In a Nutshell
The dawn chorus of birds is a natural phenomenon that invites scientific studies on various levels and complexity. In addition to answering specific ornithological questions, the dawn chorus also provides information about the current state of the surrounding environment and how it has changed over time. However, complex acoustic datasets are also valuable for computer science to develop artificial intelligence systems e.g., to train artificial intelligence to distinguish different sounds and filter out various features.
Your Dawn Chorus recordings will be a continuous invitation to scientists of various disciplines to use the material for their research questions.
Background
The dawn chorus as a scientific term describes the birds’ joint vocalizations before sunrise.
Since time immemorial, the dawn chorus has fascinated and inspired people time and again, especially during spring when the voices are most diverse.
Birds mainly sing during this time of the year to claim and defend their territories and attract mates. The time before dawn is the most crucial period for most bird species (read more).
On a scientific level, there are many questions from very different research directions associated with the dawn chorus, from questions such as “why do birds sing at the same time?” to “the physiology and learning behind bird song” and the importance of “bird song in human emotion”. There is interest from all kinds of scientific directions, ranging from physiological questions dealing with the hormonal state of the animals to questions on how birds learn their songs and how dialects, “bird culture,” spread across time and regions. The scientific interest reaches into areas of human psychology as to why birdsong is considered aesthetic and why it can foster human wellbeing.
The Dawn Chorus recordings are an excellent tool for biomonitoring. Which bird species are present in an area at any given moment and how does this change over the years?
Can warning signs be detected early or resort to protective measures? In theory, almost only birds are documented acoustically, but behind each bird is a network of environmental dependencies and this is what makes the recordings so valuable for much deeper analysis. If recordings for example show a shift from insectivorous birds to seed-eating species, then this could be a sign of a decline in insect availability. If timber is cut or if swamps are drained, this could also have an immediate effect on the acoustic features of a landscape at a given location.
The special time period of the dawn chorus marks the moment of highest bird activity and hence promises the best acoustic selection during a day.
Given this, the dawn chorus offers plenty of opportunities to engage scientifically, asking fundamental questions that have no immediate use but serve to learn about nature, but it also extends well into the applied sciences by addressing habitat health and the current dramatic species decline and measures to counteract them.
Dawn Chorus’ Variables
To understand why it is so important to get multiple recordings of the dawn chorus at any given location, one has to understand that the dawn chorus is not a statistical phenomenon, but rather dynamic and influenced by many different factors.
Already the concert on a given morning is dominated by different “orchestra members” during its course at dawn. Probably to escape acoustic competition, different species join the choir at different light levels.
Also, the progressing spring lets us hear different individuals and species. Some singers are just returning from the south, while others have already fallen silent as they have started breeding or leaving for other areas. Weather is another shaping factor, as is human activity, especially noise and light pollution, both of which can dramatically alter singing behavior.
Last but not least, changes from year to year are also of interest. Fluctuating populations, global species declines, and even local habitat changes show their influence in the dawn chorus.
All those factors must be considered in scientific models and hence it is important to get as many data points as possible to statistically determine correlations and draw the necessary conclusions.
To identify as many of these variables as possible, we would like to ask you to make several recordings at the same location on a given morning on a workday, as well as on a Sunday or holiday. (Can you hear a difference?).
For this, all recordings from habitats, even those that are seemingly boring and offer only a few species, are scientifically valuable and important!
The “jackpot” for the project is recordings from the same place on the same date and at approximately the same time of the day, recorded over many years.
Citizen Science
The Dawn Chorus project is dependent on engaged citizen scientists like you who contribute with their recordings to a solid database. But not only can you support Dawn Chorus with your recordings, you can also help us with your expertise in bird song recognition (read more).
Different from traditional studies in this research field, Dawn Chorus does not primarily collect long professional recordings from a few field sites, but rather ties in the help of many people (and their smartphones) to record many short snippets which come together like tiny puzzle pieces to form a greater picture.
No scientist alone would be able to collect synchronously repeated and over many years running recordings from all over Bavaria, all over Europe or even the whole world.
The advantage of Dawn Chorus over many other Citizen Science projects is that with relatively few instructions, citizen scientists and professionals alike can collect data without the need for advanced equipment or knowledge. These recordings are a valuable and objective resource that contributes to a database fulfilling scientific standards. This is because the recordings stay available for later reconsideration and are not dependent on the species recognition skills of the participants. To further strengthen the comparability of the data in 2021, the Dawn Chorus App was developed under scientific criteria.
Under the participation of many, a treasure of data is growing that will gain future insights not yet imagined. At the same time, the accumulating, freely available recordings attract the interest of scientists who are dependent on such data sets.
If you feel attracted to the growing set of Dawn Chorus recordings and want to draw your own conclusions from the data, you are warmly invited to integrate the collectively assembled material into your research. Please feel free to contact Dawn Chorus at info@dawn-chorus.org or via the contact form on this page. To learn even more about the project, you can find a detailed report on the 2020 project here.
Data Analysis
To analyse the data, we must be able to recognize the singing bird species unambiguously. This is a long-term, complex undertaking. In fact, some apps and algorithms exist which do recognize the species of a bird singing alone but the full Dawn Chorus with its complexity, overlapping songs of many species and undirected recordings make recognition a challenge.
Hence, the first task of research is to develop methods that help translate the recordings into datasets that scientists can work with.
Citizen Scientists can support us not only with their recordings but also with the analysis of the recordings!
Computer programs for recognizing bird songs need training to identify species correctly. In the case of the Dawn Chorus, the training is very intensive because different singers overlap and mask each other in this kind of recording. The learning happens by uploading recordings that have been curated and annotated so that each bird has an accurate timestamp of when it sings, so that there is a listing of when which species sings. The more such known snippets are fed into the system, the more precise an algorithm can predict unknown recordings. Thus, the limiting factor is not only the total number of recordings but also the number of recordings that have been annotated by informed experts and bird lovers.
That’s why we appeal to all of you who do know about bird songs to help us with the annotation!
On the one hand, this works through the Dawn Chorus App, in which you list heard species for all your own recordings, but it is also possible that other participants add or correct species. Currently, these proposed changes are collected behind the scenes and will be released later.
We also want to organise a workshop to which we want to invite all lay and professional ornithologists and bird lovers who are knowledgeable about bird song recognition to annotate together some of Dawn Chorus’ recordings. Caught your attention? Please contact us at info@dawn-chorus.org or through the contact form. We are excited to hear from you!
Scientific Projects:
Long-term studies of species composition or changes in song variation will only be possible in the future when more recordings are available, but already some projects took advantage of the Dawn Chorus’ data:
Data Challenge: Worldwide Tinkering for Science
One example from the IT field is the development and training of algorithms which should allow identifying the bird species in the sound recordings automatically.
The DCASE Data Challenge is a free competition for teams of scientists and other techies interested in data analysis which is organized yearly, since 2016.
Each year, the organisers release datasets and certain analysis tasks for participants to solve. The algorithms developed by the teams compete against each other to find the best solution to the problem. Through this open and worldwide competition, new and surprising ideas and synergies are created that advance science in this field.
In 2021, Dawn Chorus was represented in Task 5 – Few-shot Bioacoustic Event Detection, with Dr. Lisa Gill, project lead of Dawn Chorus, and a small pre-analysed subset of data. In collaboration with an interdisciplinary team, the first results got published in a scientific paper.
In 2022, Dawn Chorus is taking part in DCASE once again, and already received its first submissions (visit website).
BioWaWi: Citizen Science and Bioacustics for the Water Management
BioWaWi, a project of our scientific partner DIALOGIK, is an example from the field of applied sciences.
For the supply of drinking water, the water catchment area is an essential part of value creation.
Only areas with functional and healthy ecosystems warrant contamination-free drinking water. Biodiversity and healthy ecosystems are important for a more sustainable and climate-adapted facility management. The goal of BioWaWi is to investigate the impact of water management on biodiversity and to identify the conditions that must be met by water management in order to sustainably ensure the preservation of biodiversity. Based on the project results, the municipal utilities of Bühl are introducing an environmental management system that should also be transferable to other water supply companies in other cities.
In the citizen science part of the project, citizens are invited to support biodiversity monitoring in the collection areas. Together with DIALOGIK, the participants work out ways in which they want to contribute to the process. Reliable and easy to generate data sets are important to collect information (e.g. audio data or pictures) about the biodiversity of the target area.
Within this project, citizens are asked to collect animal sounds using the Dawn Chorus App. Afterwards, these recordings will be connected to specific landscape types using artificial intelligence.
A parallel study in social sciences simultaneously explores the opportunities and challenges of citizen science projects in biodiversity research.
More to Read
Why is there a Dawn Chorus?
The dawn chorus probably evolved because at this early time of the day, air movement is at its lowest, as the night has equalized temperature gradients, and sound can carry further than in more turbulent air.
Another theory is that the birds take advantage of the twilight at dawn, as they are less visible to some enemies during this time while exposing themselves through their song.
Acoustic Biomonitoring
Biomonitoring is the regular observation of an ecosystem, biotope or community or individual plant or animal species over time to assess environmental quality. The more regular and standardized the data are collected, the better the quality of the monitoring and the conclusions drawn from it. In the case of citizen science projects, one tries to overcome certain problems of data quality by the amount of data, still it is necessary to collect the information as standardized and carefully as possible. The big advantage of such an approach is the sheer amount of data that can only be collected with a large collective of volunteers, which statistically balances out the lower precision of individual records.
Acoustic biomonitoring has the further advantage that recordings, once uploaded, are accessible for later analysis, and the individual species can be verified by experts at a later stage of the project. The area covered by undirected audio recordings is far bigger than, e.g., the area which can be captured in photographs. In contrast to pictures, which highly depend on the camera type, optical quality and light conditions, the microphone covers a much larger area, even if places are out of sight, and these recordings can be evaluated in great detail later.