Compiling soundscapes to study global biodiversity

Sound record­ings of an envir­on­ment can be used to record sound­scapes – a pro­cess that is very use­ful for research­ers, as it enables them to identi­fy and mon­it­or the meas­ur­able biod­iversity in a giv­en a loc­a­tion. This tech­nique, which is used in eco­logy, can be applied both on land and under­wa­ter. How­ever, access to this data remains dif­fi­cult because there is cur­rently no glob­al record­ing data­base. To address this issue, Kév­in Dar­ras, a research fel­low in eco­logy at INRAE (French Nation­al Research Insti­tute for Agri­cul­ture, Food and the Envir­on­ment), has come up with a solu­tion by com­pil­ing these sound­scapes in his pro­ject entitled “World­wide Soundscapes”.

Kev­in Dar­ras. Around the world, many research teams col­lect audio record­ings to study biod­iversity. This meth­od is called ‘pass­ive acous­tic mon­it­or­ing’. How­ever, this data is rarely shared among sci­entif­ic com­munit­ies. World­wide Sound­scapes cata­logues record­ings from more than 12,000 sites – approx­im­ately 5,900 TB of data – and makes them avail­able to all scientists.

Pass­ive acous­tic mon­it­or­ing is a com­pre­hens­ive meth­od for record­ing and track­ing anim­al biod­iversity. In prac­tic­al terms, it involves record­ing sounds pass­ively using a micro­phone and a record­er – unlike son­ar, which emits sounds. Pass­ive acous­tic mon­it­or­ing can be used for all eco­sys­tems, wheth­er ter­restri­al, aquat­ic or even under­ground. It is an effect­ive, accur­ate and veri­fi­able meth­od. Depend­ing on the spe­cies and envir­on­ment, it is pos­sible to record anim­als sev­er­al dozen metres away, such as a robin in a forest, or even sev­er­al hun­dred metres or a few kilo­metres away, such as the songs of orcas, for example.

Firstly, listen­ing to the record­ing makes it pos­sible to determ­ine the pres­ence of dif­fer­ent sound-pro­du­cing anim­al spe­cies. Birds, bats, ter­restri­al and mar­ine mam­mals, insects, amphi­bi­ans, etc. can be iden­ti­fied. Using stat­ist­ic­al mod­els, it is also pos­sible to estim­ate the num­ber of indi­vidu­als present in a giv­en area. These meas­ure­ments form the basis for a sig­ni­fic­ant num­ber of decisions regard­ing the man­age­ment of nat­ur­al hab­it­ats, the mit­ig­a­tion of the harm­ful effects of urb­an­isa­tion, etc.

I used pass­ive acous­tic mon­it­or­ing to track biod­iversity for my research on trop­ic­al agroe­co­logy. Many col­leagues who study mar­ine, fresh­wa­ter or ter­restri­al envir­on­ments also use this meth­od, and I real­ised that it would be use­ful to take stock of pass­ive acous­tic mon­it­or­ing on a glob­al scale. A shared data­base provides inform­a­tion on record­ing sites and peri­ods, as well as the spe­cies iden­ti­fied. This gives the sci­entif­ic com­munity and man­agers an over­view of the regions already covered by mon­it­or­ing. The advant­age? It is entirely pos­sible to reuse a record­ing to identi­fy oth­er spe­cies. This allows sci­ent­ists to identi­fy areas that have nev­er been mon­itored and those where data already exists.

We have char­ac­ter­ised the sampling dens­ity on a glob­al scale. Of course, this was not a sur­prise, but we high­light that there is much more data in the North­ern Hemi­sphere than in the South­ern Hemi­sphere, that there are sig­ni­fic­ant gaps in Cent­ral Asia, and that spa­tial cov­er­age dens­ity is great­er on land than at sea. We observe that the data cov­ers the vast major­ity of eco­sys­tems. What’s more, in an ini­tial sci­entif­ic pub­lic­a­tion, we show – using a small selec­tion of records – that this data­base can be used to answer eco­lo­gic­al ques­tions on a very large scale, which is unprecedented.

We are still in the early stages of data exploit­a­tion and are find­ing res­ults that are already known in mac­roe­co­logy (i.e. large-scale eco­logy). For example, we observe that biod­iversity decreases as we get closer to the poles. Although this was already known, this obser­va­tion required extens­ive ana­lys­is and strong assump­tions. With the data­base, it is now pos­sible to do this eas­ily with a single stand­ard­ised method.

We also find a neg­at­ive rela­tion­ship between nat­ur­al anim­al sounds and human-made sounds. This is an indic­at­or of anthro­po­gen­ic pres­sure on eco­sys­tems. On the oth­er hand, some eco­sys­tems seem to be little affected by anthro­po­gen­ic noise: this shows that there are still many areas for research, but also pos­sib­il­it­ies for coex­ist­ence between humans and nature.

There are many ques­tions that can be addressed. We are cur­rently ana­lys­ing the sounds in the data­base: the aim is to identi­fy glob­al pat­terns of biod­iversity, or links between eco­lo­gic­al gradi­ents and a cer­tain dis­tri­bu­tion of biodiversity.

Using this new data, I would like to study the link between biod­iversity decline and cli­mate change. The glob­al scale is very appro­pri­ate, as cli­mate change affects the entire plan­et. By identi­fy­ing cli­mat­ic con­di­tions that are harm­ful to biod­iversity, it would then be pos­sible to adapt to them. Anoth­er very inter­est­ing ques­tion is the effect of human activ­it­ies on biod­iversity. By identi­fy­ing the con­di­tions that are least harm­ful to biod­iversity, it may be pos­sible to bet­ter pre­serve biod­iversity by intro­du­cing new legislation.

Legis­la­tion requires robust data, and the estab­lish­ment of an acous­tic obser­vat­ory – a net­work of sensors – would meet this require­ment. Such nation­al obser­vat­or­ies already exist in Aus­tralia, Canada and some South Amer­ic­an coun­tries. This would make it pos­sible, for example, to detect the first signs of a change in biod­iversity. Ima­gine that invas­ive spe­cies are observed in the south and begin to spread north­wards: it would then be pos­sible to take pre­vent­ive action for regions not yet affected.

Interview by Anaïs Maréchal

For more information: