Quality Control
The Split Sample
The Lakes of Missouri Volunteer Program helps our state fulfill its Clean Water Act requirements. The Missouri Department of Natural Resources (MDNR) relies on our data (around 800 samples each year) to help determine which lakes are doing well and which ones need help. To reassure data users that our volunteer data are of high quality, we have to perform some additional quality control measures. One such measure is to collect side-by-side samples with our volunteers. We call these "Split Samples". This year we collected Split Samples with volunteers at Table Rock, Clearwater, and Lake of the Ozarks. Here’s how it works.
This quality control measure is intended to verify our volunteers' ability to process water samples. Several times each season, an employee from our MU lab will accompany a volunteer as they sample. The employee will collect a sample at the same time as the volunteer and bring it back to our lab. This “split sample” is then processed by our staff and analyzed as if it were collected by an employee. Once all of the analyses are complete, we simply compare the results to one another.
By examining the differences between what should be two identical samples we can spot systemic differences between the ways our employees and our volunteers process and store samples. What we’ve found is that there is virtually no difference! When we do see differences, they are usually due to the inherent variability of what we’re measuring.
Comparing the results of samples collected and processed by volunteers with those collected and processed by University of Missouri employees
The graphs above show split sample results for four parameters (clockwise from top left): total phosphorus, suspended sediment, total nitrogen, and chlorophyll. The results from the volunteer samples are shown on the left axis while the results from MU employee sampling are shown on the bottom axis. The dashed line shows the perfect one to one ratio. If the split sample and the volunteer sample results were exactly the same, the dot will appear on the dashed line. If the volunteer value was greater, the dot will appear above the line. If the MU employee value is greater, the dot will appear below the line.
Notice that the data points are quite close to the dashed line, particularly for total phosphorus and chlorophyll. Also notice that the dots do not consistently fall to one side of the dashed line or the other. This means the methods we use do not result in consistently under or overestimating lake conditions. We examined duplicate samples from one of our in-house monitoring efforts and the results closely match what we see in the volunteer data.
This pie chart shows the number of years our volunteers have been collecting and processing water samples. Nearly half of LMVP volunteers have been monitoring for 10 to 20 years!
Our Volunteers Know What They Are Doing!
That the volunteer results closely match those of our in-house testing isn’t surprising since our typical volunteer has been monitoring their lake for 8.7 years and has 220 hours of experience collecting and processing water samples. Lynn Fair and Frank Fillo have been monitoring at Sugar Creek Lake in Moberly for 20 years, meaning that they each have over 500 hours of water quality monitoring experience.
We are proud of our volunteers and the quality of the data they help us generate. The people of the state of Missouri are lucky to have such diligent sentinels on their lakes!