MathBench > Environmental Science

Sampling

 

Collecting P/A Data

Suppose you are interested in not only the dandelions in your yard, but also the moss. (You know how a moss infestation can give your lawn a mangy look that drives lawn enthusiasts a bit crazy?) How would you sample the moss? You certainly can't pick out an individual moss and count it...

Of course, you already know that moss should be sampled using the presence/absence method of counting. When we do P/A sampling, we don't pick a grid square -- instead, we randomly pick individual points to sample. And we'll need a lot of them to get a good estimate.

Just like for abundance sampling, we need to use a random sampling technique for presence/absence data. In this case, it's a bit like throwing a dart randomly at a map.

Once we complete our study, we will have an estimate of the moss’s range in the yard (for example, it may cover 10%, 30%, or even 75% of the sample area). We’ll have a better idea of just how serious the moss infestation is.

Let’s get sampling……….!

Use the applet below to begin collecting data for the randomly generated sample points.

The 10 sampling points out of 19431 possible spots is actually a pretty low sampling size -- I picked it mainly because it would fit on one screen. Did the percent moss you calculated come out close to the the actual percent moss? If it didn't, its certainly not YOUR fault ... instead, the problem was low sample size.

The same considerations of sample size apply to P/A sampling -- and you generally need more points to get a good sample... The applet below will do all the counting for you, all you need to do is say "stop" when you think that enough sampling has been done, then write up your great Moss Quantification article.

As you can see, the estimate of the moss infestation can fluctuate quite a bit, even with 50 or 60 samples.