My MSc dissertation in Madagascar: a final report

There are many things I absolutely loved about the MSc BCM programme, which was an enlightening year. But what was the most special part for me was definitely my dissertation project. I was lucky enough to carry it out in northwest Madagascar, on frog monitoring employing very cutting-edge technology.

I can trace the motivation for this project mainly on the biodiversity technologies course during the second BCM term. This course outlined the current frontiers of various technologies that can be applied to conservation (e.g. drones, remote sensing…), and it really inspired me to do something cutting-edge, innovative, useful (and cool), and apply it to one of my favourite taxa: frogs. Globally, amphibians are threatened with extinction, and many species and ecosystems are severely understudied especially in developing countries, which halts proper conservation action. Therefore, for my dissertation I wanted to contribute to improve this situation.

After a few months of being a part-time student part-time fundraiser for this project, I flew in mid-June 2016 to Antananarivo, the capital of Madagascar, from where I travelled to the town of Mahajanga and finally to the Mariarano forest.

Only in Madagascar lemurs try to break into your car!

Only in Madagascar lemurs try to break into your car!

This area is not formally protected; however, it is rich with an outstanding and relatively understudied biodiversity. My dissertation aimed both at gathering more knowledge about the amphibians living in this area, and determine what is the most effective way to monitor them. An important part of my research was to assess the feasibility of including automated sound identification software to the acoustic surveying of frogs. To make acoustic data analysis faster, several algorithms are currently being developed to automate the identification of the vocalising species, and their accuracy is in constant improvement. In the future, hopefully it will be possible to easily deploy acoustic recorders which automatically identify the species filling in the soundscape, making the surveying of an area a lot quicker and cheaper. The programme I used is called SoundID, and it is a recently designed pattern-recognition software. For it to work, firstly the user needs to create templates from clear recorded calls of the target species. Then, once a library of templates has been created, this is used to analyse a longer recording collected from the field and look for any matches. Furthermore, I compared the effectiveness of manual acoustic surveys and capture-encounter surveys to detect the local frog species through the calculation of occupancy models.

My field work was based on repeated visits to permanent water bodies at night, which was really good fun. I would normally visit them between 8PM and 2AM, which made my average day extremely long: sleeping in, in a tent in Africa, is quite a challenge! However, full of mud, bugs and new good friends I had the best times looking for frogs and recording their calls. On an additional note, Madagascar wildlife is amazing, and the people extremely nice and helpful, which made my dissertation trip a once-in-a-lifetime experience.


Me in full field gear: recorder and headphones!

Overall, I was able to find two species which had not been recorded locally in previous surveys, (which I was extremely happy about!), counting nine frog species in total.

Heterixalus tricolor

Heterixalus tricolor, one of the local species

The use of SoundID proved to be very challenging due to the quality of recordings that I was able to collect with my equipment, a Roland-R05. Most frogs called from the middle of the water bodies sampled, and this meant that the signal strength was generally very low. However, my work showed that SoundID has the potential to be used in the field to aid the recognition of frog vocalisations, but more work is needed get better recognition rates, and better equipment is necessary to collect recordings with a higher signal-to-noise ratio. Perhaps one of the next BCMers will be keen to pick up this project and improve it! #teamfrogs

%d bloggers like this: