Scrapefin bows over a group of submerged dolphins, but how many dolphins are really somewhere in this picture?
This is not a political question, but how much did you spend on groceries last year? Oh, just give me an estimate.
Surprisingly, scientists collect data points as precisely as possible, but their numbers are also estimates. Granted, the precision of estimates vary; physics data are more precise than estimates of human and animal behavior. This truth recently swam by under windless blue skies dotted by restlessly forming and fading clouds that forecast the dolphins’ behavior we were about to see.
Survey 1 yielded six dolphins. Stick searched the seas unceasingly, apparently focused on feeding as she’s been since getting pregnant in 2010 with son Savannah. As she searched submerged, we searched the surface for her swirling yearling but he was not there. Though he could‘ve been out of our sight but within her hearing, our count was six instead of seven.
None were among the 17 dolphins we saw on Survey 2 the next day, reinforcing that the number of dolphins we collect as data points is strictly a function of our timing and their behavior. The comings and goings in the largest group showed that our count would have differed every 10 minutes!
A sleepy sextet came out of the Narrows. Mother-calf pair Valiant and Vidalia made a quick “exit stage right” though they waited a quarter mile ahead. The four remaining dolphins (bulls Schnoz and Xiphos, ladies J and Split) scattered across a broad shallow bay. Easy to count as a sleepy sextet, our count would’ve been off by half if we’d arrived after they’d scattered.
Three new dolphins joined them several minutes apart, each changing the actors and action like sequential scenes in a Neil Simon play. Local lady DD1 catalyzed a convergence of the ladies and slow southward trek, bulls Schnoz and Xiphos swirling around them like eddies. By and by, big local bull BB was suddenly there, J and Xiphos left, and the rest paraded down a slim finger of water soon dotted with bubble streams and staggering mudplumes. Whether these were the need to feed or snorting to each other, I cannot say.
“But where’s DD2?!” we cried. “BB and DD2 are always together!”
The summer of 2012 took a heavy toll on local dolphins, and even a fleeting feeling that DD2 might be among the dead triggered a decidedly unscientific pang. But he zoomed in with that bright dolphin light, cutting past the boat with the white dot on the tip of his split dorsal fin gleaming like a mast light.
His presence led to departures, passionate exchanges and a new count. DD1 sailed into the distance, leaving Split alone with the bull quartet. Schnoz courted her, his favorite female. DD2 took exception, rammed them and tackled Schnoz. BB piled on top. Split split, Schnoz dashed after her, BB and DD2 rocketed past them in an end-around, Schnoz and Xiphos responded with their own rocketing end-around, and finally something was settled.
BB and DD2 left Split to the other bulls and wandered south to Valiant and Vidalia, still waiting ahead. DD1 and J had faded away.
Estimates are based on samples. Samples are little ambassadors or reps of their parent population, so we use them as estimates of that population. I average my samples of six and 17 dolphins to estimate that the true population number of dolphins using the waters around John’s Pass on autumn mornings is 12.
How close is 12 to the true number of autumn dolphins? The hard lesson of science is that that truth is always hidden. No matter how hard Capt. John Heidemann and I try to get accurate counts, our estimates will always be off from this true number even if our sample slip-up is really small. Like the number of people counted in the recent census versus the real size of the US population, this inevitable difference between our real splash-and-dash sample and the actual dolphin population is called sampling error.
Sampling error is the difference between what you actually get and what you’re supposed to get. If you toss a coin, you’re supposed to get heads half of the time. But if you toss a coin ten times and get seven heads, the difference between the seven heads you actually got and the five heads you’re supposed to get is sampling error.
Sampling error is the soundness of a sample, which dolphins challenge constantly. Survey 3 was later that day. Ten dolphins headed into John’s Pass from the Gulf of Mexico. We’d seen six of them on one of the other two surveys: Big Schnoz with his favorite Split but also J, DD1, Stick, and Front Slash. All of these dolphin moms but Front Slash lost her calf this summer; the kids were somewhere between three months and four years old. The population, weakened without them, will require years to replace them.
When I saw the other four dolphins with them, mother-calf pair Forest and Meadow and probable bonded bulls Gem and Hoffman, I almost collapsed. Where had THESE four come from? Forest and Meadow are periodically present around here and Gem and Hoffman are occasional visitors at best.
Just when I think I’m vested in the quick-silver tenor of these intelligent animals and as aware of their rhythms as anybody, they do something unexpected like this. Their glaring unpredictability makes me wonder if we’re looking in the wrong place for their truth.
Maybe it isn’t their population estimate but in the sampling “error,” the relative constancy of their inconstancy.
In any case, as the tidal wave of political “statistics” heads your way this fall, remember that most numbers are estimates with wobble-room called sampling error. Who would’ve thought that the freedom of wild dolphins could’ve armed you so?
Dr. Weaver studies wild dolphins under federal permit 16299, National Oceanic and Atmospheric Administration. Send her an email at dazzled@tampabay.rr.com or visit her website www.dolphinsuperstore.com. Read her award-winning Dolphin Watch column weekly at www.TBNweekly.com. NOAA advises anyone who sees a stranded dolphin in the Gulf of Mexico to call 877-942-5343 or 877-433-8299.
![[Image]](/content_images/101112_out-01.jpg)
![[Image]](/content_images/070312_pco-01z_front.jpg)
