EDIT: Part of analyzing interactions depends first on finding out what you have and how much. For example, it's a little hard to analyze prey/predator interactions without first knowing what predators and what prey are available and how much. Knowing how much area it takes to support a particular population of shrews is rather critical. And it also helps to know what is available for the shrews to eat and how those numbers change over time and space. Hard to analyze behavior if you don't have any data on what lives where, when and in what numbers.
I'll give an example I've done. We wanted to look at pesticide effects on bird populations and set up a system very like the one described in the paper you cite. Over several hundred square miles and lasting about three years. We randomized our sampling areas (bird boxes, mist nets, small mammal traps) and did a lot of sheer counting. It's hard to tell if you're getting an effect from the pesticide if you don't have a good idea of normal behavior. We had teams dedicated to nothing but quality control to make sure our samples were properly randomized and all the raw data was done in special permanent ink on special paper that could not be erases w/o leaving evidence. Any corrections in field data had to be initialed and referenced (date, time, reason why). Two separate teams did data entry and if the numbers didn't check. . . not pretty. Then another team at the other end went back over the raw data and checked it against the entered data. That had to check. Your assertion that ecologists/biologists aren't going into the field is simply counterfactual. There are hundreds of permanent field research facilities (http://www.dmoz.org/Science/Biology/Ecol
". . .I have written paleo papers as well as others."
Below find a simple overview of the technique(s) being described in the paper you cite. You can also find similar techniques described in almost any type of filed research (see Maritime archaeology by Keith Muckelroy as an example of how seemingly disparate disciplines use similar techniques) Now perhaps I've misunderstood you, in which case I apologize. However it appears, to me, that you find ecological research in general and this paper in particular lacking. I'm simply asking that you be more specific in your objections. I am not suggesting that there is not a misuse of mathematics in science. There is. Nor am I suggesting that Paleontology is not heavily descriptive. However lying, by commission or omission is (IMO) common to human behavior in general (How to Lie with Maps (2nd Edition) Mark Monmonier &
How to Lie with Statistics by Darrell Huff ) and the limitations of paleontology are not uncommon nor unexpected. So my question is, what specifically do you find that differentiates Ecology or Paleontology from any other scientific endeavor in terms of bad science or bad scientists?
Or, to use your GIGO analogy, what makes their garbage any stinkier that any other garbage?
Simple random sampling in the field