In a convenient follow up to Chris’ post “replication schmeplication,” and in preparation for our Hurlbert conversation next week, I thought I’d discuss a paper from 2012 that really got me thinking about climate change manipulation experiments. I have always been interested in ocean acidification (OA), which (in case you are unfamiliar) is when excess CO2 from the atmosphere is absorbed into the ocean, leading to changes in carbonate chemistry and acidic conditions. Scott Doney from WHOI calls this “The Other CO2 Problem” – and it has recently been wreaking havoc in Puget Sound, where particularly acidic water is being upwelled into shellfish hatcheries.
An informative video can be found here:http://www.pmel.noaa.gov/co2/story/Acidifying+Water+Takes+Toll+On+Northwest+Shellfish
Despite the harmful effects of OA, the topic made it to the top of the list of important stories largely ignored by the media. But scientists haven’t ignored it. From 2000 to 2009, acidification experiments represented >60% of the marine climate change experiments, likely due to its novel and alarming implications (even though geologists have been aware of this phenomenon for decades – sorry, that’s the geochemist in me speaking).
In a similar tone to Hurlbert’s 1984 comment on pseudoreplication, Wernberg et al. (2012) wrote “A decade of climate change experiments on marine organisms: procedures, patterns, and problems” to address the challenges scientists face in marine climate change experiments (henceforth referred to as MCCEs). They compiled papers that made an explicit reference to climate change from 2000 to 2009 (total = 110), and commented on five major issues that need to be improved in order to further understand the potential effects of climate change.
One of the biggest problems was experimental design – 49% of the studies were diagnosed as having “issues with their experimental procedures”, and of those studies, 91% portrayed some form of pseudoreplication. This is a problem, in short, because it limits the inference space and the ability of researchers to extrapolate results.
So back to the OA problem: Most studies test, at the tank level, the effects of OA on say, physiology, calcification, behavior, etc. Depending on funding and resources (which doesn’t look good considering the new role of our favorite senator), it is difficult to ensure appropriate replication. It’s not easy nor cheap to manipulate the carbonate chemistry of multiple tanks (especially when costs are a factor), and so lots of studies will instead place multiple tanks into one large vat of acidified/treated water – hence, pseudoreplicate. Though a colleague of mine, who did his dissertation on OA effects on clownfish, was able to jazz up a unique system that specifically addresses this issue at a fairly low cost: http://mobile.tube.aslo.net/lomethods/free/2013/0485.pdf
I don’t want to jump the gun too much on the potential upcoming discussion in regard to Hulbert’s piece, but I wanted to gauge the class’ experience with pseudoreplication…
(And to briefly allude to factorial design, single factor experiments (e.g. testing the effect of warmth, or OA, or Sea Level Rise) accounted for 65% of all MCCEs. But as we know, there are likely concurrent, multiple stressors that are non-additive, and cannot be well-understood in isolation from one another). Interaction effects!
Wernberg, Thomas, Dan a. Smale, and Mads S. Thomsen. “A Decade of Climate Change Experiments on Marine Organisms: Procedures, Patterns and Problems.” Global Change Biology 18, no. 5 (May 02, 2012): 1491–98. doi:10.1111/j.1365-2486.2012.02656.x.