[Coral-List] USACE used bad science to rule out useful instrument for siltation monitoring
reedkc at comcast.net
Sat Jan 23 13:52:33 EST 2016
A quick dovetail anecdotal comment about the socioeconomic impact of the Miami siltation fiasco: several of us just heard from one of the stateholders at the FRRP ten year symposium (Florida Reef Resiliency Program) how Miami scuba shops can no longer realistically offer scuba or snorkeling to tourists in the local area, but have to charter trips to another county to get decent water quality and marine animals to casually observe; or, the Miami scuba shops diversify and have tourists kayak or paddle board, so they can ignore what's under the water's surface.
Dr. Keven C. Reed
----- Original Message -----
From: "Ulf Erlingsson" <ceo at lindorm.com>
To: "Coral List" <coral-list at coral.aoml.noaa.gov>
Sent: Saturday, January 23, 2016 12:15:55 AM
Subject: [Coral-List] USACE used bad science to rule out useful instrument for siltation monitoring
Before the dredging of Miami harbor (which caused a lot of siltation on the coral reefs) the US Army Corps of Engineers evaluated an instrument I invented, the SediMeter. It is capable of detecting as little as 100 grams per square meter of sedimentation (about 0.01 mm) and the accuracy is in the order of 1 mm. However, in this document (el.erdc.usace.army.mil/elpubs/pdf/doert11.pdf) the USACE used a series of flaws to arrive at the conclusion that the SediMeter is only useful for measurements "where bed-level changes are on the order of centimeters". Today for the first time I took the time to read their "study", and found that it wouldn't pass as an undergraduate paper if I had evaluated it.
The problems start small with a too small measurement tank, but that's a minor thing. Next they moved the instrument rather than the bottom, and they noticed that it led to errors but kept doing it. Still worse, they don't seem to grasp the concept of statistics, believing that all data have a Gaussian distribution if you measure enough many times. From a standard deviation of 0.84 mm they thus extrapolate that the 95% confidence interval is "±1.7mm" even if the data they present never exceed +1.1 mm or -1.0 mm in 89 samples. The last error is the worst, however. The ±1.7 mm value refers to the difference between two bed levels, but in the conclusions they state that it is an error in the bed level, and double it to arrive at "total" error for the difference. This value, 3.4 mm, they then use to argue that the instrument is only useful when the difference is several centimeters..
This is simply a junk study from the part of USACE. I will personally do a proper measurement and will be happy to provide a correct estimate of the uncertainty to any coral lister who is interested in monitoring siltation on coral reefs, especially WHEN IT COMES TO SUPERVISING THE DREDGING ACTIVITIES OF THE USACE.
Ulf Erlingsson, Ph.D.
President and CEO
10699 NW 123 St Rd
Medley, FL 33178
ceo at lindorm.com
+1-305 888 0762 office
Coral-List mailing list
Coral-List at coral.aoml.noaa.gov
More information about the Coral-List