CBER Seminar - Ben Williams, UCL

Abstract: 

Imagine a thriving forest habitat filled to the brim with biodiversity and how this might sound. Next, imagine a highly compromised version of the same habitat, cleared through logging or unnaturally intense forest fires – how would this sound different?

It can be a surprise to many that the difference in soundscape you likely have in mind between these forests can also be observed on tropical reef ecosystems. On a healthy reef habitat the fish and invertebrates produce diverse and complex soundscapes, not just useful as a monitoring opportunity, but also vital to the communication and recruitment of the resident organisms on these sites. However, we have lost over 50% of the oceans coral reefs since 1950 and on these heavily degraded sites the soundscape is much more desolate.

There’s clearly a lot we can learn from listening to reef habitats. However, this practice has only emerged in recent years meaning the recording technology and analytical methods used to capitalise on the potential of acoustics as a monitoring tool are underdeveloped.

During this seminar I’ll share work led by myself and colleagues aimed at addressing these shortcomings. This will include the testing and development of low cost audio recordings, bringing the cost of entry level equipment down from >£2000 per recorder to <£100. I will also share work we have done to advance analysis, through the implementation of automated indices, machine learning and now deep learning. These have allowed us to unlock never before seen patterns within the soundscapes of coral reefs and directly support the conservation and restoration of these.