Letters to the editor.

Dear Editor:

Recently, an article decrying the performance of aerobic treatment units (ATUs) appeared in the Journal of Environmental Health. "Aerobic Residential Onsite Sewage Systems: An Evaluation of Treated-Effluent Quality" (October 2003), which summarized and reported performance data in Kitsap County, Washington, concluded that many ATUs failed to meet specified performance standards. To use this article to judge performance of ATUs is inappropriate, and we do not believe the conclusions are supportable.

[ILLUSTRATION OMITTED]

There are two overlapping culprits, and each plays a significant role: passage of time and misinterpretation of grab sample data.

Consider the issue of time: The data reported in the Kitsap County article vary in age from four to eight years old. There is no explanation of who took samples, how samples were taken, the circumstances that surrounded the sampling event, and so forth. It is impossible to establish whether proper sampling techniques were utilized or if the circumstances brought about the sampling event.

Identifying the circumstances surrounding the sampling event is important because sampling may have come secondary to servicing or other activities that alone could have affected the results.

Interpreting the data is also critical. Grab samples are no more than indicators of probable performance at a particular moment. They are indicators because actual values can fluctuate with diurnal flow. To base declarative judgements on a single value--or a collection of single values--would be inadvisable.

The authors wrote that they expected to see better results from newer ATUs than from older ATUs. All microbial-based wastewater treatment systems require a start-up period in order to achieve maximum performance. If during the start-up period the system is not receiving design hydraulic or design organic loading (as is the case in most residential applications), it is not uncommon for the start-up process to last for several months.

The authors acknowledged that they calculated the age of the system as the time between the first sampling event and the installation date. Many new installations sit for several months until the owners move in. For this reason, the report should have examined closely the length of occupancy to make sure operation was not in start-up. A review of U.S. EPA's Environmental Technology Verification Program clearly outlines the requirements for an appropriate start-up period and states that the inclusion of start-up data in the statistical analysis of performance is optional.

Attention should be given to the application of the statistical analysis used in the article. It appears that all data were treated equally. The valid approach would have been to examine each household first. For example, it appears that over a two-year period, four samples were collected from each household. These four samples represent a unique data set because they are based on the specific treatment technology, wastewater characteristics, and flow from each household. Because grab samples were used for this evaluation, the authors would have been better served by evaluating unit performance by household, on the basis of four samples. Such an approach would have led to a discussion of whether four samples are sufficient to represent the performance of the treatment technology. Individual household data could then be added to the study data set once samples for individual households passed statistical scrutiny, on the basis of pass-fail criteria. The report would then address technology performance by examining the number of passing and failing households on the basis of statistical analyses, which would result in better accuracy and a more valid conclusion.

There are also practical issues related to analytical testing. The 20th edition of Standard Methods for the Examination of Water and Wastewater notes that a standard deviation of 30 milligrams per liter (mg/L) can result among laboratories performing a BOD test on the same 300 mg/L sample.

NSF and U.S. EPA settled the entire debate about grab samples and composite samples about 25 years ago. Dr. Nina McClelland of NSF also addressed the issue of field data from ATUs in a 1977 presentation. She pointed out that grab sample data is often dissimilar to data collected during NSF testing. She observed that performance fluctuations are common in ATUs, so grab sample data may yield results different from an analysis of a comprehensive composite sampling program.

Also, in its June 1974 publication, Wastewater Sampling Methodologies and Flow Measurement Techniques (PB 259 875, EPA 907/9-74-005), U.S. EPA makes clear two points: grab samples are inadequate and under-report treatment efficiency, and the use of low-paid, unsupervised personnel to collect samples is inappropriate.

The authors of this U.S. EPA publication compared grab samples and composite samples to reach their conclusions. They detailed their findings on page 74 of the report:

The data clearly indicate the inadequacy of relying upon a limited number of grab samples for determining wastewater characteristics or plant performance. In every case, the removal efficiencies calculated from the grab sample data were less than those efficiencies determined from the composite sample data reported by the laboratories.

The data has shown many sources of data variability and emphasizes the importance of having a professional in the field to select sampling locations, equipment, and methodology. It is obvious that those individuals responsible for surveys and sampling techniques and equipment can use any of the generally accepted sampling techniques and equipment and still intentionally or unintentionally manipulate apparent wastewater chemistry characteristics and facility removal efficiencies.

It is little wonder that there are so many disagreements regarding water chemistry characteristics and facility performance. Without an adequate monitoring program and tight controls on sampling techniques, equipment, and laboratory procedures, data interpretation can be reduced to little more than an exercise in futility.

Having noted our objections, we still believe that grab samples--as a part of a supervised, standard methodology--can be useful indicators of ATU performance, but they must be viewed in perspective. It is interesting to note that Washington standards for treated effluent from alternative onsite systems are based on 30-day averages. Even the worst grab sample generated during the study was three to five times better than average septic effluent. Although the systems did not meet their requirements, they did perform admirably in comparison with what is considered the normal onsite treatment choice.

When regulators wish to conduct serious inquiries into ATU performance, I urge them to work with industry associations, manufacturers, and academics to develop and implement study protocols all can agree upon. Developing an industry consensus is critical to being able to trust the performance claims that have been verified through the certification process.

The consensus among those who have a great deal of field experience with ATU technology (designers, installers, regulators, and manufacturers) indicates that ATUs consistently meet secondary-effluent standards when they are properly operated, maintained, and managed. Aerobic treatment units can provide a higher level of treatment and effluent quality than that which is typically achievable with conventional systems.

The industry and the member companies of the American Decentralized Wastewater Association both recognize the importance of and support independent and ongoing evaluations of performance. We want to continue to work in partnership with the academic and regulatory communities in our ongoing commitment to promote advanced treatment systems and protect public health and the environment.

Sincerely,

Don Canada, Executive Director

American Decentralized Wastewater Association

Co-author Meliss Maxfield Responds:

The samples used for this study were collected by private industry personnel certified by the county through its Operation and Maintenance Program. A local ordinance set forth the criteria for certification of individuals for onsite Operation and Maintenance, the minimum maintenance requirements, and sampling protocol for aerobic treatment systems. Certification required passing a written exam pertaining to design regulations and standards. Minimum requirements for Operation and Maintenance inspections include inspection of the septic tank, pump tanks, and trash traps (including effluent screens and pumping and cleaning if necessary, pump and pump screen inspection, cleaning if necessary, and sample collection for five-day biological oxygen demand, total suspended solids, and fecal coliform).

The initial sample set was collected three to six months after start-up, with a second sample set collected one year later unless otherwise specified by the manufacturer or the health officer. The ordinance also required a signed Operation and Maintenance contract for new-home construction utilizing aerobic treatment units within 30 days of occupancy, and sample collection followed manufacturer specifications.

Furthermore, at the time, a local ordinance required sampling for new-home construction or whenever an aerobic treatment unit was installed. As previously mentioned, initial sampling occurred three to six months after the start-up period and was not in response to failure of an aerobic treatment unit. Also, the start-up phase typically lasts only a few months; with the bulk of the effluent samples (47 percent) collected between six and 12 months after installation, the effluent quality was determined outside of the typical start-up phase and organic loading would have been within design.

Although Washington's standards for treated effluent from aerobic treatment units are based on 30-day averages, the practicality doesn't lend itself well as an effective tool for evaluation in the field. Grab sampling is what is used routinely to monitor the performance of aerobic treatment units in the state. The authors agree with Mr. Canada that 24-hour sampling instead of grab samples would provide a more definitive value for performance; however, the data for this study were pre-collected and not within the control of the authors. This limitation of the study was noted in the article.

The proposed method of analysis for the study was appropriate and statistically valid. Mr. Canada suggested that the study should have reviewed treated effluent quality on a household basis; however, the sample size was insufficient on a per-household basis for any meaningful analysis to occur. Three percent of the sites had three sampling events, followed by 10 percent with two sampling events and 23 percent with only one sampling event. The authors instead chose to analyze the differences between ATUs and the influences of season, day of week, and time of day at which the samples were collected for analysis. Furthermore, the analysis of data took into account the differences between aerobic treatment units in each of the effluent parameters (five-day biological oxygen demand, total suspended solids, and fecal coliform) while controlling for possible confounding factors associated with the independent variable.

Lastly, this was not a regulatory-sponsored study. This study was part of a master's thesis, and the viewpoints are those of the authors. In addition, the authors agree--and literature supports--that ATUs can provide better effluent quality than conventional onsite systems when they are properly operated and maintained.

Sincerely,

Meliss Maxfield