73 / 73
of this deviation is caused by the sensor drift or how much the issue reported with Winkler
sampling (sampling corrected against a reference) influences the error magnitude.
-
Profiling deployment in the Great Lakes
: In Fig. 15, the difference between Winkler titration
and ARO-USB is inside the accuracy range of the instrument, except two points where ΔDO
reached values between -0.6 and -0.8 mg/L. The reason for this larger deviation between the
methods may be related to errors associated with sample processing or the Winkler titration
rather than the ARO-USB, since the variation on temperature observed around 12:45 –
12:50am is not enough to explain the variation presented by the Winkler titration on the DO
concentration in the same period of time (15 minutes). In such a short period of time, we
cannot explain the variation presented by Winkler titration and we are considering that the
data provided by the ARO-USB are more likely to be close to the real DO concentration. In Fig,
16, although there is a larger deviation between the methods, all the values are inside the
accuracy stated by us on the instrument specifications (2% of full scale).