Have you ever wondered why singing in the bathroom sounds different from performing on a stage? The difference lies in the space available for sound to propagate. In the small space of a bathroom, sound can reverberate, thereby making it more melodious. 

Though for humans, singing is typically a recreational act, for birds and insects living in dark, dense forests, it is a way they communicate with potential mates or warn friends about predators. Singing in the forest differs from singing in an open field because of the high density of trees and plants in the former. A recent study provides insights into how the removal of these trees and plants could change the way sounds propagate in forests. 

During COVID lockdown, after receiving complaints from locals, volunteers from the Foundation for Ecological Security (FES), with the help of the forest department, began removing invasive Lantana shrubs in the buffer region of the Kanha National Park, Madhya Pradesh, to restore the forest. This ecological restoration effort provided researchers from Project Dhvani the opportunity to assess how habitat changes caused by Lantana removal can impact the red-vented bulbul (RVB), the primary disperser of the Lantana seeds.

To do that, they needed to record the song of the RVB, for which they used passive acoustic monitoring. Pooja Choksi, a collaborator of Project Dhvani and a post-doctoral researcher at the University of Minnesota, USA, says, 

Passive acoustic monitoring is when you leave a recorder out, say on a tree, and after leaving it out for a long period of time, you use the data to study these sounds.

After obtaining the recordings, the researchers manually annotated the songs of the RVB, to separate them from the other sounds within the range of the recorder. These recorders were placed in three categories of sites: unrestored sites (with a high density of Lantana), low Lantana density sites (which naturally have a low density of Lantana), and restored sites (with no Lantana).

Once the songs of the RVB were retrieved from the recordings, they were assessed for different features of the songs, such as note length, bandwidth, and so on. The researchers found that the note length of RVB songs was shorter at high Lantana density sites, whereas the songs in low Lantana density sites had longer note lengths and narrower bandwidths. Although the researchers found differences across the sites, there is a catch! 

When researchers wanted to know how density of Lantana was related to RVB vocalisation, they ran an algorithm model and found some unexpected results. ​“When we used a random forest classification model, it showed us that Lantana density was not a significant factor influencing vocalisation. Instead, the percentage of forest around the recorders and the human population nearby played a major role,” says Mayuri Kotian, research assistant with Project Dhvani. 

Samira Agnihotri, currently working with Coexistence Consortium and an expert in field of bioacoustics, comments on the anomaly, saying, 

Phylogeny is often a stronger driver of the acoustic parameters of species-specific vocalisations.

Traditionally, researchers have evaluated ecosystem health and biodiversity by studying species’ presence or absence, but they have not explored how habitat changes influence species behaviour. Passive acoustic vocalisation can serve both purposes and along similar lines, is also commonly used to assess the impact of underwater submarine noises on marine creatures.