Collecting accurate water data can be complicated: The importance of data integrity and collecting legally defensible data
Published in the Florida Specifier, October 2022
In the State of Florida, the methods used to assess water resources are covered under various Florida Department of Environmental Protection (FDEP) Standard Operating Procedures (SOPs). If you are performing compliance activities under a potable water, stormwater, or wastewater permit, or trying to determine if your source of water meets its designated use, you are going to need to follow those SOPs. But why?
In an earlier life, my FDEP district sometimes hired folks with no scientific experience to perform sampling duties. This may sound absurd, and it is, but it can be a common practice in the private sector as well. It was my team’s duty to perform water quality sampling, biological assessments, analysis, and reporting, as well as train staff on these methods regardless of the candidates’ experience.
One cohort was particularly stubborn when it came to operating within the established guidelines. Months into the process, we were still qualifying the samples these folks collected because those samples did not meet the minimum quality standards outlined in the SOPs despite them receiving the necessary training. According to my new teammates, their refusal to follow the SOPs had to do with their belief that environmental regulation was a waste of tax dollars and time.
At some point the training wheels had to come off, so I decided to couch their actions in terms of tax dollars and time. When you collect a sample, you need to consider several things before making an attempt. For example: Are the samples representative of the system or conditions you are trying to capture? Are your activities contributing noise to your samples? Are the procedures that you are using producing legally defensible data? The goal behind any sampling SOP is to standardize a set of procedures that facilitate the collecting of data that speaks for itself.
When the data suggests a water quality problem, it triggers a whole different slew of responses than if there was no problem. In the case of the State of Florida, it would have meant a multi-year investigation of whether there was an impairment, what the sources of pollution were, and studies on how to restore the system.
Thankfully, the state has a robust quality assurance/quality control program in place to flag bad data. This explanation was all it took to get my new staffers on board with following our procedures and studying the respective SOP before undertaking a new sampling mission.
I realize that some of my audience may share views that are similar to the subjects in this tale. Perhaps your only stake in resource management and regulation is as a regulated entity that wants less regulation. Perhaps you do not see your resources ever drying up no matter how much it is exploited.
Perhaps you are indifferent to any social cost to the people living downwind or downstream of your operation. So why should you care about collecting legally defensible data? You need to collect legally defensible data because bad data and decisions made using bad data is very expensive.
Bad data mixed in with good data can potentially poison all of your data and paint your organization as not acting in good faith. It would also ruin chances for your activities to remain permitted or permitted at levels that are desirable to your organization.
Before it was discovered that metals, including lead, were leaching out of pipes into drinking water, Flint, Michigan, had the third largest Legionnaires’ disease outbreak in US history. The bacteria Legionella pneumophila thrives in potable water distribution systems with poor disinfection and a supply of micronutrients like iron. L. pneumophila is just one potential bacterium that can spread in a distribution system. To combat bacterial loads, potable water supplies are disinfected with various forms of chlorinated compounds.
Water distribution systems in the US are required to maintain a residual chlorine concentration of no less than 0.2 mg Cl/L of “free chlorine”. A free chlorine test measures any chlorine disinfection agent in water not bound to anything and therefore ‘free’ to disinfect water. “Total chlorine” tests measure all chlorine in water, i.e., free chlorine and bound chlorine (i.e. chlorine already attached to something and can no longer act as a disinfectant). The most common way of testing for either is using an instrument called a colorimeter. That instrument needs to be calibrated and verified before use and then verified after use.
The difference in procedure between the total and free chlorine is the reagent packet used for each test. One packet is labeled “Total Chlorine DPD” and the other is labelled “Free Chlorine DPD”- and both packets look very similar. The lot number and expiration and type of DPD needs to be documented and associated with each value of residual chlorine measured.
Chlorine is an oxidizing agent and will bind to metals in solution as well as bacteria and dissolved organic material. If you measure the total chlorine and free chlorine of a sample and you find a huge difference between the two, you know that your disinfectant is being consumed by either dissolved organics, dissolved metals, bacteria, or a combination of all three. If you pump large amounts of disinfectant into your system and your residual free chlorine will not stay above a concentration of 0.2 mg Cl/L you have identified a problem and you need to identify the source.
To complicate matters just a little bit more, some disinfecting agents have shorter half-lives than others, so water that has been sitting in pipes of a residence that has not been occupied will have little or no disinfection unless those pipes are flushed. This is why the SOPs for collecting residual chlorine and a potable water bacteria sample stipulate that the pipe must be flushed for two to five minutes before collecting a sample.
So, what happens when a water supply entity reports total chlorine results as free chlorine because for whatever reason samplers were unaware that there are two different types of DPD and they were both in their sample kits? And what happens when a water supply entity routinely flushes water lines for as long as it takes (sometimes multiple hours) to get a >0.2 mg Cl/L chlorine residual? In both cases, the testing produces false positives painting a picture of adequate disinfection when there was little or no disinfection.
I do not know that these conditions contributed to what happened in Flint, Michigan, but I have observed regulated entities mixing up reagent packages, cooking the books when it comes to line flushing, and presenting unverified data as verified, among other things. Florida is a bit different from northern states due to different water sources, water treatment, and typically a younger and less metallic distribution system, but the potential for widespread waterborne public health problems will not be detected if the system in place to produce dependable data is ignored. The consequences could have a cascading and devastating impact on communities, economies, home values, businesses, regulated entities, public health, and future generations of Americans.
“It has been my personal experience that more often than not, samplers are poorly trained and often do not know what the purpose of their activities are, beyond satisfying the government. That attitude is often echoed by leadership, and I have even heard folks who should know better say, “a monkey can collect a sample.” I have also observed samplers having eureka moments when they receive training, and it is often positively reflected in their work product. Our system is based on some level of trust that everyone is acting in good faith but that is not enough. Your sampling program needs to follow the DEP SOPs, your staff needs to consist of trained individuals that receive regular training, and there needs to be a system of accountability and some level of independence where management oversight does not interfere with the collection of samples or degrade the quality of the data produced. Your data trail or lack thereof in any data collection program speaks for itself. “