Developing landscape-scale approaches to conserve the threatened barbastelle bat Barbastella barbastellus

The barbastelle (Barbastella barbastellus) is one of Britain’s rarest mammals, primarily due to its narrow ecological niche and historical habitat loss. Significant data gaps remain owing to the inherent difficulties associated with their detection, hindering effective population assessments across their entire range, which predominately encompasses central and southern Europe. In this thesis, a combination of ecological techniques (i.e., acoustic surveys, citizen science, radio-tracking, infrared imaging) was used to develop novel approaches for understanding B. barbastellus distribution and behaviour. The research outlines how these approaches can be applied to improve survey designs and provide new insights into bat ecology.

Our understanding of bats, especially for species that mainly roost in trees, is impeded by the difficulties of locating and studying maternity colonies. However, peaks of bat activity near sunset may be evidence that a colony is roosting nearby and therefore highlight key areas for conservation action. This ecological relationship is widely acknowledged, yet a standardised method has not been developed for surveying and interpreting these activity patterns. I therefore developed a novel approach which showed that provided a sufficient number of acoustic detectors are deployed, activity levels (number of recordings within an hour of sunset) may be used to identify sites with a high probability of supporting roosting maternity colonies of B. barbastellus. This methodology was implemented across 77 woodlands in England by citizen scientists trained in acoustic detector deployment. Follow-up trapping surveys were conducted at 13 sites identified as having high colony potential, as identified through the acoustic surveys. Radio-tracking of female and juvenile B. barbastellus bats to roost trees revealed the presence of five previously unknown maternity colonies, confirming the real-world applicability of the approach. This cost-effective method has the potential to be widely deployed and can be adapted for use with other species or in other geographical regions.

The limited availability of data on the occupancy of woodlands by B. barbastellus has hindered our understanding of the landscape-scale characteristics associated with their presence. Previous work has focused on examining local features within or immediately around focal woodlands, a scale too small to address broader landscape questions, especially when considering the large nightly commuting distances B. barbastellus are known to undertake to reach foraging areas. To understand the influence of both local and landscape features on the distribution of B. barbastellus, I developed models to predict species and likely colony presence within 77 woodlands surveyed by citizen scientists. The analysis was unable to identify any woodland or landscape characteristics that could distinguish between woodlands with low levels of recorded B. barbastellus activity (i.e., low number of individuals unlikely to be associated with a colony) and woodlands with no activity. However, both the size of the focal woodland patch and canopy ruggedness, a measure of heterogeneity, were identified as significant factors influencing the likely presence of colonies. The probability that a focal woodland patch would be predicted to be occupied by a colony increased by 4.1% for every one hectare increase in woodland patch area. The analysis could not identify any influence of the landscape-scale predictors investigated on colony presence. These results could help identify woodlands potentially supporting colonies based on focal woodland patch structure, highlighting areas for conservation action.

Whilst roads may act as barriers to movement and be potential mortality sinks for bats, both species-specific behaviour and the local environment may significantly alter bat road interactions. A comprehensive understanding of flight behaviour and movement over roads has previously been limited by the reliance on human observers, and consequent measurement errors in characterising flight patterns. Using multiple near-infrared cameras in combination with 3-D tracking analysis, I reconstructed the flight trajectories of B. barbastellus, taking into account canopy structure within a 20 m radius around each trajectory, and found that it predicted flight behaviour. Specifically, I found that canopy cover increased road traffic collision risk, with bats being 35% more likely to be at collision risk height for every 10% increase in canopy cover. In addition, flight speed was higher when there was more canopy cover, and the trajectory of flight was more often orientated along the road. Whilst I was unable to explore road avoidance behaviour, I demonstrate that flight patterns remain consistent irrespective of recent vehicle activity. To my knowledge, this is the first study to demonstrate the practical application of 3-D tracking analysis to reconstruct bat flight trajectories over roads.

This thesis has highlighted the important factors affecting the distribution of B. barbastellus, as well as the role of canopy structure in influencing individual flight behaviour over roads, and consequently collision risk. Using both acoustic and radio-tracking methods in combination with citizen science, it has also provided a new method of identifying occupancy of B. barbastellus maternity colonies within woodlands. This thesis offers novel and efficient approaches to track species’ presence and understand behaviour in an increasingly changing and anthropogenic landscape, crucial steps needed for effective conservation.