An NCEAS working group I was a relatively minor participant in has an important new paper out in Nature Climate Change (Poloczanska et al 2013 [PDF]).
Summary: We synthesized all available studies of the consistency of marine ecological observations with expectations under climate change. This yielded a metadatabase of 1,735 marine biological responses for which either regional or global climate change was considered as a driver. Included were instances of marine taxa responding as expected, in a manner inconsistent with expectations, and taxa demonstrating no response. From this database, 81–83% of all observations for distribution, phenology, community composition, abundance, demography and calcification across taxa and ocean basins were consistent with the expected impacts of climate change. Of the species responding to climate change, rates of distribution shifts were, on average, consistent with those required to track ocean surface temperature changes. Conversely, we did not find a relationship between regional shifts in spring phenology and the seasonality of temperature. Rates of observed shifts in species’ distributions and phenology are comparable to, or greater, than those for terrestrial systems.
This study is a milestone in research on how greenhouse gas emissions / ocean warming are affecting marine life. It also provides evidence that the oceans are warming. (Yeah, I know that combined, there is a bit of circular reasoning in those statements). Climate change deniers often fuss about the imperfection of temperature data; however, the biological responses of plants and animals provides equally compelling evidence of changes in temperature. We can predict how organisms will respond to warming or cooling based on a number of disciplines, including biophysics, ecology, paleontology (the fossil record), and physiological ecology.
We found that all organisms for which data is available (including marine mammals, sea birds, turtles, fishes, sharks, squid, plankton, invertebrates, mangroves, seagrasses, and seaweeds) in general, have responded as predicted to ocean warming. Responses include shifts in the geographic distribution of populations (to higher, cooler latitudes), change in phenology, organismal fitness, population abundance, and community structure.
Figure 1. Global distribution and regional location of marine ecological climate-impact studies. Observed responses (n=1,735) of marine organisms to climate change from 208 single- and multispecies studies showing responses that are consistent with climate change (blue n=1,092), opposite to those expected (red n=225) or are equivocal (yellow n=418). Each circle represents the centre of a study area. Where points fall on land, it is because they are centroids of distribution that surround an island or peninsula. Pie charts show the proportions within regions bounded by red squares and in the Mediterranean Sea; numbers indicate the total (consistent, opposite plus equivocal) observations within each region.
We did not do any primary analyses; instead we synthesized published field studies assessing the potential effects of anthropogenic climate change on marine organisms. We assessed the number of climate change impacts studies that observed biological changes consistent with warming. We got together four times at NCEAS for a week and argued, debated, collected data and studies, analyzed data, wrote the paper, analyzed the data again, etc. Add to that many thousands of emails and dozens of group video conferences.
It wasn't glamorous and was a ton of work. This is what an NCEAS working group looks like:
Yup; a bunch of scientists with laptops and an internet connection. Are you sure you want to be a marine biologist?
We spent countless hours figuring out how to do this. And there was intense debate and some unhappy participants. It was incredibly rigerous in ways that climate change deniers couldn't image and wouldn't believe. (Scientists really do LOVE to play devils advocate).
We used a “vote counting” technique to assess how many studies of different taxa, regions, and responses found effects consistent with predictions about responses to warming. We could not use a formal meta-analysis (which is often used to synthesize experimental studies) because there was no way to determine what proportion of observed responses in the component field studies was due to anthropogenic warming and other factors. Each component study combined physical data (generally ocean temperature) with a biological time series at least 19 years long that extended beyond 1990. Response variables included population range limits and density, individual fitness (e.g., growth and reproduction), and the timing of organismal development and important life history events like spawning. We only scored the authors interpretation of the results and we did not reanalyze, reinterpret, or weight studies by duration, “quality” etc. We spent many hours debating this aspect of the design in particular: being scientists we were not always in agreement with the authors interpretation and were at times critical of the analyses.
Response was ‘consistent’ with direction of change expected under climate change, 2. Response was ‘inconsistent’ with direction of change expected under climate change or the consistency with or against climate change was unclear, that is no clear expectation of change was given or was available post hoc and 3. ‘No change’ indicating no response was found. We used on the authors’ expectations of direction of response to climate change.
We were mindful of a variety of potential pitfalls like duplication of observations and we obsessed about publication biases that could affect our results (e.g., if studies that found significant climate change effects were selectively submitted or published).
Download the manuscript PDF here, the supplementary information PDF here, go here for a nice summary by the lead authors, and here for the NCEAS press release.
Posted by John Bruno on Wednesday, 11 September, 2013
 |
The Skeptical Science website by Skeptical Science is licensed under a Creative Commons Attribution 3.0 Unported License. |