Comparison of YSI 6136 Turbidity Probe Data to Hach 2100AN Benchtop Turbidimeter Data

1
Comparison of YSI 6136 Turbidity Probe Data
toHach 2100AN Benchtop Turbidimeter Data

• Bill Romano and Christopher HeyerMaryland
  Department of Natural Resources

2
(No Transcript)
3
Purpose of Comparison

• To determine if the YSI 6136 turbidity probe
  accurately measures in situ turbidity.

4
Assumptions

• No significant difference between in situ and
  whole water methods.
• Data from Hach 2100AN are true sample values.
• Error associated with Hach 2100AN and whole water
  sample method is negligible.

5
YSI Data
6
DNR Data
7
Relationship Between Methods

• Based on this relationship and its
  non-significant intercept term, we would expect
  the YSI 6136 probe to accurately measure the true
  turbidity of the sample.

8
Percent Difference

• About 75 of the observations are within one
  standard deviation of the mean, indicating a
  fairly normal distribution.

9
Relative Percent Difference

• The randomness of this relationship provides
  evidence that the differences in observations
  between the two instruments are random and are
  unrelated to any one instrument.

10
Inherent Variability With Methods

• Light scatter is being measured in different
  environments.
• In the lab, light scatter is effectively measured
  in a box, while in the field there is any number
  of environmental effects that could impact any
  given turbidity measure.
• Small three-dimensional spatial differences (x, y
  and z coordinates) can exist between where the
  YSI probe in situ measurement is observed and
  where the whole water sample is collected.
• Spatial difference may be insignificant at high
  turbidities, but can be highly significant when
  turbidity values are low and turbidity is patchy.
• For example, at a true turbidity value of 10 NTU,
  a difference of 3 NTU in the turbidity readings
  between the in situ measurement and the whole
  water sample measurement would result in a 30
  difference. However, that same difference at a
  true turbidity value of 60 NTU would only be a 5
  difference.

11
Temporal Variability

• The relatively large variability we observed over
  small time scales (e.g., 15 minutes) could easily
  be the result of patchy turbidity.

12
Conclusions

• Strong relationship between the YSI 6136
  turbidity probe and the Hach 2100AN benchtop
  turbidimeter.
• Normal distribution of the percent differences
  between the two instruments turbidity
  measurements.
• Apparent random error associated with the Hach
  2100AN instrument and in the whole water sample
  collection and processing.
• Patchy distribution of turbidity was evident and
  caused dramatic differences in values at low
  turbidities.

13
Recommendation

• We believe that the YSI 6136 turbidity probe is
  accurately measuring in situ turbidity and
  recommend that whole water turbidity sampling be
  discontinued.