17 / 59
Ref. No. [UMCES] CBL 2015-008
ACT VS15-01
17
comparison, but because they were calibrated directly in the test tank water against the dye
measured pH they maintained direct agreement throughout the test (Fig. 2). Linear regression of
the data has a slope of 1.006 with an r
2
of 0.999. Metrohm results are included within some of
the time series plots below to provide a more continuous record of test conditions, but no
quantification of instrument accuracy was conducted with these data.
Figure 2.
Cross plot of matched Metrohm pH results that measured continuously within the test tank at
15 minute intervals versus dye measured reference pH results. Three-hundred, fifty-one dye based pH
measurements were made over the ten week test.
Once a week multiple reference samples were collected from the test tank in a burst, i.e.
sampled as rapidly as possible (including analysis time) to examine heterogeneity within the test
tank and repeatability of the reference pH results. The average standard deviation of the
reference pH value for ten burst sampling events was 0.0025 with a range of 0.0055 to 0.0005
(Table 1).