[Coral-List] Data quality of any repeated reef survey including Reef Check

Gregor Hodgson gregorh at reefcheck.org
Wed Jan 9 13:18:07 EST 2013


Don,

I'm sure I am not the only coral lister who is wondering why an experienced
reef guy like you would be wondering why a repeat transect survey might show
differences in data on giant clam numbers.  Why would you expect them to be
the same? Using most survey techniques including Reef Check, there are
several reasons why there might be differences between repeated surveys of a
given transect that is marked at one or even both ends with an accurate GPS..
Unless the track of a long transect line is marked every 2 to 3 meters it is
HIGHLY unlikely that it will follow the exact same track each time,
therefore the attached benthic organism counts may change each time. Few
research programs are willing to take the time and effort to permanently
mark a long transect track, and there are also issues with using permanent
transects. So all researchers face this variability and use larger sample
sizes as one way to reduce it. Changes in giant clam numbers between surveys
are most likely caused by a different location of the transect than a
misidentification or missed observation of an easy to identify large animal
like a giant clam. (But there were those transects in Vietnam with over 1000
small clams per transect Moshira Hassan will not forget.)

In the case of Reef Check, we use a 100 m long transect line and monitor two
depth ranges for a total survey area of 800 m2 actually surveyed. This is a
pretty big patch of coral reef. But whether the data recorder is a trained
volunteer or a PhD coral reef ecologist, the transect will be in a slightly
different place ‹ more so when waves and currents are strong/variable.

The question of reliability of trained certified volunteers which you seem
to be questioning was answered long ago in several comparison papers in
tropical and temperate environments. In fact, so many researchers rely on
Reef Check methods that over 50% of Reef Check teams are comprised of
research scientists with Master's degree or PhD in coral reef ecology. We
have a standard automated error checker and 3 levels of human error checking
as part of our QA system. We throw out about 30% of data received due to
unresolvable issues. In my experience, most PhD marine scientists tend to
specialize in one taxonomic area and are just as prone to errors in
identification outside their "comfort zone" as trained volunteers. Its also
worth bearing in mind that Reef Check only uses about 20 indicator organisms
mostly at the family level and requires each volunteer to be trained, pass
ID tests and be certified in Reef Check. Bear in mind that many academic
institutions rely on "beginner" undergraduates to collect data for their
research programsŠ.so the point is we are all in the same boat. There are
going to be differences and errors in any repeated survey of a given reef,
but we have put more effort and thought into reducing those errors than
most. 

Reef Check data have been used at a global, regional, national and local
reef scale to track reef health in about 90 countries and territories and in
numerous peer-reviewed publications -- see our publications page:
http://reefcheck.org/about_RC_Reef/Publications.php

Because of these sampling issues that are common to all sampling techniques,
we recommend a sample size of 3 to 5 sites when the data are going to be
interpreted at a local scale such as an individual reef or small MPA.

Hope this helps.

Gregor

 Harding, S., C. Lowery and S. Oakley (2000) Comparison between complex and
simple reef survey techniques using volunteers: is the effort justified ?
Proceedings 9 "' International Coral Reef Symposium, Bali, Indonesia 23-27
October 2000, Vol . 2. <http://reefcheck.org/PDFs/Harding%202000.pdf>

 Gillett G, Pondella D,  Freiwald J, Schiff K, Caselle J, Shuman C, Weisberg
S. (2011) Comparing volunteer and professionally collected monitoring data
from the rocky subtidal reefs of Southern California, USA. Environmental
Monitoring and Assessment: 0167-6369: DOI: 10.1007/s10661-011-2185-5
<http://www.springerlink.com/content/ar657762m0m62535/>

Gregor Hodgson, PhD
Executive Director
Reef Check Foundation
PO Box 1057 (mail)
17575 Pacific Coast Highway (overnight)
Pacific Palisades, CA 90272 USA
T: +1 310-230-2371 or 2360
Gregorh at reefcheck.org
Skype: gregorh001

  

------------------------------

Message: 4
Date: Wed, 9 Jan 2013 11:05:38 +0800 (SGT)
From: Don Baker <reefpeace at yahoo.com>
Subject: [Coral-List] Reef Check - Data Useful or Subjective to Human
Error or Bias Perception?
To: "coral-list at coral.aoml.noaa.gov" <coral-list at coral.aoml.noaa.gov>
Message-ID:
<1357700738.99631.YahooMailNeo at web194701.mail.sg3.yahoo.com>
Content-Type: text/plain; charset=us-ascii

Dear Coral-Listers,

This may have been covered in the past but to get some current feedback on
Reef Check data and methods.... and if such is useful and how useful?

For example I have noted in a Reef Check data collection on a set survey
location [(via marker buoys) & GPS coordinates as backup for lost / stolen
buoys] for eight (8) consecutive monthly surveys, whereas, giant clams have
'appeared' and then 'disappeared' from month to month. Fish swim in and
out...but giant clams do not.

And while I am at it....to address the usefulness of Coral Watch as well
within the same Subject heading herein.


Comments?

Regards to all,
Don

------------------------------

_______________________________________________
Coral-List mailing list
Coral-List at coral.aoml.noaa.gov
http://coral.aoml.noaa.gov/mailman/listinfo/coral-list

End of Coral-List Digest, Vol 53, Issue 9
*****************************************





More information about the Coral-List mailing list