Sounding out the Lay of the Land- An Interview with Alison Fairbrass by Charlie Hearst
Alison
Fairbrass is a PhD student at University College London (UCL) researching within
a new and exciting branch of science; soundscape ecology. Although using sound
to study nature is nothing new, as anyone who’s surveyed bats or listened to
bird songs can attest, this field is novel in the sheer scale it operates on. Whilst
acoustic ecology tends to be more species focussed, soundscape monitoring takes
in the orchestration of all the different sounds (biological, geophysical and man-made)
within a landscape in order to understand the environment on much broader
terms. Although there has been recent interest in soundscape monitoring, there
has not been much effort to apply this method to urban ecosystems. Attendees to
this year’s National Bat Conference may have seen Alison present her work on
developing technology for monitoring urban soundscapes; fortunately for those
who didn’t we were able to catch her for a few words about her work.
What
was your first introduction to bats?
After graduating I worked for an
urban ecology research group in Birmingham. One of the projects they were
looking at was the persistence of bats in the city; how the connectivity of the
urban green spaces affected where they foraged. So I worked on that project,
first of all analysing quite a large data set of bat recordings, processing
that to identify what species they recorded and then spent a summer running
around gaps and tree lines filled with different intensities of light to see
whether light and gaps between tree lines caused by road constructions affected
bat movement through the city.
Can you tell us a bit about your work?
Biodiversity in cities is incredibly
important. It provides plenty of services for populations. But it’s really
difficult to know what we have. Monitoring biodiversity anywhere is tricky, but
it’s particularly hard in an urban setting; the land is divided into tiny
little parcels owned by different parties; there are plenty of safety issues;
and the equipment gets stolen. I’m working on new technology for monitoring
biodiversity in cities. The aim is to make it easier to monitor urban nature
over large spaces and time periods by using acoustic recordings. So obviously
I’m interested in bats as you can survey them by their echo location calls in
the ultrasonic landscape. But there are other animals in the lower frequencies
such as birds, invertebrates and land mammals. So I’m working out whether
there’s a way you can stick a recorder out in your garden or park and use that to
understand what you have there in terms of biodiversity.
What
advantages does acoustic surveying have over more traditional visual ID
methods?
One of the selling points is the
reduced resources involved in monitoring. If you can make that initial investment
of sticking up recorders in the first place, leave them for long periods and if
you have the technology to process that data in an automated way, then you will
make huge savings in the long term in getting what is a massive amount of information.
The issue is it’s not as if people are doing that now with huge costs. No one’s
doing it anyway. It’s not happening. So there’s been recognition that we need
more understanding of ecological populations to conserve them, without that
data we’re blind.
Another advantage is that you
effectively have a historical snapshot of what a place sounded like
acoustically. If you store that data in somewhere like the biological records
centre, then it can be used by others down the line. There’s loads of things
I’m sure I’m never going to do with the soundscape data I’m collecting now, but
when it’s archived and hopefully publicly available, there’ll be tonnes of
questions that other people will be interested to use it for.
So
it’s a quicker and cheaper method of surveying. Is it more efficient and
accurate?
There are a few studies where people
have tried to compare the data from human surveyors and from automatic
recordings; it’s as good, if not better. There’s an awful lot of human error
that can be involved in doing a survey, what one person sees, hears and thinks
is one thing, but that can vary a great deal between surveyors. So using
technology could be a way of removing that subjectivity from monitoring.
Any
horror stories?
Not me personally, but I have had the
police call on me once in Birmingham. I was just doing a bat survey and
someone saw me when looking out of their window and got suspicious. I guess it
can look pretty suspect when lurking around a neighbourhood hiding non-descript
black boxes in places.
Could
this work feed into how human generated sounds such as traffic affect
biodiversity?
That’s one thing I’m quite interested
in looking at from the data that I’m recording in London at the moment; how the
presence of particular human sounds at a site relates to the biotic sounds that
we record. Maybe there are certain anthropogenic sounds that can be used in
complement to biotic sounds to tell us what the nature of the environment is as
well. For example there are a few interesting studies where it’s been shown
that birds and some insects change the way they call due to anthropogenic
sounds.
Are
there ways around measuring the presence of species that aren’t typically
noisy?
That’s one question I will be working
on over the next year; what the noisy species can tell us about the quiet ones.
The idea is that you should be able to use the noisy species and everything
else in the soundscape to understand what the environment is like including the
quiet things like plants and small invertebrates. So that’s exactly the kind of
relationship I’ve been looking at everywhere I’ve made recordings by also
surveying the local environment. Hopefully, I can try and understand the
interaction between the two to find if sound can be reliably used as a measure
of an environment as a whole.
Back
to developing your software, what does that entail?
A lot of programming. I had to
improve my ability to manage large data sets and work with them to pull out
information. Luckily I’ve had some help with computer scientists here at UCL. I
guess it’s about understanding how you can characterise different types of
sound; identifying and differentiating between natural and anthropogenic
sounds; geophysical sounds like rain and wind; sounds between and within a
biodiversity group and see how you can get to a more taxonomic group level.
Trying to work out how I can do that is really what I’m tackling at the moment.
The next stage from that, and I’m
sure this is what people are usually interested in, is getting species ID. I
wouldn’t really do that, that’s a whole PhD in of itself, but my work can feed
into existing tools such as iBat which can do pretty good European bat species
identification. What I’m aiming for is getting index measures of
biodiversity in an urban environment, like a kind of summary or quick snapshot
of the ecology of the habitat on a community scale. Those kinds of measures can
be used over long periods of time to measure trends in biodiversity.
Any
surprises from your study?
What was interesting is that every
location I surveyed in London, even the most central locations like right on
Tottenham court road had bats. Even at sites where no birds were recorded for
an entire week, we still got bats; they’re pretty amazing, almost like an urban
adapted taxonomic group. Another was just how much anthropogenic sound there is in the ultrasonic frequency range, of course we dominate the lower ranges, but something like a breaking vehicle can go really high in pitch. So I think ultrasonics should also be considered when discussing conservation and noise pollution, as animals such as bats use those frequencies.
Posts
