[Coral-List] [EXTERNAL] USACE used bad science to rule out useful instrument for siltation monitoring

Ulf Erlingsson ceo at lindorm.com
Sat Jan 30 10:50:33 EST 2016


On 2016-01 -23, at 12:24 , Jordan-Sellers, Terri SAJ wrote:

> Dr. Erlingsson - I am surprised that it took more than 2 years for you to review the two assessment reports conducted on the Sedimeter by the USACE Research and Development Center (ERDC) released in 2013 and 2014. It is my understanding from the researchers I spoke with that you were aware of the faults identified in their reviews and were working to remedy them in a newer version of your sedimeter. 

Yes, we were aware of their comments in general terms but until now I hadn't read the analysis of the accuracy and it turns out that it is 2 orders of magnitude wrong because of flaws in the mathematics, and that made me upset because I have actually believed in their results until now. 

Let me make it clear to all on this list that the USACE guys measured reasonably well, they just interpreted the data wrong, and there is no sign that there was malice involved. They just interpreted following a normal procedure, and ended up 2 orders of magnitude wrong because the sensor design of the SediMeter does not behave like your typical single-value sensor (imagine a ruler with numbers only for cm so you have to interpolate in between; but it has 1,000 divisions between the cm lines so you can see if it changes 1/1000 cm, you just don't know very well if it is 433/1000 or perhaps 448/1000). Thus, the error is systematic and cyclical, and also, we left the possibility open for the user to fine tune virtually everything in the bottom estimation procedure including the correction for this cyclical error. Doing so is optional, but even without doing so the error is less than 1 mm. 

As to the field test, it was the cleaner add-on that failed. They never tested the SediMeter SM2 without cleaner. It has been used in the field for years. The new model was released in October, 2013 (http://lindorm.com/pr/pr20131017.php) after one year of corrosion testing of the cleaner.


More information about the Coral-List mailing list