8 / 59
Ref. No. [UMCES] CBL 2016-013
ACT VS16-04
8
following the approach described in Bittig et al. 2014. The reservoirs of the thermostat baths were
constantly bubbled with either N
2
gas or air to maintain discrete DO levels. A submersible pump
was added to each bath to ensure uniform flow and oxygen conditions and instruments were
mounted at a fixed position within the baths to minimize variance due to instrument manipulation.
Instruments were programmed to measure every 10s continuously (or their fastest possible rate) for
eight minutes following the exchange. For instruments with the capability, real-time monitoring of
instrument output was monitored to verify a steady state reading had been obtained. Instruments
were moved from the high DO concentration to the low DO concentration and subsequently
reversed to check for response hysteresis. During transitions, care was taken to minimize
carryover by shaking off residual water. The sensor was then carefully inserting into the new
bucket and mixed by hand to ensure no bubble entrapment and full exposure to the new solution.
Reference samples from each reservoir were taken at the beginning and end of the exposure. The
test instrument was equilibrated in the high DO reservoir for at least 30 min prior to the exchange
to ensure temperature equilibration.
Lab-based Stability Test
A laboratory stability test was conducted to examine potential instrument drift in a non-
biofouling environment. These results are contrasted to the stability of measurement accuracy
observed in the long-term field mooring deployments. The test occurred over 56 days, with daily
temperature fluctuations of approximately 10
o
C, achieved by alternating the set point of the
recirculation chiller. Reference samples were collected at minimum and maximum temperatures at
least 3 times per week. The test was conducted in deionized water at saturated air conditions.
Tanks were well circulated and open to the atmosphere. Water in the test tank was exchanged as
needed if there was any indication of biological growth. Instruments stayed continuously
submerged and were not exposed to air during any water exchange. The goal of
comparisons of
accuracy over time between the field and a sensor deployed similarly in the laboratory is intended
to provide insight into drift and reliability intrinsic to the instrument relative to changes that may
result from biofouling.
Moored Field Tests
Field Deployment Sites and Conditions
A four month moored deployment was conducted at Michigan Technological University’s
Great Lakes Research Center dock in Houghton, MI. Instruments were deployed in January and
kept under ice cover until April. Instruments were programmed to sample at a minimum frequency
of once per hour. ACT collected reference samples twice per day for 4 days per week during the
entire deployment. Instruments were moored at approximately 4m depth and surface access
through the ice was maintained by gentle circulation with a propeller to allow deployment of the
Van Dorn sampling bottle. The goal of this test application was to demonstrate instrument
performance (reliability, accuracy, and stability) in winter-time environmental conditions and to
demonstrate the ability to operate continuous observations under ice.
A three month moored deployment was conducted at the Chesapeake Biological Lab Pier,
Solomons, MD. Instruments were deployed between May and August during a period of warming
temperatures and high biological production. Instruments were moored at fixed depth of 1m on a
floating dock. Instruments were programmed to sample at a minimum frequency of once per hour.
ACT collected reference samples twice per day for 3 days per week and collected six samples on