From banding to Bayesian analysis
Microbats are important but poorly understood
To improve management decisions for wildlife populations we need better estimates of survival rates
We provided these estimates via a literature review of microbat banding studies, combined with monitoring data and Bayesian analysis
Microbats are fascinating creatures. They make up 17% of mammal species. They occur in a broad range of habitats, from deserts to rainforests to our cities, and are incredibly diverse in their ecology, feeding on everything from nectar to insects to fish and frogs. In doing so, microbats provide us with a range of vital services including pollination and pest control (see the box on ‘bat service’.). Despite their ubiquity, variety and value, microbats remain poorly understood. To most people they are invisible – tiny, nocturnal and with calls at frequencies that are mostly beyond human hearing. Maybe that’s why so many species of microbats are currently threatened by a range of human impacts. These include the devastating impacts from the introduction of whitenose syndrome, wind-energy development, cave disturbances, sensory pollution, and loss of roosting and foraging habitat. There is now a pressing need to understand how these threats are affecting the viability of bat populations. Population viability analyses (see Decision Point #68) can help us predict how bat populations might respond to management interventions, but their application is limited by high levels of uncertainty about vital rates, especially survival.
It has long been recognised that survival rates and associated longevity for many mammals follows rate of living theory: if you’re a small mammal, you tend to live fast and die young. Our knowledge of this relationship can help us predict what the survival rate of a mammal species may be. Not so for bats, the oddballs of the placental mammal world, which live for decades despite their size. One particularly stoic species, Brandt’s bat (Myotis brandtii), is known to live for over 40 years in the wild and only weighs around 6 grams. Because mass alone was unlikely to be the primary driver of differences in survival rates, we set out to identify alternative traits which could be informative for bat population studies. Bats have been banded since the 1910s, so there were a substantial number of mark-recapture studies for us to base our analyses on. We searched the literature for published annual survival estimates of wild bat populations, and extracted 193 survival estimates which covered 44 species and seven families (Lentini et al, 2015). For each of the 44 species, we used databases and reference literature to characterise the traits we suspected could be affecting longevity or survival. This included body mass, what they were feeding on and how, where they tended to roost, latitude, age at which females reach reproductive maturity, and the average number of young born per female each year (bats generally breed once a year).
So what did our resultant model tell us? Well, if you’re a bat then it’s not how big you are that matters when it comes to annual survival, but how many young you have per year. This is probably unsurprising for anyone that’s ever seen a newborn bat – they are massive (relative to the size of the mother). For some species it’s up to a third of the mother’s mass after she has given birth. So, having two of these huge things at a time instead of one is likely to be associated with great physical stress and energetic expenditure. What you’re feeding on is also important – species that feed on fruits and nectar, that tend to occur in tropical regions, experience higher survival than species that hawk insects on the wing, possibly because they are less exposed to predators. Finally, and unexpectedly given our note above about young and birth, males had lower survival rates than females. This could be explained by the fact that the majority of the species in our trait model were vespertilionids or ‘evening bats’, and for these species it tends to be the males that disperse from where they are born. During this time there is increased mortality as a result of predation, the energetic costs of movements, or lack of familiarity with habitats. Armed with our model, we were now able to predict the survival rates of species for which we had no data, but their traits were known.
There are many factors which may influence survival that will not be captured by trait models such as ours, so the existence of these models does not in any way diminish the need for good empirical data. Rather than blindly accept the model predictions, we demonstrated how to combine the ecological knowledge we gained from the trait model with real-world survey data, using an eight-year monitoring program conducted on two species which occupy bat boxes in Melbourne’s north: Gould’s wattled bat (Chalinolobus gouldii) and the white-striped free-tailed bat (Austronomus australis). Bat boxes can serve a number of important functions. The obvious one is to provide a roost for bats in areas which lack natural hollows due to the loss of large dead trees in the landscape. Within Melbourne however, monitoring programs are showing that the boxes are only being used by the most common species, so their role in conservation of threatened species is questionable. However, boxes provide researchers with direct access to study populations, helping us gain insights into the ecology of some species which are difficult to capture. The boxes also allow for community education opportunities, where members of the public are able to see and hear about the bats, often for the first time. In the process, we are able to dispel some common myths and misunderstandings.
Our model predicted that survival of the white-stripe free-tailed bat (average range of 0.57-0.79) would be higher than Gould’s wattled bat (average range of 0.44-0.69), because the latter bears twins. We used these ranges of values as prior information in a Bayesian analysis of survival for the bat box data, a powerful advantage of Bayesian approaches (see Decision Point #58). By constraining our analyses to a range of values (the prior distribution) which we know are reasonable based on our ecological knowledge (the trait model in this case) we can reduce the amount of survey data needed to reach a given level of precision in our final survival estimates. In a time when resources for ecological research are limited and responses to threats must be rapid, it is surprising that we are not making better use of these types of approaches. Traitbased modelling provides insight into the processes driving communities and populations, and in using the learnings of past studies instead of starting from scratch we are able to better inform viability analysis and the range of management applications it serves.
Prior to working on bats around Melbourne I had the opportunity to carry out some research on bats in agricultural areas as part of my PhD (see Lentini et al, 2013, for what we discovered).
Most people don’t think of bats around farms but they provide a very valuable ecosystem service in terms of insect control. Bats are relatively hardy little creatures, and often make up a large proportion of the mammalian fauna in agricultural areas when other species have become locally extinct. As insectivores, they provide a vital ecosystem service to farms by controlling insect pests, and consume 40-100% of their body weight in a single night.
Lentini PE, P Gibbons, J Fischer, B Law, J Hanspach & TG Martin (2013). Bats in a farming landscape benefit from linear remnants and unimproved pastures. PLOS One. http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0048201
More info: Pia Lentini firstname.lastname@example.org
And for more info on Melbourne’s Bat Box Monitoring Program see https://batboxes.wordpress.com/
Lentini PE, Bird T, Griffiths SR, Godinho LN & BA Wintle (2015). A global synthesis of survival estimates for microbats. Biology Letters DOI: 10.1098/rsbl.2015.0371 http://rsbl.royalsocietypublishing.org/content/11/8/20150371#sec-15