How the BirdCast Team Created Super-Accurate Bird “Forecasts”
[Lisa Kopp] So welcome to today’s webinar from the Cornell Lab of Ornithology today we’re going to be talking about migration, an important and exciting topic for this time of year. But before we get started with today’s webinar, I do want to read the land acknowledgment that is representative of where Cornell University is located in Ithaca, New York.
Cornell University is located on the traditional homelands of the Gayogohónǫ’, the Cayuga Nation. The Gayogohónǫ’ are members of the Haudenosaunee Confederacy, an alliance of six sovereign nations with a historic and contemporary presence on this land. The Confederacy precedes the establishment of Cornell University, New York State, and the United States of America. We acknowledge the painful history of Gayogohónǫ’, dispossession, and honor the ongoing connection of Gayogohónǫ’, people, past and present, to these lands and waters.
For those of you who aren’t familiar, the Cornell Lab of Ornithology is home to a community of researchers and supporters from around the world who appreciate birds and the integral role that they play in our ecosystems. And our mission is to advance leading-edge research, education, and citizen science that helped solve pressing conservation challenges. And we’re to be talking about all of that today.
This work, including today’s webinar, is primarily funded through donations from people like you, so if you enjoy today’s webinar, I hope you’ll consider being a donor. You can do that by visiting birds.cornell.edu. I am Lisa Kopp. I’m on the Visitor Center team at the Cornell Lab, and I’ll be facilitating today’s conversation. And with us is Andrew Farnsworth. Andrew–
[Andrew Farnsworth] How are you?
[Lisa Kopp] Good– so we are about to get into talking about Andrew’s work, but I just want to run through a couple of quick tech pointers for today. So for our Zoom audience, we have live captioning. If you go to the bottom of your Zoom screen and click on Transcript or the three dots to view or hide transcript, you’ll be able to view those captions.
I will be asking Andrew a couple of questions to start us off. We’re going to talk a little bit about what BirdCast is, how it’s useful this time of year, but we’re also going to be looking to you all to help guide our conversation. So please use the Q&A tool in Zoom to ask your questions, and I will do my best to relay the most popular and pressing ones to Andrew. But we also do have some team members who are behind the scenes also helping to answer some questions, if we don’t get to everything verbally.
We also have the chat function in Zoom, but that is only going to be used for tech needs. So if you’re experiencing video or sound issues, shoot us a message in there, and we will do our best to help you out.
And then we are also streaming live on Facebook, so hello to our Facebook audience. We don’t want to leave you out of the fun, so if you have questions or are experiencing any issues, please use the Facebook comment section. And again, we’ve got two members who are helping to monitor those and pass along questions to me and to Andrew so that we can get everybody’s– as many questions as we can answered. So last thing about Facebook is that we have had some spam and bot issues, so don’t click on any links that aren’t coming directly from the Cornell Lab of Ornithology.
So I think that those are all my announcements. We can go ahead and get going. So thank you again, Andrew, for being with us today. I’m really excited to talk about BirdCast. But before we talk about that, can you give us a little bit of background information on who you are and what you do at the Lab?
[Andrew Farnsworth] Sure. Thank you, Lisa, and thanks, everybody who is watching and listening in the audience. I know that there are many of you from all around the US and around the world. I’m thankful that you’ve taken the time to tune in today. So as I said, my name’s Andrew Farnsworth. I’m a senior research associate at the Cornell Lab of Ornithology in the Center for Avian Population Studies.
I direct the BirdCast project, which really is all about migration ecology that we’ll be talking about shortly. And just so you know, I started this from the perspective of a birder– 40 some odd years ago– and came to it as a scientist second. So hopefully we’re going to cover all of those bases today.
[Lisa Kopp] Yeah. That’s fun to hear what drew you into this work. I know around the Lab I sometimes hear people referring to their spark moment, the minute that something clicked as, wow, this is incredible, or this behavior is just sending me on this journey to explore. Do you have one of those moments that you want to share?
[Andrew Farnsworth] Yeah. There are a lot. I have a few in particular that I’ll share quickly. The first one starting birding– a male Wood Duck for me. One January morning in Rye, New York, seeing that spectacular array of unimaginably combined colors– that was the kickstarter for me.
When it comes to thinking about migration, hawk migration was the thing that did it. As a seven or eight-year-old, seeing thousands of Broad-winged Hawks near where I grew up in Rye, New York– in particular, end of the day migration, where birds are kettling and swirling around in these incredible circling flocks, and the sun is flashing off of them, and they’re low enough you can really see them– that really did it for getting me hooked on migration.
And then, thinking about nocturnal migration– which I’ll tell you about shortly– is really what happens most for birds that are on the move, in terms of their migration strategies. That was a little bit later, but flight calls are what really drew me in– these calls that birds use at night to keep in touch, to communicate, also to signify disorientation. These kinds of experiences crystallized what became a– thankfully, a career in science that I’m really fortunate to do this thing I love, and study migration in the Lab.
[Lisa Kopp] That’s so fun to hear the different stages that brought you to where you are today and what you’ll be sharing with us. So BirdCast– it’s in the title. We know that it’s, as you just said, the project that you lead. Could you talk a little bit about what that is, what that means?
[Andrew Farnsworth] Sure. At its core, BirdCast is a project that really seeks to understand and inform all of you where, when, how many birds are migrating and will migrate. It’s focused, in the last really 10 years, heavily on using radar data that we’ll talk about today, but it does cover a lot of other bases, including where and when species are on the move– from citizen and community science data– also does occasionally cover flight calls.
But it’s really about birds on the move, and their lives on the move, and understanding that, and characterizing it in the best ways we can, and then taking that really cool science and turning it into conservation action, education, opportunities to spread the word and get you excited about migration.
[Lisa Kopp] I think that this project is so fascinating, and I’m excited to talk through all those different layers that you just mentioned. There’s the– how this can help you see birds in your backyard, but also how it’s really helping to conserve birds and provide information for researchers and scientists too. So starting at the most brass tacks level, how did the idea of using weather data to track birds come about?
That’s not something that I, prior to working at the Lab, really understood was a possibility. I just thought that those radar– the radar was for cloud activity and told you when it was going to rain, and it was what you saw on the Weather Channel app when it was green, or really heavy rain was red. How is it being used for tracking bird migration?
[Andrew Farnsworth] Well, the concept is actually really 70 or 80 years old. It actually really just predates World War II, when the first radar technology was deployed to track planes. And early in that experience, some of the really highly qualified radar operators were noticing these patterns on radar that clearly were not planes, and astutely observed that they were birds.
So this idea of thinking that you could detect birds on radar is actually quite old. As the military use of radar transitioned to more civilian use, weather surveillance radars and networks of those evolved. And using the same basic physical properties, same microwave pulses of energy from an antenna to survey the atmosphere, those detect meteorological and biological things that are up there.
And birds are a huge, huge part of that activity in the atmosphere. So it’s been a long-standing application potential for radar, but it’s really only been, say, in the last 50 years that it’s sort of come into its own– and truly come into its own, I think, since the BirdCast project has taken off and started to allow us to look across from the individual tens of kilometer scale all the way up to what’s happening across the continent.
[Lisa Kopp] Yeah, the history of it is fascinating. So if radar can track a whole lot of things, how– and actually, someone’s already asked the question that I knew I was going to ask you– is, how is it able to differentiate what’s what? How do you know what’s rain versus butterflies, as one of our participants asked, versus bats or– and birds?
[Andrew Farnsworth] That’s a really good question. Let me share my screen with you right now and show you just an example of what we’re talking about, when it comes to weather surveillance radar data. So I’ll frame it up in terms of the network in the US, and then talk a little bit about the data, and get into how we do this at a very general level.
So picture the continental US, and then each one of these dots is a weather surveillance radar in the current network that we use now. Some people refer to it as NEXRAD, the next generation radar. So it’s really the second weather surveillance radar in the US. Some people will refer to it WSR-88D, Weather Surveillance Radar 88 Doppler, which identifies what the sensor is.
And there are 143 of them in the continental US, and what they do is they survey the atmosphere, and they produce– each one of them gathers data, sensing the atmosphere around it to several kilometers away. And we can produce a mosaic like this of the radar data. And what you see here covers all of those different kinds of targets, if you will, the things– the objects that are up in the atmosphere that you asked about.
There is biology here. There is meteorology. So in that biology, there are birds, there are bats, there are insects. And in the meteorology, there is precipitation of various types and intensities that highlights where different meteorological phenomena are happening, where boundaries between air masses are, where there’s rain.
So just to orient you quickly, you’ll notice that there are some really blocky areas of very intense reds and yellows on this map, and that corresponds to really intense precipitation. You’ll also notice this much more uniform pattern of greens and blues, and that generally corresponds to the biology. So if we’re to remove those, highlight them, and take them away, you get the sense of what taking out the meteorology and leaving the biology looks like.
Now, the complexity of identifying first between meteorology and biology– well, that takes a lot of looking at images and a lot of labeling them from experts that understand, oh, this is precipitation, this is biology, labeling those images, and then training computer models to recognize those different cases. When it comes to the biology, it can be a little bit more challenging, because insects, and birds, and bats all overlap in particular places.
So we use a few different techniques to try to identify those, including very simple ones– for example, understanding that bird migration occurs in very particular times of the year, and understanding, when there are birds aloft, that, because of the characteristics of the radar, they really dominate what the radar samples. So when birds are present, for example, and insects and bats are present, birds often will dominate that signal.
Can you recognize them in other ways? Well, yes. If you look at some of the other kinds of products that the radar can produce, birds and insects occasionally move differently in the atmosphere. Many insects drift with the winds, while birds can really fly across the wind, or even into the wind. So there are different patterns like that that we can use of their speeds and directions that also help us identify what’s happening– and so too the patterns of, for example, bats emerging from caves, where they expand in these incredible numbers from very particular locations on the landscape, where we know there might be millions and millions of bats– like coming out of Bracken Cave in Texas in the spring, for example. So we can understand where those things are happening too. So it’s really a combination of all of these different things that we’re doing to identify what’s there.
[Lisa Kopp] Wow. Yeah, so it sounds like a combination of people really studying and understanding these things, and then training machines to help understand them over time? Is that accurate?
[Andrew Farnsworth] That’s right. That’s definitely accurate– absolutely.
[Lisa Kopp] We’re getting some really great questions. I just want to mention to everyone that some of this is going to– we’re going to get to those questions later. I’m not going to ignore your questions, but we’re going to talk about some of the great things that you’re asking already. So obviously, it’s fascinating to be able to just see the movement of birds to understand where they’re going and when they’re going, but what’s the mission or what’s the goal of what BirdCast is doing by reading these maps and understanding this data?
[Andrew Farnsworth] So we really want to do three different things. There’s science that we want to learn. There are things we don’t know about where and when birds migrate; how many birds are migrating, when they’re on the move, with what those kinds of movements correlate; and then really be able to understand, how are those patterns changing? Why are they changing? The science– there’s no question– there’s one avenue that we’re traveling there.
The other two are equally important to us. One is outreach to– from, again, the birding perspective, the observation, the natural history, the engaging people to understand the spectacle of migration, and taking this kind of a tool– which is very much indirect, in terms of experiencing what’s out there– and helping to guide people to– well, how might they experience it, and really understand what the spectacle is from an in-person perspective.
And then finally, the– taking that science and that engagement, and promoting the excitement, and applying that to conservation– so conservation actions. How can we change our behaviors to protect birds that are on the move, since birds are these incredible indicators of ecosystem health and a way to gauge in a very quick manner, how is our planet, what’s the health of our planet? Taking advantage of that science to inform actions that can protect birds, and then scale up from there– that’s a really key component of the work as well.
[Lisa Kopp] Yeah. So it sounds like there’s sort of a three-pronged approach. There’s the monitoring, there’s the forecasting, and then there’s the science to action or the conservation piece to it also.
[Andrew Farnsworth] That’s right.
[Lisa Kopp] That’s a lot. That’s a lot you guys are trying to work on right now. Yeah. Can we talk a little bit about those predictions, and how the radar is allowing you to predict what will happen next, and what that means for migration or people who are curious about what migration might be happening in their area?
[Andrew Farnsworth] Definitely, yeah– let me share some more with you to– so you can see what we’ve got here, in terms of thinking about these forecasts. So one of the components of making these forecasts is obviously that we have these radar training data. So we have information from 25 years of this sensor network operating in the US, and collecting and archiving those data.
And they’re freely available to the public. It’s a federally funded agency that has collected them. And so we take these data, and again, we’ve got the humans in the loop helping us understand what’s birds, what’s precipitation, et cetera. So we take those radar data and then we start to learn associations with when birds are on the move and different kinds of weather information.
So you all are familiar with weather forecasts and the kinds of information you can get out of weather forecasts, including what the temperature’s going to be, whether there’s going to be rain, what direction at speed is the wind? And this kind of information we can then relate to these patterns that we see on radar, and that is what produces our migration forecasts. So taking all of this information, both from the radar perspective and– nearly 25 years of it– and co-collected weather data, all these meteorological variables that we have at our disposal, and relating those two things, and then looking at how those predict certain kinds of conditions when high intensity migration happens or when no migration happens.
And the model, it turns out, is very good at that. It was published a few years ago. Benjamin Van Doren and Kyle Horton, who published this paper, really evaluated this model, and it’s the one we currently use today. We’re quite confident in it. It’s very much reliable across years– so a really exciting opportunity. And the kind of information it produces, the kind of map you’d see– what does the forecast look like?
Well, here’s the forecast for tonight. And what you see here is obviously a map of the contiguous US. And before the question comes up, we don’t show Canada on here and we don’t show Mexico on here, because we can’t access those data yet in the same way that we can access the US data. We would really like to do those areas, and hopefully in the future– but so what you see is the contiguous US.
And you’ll notice this map that changes from the very dark and purplish colors ranging up to very bright yellows and whites. So that scale ranges from very low migration up to very intense high migration. And in this particular map, for tonight, you’ll notice the upper Midwest, and the Eastern Great Plains, portions of the coastal plain, and the Southeastern US really light up with moderate to high intensity migration.
And I also want to call your attention to the fact that we do predict a number of birds that we expect to be aloft, and that’s in the bottom left of this image. That is a correct number that you see. That is in the millions, so just over 400 million birds aloft, are forecast for tonight. And we can do this for 24 hours in advance, 48 hours, 72 hours in advance. We do have a little secret that I can share with you that also allows us to go farther out into the future.
I highlight this, even though I probably shouldn’t. But I’m going to let you in on the secret. In the beginning of October, we’re expecting a really massive movement. This is over 700 million birds. Because we have weather data that go out two weeks or three weeks in advance almost, we can do forecasts that far. But generally, as you know from weather forecasts, the information and the quality of what we know about the weather two weeks in advance– well, sometimes we’re not so good at it. So I’m not supposed to show these, but I thought you might like to see what’s going on.
But back to tonight, just to orient you from the perspective of, what kind of information can you glean from these, other than just, whoa, there’s going to be really big migration in particular areas– well, thinking about how you might go birding, for example, as a function of this, those areas in the upper Midwest, especially in the southern reaches– really central Mississippi River Valley, Southern and Eastern Great Plains– birding tomorrow morning in some of those areas could be really interesting.
Checking whatever green habitat you have, visiting places where you see birds continuing their nocturnal movements into the day– really good opportunities to visit– and so too along the Atlantic coast of the US as well, where there’s going to be nice movement– thinking about areas where you might visit in the coastal Carolinas or in the Delmarva Peninsula. You guys, and those in the upper Midwest and central Mississippi River Valley, you should be in for a treat tomorrow.
And I should highlight that this map is about when peak migration usually occurs at night. So this is about three hours after sunset. So when you think about what’s on this map, and where you are, and where the birds end up– because they’re still actively migrating at this snapshot we’re predicting– a few hours beyond that, a few maybe hundreds of kilometers to the south or east of where you’re seeing these pulses, those are the areas that are going to get the inputs of birds.
[Lisa Kopp] It’s really exciting to see that kind of tool available for people who want to be able to go out at night and observe what’s happening above them. We’re getting some questions about out west, how it looked like on that first map you showed that there’s less radar stations out west. Does that mean that it’s harder to understand what’s happening out there? Is there a lack of data, or is it made up for in other ways?
[Andrew Farnsworth] That’s a really good question. So one of the challenges that we experience with radar, in particular in the Western US, is there are things that block the radar beams. In particular, in the West, dealing with varied topography and mountains, there are a lot of places where the radar just cannot survey. And so too there are places where the radars are, of course, above those mountains so that it can survey the atmosphere.
And as you get higher and higher, you very quickly get out of the range where birds are occurring. So any radars that are elevated tremendously may be only sampling the very, very top of where– or the very bottom of where migration occurs, because they’re so high, and birds are really not up higher. Or there may be beam blockage, and the radars are actually not covering so well.
So part of the reason that the west often looks a little bit different than the east has to do with that. Part of it does have to do with the numbers of birds that are aloft too. Generally, east of the Rockies– and really, in the center of the country– that’s the thoroughfare through which most of the bird migration in the country passes. So there is also a bias from that perspective, just in terms of the numbers, that, if you– as you get east of the Rockies, for various reasons, there are many, many more birds as well. So it’s a combination of those things for why we get that pattern on radar. And we can account for that a little bit, but– great question.
[Lisa Kopp] Another question that’s popping up related is, what altitude are birds usually migrating at? Or are the radar stations measuring between the range there?
[Andrew Farnsworth] That’s a very good question too. So these radars can sample at various different elevation scans. They can turn the antenna a few degrees up, and that allows very easy sampling of up to 10,000, 15,000, 20,000 feet– even higher. So think about thunderstorms and the tops of some thunderstorms, 20,000, 30,000, 40,000 feet above the ground.
That’s way above where most bird migration occurs. Yes, some do occur that high. Some birds are migrating at 10,000, 12,000, 15,000 feet above the ground, but really the bulk of it that we’re talking about is in that below the 2,000 or 3,000 foot above the ground range. So that’s really where a lot of it occurs. It does vary– it varies by night, it varies by location, and it varies by species. But really, that big pulse, thinking about the lowest 3,000 feet above the ground, is really what we’re talking about.
And the radar can do a pretty good job of sampling that. It does not sample perfectly across the entire US. There are, as I mentioned, 143 of these. So really, there are almost 143 donuts where we can sample what’s going on, and then we have to do some interpolation in between to understand, well, what might it be, and project what it might be, given the speeds and the directions birds are moving. But nonetheless, it’s samples well in that 2,000 to 3,000 feet above the ground range.
[Lisa Kopp] And I would guess that some of that changed– bird behavior’s going to change based on what’s happening in the atmosphere. So we’re getting a lot of questions about– you talked about separating out weather from the biology, but I would also assume that weather affects the biology, or fires affect the biology. So are you able to pick up that kind of information on the maps too, when there are extreme things happening in the atmosphere?
[Andrew Farnsworth] Yeah. So we can certainly look at cases where we see intense migration occurring either in advance of or after major frontal systems that pass, depending on the season. So too with hurricanes and tropical cyclones, when they come into the country we can see– even though it’s a little bit different than migration, there are times where we can see biological activity in the center of the circulation of those storms.
So absolutely, there– we do, in many respects, want to remove the weather, when it comes to visualizing, well, where and how many birds are occurring, and how does that change? But we’re also very cognizant of the need to maintain the weather information to understand how birds are, one, using the conditions through which they’re flying, and also how those conditions– especially the extreme conditions– may affect them. And not that goes for the very local all the way up through thinking about climate, and how changes in climate may affect migration.
[Lisa Kopp] Mm-hmm. That makes sense. We’re getting some questions about more specific birds or just understanding the specific details. And what you’re talking about is scale. We’re looking at massive scale with BirdCast. And I can mention that, on Wednesday, we did a webinar with the eBird Status and Trends team. And, actually we talked about BirdCast in it, and how the– these two projects are really closely linked, because you can put the information together and get some really very valuable information to understand what you might see and when you might see it.
So for those of you who didn’t tune in on Wednesday, I would encourage you to check out the link that we’ll put in the chat that will take you to the archived video page, and you can watch that and understand how eBird Status and Trends maps can allow you to drill into the data to find specific species and the likelihood of them being in specific areas at high levels during certain times of the year. And Andrew, I know you’re a part– the Center for Avian Population Studies, your big department, houses eBird also. I don’t know if there’s anything else you want to speak to you about the interplay between those two.
[Andrew Farnsworth] Sure. They’re critically important. One way we like to think about monitoring bird migration and studying it is that there’s no one method that tells you everything you want to know. So there are these different complements that give information and provide data that are really essential to understanding the whole. So from the perspective of BirdCast, our dominant research has been using radar to study the aeroecology of migration.
And radar, as you’ve gotten the sense, is just talking about birds. It is not species-specific. Identifying species on and from radar data is virtually impossible, unless something very specific about where you are. For example, I mentioned bats coming out of caves. Purple martins and tree swallows coming out of roosts have very specific patterns, and we can localize those on the ground.
But for the most part, when we think about radar, we’re talking about bird biomass, so the numbers of birds, per se– we relate that, because that’s really talking in particular adding value at night, when we can’t do any kind of visual surveys for the most part, relating what we have from the status of trends data from eBird, knowing where birds are on the ground during the day, and how that changes a time– that’s essential information to complement what’s happening at night.
So really, they’re telling us different things, but they’re integrally connected. And the eBird data can inform the radar data in a number of ways– one, thinking about the species that are on the move and where the pulses of birds on the move may come from, what assemblages might be, and so on. So too could acoustic information potentially tell us something about those assemblages.
But perhaps even more important, being able to know what birds generally on– are on the move allows us to tune our radar data little bit more to get the bird numbers a little bit better, in the sense of that radar capturing information about waterfowl movements in February and March, and primarily songbird movements in April and May, and then shorebirds and some other waders later in May. That allows us to basically control for some of the sizes and the mass of these birds and think about tuning our numbers a little bit– so really important complements, but definitely different kinds of data sets– no question.
[Lisa Kopp] Well, and you bring up a really good point, which is that it’s really hard to see birds migrating at night. But when you were talking about what initially got you into this, you mentioned night calls. And so we’ve gotten some questions from people. One person said, if birds are flying at 2,000 feet at night, how can we observe them or how can we see them? Is it worth talking about night calls? And maybe, actually, we could talk about Merlin Sound ID and how there is a way to then understand what’s happening up above in the dark.
[Andrew Farnsworth] Yeah. The flight call monitoring is a really interesting and pretty long-standing area of interest of– area of investigation at the Lab and a few other places around the world. And so these flight calls are quite unique to periods of migration. And in some cases, they are inaudible above even a few hundred feet above the ground, especially for smaller birds.
So the idea of monitoring them is challenging– one, because recording them properly can be difficult. Two, they tend to be very short in duration. So some of them may be as little as 40 to 50 milliseconds long. So this whole idea of trying to keep humans in the loop with technology– flight calls is a prime example where we have to use the birding skills of really experts that understand what these flight calls sound like and their characteristics to train models– very challenging, because the signals vary, and obviously where birds are in altitude, and the volume that they produce of information.
So too, there’s a ton of background noise– so lots of challenges. That said, there are a few really exciting paths at the Lab right now, I think, to address some of that and to really take flight call monitoring hopefully to the next scale. Sound ID with Merlin is one of them. Hopefully some of you in the audience have had a chance to explore that. Right now, it’s most effective in bird songs and calls you hear during the day, but we’re working on training that with flight calls. And we expect some really excellent results.
BirdNET is another project that’s been really working on the machine listening, if you will, and training machines to do the kind of listening we do. And BirdVox is another project. So these paths to get to the automatic processing of information and trying to extend our capabilities as humans to identify these, including up to as far as we can record sound in the atmosphere, that is– flight calls really– that’s the reliable method for knowing what’s on the move at night when it’s moving– at least for these birds.
So it’s a critical piece of the puzzle that, again– it’s a complement to the radar data that are telling us where, when, and how many birds are moving; the eBird data that tell us about where birds are distributed on the ground; and then the acoustic monitoring that actually tells you, well, what species you’re calling at night as they’re actively migrating?
[Lisa Kopp] Yeah. There’s so much information that is being learned from this, and it just opens up more possibilities, right? You just want to understand more and understand things better. Related to that, I don’t feel like it can be overstated how fascinating just having this information about bird behavior is, but are there things that we’re learning about birds and their– in terms of scientific findings or research that is coming out of this data that’s showing us new things about birds being at risk, or helping us make decisions from a conservation perspective?
[Andrew Farnsworth] Yeah. I think we’re learning quite a lot. Let me share a little bit with you. What I’ll do is share a little bit about some of the things we’ve learned from the radar data, and then maybe we can talk, if the questions come up a little bit more, about the conservation application. So one of the things I’d like to highlight is, because we can look across all of these years and average when migration occurs, we can create maps like this, where we’ve taken out the weather data, applied it to only bird migration data, and smoothed it out to try to visualize what’s happening.
So what you’re seeing here is the pulse of spring movement and the return pulse of fall movement. The size of the circles represent the migration intensity, and the arrows you see represent the direction to which birds are moving. So in the spring here, you see this big pulse going up the center of the country, and then, at some point in the middle of the summer, you see all those arrows turn around and you see that pulse return.
So this idea of thinking about migration, where it’s distributed in the country, where the intensity occurs, and when it occurs, what direction it’s moving– this is all possible now over really extended period of time. So we can also dig a little bit deeper and start to think about, well, what does that mean for particular areas, how many birds are moving during these peak times, and really characterizing when the peak times are.
So on this rather complex graphic, the purple circle is around Texas. And Texas in the spring is the migration hotspot probably on the planet– talking about probably a billion– I said a billion– birds passing over Texas each spring. And the movement there– it’s migration central. And being able to look at a radar like, in this case, the Brownsville radar, and look at the pattern of the numbers of birds and when those occur– so on the bottom scale, you’ll see months from March up to June, and on the y-axis, the vertical scale, you’ll see the numbers of birds.
So if you follow along, you’ll see that there’s a 19-day period in May– April 19 to 7 May– that, on average, captures about 50%– that 19-day period captures about 50% of that movement of the total migration traffic through Texas. So we can define that peak window of when most of the migration is coming through. And we’re learning this for four radar stations around the country.
And then we can think about, well, what does it look like in the spring when we think about, what are the average peak migration dates based on the last 20 plus years of data, and starting to create a map like this to think, well, you can follow the wave of peak movement to the north. And because we can do these things and we can look across this large time period, we can also think about how the timing of peak has changed.
So in this graphic, which comes from a paper that was published a couple of years ago by Kyle Horton and a few others in our group, what we’re looking at is the change in time. And the change in time, the– when the peak migration occurs in the spring is advancing, so it’s getting earlier. In this case, we’re talking about between one and two days per decade. It seems like not that much, but let me come back to that in a minute.
So it’s advancing, so the peak is getting earlier every spring. And it’s changing at different rates. It’s happening faster at higher latitudes in the US. So when we think about climate change and its disproportionate effects, and where temperatures are rising faster, it’s not happening equally across the landscape. And in the case of temperature, which is one of the correlates that really strongly tracks this timing of peak change, we can investigate that further.
And to come back to the days per decade, think about one to two days per decade not from the perspective of, well, if I count up the total number of days in a decade, that doesn’t seem like much. But think about, over a course of decades, and a particular date you might be interested in– your birthday or a holiday– that every 10 years, that day is shifting a day or two earlier; that, after 100 years, you’re talking about 10 to 20 days earlier; and then, beyond that, you may be in a different month.
So from an evolutionary perspective, we really start to be talking about some major change. So seeing that kind of thing is also something we can do. And digging down a little further, how many birds are aloft every year in the spring and the fall?
These greenish blue arrows tell you what’s coming across a transect of our southern radars, and then also a transect of our northern radars– so 3 and 1/2 billion birds coming out of the tropics every spring and 2.6 billion birds continuing into the boreal forest and the prairies of Canada, and the tundra, and in the fall, these orange arrows, 4 billion birds coming back into the US, and then 4.7 billion coming out of the US, moving to– back to the tropics. So being able to put some numbers like this has been incredibly powerful for us.
[Lisa Kopp] Wow, yeah. I can imagine it’s alarming, but of course, it’s important to know that hopefully we can put some changes in place to prevent these trends from continuing. I know you mentioned it at the beginning that Canada and Mexico are not on the map. We don’t have access to that radar data.
Theoretically, every continent, most countries, have their own radar data. Are there any sort of international attempts to do this kind of work elsewhere, or any sort of partnerships that might form to be able to see what– because there is so much connection between what’s happening in Canada, and Mexico, and beyond in Central and South America, and the birds that we see for part of the year here.
[Andrew Farnsworth] That’s a really good question. There definitely are other groups around the world with which we collaborate. We are working closely with the Canadians right now to hopefully integrate their radar data and integrate our radar data into their systems as well so that we have the capability of expanding the BirdCast models just on the North American continent.
We also work really closely with groups in Europe, and there are some really tight relationships there. We have some ongoing projects with them. There is an extensive weather surveillance radar network in Europe. It’s a little bit of a challenge, because the data– dealing with national meteorological offices for each country– of course, the European Union has a diverse array of countries, and because of those political boundaries, federating all of those data can be a real challenge, especially over long periods of time, when there have been different approaches to archiving data, or perhaps different ways of making them available.
But we are working closely, and absolutely, some of the studies, some of the kinds of work that we’ve done here is applicable in Europe, and will happen there, we hope, at some point soon. And we’re working closely. So too, there are opportunities in China. China has an extensive– in fact, probably the biggest weather surveillance radar network. And we spoken with Chinese scientists and meteorologists.
And so too, Australia and India have networks. The notion that we could really study the atmosphere and movements of birds and other biology as well at a global kind of perspective, thinking 5, 10, 15 years out– I think it’s a real compelling and big opportunity– a lot of challenges, but absolutely something that could happen, no question.
[Lisa Kopp] Wow– yeah, sounds exciting. So we’re getting some questions about what we can do to help birds that are migrating at night. I’m already seeing some comments about Lights Out, so would you mind talking a little bit about some of what we’ve learned is really a hindrance to birds migrating at night, what can really make it more– make their journey harder, and what we could do to help?
[Andrew Farnsworth] Absolutely– so let me talk a little bit about Lights Out and how we connect the science to conservation action from that perspective. So for those that have been asking the question about, what are some of the hazards, thinking about birds migrating at night, and then thinking about the human footprint, for example– in particular, structures that we erect, whether they be communications towers, or electrical and energy grids, or buildings and houses– these structures– they interfere with the airspace.
And really, in the last decade or so, we’ve started to think very much about the night sky, and the sky in general, as airspace as habitat for birds– so thinking about the incursion of these structures. Unfortunately, it has ramifications– collisions in particular, birds colliding with all of these types of structures. And it’s a serious, serious issue, talking about– probably between 400 million and a billion birds every year, a lot of them migrants, are killed in collisions in the US annually.
So it’s a large number. It’s something that really has two major components. One of them absolutely has to do with attraction to light. The other has to do with reflective glass surfaces and places where birds collide that they’re not perceiving that there is something there with which they could collide. So just to highlight here a couple of the components of what this means, think about light from the perspective of– in the 17th and 18th century, oil burning lighthouses as a means to guide ships at night through dangerous passage– those attracted birds.
There’s pretty detailed literature in some cases of the species and the numbers involved– massive, massive collisions to those lighthouses. Fast forward 100 years to the end of the 19th century, and for example, the first World’s Fair that was electrified in Chicago, 1893– that was illuminated. There were bird collisions there. With that moment in history, suddenly this change of light at night and light pollution at night began an ever-rising and continuing increase that is a fundamentally new experience that birds are having.
Birds are attracted to light. They are disoriented by it, and unfortunately, when it comes to where the– most light is, it’s often centered in places where there are these human structures. So birds may collide at night directly with these structures as a result of light. They also may be put into hazardous situations, because they’re attracted, and disoriented, and stay in these areas– for example, in a city– for example, New York City earlier this month, where birds collide during the day with glass and reflective structures, because they’ve been attracted and disoriented at night by the light.
And these kinds of situations are– it’s awful. If you’re a birder, if you’re a scientist, if you’re a conservationist, if you’re a human that cares about your environment, you don’t want to see this kind of mortality. So yeah, there absolutely are things you can do about this. I want to talk quickly about the Lights Out piece, so– and just the simple change of being able to see what happens when you turn lights on and off.
Some of you may recognize this as the Tribute in Light in lower Manhattan. And these very bright lights attract and disorient birds. What you’re seeing in this video are not insects or snowflakes. These are migrant birds. They’re dominated by small songbirds, four species of warblers– American red star, oven bird, black and white warbler, Northern parula in these cases. This video from 2015 typifies the level of attraction.
At any frame of this video, we’re talking probably about 2,000 birds in the frame. So there’s a lot of bird activity here. And one of the things that we were able to show, working closely with the producers and organizers of the event, is that, when large numbers of birds are in the beams, they allowed us to turn the beams off to protect bird migration.
Despite all of the intense political, and emotional, and societal charge, they said, we’re willing to take this compelling step and work with you so you can understand what’s happening, and also protect birds. So in this graphic, just think about it as green on the graphs is where the lights are on. Gray is where the lights are off. On the top graphic– number of birds that are aloft that we can identify with radar. The speed that they’re moving is the middle, and the vocal activity– the flight calling– is the bottom.
And the important thing to note here is the periodicity of this. When the lights are on– large numbers of birds concentrating. They’re flying at slower speeds and circling. They’re calling a lot more because of this disorientation. As soon as the lights go off, the bird numbers drop dramatically. The speeds pick up, because birds start flying more typical nocturnal migration strategies, and stop circling. And the calling goes down, because birds can reorient.
So this notion of how quickly it can happen that birds can respond to turning off lights is something that we think is really powerful. And it’s not just about these really bright building– exterior lights. It’s also about individual lights too. So this comes from a study that was just published earlier this year from a long-term data collection effort in Chicago at a particular building, McCormick Place, where they were collecting collisions. And we related that to the bird migration traffic we can identify from radar, and also weather conditions, and understanding also how much light was on in the building.
And from the light perspective, if we had no– or if we moderately decreased the number of windows to half, we see a dramatic drop in the number of bird collisions. If we increased the building, for example, to full illumination, we would see tremendous increase in numbers of collisions. So the idea that every light– every interior light, every office light– turning that off matters is a very real one.
And we have these kinds of tools now to help understand, well, if you want to take that kind of activity and turn off your lights at a particular moment in time, we can use the migration forecast, for example, to highlight, when is there going to be an intense area or an intense movement on a particular night, and where is that going to occur? What you see here are four different accounts, from left to right– San Diego tonight, New York tonight, Portland, Oregon tonight, and Minneapolis, Minnesota tonight. New York, Portland, Minnesota– high migration forecast– great night to turn lights out, an opportunity to make a change that’s very simple. Now, of course, we want as little light illumination and as little light pollution as possible, but starting small, where we can, where it’s not happening already, for example, turning off individual lights can make a huge, huge difference.
And we’ve been exploring that in Texas in our Lights Out Texas program. And it’s been picked up in the most tremendous way there. Dallas, Houston mayors have proclaimed Lights Out nights. And a number of different cities have adopted these kinds of proclamations. Buildings, including entire fleets of buildings for particular managers and owners, have bought into it as well, and have agreed to turn off non-essential lights.
There are other cities around the US. Chicago in particular as a leading example that’s been really excellent about starting to really consider seriously legislating this and making it into something that just has to happen at a broad scale over long periods of time– great Lights Out programs in Portland, and Baltimore, and New York. So the idea of taking this kind of science and then applying it to the conservation action, and connecting that dot from things we’re really interested in knowing about how birds behave and why with, how can we make changes to the things that we do– in fact, very simple changes– that’s a very compelling connection for us, and something we try to do often at the Lab. But I think this in particular for BirdCast exemplifies that behavior of taking science to some kind of conservation action.
[Lisa Kopp] Well, and it’s just so encouraging to hear of something real that you can do. Sometimes these problems feel so huge that it’s hard to feel like one person or one little thing that you do can really have an impact. And this is showing that something as simple as turning off your lights really can, and does. And if it picks up steam in the way that it sounds like it is, it’s really– again, it feels like a rare win for the environment, because these days, they’re hard to come by, it seems.
[Andrew Farnsworth] Yeah. And it’s an opportunity to engage even further and say, OK, starting with the simple activity of turning off the light, and then expanding that to thinking about, well, how do I reduce this at a much greater scale, locally, regionally, globally. And that also then begets the discussion of well, OK, light is part of it, as we talked about earlier.
Building design is a huge part of it as well, getting there into– terms of buy-in from the architects and from cities, bird-friendly ordinances– for example, like the one that just passed in New York earlier that really mandates bird-friendly glass, bird-safe glass where there are patterns that birds can perceive, but humans not– maybe not so much. That’s got to be part of the discussion too. And that’s one of the directions absolutely to take this conversation is act very precisely and at a particular moment in time, grow that action, and then really start to think about, well, what also is a significant component, perhaps even more of a component of this problem that we might be able to address as part of this discussion? Those two things in particular are both active areas of research and conservation action, and clearly a lot of interest around the country– and around the world too.
[Lisa Kopp] Mm-hmm. Yeah. And I should mention, on our archived webinar page, we have two talks that people might be interested in that dig into these topics specifically. We did a webinar specifically with the Lights Out team– who work under you, Andrew– where we get into the details of that. So that’s worth checking out. We’ll put that in the chat.
And then also, last Monday, we hosted a webinar with Christine Sheppard of the American Bird Conservancy, who is really at the leading edge of bird collision prevention work. So if you want to hear from the– really, the expert on that, that’s a great one to tune in to as well. We’re getting a couple of questions here about non-physical things in the air that might affect birds– so wavelengths or increasing frequencies in the air. Are any of those things being seen as hazards or issues for birds as they migrate?
[Andrew Farnsworth] That’s a good question. Radio frequencies certainly do interfere, in some cases, with birds ability to use some of their sensory cues, if you will, to orient and to navigate. So in a similar sense to light in different wavelengths of light influencing birds’ capabilities, so too some radio frequencies can. We are seeing those kinds of patterns.
But the research into some of the interference– it’s evolving, and it’s evolving quickly. I think there are some big opportunities now, given the advances in technology, to think about how we can really understand, well, OK, what is it about particular aspects of this either radio frequency or light wavelengths that are influencing birds?
Is it something about the way that they are specifically either disorienting or interfering with their magnetic capabilities, the– whether it’s the molecule in the eye that we think helps birds orient, for example, to magnetic fields, or is it something that’s a little bit broader? It’s really active area of research trying to understand the kind of conflicts that evolve from birds that have evolved these sensory capabilities to perceive the magnetic field, to use celestial cues and orient to star fields, to calibrate their kinds of compasses based on sunset– and then understanding, well, how is human activity influencing that?
There’s a lot we still don’t know, including, exactly why is it that birds are attracted to light? We have some hypotheses out there. And obviously, it’s a pattern that we see that goes far down the evolutionary tree to some very much ancestral organisms that are much simpler, if you will, than birds. But still, the understanding of exactly why that happens– we’re still looking for it.
[Lisa Kopp] Wow, that’s always the exciting, and probably incredibly frustrating thing about science is that, as soon as you solve one thing or get an answer to one thing, it leads to 10 more questions or 10 more things to explore. It is incredibly hard to believe, because has been such a wonderful conversation, but we only have about a minute left. So Andrew, is there anything that you want to share about the future of your work or anything, any last-minute tips for migration season, as we’re heading into it, for people?
[Andrew Farnsworth] I can share a couple of both. So tips for migration season– biggest tip of all is you should get outside, and you should look around and be observant. All this technology is wonderful at helping us understand really when we should go out and observe ourselves. So always keep that in mind. I think the human in the loop, as we were talking about– it’s absolutely essential, both from the perspective of experiencing nature and what that means when we connect to it, but also informing this kind of capability that we now have to apply technology.
We still need be human, and we need that experience, and it’s good for us to have that experience. So go outside. Go birding. And tell us what you find in eBird, please, especially during migration season. And from the perspective of BirdCast and where we’re going with this, you saw a little bit today that we’re really trying to continue what I think is high-quality research. And it’s really important for us to provide whoever would like it the science that explains what’s going on or the patterns we can characterize.
And that’s a fundamental goal of the project. But really, it’s also about inspiring people to go out and look, to think about migration and what it means to the bigger scale of the planet, and then be able to affect some change. So if you’ve heard something that is interesting to you, visiting the BirdCast website to find out a little bit more about it– birdcast.info– that would be great.
You can always reach out to the team if there are specific questions. I know, from talks like this, there’s a lot of information that we covered and a tremendous amount that we did not cover, so you can reach out to us any time and we’ll happily answer your questions.
[Lisa Kopp] That’s great. Thank you. And we’re putting contact information in the chat right now for people. So for all of those of you who are on Zoom, you’ll actually be getting a link emailed to you in the next few days with the archived version of this talk. And we will also include some of the helpful links that we’ve been chatting out.
We can’t, unfortunately, include the chat transcript or anything like that for privacy purposes, but we’ll include some of the top links that you all are interested in. You’ll also have access to all of the past recorded webinars. Those live on our Bird Academy page. And we mentioned a few that talk about some of the things that Andrew touched on, so we hope that you enjoy those.
And Andrew, thank you so much. It’s always so fascinating to hear about BirdCast. Yeah, it’s sort of mind-blowing and inspiring all at the same time. And the way that you’re able to take these really complicated concepts and make it totally approachable is– it’s so fun to talk to you about this.
[Andrew Farnsworth] Well, thank you so much. Thanks to all those in the audience who spent their lunch hour, or breakfast hour, or whatever hour it might be around the planet listening. I really appreciate it. So does the entire BirdCast team. We’re very thankful for your interest.
[Lisa Kopp] Well, thank you again, and I hope everyone has a great rest of the day. And as Andrew said, get outside this weekend and enjoy some birds. All right, thanks, all. Bye.
End of transcriptSome 80% of North American migratory birds travel under the cover of darkness, adding to the mystique of migration season. In recent years, the advent of BirdCast has made it possible to forecast when and where birds are flying en masse—making birdwatching more fun and offering chances to take quick action and save birds’ lives. Join Dr. Andrew Farnsworth to learn how BirdCast merges weather radar and data modeling to assist birders, scientists, and conservation efforts around the country.
This event is part of our virtual Migration Celebration. Visit the Migration Celebration webpage for the full schedule of events, migration resources, and family-friendly activities.
- Stay up to date on the latest from BirdCast, including daily migration forecasts.
- Keep birds safer on their migration journeys by going Lights Out.
- Discover other ways to keep birds safe and happy during migration season.