Of beginnings and biases

So it’s the summer holidays for me – or at least it is for another week, then it’s back to Southampton University for my last year as an undergraduate student. As part of the summer break I have been doing some (though not nearly enough) work on my Masters of Science project which I explained not very well in my post “My Msci Research Project – Macroecological changes in offshore shelf seas” and now seems as good a time as any to review what I’ve done; and to do so publically in the name of open access science!

Essentially the upshot of the work I have been doing has been to look at the project as described in the previous post and thinking about the methodology and the approach that I will use in undertaking the research. The result of this? I now have a two page list of biases for which I will have to find solutions, adjustments or reasons to discount them and options as to my methodology.

Geological time-scale represented in spiral form

Geological time-scale, the past being at the base of the spiral, image from USGS

Scope

The scope of the project – if taken to it’s pedantic conclusion – is epic, and far beyond either my research experience and the time-frame in which I am able to undertake it. Taking a census of every publically owned fossil from every one of the deposits that I could study would be not only prohibitively expensive due to air fares etc, but very, very time consuming. The solution? to do what other researchers have done when studying palaeoenvironment, take the smallest reasonable stratigraphic unit, in some cases this will be at the level of the formation (such as the Blue Lias) and for others the member (such as the Oxford Clay, where the Peterborough Member is promising), this will also hopefully give a more accurate set of size structure profiles for the study as it will be closer to taking a snap-shot of the palaeo-ecosystem at any one time.

Resolution

The second and last problem for which I have a definitive solution at present is how best to get a suitable resolution for the project, there were two approaches to this and both had their merits:

The pragmatic approach is to look only at the most complete fossil records within palaeontology’s current appraisal of the fossil record.

The idealistic approach is to take the 200 million year scope of the project and split it into “sensible” intervals, and take a sampling at each… say… 40 million year step.

Now if the fossil record was perfect (or even near representative) the decision would be a no brainer, the idealistic approach would be best, and the most scientifically defensible. The idealistic option is certainly what you would do if you were undertaking a study whose study period started tomorrow as it gives the best temporal resolution (or view of changes over time). The record is demonstrably inappropriate for the task, as anyone who is familiar with the palaeontological importance of lagerstätten to the field will understand – most animal carcasses never preserve, and most fossil sites are heinously misrepresentative of the original ecosystem. So in this instance it is more sensible to forego temporal resolution in favour of the best fossil records, with the hope that it will produce a more accurate result in terms of the ecosystem. The risk being that there may be large gaps which make the study less useful in studying changes over time.

Now to move on to some of the problems for which I have no solution as yet…

Scaling of partial specimens

How does one go about using a fossil for a body size metric if you only have a skull? Well in practice when it comes to skulls it is quite simple as most animals have a complete fossil somewhere and the scaling is simple as a proportion of body length – so long as you find the right paper! What about Ichthyosaur paddles, fish ribs or other less simple fossil body parts?

One option is obvious… ignore partials… but by heaven that is the worst solution and would discredit the study before anyone got beyond the abstract! So what else can I do? some people have said that if you know the number of a vertebrae (i.e.: where it comes from along the vertebral column) you can work out a vague size… is that good enough? What about anything other than skulls and vertebrae; can I ignore them?

Prominent British geologist and palaeontologis...

William Buckland | Image via Wikipedia

Collection Biases

Many of the deposits I’ll be studying have the same problem, their fossil collections were mostly excavated in victorian times, when science was done by country gentlemen or rich folks who wanted the biggest, most interesting vertebrates or even other fossils.

Chris Andrew (Lyme Regis fossil expert) related to me a story about William Buckland‘s coprolites (fossilised excrement). When they were surveyed approximately 1 in 20 had Ichthyosaur remains in them. The survey suggested that this is the proportion in reality within the blue lias, yet any of the miriad fossil hunters on the coast can tell you that this isn’t even close to true, finding many coprolites and maybe being lucky to have one a year with such remains within them. The cause of the disparity? The collector had only purchased coprolites that they thought interesting… ambiguous ones with nothing “interesting” were not bought, so biasing the result.

Similarly institutions are going to be more willing to buy specimens that show something different to what they already have in their collections, because if they bought everything they would run out of space! Should I consider all museum collections biased? How does one adjust one’s research outcomes to counter this? Should I even try?

Preservation Biases

On a very similar (but non-human) level and as already stated, not all animal carcases preserve. Some types of animal preserve better than others (e.g.: Ichthyosaurs compared to fish), how can this be tackled?

Actually there is a lot of research on this, it has it’s own field within palaeontology – “Taphonomy” –  but I haven’t researched it yet, and even then the aforementioned Chris provided me with another option which I may describe elsewhere if I use it, or even if I don’t.

In closing, this just gives you an idea of the miriad of problems anyone would face when they are trying to come up with a piece of research, and by golly I can understand why researchers spend so very long on research applications, justifying every decision you have to make is a real eye-opener. Long gone are the days of GCSE science classes when I could say I thought it was the best option because the teacher had advised it!

Ben Brooks

My Thanks to Dr Tim Ewin (Natural History Museum), Mr Paddy Howe (Lyme Regis Museum) and Mr Chris Andrew (The Fossil Workshop, Lyme Regis) for being willing to discuss my project with me and helping me identify the many problems that I now know of, though I’m sure there are more to come!

Short-Link for this post: http://wp.me/pFUij-7R

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s