Now, continuing on my previous train of though. There are many ways in which ecology gets complicated and tends not to look as pretty as the previous post shows.
Let's look at some of my MS (Master's of Science) work as a starting point.
I worked in a field called paleoecology in order to tie in aquatic species with water quality to look at pollution affects on a lake over time. Diatoms (see Fig. 1) are a form of algae that is made up primarily of silica, and therefore fossilizes in the sediment of the lake they live in when they die. What we can do then is take a core of the lake bed. Each year a new layer of sediment is added to the bottom of a lake. We can take a core of a lake, date it using chemistry (such as Lead 210 dating) and compare what things are in each layer of the core with the associated date. This allows us to track ecological community changes (the relative number of each species or population of diatoms in a year) through time. NOW in addition to all that knowledge you can also run chemistry on the surface of the lake bottom (before the chemicals mix or breakdown). You can compare the water chemistry (things like nitrogen, phosphorus, oxygen, pH, or temperature) with the species found and their relative numbers. This can tell you the preference and tolerance of these species to a given water chemistry factor. So species that have a lot of individuals fossilized at the same level as the water chemistry you tested do very well in those conditions (say high nitrogen, which remember is basically a fertilizer). Species that are low in numbers (or not in the sample) do poorly under those conditions. With me so far?
Fig. 1: Diatoms
Fig. 2: A core from the bottom of a lake (depth of mud at bottom)
Okay then so here's the kicker, with the relationship of species to water chemistry, you can look back at older layers of the lake sediment and count the number of species and individuals that are fossilized. Once you know the community structure, you can use that information to estimate what the water chemistry was likely like! You can start to understand how a lake has change over hundreds or even thousands of years!!
Now of course, is where is starts to get dirty. Remember this is still ecology. So what happened with me is, after 2 years (normal length of a MS program), it didn't work. But in a way that's what made the project really interesting (and of course at the time very frustrating). When I ran my estimates and compared them to known emission rates of nitrogen from the area they were exactly inverted! My model checked out well against expected ecological relationships (we call it r squared, or how tightly the relationship of your data is). For something more linear like chemistry or physics you would never accept and r squared value below 0.8 or so (on a scale of 0-1). In ecology? Remember 0.4 is fantastic and numbers like 0.1 are not even unusual. The reason is that there are many factors that effect an organism and its habitat and the variable you are testing is likely only one piece of a much larger puzzle. That doesn't mean it's not important, it just means it's complicated and intertwined.
In my case what we think may be happening is one particular species tends to completely dominate the older sediments, and the problem it this species is a known generalist (often can do well in many different conditions and it quite good at competing with other species for a limited resource). I proposed that the model needed to be able to ignore such species. More in depth later. As a result I have a very controversial paper in lieu since the model has been used for 30 years. So it's been hard to publish. See previous Science post for likely reasons why...
"Science is facts; just as houses are made of stones, so is science made of facts; but a pile of stones is not a house and a collection of facts is not necessarily science." Henri Poincare
