lechowicz 1984, Roberts et al. 2015). Our analyses also showed that proximity to some plants, in particular dog’s mercury M. perennis , could reduce occupancy. Whilst there are many ground flora plants which are not utilised by dormice for food or nest material, dog’s mercury is a poisonous plant toxic to many mammals and, dormice might specifically avoid it because of this (Rugman et al. 1983). Further research into its toxicity would be valuable. In addition, dog’s mercury is very prolific and can completely cover large areas of the field layer, competing with other species which are suitable for dormice.
Our results also reveal the importance of tree abundance and tree canopy cover for hazel dormice, predicting that the probability of occupancy should double when the relative abundance of trees goes from 10 to 30 within 10 metres of a nestbox. Dormice are arboreal and can travel up to 152 metres from their nestbox in search of food, this movement is facilitated by an abundance of trees with suitable branching structures (Bright & Morris 2009). Indeed, previous research indicated dormice prefer nestboxes in forest stands with higher cover or denser (Bright & Morris 1990, Juškaitis & Augutė 2008, Mortensen et al. 2022). However, we show here that very high tree canopy and understorey cover (>85%) is likely not ideal, with an apparent optimal around 80-85%. An effect that may not be detectable when using coarser density indices (Mortensen et al. 2022). Very closed canopies can prevent solar radiation leading to lower temperatures and may also limit sunlight reducing plant growth below the tree canopy. On the other hand, open canopies allow too much radiation and occur under limited tree cover that limit arboreal movement.
Higher occupancy was also associated with lower human disturbance. Occupied nestboxes were more likely to be located further from woodland edges and footpaths. Hazel dormice do not completely avoid disturbed sites, and have been reported at roadside habitats in Germany (Schulz et al. 2012). However, our study site is a well-visited National Trust property with high footfall of people, especially at weekends and in the Spring and Summer when dormice are active. Nestboxes located closer to the footpaths and woodland margins are likely exposed to higher noise levels and potentially people could disrupt dormice (trying to look inside nestboxes) if these are visible from paths.
Collectively, these results lead to management recommendations for the placement of nestboxes and site management that build on previous research that focused on wider habitat and nestbox design (Morris et al. 1990, Juškaitis 1997, Madikiza et al. 2010). In particular, given footpath effects, larger and less visited woodlands should be preferable sites for nestbox placement. Within those, nestboxes should be placed preferentially in core woodland areas with high abundance of hazel, hawthorn and honeysuckle, good tree abundance and a late-spring tree canopy and understorey cover around 80-85%. If these conditions are not present, management to promote them should be implemented through felling or coppicing. Coppicing is often employed as a management strategy but the planting of honeysuckle, a fast-growing species, is not generally considered and based on our results could improve dormice occupancy. Management of trees can also be important to avoid fully closed canopies. Finally, the spacing between nestboxes should be considered. Our results show higher occupancy for nestboxes within 5 metres of each other and when located around 50 m away. The first result may reveal individual dormice moving among nestboxes located in very closed proximity, which is an optimal outcome if the aim is to increase population size (i.e., maximize the number of distinct dormice using nestboxes). While additional research is necessary and may not be practical in smaller settings, we tentatively suggest placing nestboxes around 50 metres from each other if possible and within the optimal microhabitat conditions described above.
In conclusion, our study addressed a knowledge gap to understand the role of microhabitat on nestbox occupancy by hazel dormice. However, additional information is still needed to facilitate the recovery of the hazel dormouse. For example, despite collecting data on dozens of plant species during our vegetation surveys, dormice occupancy seems to be influenced by just a handful of key plants. Surveys required working closely to nestboxes, and thus, to minimize disturbance we completed these during the scheduled monthly monitoring by a Natural England dormouse class licensee. More frequent surveys may identify rarer but potentially important plants or seasonal changes we were unable to monitor. In addition, our occupancy time-series did not allow analysis of temporal patterns, but it would be interesting to consider how past occupancy influence future use. Research on variation among individual dormice in their preferences will also be valuable. Marking dormice using pit-tags and camera traps could be used to understand temporal and individual patterns of nestbox use. While we wait for this additional understanding, our results reveal microhabitat variables that influence hazel dormice occupancy of nestboxes offering advice to placement and local scale management to promote conservation of this little mammal.