Performance Verification Statement for theIn-Situ Troll 9000 Rugged Dissolved Oxygen Sensor

September 26, 2016

Dr. Tom Johengen

ACT Chief Scientist

CILER-University of Michigan

4840 S. State St.

Ann Arbor, MI 48108

Dear Dr. Johengen,

In-Situ was pleased to again participate in a study conducted by ACT. Throughout the study,

we have been impressed by the thoroughness, the comprehensiveness and the adherence to

experimental protocol, procedures and quality as assurance demonstrated by the ACT

personnel and its partner sites. We support the aims of the ACT instrument evaluations and

appreciate the clear, unbiased summary of instrument performance in both the laboratory

and coastal applications.

A thorough investigation of the 9500 sondes used in the lab testing and field testing was

conducted pre-deployment and post deployment. The data showed the sondes operating

within specified parameters during all tests at In-Situ and showed no evidence of impending

failure prior to deployment. Post deployment, the sonde used in the Kaneohe field test

exhibited the same behavior in the factory. It is believed that this instrument experienced

water infiltration that caused the failure.

Lab Results: The RDO sensor performed well in the lab testing across all temperature and

salinity levels. Minor offsets from the reference are believed to be due to changes in salinity

and lack of compensation for such changes.

Houghton, MI Field Test Results: The pre-deployment check of the RDO conducted at the

MTU site showed a positive offset which continued into the deployment. The positive offset

is believed to have been introduced during the user calibration. Post correction of this offset

should bring better alignment between the RDO and the reference instrument.

Chesapeake Biological Labs Field Test Results: Approximately half way through the

deployment, the instrument showed significant drift and reported values near zero mg/L

for remainder of the deployment. The deployment photographs show significant fouling on

the surface of the sensing material. As the sensor becomes completely covered with

biofouling, it is no longer able to take measurements and will drop to values near zero.

Because the instrument was still reporting values and there was a high degree of fouling