Compiling soundscapes to study global biodiversity

Sound record­ings of an envi­ron­ment can be used to cre­ate 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”.

Kévin 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: