Biodiversity can be defined as the variety of plant and animal life in particular habitats of species. Ecologists have been trying to measure this for decades, through numerous different methods. As technology has become more advanced, new ways of measuring biodiversity have become apparent, one of which is through acoustic monitoring. Measuring the biodiversity of certain landscapes is crucial for numerous reasons: it can be used to assess the ecological health of an area, give information to conservationists on how much work needs to be done to protect it; as well as providing information on the abundance of certain threatened species' rarities.
After reading several papers on the use of Acoustic Monitoring to measure Biodiversity, I became interested in this topic and took recordings of forests that I visited, with the aim of analysing them to measure their biodiversities relative to each other. It is important to note that formal research done by companies to provide extremely accurate results uses thousands of hours' worth of recordings before analysing them, whereas my recordings averaged around 1 minute. This was due to limited time as well as increased ease of analysing shorter lengths of recordings by hand, instead of complex algorithms that formal research companies use. In investigations carried out, I focused on two distinct forests, both situated in Slovakia. The first recording was taken in Dubová, Svidník District, and the second in Banská Bystrica. As both are found around 25km from each other, at mid-latitudes, these are similar temperate forests. Although animals that dominate these forests include the brown bear and several species of deer, the use of video recording, along with longer periods of recording, would have to be used to spot these animals. Thus the following analyses of recordings are focused primarily on bird calls.
To complete my research, I used MAZTR, an online program that converted my mp3 recordings into spectrograms, which are essentially a visual illustration of a voice recording. The horizontal axis shows time, the vertical axis shows Frequency in kHz, and amplitude is shown through varying colours.
The following spectrogram shows a visual representation of the recording taken in Dubová, Svidník District.
There are several pieces of information that can be collected from the spectrogram shown above. Firstly, the peaks in Frequency at 1, 2.5, and 65-70 seconds are all due to human activity, i.e. myself moving away from the phone after starting and before stopping the recording, and thus should be discounted from the analysis. Furthermore, the streak at 0-0.5 Hz that is constant throughout the recording is some sort of background noise / wind, and should thus be ignored. However, the intensity of this noise, in this case, 40-60 dBFS (decibels relative to full scale), could suggest increased biodiversity within plants, as part of my previous research (also in this website under Featured paper) showed that increased wind speeds led to increased pollination within wind-pollinated species. Finally, the long streak (colour blue) ranging from 10-13kHz and from 10-60s is also background noise and should not be included in the analysis.
Due to the fact that the bird calls of one specific bird are at similar frequencies and do not vary substantially, this can be used to identify the number of species that can be seen in the spectrogram above. Primarily, there are at least two birds present which are singing in the range of 14-16 kHz for the majority of the recording. This is evident as one bird sound is usually in the form of a speck/dot, whereas at ~15 kHz at 35s, it is seen that there are numerous dots at the exact same time period. Although often calls can show up as up to 2 dots, the fact that there are 3-4 layered on top of one another suggests there are numerous birds of this species calling at the same time. This is highly unusual, as most birds sing in the range of 1-8 kHz, and only a few can reach up to 16 kHz. There is another bird singing at around 8 kHz, shown through 4 calls that it makes throughout the recording. Finally, the calls made at 25-45s at around 10 Hz signal another bird.
Assuming that the recording took in bird calls from a 100m radius (although birds can hear calls further than this), and 4 separate bird calls were noticed, this means that, on average, there is at least 1 bird per 0.0025 km squared, and 1 bird species per 0.0033km squared in the forest recorded in. Furthermore, there may be birds that were not vocalising during the recording that were not taken into account.
The following spectrogram shows a visual representation of the recording taken in Banská Bystrica. The quality of this spectrogram is much higher than the first one taken in Dubová. The reason for this is unknown, as the same recording tool was used (iPhone), downloaded into the same format (mp3), before being converted into a spectrogram with the same settings (2400 px width of image and 20kHz frequency range) using the same program (MAZTR).
Once again, the spike in frequency at the beginning and end of the recordings, the wind, and the background noise must all be ignored. As the layout of this spectrogram is different (e.g. one cannot view each dot at a similar frequency as an individual) increased densities of dots, or 'clumps' at similar frequencies over a period of time represent an individual organism. One example of this is at 5 kHz, where from 0-25s there is an increased density of noise. A shorter example of this is from 21.5-22.5s, which when listening back to the recording is not a bird, but an insect, perhaps a honeybee, that flew very close to the recording device, close enough to be heard. Finally, an individual can be seen calling at around 15 kHz, where throughout the period of the recording, numerous calls are made at this frequency. This shows that there are at least 3 species (or 2 bird species) in a 100m squared radius, leading to a calculation of 1 bird species per 0.0050 km squared, or one organism per 0.0033 km squared.
This preliminary investigation shows that the forest of Dubová (forest 1) may be more populated than the forest of Banská Bystrica in terms of bird species, which may suggest that this forest is more rich in biodiversity, as the species count of a specific area is weakly positively correlated to biodiversity. However, as the sample size is so low, the accuracy of this statement could be greatly increased with increased hours of footage, more recordings taken in different areas of the forest, video footage, and data represented in the same format.
To the left is an image taken of the forest in Banská Bystrica at the time of recording, and below is an image taken of Dubová forest at the time of recording. At first glance, a viewer of these two pictures would suggest that the Banská Bystrica forest is more dense in foliage/number of trees and thus would have a higher chance of having more biodiversity in terms of animal variety, as there are more areas to act as habitats for organisms. The Acoustic monitoring completed above suggests the contrary, that The Dubová forest is, in fact, more biodiverse. Although both statements take in a lot of assumed
factors, for example, soil pH could be different in the two forests, trees could be more habitable (i.e. easier to build nests on by birds) and so forth, the fact that both are reasonable statements to make yet are opposing shows the importance in creating new ways to tackle the complex matter of how to measure biodiversity. Once numerous different methods, which are of a certain accuracy, can be relied on to measure biodiversity, they can all be used together to visualise a picture of one area's biodiversity in contrast to another.
Thank you for reading this post on biodiversity, and feel free to list another way of measuring biodiversity that you would like me to research as above, in the comments!
Comentários