Cardinal bird in tree

Biodiversity, Birds and BirdNET (and Why They All Matter)

  Estimated read time: minutes

In the context of sustainable agricultural production, biodiversity is important in several ways. As biodiversity increases, ecosystem functions improve, ecosystem stability improves and overall production increases.

But how do we know if these things are increasing and improving?

While it may be hard to fully understand and quantify ecosystem functions and stability, that’s not the case with production. Production is where we get all our food and fiber and is therefore the ultimate goal of agriculture in general. We’re fairly skilled at weighing, counting and otherwise measuring production.

Production, however, doesn’t tell the whole story, and is a lagging indicator — the answer comes after all the dust has settled. There may be lots of good things happening in terms of biodiversity, but they won’t necessarily be reflected in production — at least not for a while, maybe even for several years. Production may actually decline while we transition to a more sustainable production system.

Measuring production, therefore, while informative for other purposes, doesn’t tell you much about biodiversity, especially in real time. Measuring biodiversity is hard. It requires considerable effort, some training, and, if really detailed data is needed, advanced training.

bird on fence with grasshopper

We need a shortcut to measuring biodiversity

Some organisms tell you more about the environment than others because they are more particular about their habitat. I can tell quite a lot about the health of your pasture and your management by looking at what species of plants are present and their relative abundance. I can identify what the available habitats are, and which are preferred by the different species. That skill takes training and experience to develop, though.

Most animals can do the same thing. They can quickly analyze a habitat and make a call as to whether it supplies their needs. Much as, when I look in the fridge, I can tell whether my wife has been to the grocery store, or whether I’m going to be taking her to town for a burger for dinner.

If our management objective is a healthy grassland, why not consult some on-ranch experts on healthy grasslands? In any habitat there is probably a group of birds that tend to occur together because they have similar habitat requirements — a guild.

Each species within the guild may prefer a particular extreme of that habitat, which can minimize intra-guild competition. Northern bobwhite prefer some brush, while western meadowlarks won’t necessarily tolerate it. These preferences may differ regionally as well. The key though, is that in any specific region, there should be a fairly well-defined guild of grassland birds.

So instead of asking the chucklehead who looked at your pasture for a minute and then went home and got his dinner out of the refrigerator, why not ask the real experts who make a living there 24 hours a day — grassland birds should be the experts on grasslands.

If your intent is to manage for a healthy grassland, (assuming the correct context – you’re not trying to do it in a mountain-side forest or a swamp) and you have birds that belong mostly to shrub or forest guilds, you have a problem.

For a moment there, I’d thought I’d written myself out of a job, but this issue saved me: The birds can tell you that you have a problem, but not how to solve it. Fortunately, at least for my sake, I can help you out there. If you’ve got the wrong birds, you probably need to investigate prescribed fire and growing-season burns. Those birds don’t know everything!

dove perched on fence

Using technology to document avian diversity

What about that shortcut I mentioned?

Ornithologists have been using bird calls to identify and estimate populations of birds for a very long time. There are a few drawbacks to that method, however.

One hurdle to estimating bird abundances is related to ornithologist abundances. A trained ornithologist is a rare “bird” – there’s seldom one around when you need them. Beyond their scarcity comes the fact that the most experienced ornithologists tend to have poorer hearing (we lose our ability to hear some frequencies with age) and are also somewhat less inclined to traipse through the pasture when it is still too dark to see the green briars lurking, waiting to trip and strangle them (wisdom comes with age).

Enter BirdNET, a neural network available from the K. Lisa Yang Center for Conservation Bioacoustics at the Cornell Lab of Ornithology. It can detect and identify birds by their calls from recordings.

BirdNET does a pretty good job of detecting calls in recordings. It is limited, however, when it comes to larger data sets like what would be collected over several days in a field sampling environment. There would be too much data collected to upload to BirdNET. There is a desktop version, but now we run into technical issues that most folks would have trouble overcoming.

BirdNET-Pi is a version of BirdNET capable of running on a Raspberry Pi, a small, single-board computer that, when connected to a mouse, a keyboard and a monitor, can act as a mini personal computer. This is a game-changer. Now, by adding a soundcard and microphone, we can record bird calls and analyze them in almost real time, without storing terabytes of data, and do all of it onboard the local device. The result is a list of names of the birds detected. Comparing that list to a list of what birds should ideally be present would be quite useful.

Each region might have a slightly different list, so enlisting the help of a local ornithologist would be advised. We still don’t know how many of each species there are because some individuals vocalize more than others but knowing which are present is a useful measure of habitat quality and therefore soil health. The question to be answered is “How many of the grassland guild species are present?”.

Using BirdNET-Pi we can monitor grassland birds, which are closely linked to grassland health, which in turn, is a direct reflection of soil health. You can’t have healthy grassland bird populations without healthy soils.

BirdNET-Pi has the potential to easily collect biodiversity data one can use, with minimal effort and a reasonable cost.

That’s a pretty good shortcut.

The components for a BirdNET-Pi system can be purchased readily off the internet. A well-written guide is available online that can be followed to install and configure the operating system and BirdNET-Pi. Deploying the system outside of the range of electricity and Wi-Fi is a bit more challenging and requires additional hardware and configuration, but it can be done.

Want more details on my adventures building and deploying a BirdNEt-Pi in the field? Visit my blog.

Mike Proctor

Mike Proctor serves as a senior research associate at the Noble Research Institute, which he joined in 2010. A field botanist by training, Proctor is especially interested in the role of fire and herbivores (including cattle) in maintaining diversity in rangelands. Follow him on Instagram @rangeroaster.

Article Reprint

For article reprint information, please visit our Media Page.


Leave a Reply

Your email address will not be published. Required fields are marked *