How do common research scales match up across ocean science disciplines?

Apparently not very well.  A couple weeks ago I came across a really cool paper by some current and former students at the University of Washington School of Oceanography and the UW School of Aquatic and Fisheries Science.  The authors are/were participants in the NSF IGERT* funded Intergraduate Program on Ocean Change (iPOC).  The paper, Disciplinary reporting affects the interpretation of climate change impacts in global oceans, was published in Global Change Biology, a high impact journal and no mean feat for a grad-student only publication.  I have some sense of what goes into that sort of project; I was involved in two grad-student led review studies at the UW.  The first, thanks largely to the perseverance of lead author Eva Stueeken, eventually saw the light of day after two years and one rejection.  The second was abandoned after one year, when we actually had a draft in hand and realized that it didn’t go anywhere (or maybe it tried to go everywhere).

In this paper lead author Donna Hauser and colleagues address the question of whether the spatial and temporal scales commonly studied by the different ocean science disciplines are comparable.  They quantified the temporal and spatial extent of 461 top papers split among three biological domains (benthic communities, vertebrates, plankton communities) and a physical/chemical oceanography domain.  That in and of itself reflects some bias, but the authors are biologists (and may therefor be forgiven for the more detailed look at biology), and there are reasons to suspect that biology, as a sub-discipline of the ocean sciences, functions a bit different from chemical and physical oceanography.

Not surprisingly the authors found that there are big differences in the spatial and temporal scales favored by these domains.  The above figure captures the spatial differences.  Very few of the biology studies selected by the authors were global in scale, though there were many that were multiregional.  There were also strong regional differences between domains.  Surprisingly there were no benthic studies in the Arctic (with a broad continental shelf the Arctic is known to host a rich benthic ecoystem – hence the walruses and other large bottom feeders), and no physical or chemical studies in the Mediterranean.  Plankton studies were lacking in the Indian (again, surprising; there are huge planktonic ecosystem shifts happening in the Arabian Sea).  A similar pattern is observed across temporal scales, with physical and chemical studies tending to address longer time periods:

 

So why the difference in temporal and spatial scales across the different domains, and is it a problem?  It isn’t that hard to imagine why the spatial scales differ.  Hausser et al. spend some time addressing motivations for undertaking studies at different spatial and temporal scales, but I think a big difference between sub-disciplines is simply in how each collects data.  The bread and butter of physical oceanography, for example, are temperature and salinity measurements.  These measurements are collected en mass  by hydrography cruises and by autonomous sampling platforms (e.g. floats and gliders).  The methodological challenges lie in data analysis and interpretation.  It’s a bit more complicated for chemical oceanography, and the study’s authors didn’t break that sub-discipline into specific domains, but data acquisition is also high volume for chemical oceanographers studying things like pH, dissolved oxygen concentrations, nutrients, and anthropogenic tracers.

For many biological oceanographers data acquisition remains the bottleneck.  The use of satellites to identify ocean color and, more recently, underway flow cytometry to quantify phytoplankton are important advances toward high-throughput sampling, but they remain the exception (and even these advances have major limitations).  Most biologists working in the ocean still 1) haul nets to laboriously count zooplankton and conduct experiments, 2) filter volumes of water for the exceedingly tedious job of extraction and analyzing RNA, DNA, and protein, or 3) actually jump in the water to count things.  None of these approaches are going to get you very many datapoints, even after a lifetime of research.  Thus spatial scales are inherently limited.  It’s a little less obvious why biological studies should be spatially and temporally limited.

The figure above breaks things down even further into topical areas and suggests to me that the temporal constraint, as for spatial scales, is methodological.  The techniques commonly applied to the study of archaeoplankton and bacterioplankton, for which a cutting-edge study might evaluate change across just a season, have really only been around since 2005 (for a very cool exception, or rather work-around, check out this paper).  And the techniques have evolved so rapidly since then that the 2005 stuff is hardly comparably to what’s sampled today.  This is compounded by some institutional laziness among biologists regarding record keeping, and inertia regarding the adoption of standards and high-quality databases.  The situation gets better as you move up the food chain toward topics that rely on more classic, records-based ecology.

So is this a problem?  Well, the authors probably wouldn’t have gotten published in Global Change Biology if it wasn’t.  There are (at least) two reasons this is a problem.  The first, which is really the thrust of the paper, is that you can’t integrate across disciplines and develop a comprehensive understanding of ocean change if everyone is studying things at different scales.  This isn’t just a scolding to biologists to upscale their studies spatially and temporally.  As the last line of the abstract notes, there is a need to measure biological responses at biologically relevant timescales.  Often the biologically relevant scale is just different from the scale that attracts expertise and funding to physical and chemical oceanography.  Biologists do need to break though their malaise and bottlenecks and generate analysis at broader temporal and spatial scales, but the funding agencies might also need to think creatively about motivating collaborative work at the most relevant scales.  That scale might not be global, and it might not be decadal, even if that’s the surest way to get papers published in Science and Nature.  The second reason this is a problem is a subset of the first, and has to do with the lack of historical records in biological oceanography, something that we aren’t in any hurry to fix.  We have the technology to conduct spatial and temporal analyses of microbial community structure and function at a moderate resolution, for example, but very few people are trying to implement those analyses into existing continuous ocean sampling programs.  At some point that’s going to have to change.

*In classic fashion, and following the GK-12 program, IGERT become popular and successful, and so was cut by NSF.  Funding will continue for existing IGERT programs until the end of their funding period, but there will not be new calls for IGERT programs.  As a graduate student I benefited from 3 years of IGERT funding through the UW Astrobiology Program, which not only paid my stipend but provided some small funds for research and travel.  IGERT was a unique and highly effective program (as was GK-12, and EPA STAR, and…) and I’m sorry to see it go.