Monday, August 29, 2011

AGoogleADay-inspired research

 One of the pleasures about writing questions for the AGoogleADay (AGAD) series has been to see reader reactions.  Most often people write in with thanks for a question they hadn’t thought about, or a perspective that searching to solve a question gave them. 

You’re wrong!  On the other hand, some people write with quibbles or questions about the AGAD daily challenge.   About a third of the time they have a legitimate complaint about a correct answer not being accepted by our automatic answer verification code.  (“The answer is 10,000,001 and YOU marked it wrong!”  True.  We should have allowed for a little more leeway in the answer range…) 

We try to fix these as quickly as possible, which can be tricky.  In the question about the “southern-most temple at Machu Picchu” we had folks saying we should also accept answers in Spanish (“Templo del Sol” in addition to “Temple of the Sun”), although my favorite complaint was that we should have also accepted answers in the language of the Incas since it was an Incan site!  If only I knew what that temple was called in classic Quechua…

I’m right!  Most often people write to let us know that their answer was correct, when in fact they hadn’t read the question fully.  This is probably the most common email we get, which worries me slightly.  The people playing AGAD are for the most part pretty careful readers and searchers—such comments always make me review the question to see if we couldn’t have phrased it better.  But often, no, we couldn’t have… they just misread the question and found an answer to a question THEY had in their head, not what we actually asked. 

More recently we ran a question (Aug 28, 2011): “One particular kind of bird migrates farther each year than all others. If you added all the miles an individual bird of this kind flies in an average lifetime, how many times would it have circled the Earth?”

There were three interesting complaints about this question...

1.  Your answer is off by a huge margin!  We had a fair number of people write in to say that the arctic tern (the bird in question) actually flies around 24,000 miles in an annual migration, and not (as we say in the answer, 44,000) .  This is a great question about research on the web.  Why? 

When you do the obvious search for [ arctic tern migration ] you’ll find several sources agree with the 44,000 number: -- 44,000 miles / year

But some give a different answer -- 22,000 miles per year:  -- 22,000 miles/year

What's going on with the variation?  (Looks to me as though someone mistakenly copied a km measure for a mile measure...)  

To find an authoritative result, I turned to the Proc. National Academy of Sciences (a very well-known, very authoritative journal in biology).  In the article:

     Tracking of Arctic terns Sterna paradisaea reveals longest animal migration
     PNAS 2010 107 (5) 2078-2081; doi:10.1073/pnas.0909493107; February 2, 2010
     Authors: Carsten Egevanga, Iain J. Stenhouse, Richard A. Phillips, Aevar Petersen,
            James W. Fox, and Janet R. D. Silk

The authors write:  "Our tracking of 11 Arctic terns fitted with miniature (1.4-g) geolocators revealed that these birds do indeed travel huge distances (more than 80,000 km annually for some individuals)."

Which is actually around 49,700 miles / year round trip. 

Why such a long distance?  Not only do they fly pole-to-pole (which WOULD be around 24,900 miles/year), but they also fly back and forth in a daily hunt for food, an activity that basically doubles the distance they fly.  It's a remarkable feat!

The moral (from a SearchResearch perspective) is that errors happen.  They happen all the time.  In this case, it looks like a units conversion problem that happened at one place and was copied along.  (And, FWIW, the Wikipedia entry on this is correct!  Many other sources that look decent are in fact completely wrong.) 

2.  How can you ask questions that need precise answers given that the information is approximate? 

 A few other writers worried about trying to do a computation of “average lifetime migration distance” (which is what we’re asking you to do).  How’s that possible?

Here’s what I did:  If you search for the bird’s lifetime and find it varies (from 20 – 25 years) and then search for estimates of the birds migration distance (44,000 – 49,700 miles / year) you can compute upper and lower bounds. 

Upper bound (that is, maximum distance traveled) = 25 * 49,700 = 1,242,500 miles
Lower bound (minimum distance traveled) = 20 * 44,000 = 880,000 miles

If we take the earth circumference as = 24,901 miles (measured at the equator—note that you get a different number if you measure around the poles)

Upper bound:  1,242,500/24,901 = 49.8 times around the globe
Lower bound:  880,000 /24,901 = 35.3 times around the globe

BUT… unlike other game shows, AGAD accepts a range of answers.  In this case, anything between 35 and 50 would have been okay.

3.  What about….?  

Several people wrote in with other long-distance fliers.  The muttonbird and the sooty shearwater were also posited as the winner in the “longest migration” flights. 

Well, the term “muttonbird” refers to any of a number of seabirds that are commonly eaten.  One such bird is—surprise surprise!—the sooty shearwater, Puffinus griseus

If you look at the migration routes for both the shearwater and the arctic tern, both birds essentially migrate from pole-to-pole over the course of a year.  So I suspect they both travel essentially the same distance.  It’s possible that shearwaters actually travel farther, but the overall distance will be about the same. 

If we look up a comparable report from the PNAS about shearwater migration distances, we find:   

        PNAS  103(34): 12799-12802. (2006) doi:10.1073/pnas.0603715103
          Authors: Shaffer, S.A.; Tremblay, Y.; Weimerskirch, H.; Scott, D.; Thompson, D.R.;
               Sagar, P.M.; Moller, H.; Taylor, G.A.; Foley, D.G.; Block, B.A. & Costa, D.P.

These authors write:  “… shearwaters fly across the entire Pacific Ocean in a figure-eight pattern while traveling 64,037 km (+/-9,779 km) roundtrip, the longest animal migration ever recorded electronically.”    (Or:  39,790 miles, +/- 6076 miles)  Much less than the arctic tern. 

Note that this was in 2006.  The previous paper about the arctic terns was published in 2010.  Darn it.  Scooped! 

Bottom line for search/research:  As we know, research is a tricky business that needs constant digging to get deep enough to the bottom.  And, of course, success is often temporary.  I’m sure that authors Shaffer, et al., thought they’d found the longest migratory bird, but then the arctic tern paper comes out a few years later with a longer migration path.  It’s possible there’s still another bird out there that does a longer commute.  But that’s the great thing about science—we constantly update our knowledge as we learn more about the world.  As John Maynard Keynes has been said to have said,  “When the facts change, I change my mind. What do you do, sir?”  (I’ll let you do the research into whether or not he actually said that or not.) 

No comments:

Post a Comment