[Coral-List] Reliable coral reef stats

[John Bruno]

Re, David Fisk's comments about Reef Check:

My understanding is that RC was born during the infamous "Gathering of the tribes" in Miami specifically as a means to monitor reef state and test whether or not they were "declining".  To my mind, RC has successfully done that. Without the data generated by RC and its thousands of volunteers, we frankly would know very little about global trends in reef state, in particular coral cover.  I have published a number of papers using the RC data, e.g. (all of which are available here:  http://johnfbruno.web.unc.edu/research/ );  


Selig, E.R., K.S.Casey and J.F. Bruno. 2012. Temperature-driven coral decline: the role of marine protected areas. Global Change Biology 18:5 1561-1570









Butchart, S.H.M., M. Walpole, R. Almond, B. Bombard, J.F. Bruno, et al. 2010. Global biodiversity: indicators of recent declines.  Science 238:1164-1168






Schutte, V.G.W., E.R. Selig and J.F. Bruno. 2010. Regional spatio-temporal trends in Caribbean coral reef benthic communities. Marine Ecology Progress Series 402: 115-122






Selig, E.R. and J.F. Bruno. 2010. A global analysis of the effects of marine protected areas on coral loss. PLoS One 5:e9278





Yes, the RC methodology and data are imperfect. But so is that of AIMS, NOAA, AGGRA, etc. I've worked with all these datasets and they all have limitations. There is no perfect survey methodology-there are always trades offs.  I only use the RC "total coral cover data" for which I think their methodology is more than adequate. Their surveys could be somewhat less precise than those by scientists, but I don't think it really matters and I think their data are as (or nearly as) accurate.  The huge advantage of RC data is the gigantic sample size, even at regional and subregional scales for which many hundreds of surveys are performed annually. This is the BIG benefit of citizen science in other fields as well, e.g.; the Christmas Bird Count, Bob Cowen's new Plankton Portal http://www.planktonportal.org/, Snapshot Serengeti, REEF, etc. (see this important paper by Chris Stallings based on REEF data: http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0005333).






For coarse-grained survey work, many of us are happy to trade off a bit of precision for a big increase in spatial coverage.  






I have seen several manuscripts over the years comparing RC data with scientists (e.g., AIMS) surveys; the results were essentially the same. Unfortunately, none of this has been published or is really available (as far as I know). I think getting these comparisons out there would be a big help (hint, Gregor).  






I have done some similar comparisons, e.g., for Bruno and Selig (2007)  


http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0000711






"It is possible that the estimated recent and current coral cover in subregions with substantial Reef Check data was inflated by the Reef Check site selection protocol.  Reef Check instructs team leaders to survey the healthiest local reefs: “Teams were instructed to survey outer slopes on exposed reefs that were considered to be the ‘best’ sites in their area – those believed to be least affected by human activities and having the highest percentage of the seabed covered by living hard coral" [6].  From 1997 to 2004, 39% to 75% of the total annual coral cover estimates [for Indo-Pacific Reefs] were based on Reef Check surveys.  Therefore we may have underestimated the rate and extent of regional coral decline. Our estimate of current coral cover is possibly higher than the true average because 259 of the 390 surveys performed in 2003 were based on the Reef Check methodology.  Additionally, this could have confounded comparisons with subregions in which relatively few Reef Check surveys were performed.  The Reef Check cover estimate for 2003 of 21.1% ± 0.62 (n = 259) might be considered the average of the healthiest reefs in the subregions surveyed extensively by Reef Check (mainly the southwestern Pacific, Indonesia, mainland Asia, and the Philippines).  However, the 2003 Reef Check cover mean is slightly lower (t = 2.0, p = 0.04) than the non-Reef Check estimate of 24.1 ± 1..6 (n = 133).  In fact, our analysis of the effect of the data source on measured coral cover suggests that the site selection bias in the Reef Check methodology might not overestimate subregional or regional coral cover.  A robust quantitative comparison between Reef Check and non-Reef Check cover estimates is not possible because few subregions have been adequately surveyed in a given year using both methods.  But we cautiously compared the annual mean cover estimates for subregions surveyed by both Reef Check and non-Reef Check teams between 1997 and 2004 with a paired t test: the Reef Check mean (20.3 ± 0.8) was significantly lower than the non-Reef Check mean (32.1 ± 2.3) across the 21 subregion/year pairs for which data are available (t = 4.5, p = 0.0002).  These results suggest that the Reef Check site selection protocol might not bias estimates of subregional coral cover by preferentially surveying the highest cover reefs.  One reason for the lack of a significant bias is that there are very few remaining high cover reefs that could appreciably influence subregional averages (i.e., because there is little difference between the ‘healthiest’ local reef and the average local reef).  It is also possible that Reef Check volunteer teams lack the resources to adequately pre-survey local areas to a degree that would enable them to effectively identify non-representative high cover reefs."


  


6. Hodgson G (1999) A global assessment of human effects on coral reefs. Marine Pollution Bulletin 38: 345-355.