17 / 47
Ref. No. [UMCES] CBL 2015-009
ACT VS15-02
17
Two separate sondes were used at this deployment, one containing antifouling copper
mesh within the sensor guard and one with just a plain guard. The Manta2 with antifouling
operated for 23 consecutive days before experiencing a power failure, corresponding to 2128
observations and a data return rate of 83% for the total deployment. Time series results of the
Manta2 with antifouling and corresponding reference pH results are given in figure 2. Ambient
pH measured by the Manta2 ranged from 7.95 to 8.63, compared to the range measured from
reference samples of 7.933 to 8.077. The bottom panel presents the time series of the difference
between instrument and reference pH measurements for each matched pair (N=64 observations).
The average and standard deviation of the measurement difference (Manta2 – Reference) over
the total deployment was 0.258 ± 0.181 with a total range of -0.014 to 0.551. As noted in the text
on pH scales, a difference of approximately 0.13 pH units could be expected because of the
difference in calibrating with an NBS buffer versus our reference measurement based on the total
pH scale.
A cross-plot of the matched observations for the Manta2 with the Copper Screen is given
in figure 3. The linear regression of instrument versus reference pH values was statistically
significant but with a very low goodness of fit and high slope (p=0.001, r
2
= 0.16, slope = 2.89).
In general the instrument measured variation did not meaningfully correspond to the variation in
ambient pH as measured from the reference samples.
The Manta2 with no antifouling operated continuously for the entire 28 days of the
deployment, resulting in 2579 observations at 15 minute intervals. Time series results of the
Manta2 without copper mesh and corresponding reference pH results are given in figure 4.
Ambient pH measured by this Manta2 sonde ranged from 8.28 to 8.60, compared to the range
captured by the reference measurements of 7.933 to 8.077. The bottom panel presents the time
series of the difference between instrument and reference pH measurements for each matched
pair (N=84 observations). The average and standard deviation of the measurement difference
(Manta2 – Reference) over the total deployment was 0.512 ± 0.059 with a total range of 0.333 to
0.584. As noted in text on pH scales, a difference of approximately 0.13 pH units could be
expected because of the difference in calibrating with an NBS buffer versus our reference
measurement based on the total pH scale. The range in pH values measured by the Manta2 was
nearly three times greater than that measured by the reference samples.
A cross-plot of the matched observations is given in figure 5. The linear regression
between the instrument measurement and reference values was not statistically significant
(p=.024, r
2
= 0.06) and again the instrument measured variation in pH did not closely match the
variation in ambient pH measured from the reference samples.