Hacker News Reader

Comment by alphakappa | Hacker News Reader

(Note, this is just a critique, not criticism of some obviously painstaking research)

It would be good to see the results after these experiments get peer-reviewed and replicated. Without just the information presented, it's possible to imagine issues. For example, data is collected every four minutes (We don't know for how long, but let's assume that it's for a few seconds to conserve power). Is it possible that when they are on the ground, they aren't completely still - will their motion on the ground be mischaracterized as flight?

It would be great if an altitude sensor could be practically added to the sensor package.

View on HN

My first thought would be to scoff at your criticism. Would the birds' travel patterns be any less impressive if they were resting for very short periods during these 200 days? The sensor measures acceleration, so the only potential confusion might be with light movement near ground / trees.

But according to another study, some migratory birds rest for only seconds at a time during their flights:

http://www.livescience.com/1045-migrating-birds-hundreds-dai...

At 4-minute intervals over 200 days, you have 72,000 datapoints. There are 1,920,000 9-second intervals (avg nap period from other article) over 200 days, so given their data collection spans only 3.75% of this time there's a chance they missed one of these naps.

Nevertheless, this is still very interesting.

That got me curious, so I ran the numbers on the chances they'd miss all of the naps; assuming these birds take only one 9-second nap a week (28 naps total over 200 days) there's a 35% chance the researchers would have missed all of them ((1 - (28 / 1920000)) ^ 72,000), which is pretty reasonable. But that chance goes down to 8% for one nap every 3 days, and to 0.055% for one nap a day. I'd say maybe one 9-second nap every few days is the lower limit of what I'd find plausible (assuming this was run for only one bird).

EDIT: Ah, missed this in the article - they were three birds. So the chances they'd miss all naps for all three birds goes down to 0.055% for one nap every three days, and 4% for once a week. So 9 seconds once a week is barely believable, but if they can reproduce this with another couple of birds that becomes really unlikely.

Those thrushes don't land to just nap for 9 seconds and then take off. They sleep for 9 seconds during a longer rest period, because they have to stay alert. They can't fly during the day because of airborne predators.

"My first thought would be to scoff at your criticism. Would the birds' travel patterns be any less impressive if they were resting for very short periods during these 200 days?"

The criticism isn't aimed at the birds; it is aimed at the logic of the researchers. The critic claims their conclusion does not follow from their data.

> Is it possible that when they are on the ground, they aren't completely still - will their motion on the ground be mischaracterized as flight

Well, actually swifts can't even walk and have a very hard time taking off from ground. I don't know their behaviour once in Africa, but while they're in Europe you only ever see them in flight, on phone cables or clinging on cliffs. They fly straight to and from their elevated nest. I doubt very much that these situations would look the same as flight.

Also... I think you guys could just try to have a little bit of trust in the knowledge of people who obviously know swifts better than you do. I think you can trust them for having studied their data a bit.

Dude. Asking questions and offering constructive criticism is how we understand good science, improve bad science, and discern the one from the other. "Having a little bit of trust" instead of asking questions is exactly what scientists should never do.

The people in here who ask questions aren't scientists. Asking "but are you sure the sequence shows flight and not moving on the ground" for a bird that is widely known to never touch the ground is not constructive criticism, not good science, it doesn't improve anything, it just distracts from pertinent questions from knowledgeable people.

Good science is criticism from people who have some knowledge in the field. Otherwise you're no better than the client who tells the designer that making the page title blink would add a nice touch to the homepage.

> Asking "but are you sure the sequence shows flight and not moving on the ground" for a bird that is widely known to never touch the ground is not constructive criticism

"Widely known"? Are you reading the same article? No one had any idea this happened until this study.

(Yep, double post. Too late to delete. Sorry.)

> Asking "but are you sure the sequence shows flight and not moving on the ground" for a bird that is widely known to never touch the ground is not constructive criticism

"Widely known"? Are you reading the same article? No one had any idea this happened until this study.

Minor point -- swifts don't perch on phone cables. That would be swallows and martins (unrelated). Swifts only land on the sorts of sheer rock faces / buildings that they nest in AFAIK.

Final random swift observation: you can see some swift species hovering very briefly to take nectar from flowering trees which is cool, seeing as molecular and other evidence shows hummingbirds to be their closest extant lineage.

Pressure sensors don't really give a good measure of altitude from one sample. For starters weather can drastically change the pressure in a region. If a sensor package could somehow filter that out with a 4 minute sample rate (doubtful) then the sensor is still giving you altitude relative to sea level which means the data would have to have a GPS component to differentiate between on the ground on a tall hill or flying over a flat lowlands area.

You're right. So how else could this problem be solved given the power constraints?

(P.S. I do have to mention - It's very likely that my critique is completely invalid because they may have performed measurements of these birds in flight Vs. these birds on ground, and the accel data may show completely different characteristics for both. It's also likely that they may have chosen the sampling rate and data duration based on these experiments, but since this is HN, it's fun to speculate on how this experiment could be improved.)

Somebody else posted a link to this graph from the publications that shows results for different movement patterns actually observed while the birds were in the summer colony: http://www.nature.com/ncomms/2013/131008/ncomms3554/carousel...

how else could this problem be solved given the power constraints?

I wonder if gravity changes enough at a typical flight ceiling to be detectable by a very accurate accelerometer?

At 5km up the difference is miniscule, back of the napkin says the difference is ~-0.01 m/s^2. Which is pretty small to use as a classification filter. All the way up to 40km it's still only ~1.2% weaker.

With fish they have tags that record light intensity, pressure and temperature which is enough to give a low res position estimate [1]. Not sure how well this would transfer to this problem but it is cool tech.

[1]: http://www.gtopppublications.org/repository/254_07-11-11_Blo...

Mostly a nitpick, but pressure sensors don't give a good measure of absolute altitude ever, regardless of the number of samples. You must calibrate it using pressure measurements from a ground station nearby in space and time, or else it tells you roughly nothing.

Weather matters a lot, as you say. The variation caused by weather can be equivalent to several thousand feet of altitude. Worse, changes due to weather are long-term. You can't filter them out, period. Samples taken four minutes apart can tell you changes in altitude, since the weather won't change that much in such a short time, but there's no way to backtrack to absolute altitude without calibration, no matter how many samples you take.

True. However if we're looking for very brief landings followed by periods of flight you might see that in the pressure data.

You'd expect some flat "flight level" pressure over a day, with V pressure changes to correlate with landings (or swoops in mid-air). You would need to correlate these with other sensors for it to mean anything concrete but it could corroborate or disprove other theories.

Right, but you couldn't distinguish a V that was a brief landing from a V that was a dive and climb, nor could you distinguish a flat line that was cruising from a flat line that was sitting on the ground.

I don't doubt that pressure data would be useful, it just doesn't give you altitude data.

I think it has been peer reviewed? http://www.nature.com/ncomms/2013/131008/ncomms3554/full/nco...

What is really needed is access to the raw data.

Reply on news.ycombinator.com