An Experimentally Determined Persistence-Rate Correction Factor for Scat-Based Abundance Indices

1. What they did

Counting scat is a common way to compare how many animals live in different places. But scat does not last the same time everywhere: dung beetles, rain, drying and other factors can make it disappear faster in some habitats and seasons than others. Fewer scats found at a site may then reflect faster decay, not fewer animals.

The author proposed a simple way to measure how long scat lasts and use it to correct scat counts:

• Collect pellets known to be fresh from a place animals use heavily.
• Air-dry and mix them.
• Set out pellet piles at random or systematic spots in the study area and mark them.
• At the time of the scat survey, check what proportion of these piles is still visible (the persistence rate).
• Divide the observed scat counts by the persistence rate to get corrected counts.

He tested this on sambar, a large deer, in monsoon forests of northern Thailand, as part of a wider study on how poaching affects seed dispersal. He searched four half-hectare plots in each of four national parks, but found sambar scat in only six: four in Khao Yai and two in Nam Nao. In each of these plots he cleared all existing pellet piles from four belt transects (50 by 4 m) in mid-July 2003 and counted new piles in mid-October. On the same days he cleared the transects, he placed 10 piles of 10 sun-dried pellets at random spots in each plot, collected fresh in early July from a campground lawn heavily used by sambar. In October he scored each pile as visible or not visible.

2. What they found

• Correcting for scat loss changed the order of the sites. Uncorrected, the order was KY1, KY2, then KY3 and KY4 equal, then NN3 and NN1. Corrected, KY4 moved up from joint third to second: KY1, KY4, KY2, KY3, NN3, NN1.
• If scat lasted equally long everywhere, the ratio of corrected to uncorrected counts would be about the same at every site. Instead it varied by a factor of 2.7 across the six sites (roughly 1.25 at NN1 to about 3.3 at KY4, read from the figure) and by a factor of 1.6 between the two parks.
• The author speculated that the differences came from variation in dung beetle numbers, local weather and habitat; drier sites probably kept scat longer because less rain fell directly on it.

3. What they concluded

• Differences in how fast scat decays between places can be an important source of bias when scat counts are used to compare sites.
• Measuring a persistence rate with experimental pellet piles should make scat-based abundance indices more accurate, and gives a necessary value for models that turn scat counts into population estimates.
• The correction matters most in “standing crop” surveys, where plots are not cleared first and the age of scat is unknown, but it should also be used in “clearance plot” surveys, to account for scat that disappears between clearing and recounting.
• Setting out experimental piles avoids two problems with marking naturally found fresh scat: it can be hard to tell whether scat is fresh, and enough fresh piles can be hard to find.

4. How tracking was or was not used

The study is a sign survey: sambar were counted through their scat pellet piles on belt transects, not seen directly. Sambar scat was identified and told apart from muntjac and mouse deer scat by its size. Fresh scat was collected from a site of heavy sambar use, and experimental piles were scored simply as still visible or not.

5. Tracker or observer certification

Not reported. The author says he conducted all surveys himself to reduce observer bias, but nothing is said about his experience in finding or identifying scat.

7. Implications for conservation (from the paper)

• Scat counts are widely used to track mammal populations without seeing the animals. Without a correction for scat loss, differences between sites or years may reflect decay rates rather than real differences in animal numbers.