Microkhan by Brendan I. Koerner

Counting the Jumbos

October 6th, 2009 · 2 Comments

While perusing this AFP piece about a poaching bust in the Central African Republic, we stopped and mumbled “hmmmmm” upon reading this hard-to-swallow stat:

Experts say some 38,000 African elephants are killed each year for their tusks.

Really? That seems like such a ridiculously high figure; at that clip, wouldn’t the species (or, to be precise, the genus) be entirely extinct before the next Summer Olympics, if not sooner?

As it turns out, African elephants are seemingly much more plentiful than their Asian brethren. But we must take the estimate of 600,000 survivors with a grain of salt, because counting pachyderms is a tricky, tricky business. The animals not only hide from aerial surveyors but, as we’ve noted previously, may conceal the corpses of their deceased relatives. So what’s the solution? Using sound:

Geophones were used to record the footfalls of elephants and other large mammal species at a water hole in Etosha National Park, Namibia. We were able to discriminate between species using the spectral content of their footfalls with an 85% accuracy rate while only using a single geophone. This was done using correlation coefficients comparing the shape of the spectra for various species. An ANOVA found significant differences between these correlation coefficients (F4,1785 = 147.78, P = 0.000). An estimate of the energy created by passing elephants (the area under the amplitude envelope) can be used to estimate the number of elephants passing the geophone.

We do wonder, however, if such a method is able to account for the footfalls of babies—or, for that matter, how it differentiates between the sounds made by elephants, hippos, and rhinos in areas where the giant mammals co-exist.

(Image via Sideshow World)


Tags: ······

2 Comments so far ↓

  • Jordan

    I would think that the ability to discriminate between species is going to be dependent on the overlap between the mass distribution curves of each species. It looks like the lower end of elephant’s average mass (3600) overlaps with the upper end of white rhinoceros average masses (up to 3500-4600 kg on the extreme side) and usually exceeds the upper end of black rhinoceros average masses (850-1800 kg) and hippopotamuses average masses (1500-3200 kg). Given the extremely small white rhino population, they probably won’t get too many false positives that way. Also, it may be that elephants have a particular gait, which could be used for further discrimination in the signal.

  • Brendan I. Koerner

    This is def. one of those cases where I’d have loved to see the full paper, rather than just the detailed abstract. But the Microkhan budget (c. $0) doesn’t really allow for ScienceDirect access.

    Good point re: how differing weights between those animals would make for semi-distinctive sounds. The paper does seem to admit a rather large margin of error, but I reckon it’s still an improvement over the aerial spotting method (esp. in thickly jungled nations).