Raptors, Aphids, Dragonflies And Robins
September 24, 2018 11:57 AM Subscribe
“So, fun fact: birds and insects show up on radar. Often. As in, pretty much every day.” (link to Twitter thread; you can also read the same thing on Thread reader)
Having recently contended with a mild red aphid infestation on my tomato plant- the fact that they can show up on radar is causing me to have a mild existential crisis.
Besides the fact that I am hugging my knees and weeping, this is really cool.
posted by Homo neanderthalensis at 12:31 PM on September 24, 2018 [1 favorite]
Besides the fact that I am hugging my knees and weeping, this is really cool.
posted by Homo neanderthalensis at 12:31 PM on September 24, 2018 [1 favorite]
Thanks. As a birder and a meteorologist, this is right up my alley. I’ve been studying this on my own for years, but this provides some useful information that is new to me. It does a great job of making it understandable.
posted by jkent at 12:46 PM on September 24, 2018
posted by jkent at 12:46 PM on September 24, 2018
The national weather service has a kind of best of page for Mayfly radar observations which has some fun photos too if you like your planet covered in bugs.
posted by mattamatic at 12:58 PM on September 24, 2018 [2 favorites]
posted by mattamatic at 12:58 PM on September 24, 2018 [2 favorites]
I’m looking forward to seeing the results from Sky Net!
posted by monotreme at 1:05 PM on September 24, 2018
posted by monotreme at 1:05 PM on September 24, 2018
This was fascinating to learn and see illustrated. Doubleplus thank you OP for including thread reader link.
posted by GoblinHoney at 1:41 PM on September 24, 2018 [3 favorites]
posted by GoblinHoney at 1:41 PM on September 24, 2018 [3 favorites]
Fun stuff... in June we has a closely related post here
posted by exogenous at 1:49 PM on September 24, 2018
posted by exogenous at 1:49 PM on September 24, 2018
The University of Washington's Cliff Mass out here in the Pacific northwest has covered this several times over the years on his blog. Most the entries have a lot of screencaps that illustrate the phenomenon; its' always fascinating to read on it.
posted by los pantalones del muerte at 1:53 PM on September 24, 2018
posted by los pantalones del muerte at 1:53 PM on September 24, 2018
!!??What??!!! This is my favorite thing I've learned about in a long time...and it's competing with my Dino 101 class so it's that cool.
posted by lemonade at 2:20 PM on September 24, 2018
posted by lemonade at 2:20 PM on September 24, 2018
I fucking love it when one of these nerdy-as-fuck twitter threads comes around and teaches me cool shit.
posted by rmd1023 at 2:25 PM on September 24, 2018 [3 favorites]
posted by rmd1023 at 2:25 PM on September 24, 2018 [3 favorites]
I have friends who are using radar to find nesting spots of critically-endangered black-capped petrels. These birds live and forage out at sea. They nest on land, of course, but only come back to their nests at night. Because so much of the Caribbean has been developed, they now nest up in the mountains. So they basically mount a marine radar unit in the back of the truck and scout likely pathways for radar images that look like these birds. Doing this they found the first evidence that they are nesting on the island of Dominica in 150 years. Pretty cool stuff.
posted by snofoam at 2:37 PM on September 24, 2018 [6 favorites]
posted by snofoam at 2:37 PM on September 24, 2018 [6 favorites]
Future radar mapping will be so precise that it will allow for discrete counts of bugs. So you'll have comfort in knowing there are 10 billion of them swarming in the backyard right this moment.
I am torn between hating you forever for this comment and actually liking the scientific implications of this from an ecological standpoint. I’m so conflicted!
posted by Homo neanderthalensis at 3:10 PM on September 24, 2018 [1 favorite]
I am torn between hating you forever for this comment and actually liking the scientific implications of this from an ecological standpoint. I’m so conflicted!
posted by Homo neanderthalensis at 3:10 PM on September 24, 2018 [1 favorite]
From the NYT last week: radar images show what happens when migratory birds congregate around the 9/11 tribute in NYC, and what happens after the lights are turned off for a few minutes. They have volunteers on the ground counting birds but I do wonder if they could just rely on radar.
posted by fedward at 3:41 PM on September 24, 2018 [2 favorites]
posted by fedward at 3:41 PM on September 24, 2018 [2 favorites]
I've noticed these radar patterns for years and assumed they had to do with changes in air temperature or something since they always happen around dawn and dusk.
I, like many others in this thread, am not sure how I feel about the knowledge that these are actually the result of so, so many bugs. And bats. And birds too I guess.
posted by Mr.Encyclopedia at 5:19 PM on September 24, 2018 [1 favorite]
I, like many others in this thread, am not sure how I feel about the knowledge that these are actually the result of so, so many bugs. And bats. And birds too I guess.
posted by Mr.Encyclopedia at 5:19 PM on September 24, 2018 [1 favorite]
So what I'm taking away from this is that Skynet will be developed "to investigate variability in species composition and phenology," with the ultimate goal "to develop better radar classification algorithms for detecting and monitoring insect populations." Because there's no chance at all that this might have a side effect of helping robots develop better radar classification algorithms for detecting and monitoring, uh, other populations, no sir.
posted by limeonaire at 6:28 PM on September 24, 2018
posted by limeonaire at 6:28 PM on September 24, 2018
There are commercial portable radar systems that wind farm developers use to survey bird and bat migration paths for planning and also real time monitoring to reduce mortality. Developers can either avoid building in high migration paths or else shut down the turbines temporarily if they detect an approaching swarm.
posted by JackFlash at 7:03 PM on September 24, 2018
posted by JackFlash at 7:03 PM on September 24, 2018
I've noticed these radar patterns for years and assumed they had to do with changes in air temperature or something since they always happen around dawn and dusk.
You were probably right the first time!
Oftentimes, birds' radar signatures are mistaken for temperature inversion-related radar phenomenon, and vice versa. This was recently the case in a viral news story about radar that allegedly showed birds taking to the air en masse, shortly before an earthquake in Oklahoma. More in this reddit thread and elsewhere online (ie fb stories about it).
posted by Perko at 8:03 PM on September 24, 2018 [2 favorites]
You were probably right the first time!
Oftentimes, birds' radar signatures are mistaken for temperature inversion-related radar phenomenon, and vice versa. This was recently the case in a viral news story about radar that allegedly showed birds taking to the air en masse, shortly before an earthquake in Oklahoma. More in this reddit thread and elsewhere online (ie fb stories about it).
posted by Perko at 8:03 PM on September 24, 2018 [2 favorites]
I live in Austin and the local weather radar often shows what looks like light rain but is actually the bats coming out at sunset from under the Congress Ave bridge and flying along the river.
posted by a humble nudibranch at 10:12 PM on September 24, 2018 [2 favorites]
posted by a humble nudibranch at 10:12 PM on September 24, 2018 [2 favorites]
I remember an old discussion on Usenet about a UFO case, and somebody suggested the weird radar returns mentioned could be a flock of birds or bats. There was a general dismissal by the other participants that birds or bats could show up on radar. WELL.
posted by happyroach at 11:57 PM on September 24, 2018 [4 favorites]
posted by happyroach at 11:57 PM on September 24, 2018 [4 favorites]
That's really cool. I was sort of surprised that they would show birds and insects so clearly, but then I realised they're designed to pick up raindrops, so.
posted by lucidium at 6:51 AM on September 25, 2018
posted by lucidium at 6:51 AM on September 25, 2018
Most fascinating bit for me: The can tell which way the insects are facing because the large radar pattern is in a bowtie shape that's created by... the bowtie, skinny-waist shape of the insects themselves?? How does that work?
posted by clawsoon at 7:38 AM on September 25, 2018
posted by clawsoon at 7:38 AM on September 25, 2018
Radar hits long skinny insect on its long skinny side: more reflection back to receiver, shows up on radar as the lobes of the bow tie.
Radar hits long skinny insect on its very tiny head: less reflection back to receiver, doesn't show up on radar.
The insects themselves don't necessarily have narrow waists; they only need to be pointed in more or less the same direction, and have more surface area to reflect radar signals in some dimensions than in others, e.g., insects with long, rod-shaped bodies like dragonflies, mayflies, or grasshoppers.
A more spherical insect (June bugs?) would reflect more or less the same amount no matter which direction the radar signal came from, so the radar image would be circular even if they were all facing the same direction.
posted by Spathe Cadet at 8:23 AM on September 25, 2018 [3 favorites]
Radar hits long skinny insect on its very tiny head: less reflection back to receiver, doesn't show up on radar.
The insects themselves don't necessarily have narrow waists; they only need to be pointed in more or less the same direction, and have more surface area to reflect radar signals in some dimensions than in others, e.g., insects with long, rod-shaped bodies like dragonflies, mayflies, or grasshoppers.
A more spherical insect (June bugs?) would reflect more or less the same amount no matter which direction the radar signal came from, so the radar image would be circular even if they were all facing the same direction.
posted by Spathe Cadet at 8:23 AM on September 25, 2018 [3 favorites]
You think radar is cool? Let me tell you about LiDAR!
RaDAR as I will remind you all is the abbreviation for Radio Detection And Ranging, relying on radio waves. It's a fairly macro tool, meaning that it works well - even in inclement weather conditions, to track the movement of fairly large objects, or in this case the movement in tandem of many small objects such that they mimic the movement of one larger object.
LiDAR stands for Light Detection And Ranging, which will be at least probably the go-to technology for self-driving cars in our lifetime (probably augmented with RaDAR still for the aforementioned weather reasons). As implied by it's name, it uses light "waves" - laser "pings" if you will to return a 3D image in orders of magnitude more detail than radar can (see here).
Lest you be of the impression that it's a "new" technology - we first started using it in the 1950's to measure the distance to the moon. (around 2:30 in this video there's a short explanation of how each LiDAR ping is calculated - just imagine the massive amount of data from one flight of an airplane with a sensor constantly measuring 30 pings per square meter it views)
I could spend a lot of time going into the nerdy details of how it works, but at a high level, a system operator determines the level of detail specificity they need in their output, which determines the number of "points" per square meter that they will ask the LiDAR sensor to return. The sensor, typically vehicle or airborne-vehicle mounted, then does return its slew of pings, in multiple "swaths" of what you could call something akin to camera pans, the results of which require high computational power to grok a massive data set of points to render 3D images down to a centimeter level of detail, when the sensor itself was anywhere from 20 to a few hundred meters away from the object. (!!!)
I can geek out about this because currently I'm working with an energy provider to build advanced analytics using machine learning to scan massive amounts of LiDAR data sensed from the service territory across the state that this provider operates in. They fly fixed-wing and helicopters to gather LiDAR from the sky, while the minions from Silicon Valley are gathering it from the ground with their street-view vehicles. The results allow for a 3-D rendering of things like power poles and the conductor lines between them (as well as all of the other above ground electrical assets), the implications of which are massive. Theoretically, in the very near future, I will be able to tell you out of ~2.5 million power poles, which ones are currently laying on the ground, vs. standing straight up. I will be able to tell you which ones have different types of assets attached to them. Which ones have a wire down in between them. Which ones have a tree branch intruding within a given radius of a conductor line, posing a risk to bring that line down in a high wind event.
The ironic thing is, with LiDAR, it's SO accurate that we have to scrub the data for thing like birds, which are considered "interference" in the data because the bird may be blocking my view of the asset on the power pole that I want to see, if it's sitting on a wire or cross-bar above that asset, blocking my view of it from the sky. Or even - get this - if it happens to be flying through that exact square meter that the LiDAR is sensing at the moment it passes over that square meter - a big enough bird flying over it creates an anomaly in the data that we have to teach the computer to first learn, and then correct for, because this happens much more often than you would perhaps expect. When you think about it, birds frequently land, sit, or even nest on power lines and poles, so it's not all that exceptional that you would return higher quantities of them if you are specifically scanning environments that contain those energy assets.
The data sets we are getting are so massive that the main problems with them that we now are running into are storage space (how big can these bytes get amirite?) and the energy required to power the computational infrastructure required to render images from so many ultramuchbazillion points of laser data that LiDAR returns.
tl;dr - self driving cars will probably use cameras, radar, and LiDAR for the foreseeable future, but LiDAR is the not-exactly-new kid in town and he's super cool. And you'll probably learn about it in application the next time you get pulled over for speeding.
posted by allkindsoftime at 2:50 PM on September 25, 2018 [4 favorites]
RaDAR as I will remind you all is the abbreviation for Radio Detection And Ranging, relying on radio waves. It's a fairly macro tool, meaning that it works well - even in inclement weather conditions, to track the movement of fairly large objects, or in this case the movement in tandem of many small objects such that they mimic the movement of one larger object.
LiDAR stands for Light Detection And Ranging, which will be at least probably the go-to technology for self-driving cars in our lifetime (probably augmented with RaDAR still for the aforementioned weather reasons). As implied by it's name, it uses light "waves" - laser "pings" if you will to return a 3D image in orders of magnitude more detail than radar can (see here).
Lest you be of the impression that it's a "new" technology - we first started using it in the 1950's to measure the distance to the moon. (around 2:30 in this video there's a short explanation of how each LiDAR ping is calculated - just imagine the massive amount of data from one flight of an airplane with a sensor constantly measuring 30 pings per square meter it views)
I could spend a lot of time going into the nerdy details of how it works, but at a high level, a system operator determines the level of detail specificity they need in their output, which determines the number of "points" per square meter that they will ask the LiDAR sensor to return. The sensor, typically vehicle or airborne-vehicle mounted, then does return its slew of pings, in multiple "swaths" of what you could call something akin to camera pans, the results of which require high computational power to grok a massive data set of points to render 3D images down to a centimeter level of detail, when the sensor itself was anywhere from 20 to a few hundred meters away from the object. (!!!)
I can geek out about this because currently I'm working with an energy provider to build advanced analytics using machine learning to scan massive amounts of LiDAR data sensed from the service territory across the state that this provider operates in. They fly fixed-wing and helicopters to gather LiDAR from the sky, while the minions from Silicon Valley are gathering it from the ground with their street-view vehicles. The results allow for a 3-D rendering of things like power poles and the conductor lines between them (as well as all of the other above ground electrical assets), the implications of which are massive. Theoretically, in the very near future, I will be able to tell you out of ~2.5 million power poles, which ones are currently laying on the ground, vs. standing straight up. I will be able to tell you which ones have different types of assets attached to them. Which ones have a wire down in between them. Which ones have a tree branch intruding within a given radius of a conductor line, posing a risk to bring that line down in a high wind event.
The ironic thing is, with LiDAR, it's SO accurate that we have to scrub the data for thing like birds, which are considered "interference" in the data because the bird may be blocking my view of the asset on the power pole that I want to see, if it's sitting on a wire or cross-bar above that asset, blocking my view of it from the sky. Or even - get this - if it happens to be flying through that exact square meter that the LiDAR is sensing at the moment it passes over that square meter - a big enough bird flying over it creates an anomaly in the data that we have to teach the computer to first learn, and then correct for, because this happens much more often than you would perhaps expect. When you think about it, birds frequently land, sit, or even nest on power lines and poles, so it's not all that exceptional that you would return higher quantities of them if you are specifically scanning environments that contain those energy assets.
The data sets we are getting are so massive that the main problems with them that we now are running into are storage space (how big can these bytes get amirite?) and the energy required to power the computational infrastructure required to render images from so many ultramuchbazillion points of laser data that LiDAR returns.
tl;dr - self driving cars will probably use cameras, radar, and LiDAR for the foreseeable future, but LiDAR is the not-exactly-new kid in town and he's super cool. And you'll probably learn about it in application the next time you get pulled over for speeding.
posted by allkindsoftime at 2:50 PM on September 25, 2018 [4 favorites]
Spathe Cadet: The insects themselves don't necessarily have narrow waists; they only need to be pointed in more or less the same direction, and have more surface area to reflect radar signals in some dimensions than in others, e.g., insects with long, rod-shaped bodies like dragonflies, mayflies, or grasshoppers.
So does that mean they have to be flying as a group in a giant bowtie swirl in order for the bowtie to show up on radar? I'm confused.
posted by clawsoon at 3:51 PM on September 25, 2018
So does that mean they have to be flying as a group in a giant bowtie swirl in order for the bowtie to show up on radar? I'm confused.
posted by clawsoon at 3:51 PM on September 25, 2018
They are flying in a vaguely round cloud; the physics of radar (the bugs are more visible from the side than front or back) makes the return appear bow tie shaped.
posted by Mitheral at 8:54 PM on September 25, 2018 [2 favorites]
posted by Mitheral at 8:54 PM on September 25, 2018 [2 favorites]
Another example that popped up on Twitter, from Finland this time:
Massive burst of chaffinch migration seen by the weather radars today morning. The start of the migration follows the sunrise from east to west.
posted by Bloxworth Snout at 9:31 AM on September 26, 2018
Massive burst of chaffinch migration seen by the weather radars today morning. The start of the migration follows the sunrise from east to west.
posted by Bloxworth Snout at 9:31 AM on September 26, 2018
« Older Breaking A Sacred Trust | Arthur Mitchell, Dance Legend, R.I.P. Newer »
This thread has been archived and is closed to new comments
posted by Orlop at 12:28 PM on September 24, 2018 [1 favorite]