Aerobic residential onsite sewage systems: an evaluation of treated-effluent quality.

Introduction

Aerobic treatment unit (ATU) systems are becoming more prevalent in the state of Washington because the better-drained lots have been developed, leaving only lots with more poorly drained soils available for development. In addition, amendments made in 1995 to the State Board of Health On-Site Sewage Regulations, Chapter 246-272 WAC, restrict the installation of standard gravity onsite sewage systems in the state to lots that have a minimum of 2 feet of soil available above the water table or an impervious layer (e.g., clay or hardpan). Much of the undeveloped area of Kitsap County has either a high water table or a clay or hardpan soil that prohibits installation of conventional gravity onsite sewage systems, leaving alternative onsite systems as the only choice. The most common reasons aerobic systems are used are that there is not enough land area to accommodate the soil absorption, that soil conditions are not appropriate, or that the water table is too high to allow adequate treatment of the wastewater in the drainfield before the water is released to the groundwater. Another site-related concern is the existence of lots close to surface water. One of the most common reasons that aerobic wastewater treatment units are chosen is to replace failing septic systems.

In Chapter 246-272 WAC, the Washington State of Board of Health On-site Sewage Regulations allow placement of ATUs within 100 feet or less of domestic drinking-water supplies (Washington State Board of Health, 1995) because ATUs can treat sewage effluent to a higher standard than a conventional gravity onsite sewage system. In Section 15501 of 246-272 WAC, the regulations require that the local health officer provide operation and maintenance information to homeowners of onsite sewage systems, and initiate periodic monitoring of each onsite sewage system. This mandate, however, is unfunded (i.e., though it is a state regulation, no funding mechanism is provided to carry out the requirement). Many local health jurisdictions only provide informational material and do not require homeowners to submit results from effluent testing.

Methods

The study sample consisted of 184 residences with onsite ATUs in Kitsap County, Washington, with at least one effluent sample reported to the Bremerton-Kitsap County Health District during 1995 through 1999. A total of 781 effluent sample results were reported from these onsite ATU sewage systems. The three brand-name ATUs of interest were: MultiFlo[TM], Norweco[TM], and Whitewater[TM].

The data for this study came from annual operation and maintenance reports that were submitted to the health district. These data included name and address of property owner, type of ATU, tax parcel number, installation date, and results of five-day biochemical oxygen demand (BO[D.sub.5]), total suspended solids (TSS), and fecal-coliform testing.

The underlying population comprises all the residences with an ATU in Kitsap County, including those for which no effluent test results were reported to the Bremerton-Kitsap County Health District. The extent of non-reporting is not known, and the size of the population could not otherwise be determined.

Analysis of Data

The analysis of data included descriptive evaluation of the effluent measurements and independent variables (i.e., type of ATU, as well as season, day of week, and time of day of sample collection). A comparison of effluent measurements for the ATUs followed, using group means, t-tests, and one-way analysis of variance (ANOVA). The authors used analysis of covariance to compare ATUs while controlling for possible confounding factors associated with the independent variables. Finally the sample results were compared with both state and national standards to assess compliance with these requirements.

Results

Almost half (45 percent) of the ATUs in the study sample were Whitewater units. MultiFlo units were the next most common (31 percent), followed by Norweco (22 percent). Most (77 percent) of the ATUs were no more than 12 months old at the time of sampling. The Multi-Flo and Whitewater ATUs were about twice as likely (42 percent and 31 percent, respectively) to be very new (i.e., newer than six months at first sampling) as were the Norweco ATUs (18 percent).

The installation dates were not documented for all sites; dates were readily available, however, for 102 (83.7 percent) of the 184 sites. The age of each system was calculated as the time between the first sampling event and the installation date, if an installation date was known.

Effluent Samples

Several state-certified laboratories were involved in analysis of the effluent samples. The samples were analyzed in accordance with methods described in Standard Methods for the Examination of Water and Wastewater Analysis (Lenore & Greenberg, 1998).

A total of 781 non-chlorinated effluent samples were collected, of which 38 percent were analyzed for both BO[D.sub.5] and TSS, and 24 percent were analyzed for fecal coliforms (FC). For all three effluent parameters, some results were high; however, county laboratory personnel indicated that these results were within the range of plausibility. Because the effluent data were skewed toward higher values, the data were transformed logarithmically to yield a more nearly symmetric or normal distribution when necessary for data analysis.

Relationship of BO[D.sub.5], TSS, and Log FC

Bivariate associations were examined with the Pearson correlation coefficient to see if there was any linear association among the effluent parameters. A high degree of correlation between log BO[D.sub.5] and log TSS was found (r = .8, p < .001). Each of these parameters showed a weaker but still statistically significant correlation with log FC (for each pair, r = .5, p < .001).

Differences in Effluent Quality, ATU Type, and Sample Timing

On average, effluent from Multi-Flo units was of better quality than that from units of the other two types (Table 1). One-way ANOVA found significant differences among the three ATU types for each of the three effluent parameters (for each parameter, p < .001). Some of the effluent parameters also showed significant differences depending on when the samples were collected. The values for log BO[D.sub.5] and log FC differed significantly by the season, with lug BO[D.sub.5] tending to be lowest in samples collected during the winter and log FC tending to be highest in fall samples. Values for log BO[D.sub.5] and log FC also tended to be lower in samples collected in the first half of the week. No pattern relative to season or day of the week was evident for log TSS values. In the small number of samples for which time of day was noted, however, log TSS values generally were lower for samples collected in the early afternoon than for samples collected in the morning.

Multivariate Analysis 

The authors used analysis of covariance to evaluate the differences among the ATUs for each of the three effluent parameters while controlling tot possible differences arising from when samples were collected. The analysis did not include the time of day, because this information was available for only a fraction of the samples. The B values in Table 2 represent the difference between the average value for a specified category (e.g., Norweco ATU) and the average value for the corresponding reference category (i.e., Whitewater ATU) on a logarithmic scale, with the other variables shown in the table being controlled for. The reference categories were chosen arbitrarily and were assigned a relative value of 0 (because a reference category does not differ from itself). Table 2 also shows 95 percent confidence intervals (95 percent CIs) in parentheses for each value; the difference between a B value and its reference value is statistically significant (i.e., p < .05) if the 95 percent CI does not include zero.

The effluent samples were collected between spring 1995 and fall 1999. Over a quarter (28 percent) of the samples were taken during the months of December and January, and samples were taken next most frequently in June and July (22 percent). The majority of the effluent samples were collected during four time periods: summer 1997 (13 percent), winter 1998 (15 percent), fall 1998 (10 percent), and winter 1999 (11 percent). Overall, more than one-third (37 percent) of the samples were collected during the winter season, and less than one-quarter (17 percent) were collected in the spring. The most common sampling day was Thursday (60 percent), followed by Tuesday (21 percent). Very few samples were collected on Saturday or Sunday (0.6 percent and 0.9 percent, respectively). Sample collection times were available for only 85 samples; of these, 81 percent were taken in the morning.

The multivariate analysis delineated in Table 2 found that the Multi-Flo ATU had significantly lower log BO[D.sub.5], log TSS, and log FC values than did the Whitewater ATU, while controlling for possible differences arising from sample collection times. The quality of effluent from the Norweco and Whitewater ATUs did not, however, significantly differ. Regardless of ATU type, the log BO[D.sub.5] values generally were lower in samples collected during summer, fall, and winter months than in samples collected in the spring. These differences, however, were statistically significant only for winter samples compared with spring samples. Conversely, the log FC counts generally were higher during the nonspring months, although the difference was statistically significant only for samples collected in summer and fall.

The seasonal variation in effluent quality was not uniform across the different types of ATUs (Table 3). Winter was chosen as the reference season for the model examining seasonal variation because samples collected in the winter tended to show the least variation by ATU and the three sewage parameters. When only the records for samples collected during winter were considered, no significant difference was found between Whitewater ATUs in the regression models for log BO[D.sub.5], log TSS, and log FC. The Multi-Flo ATU, however, did have log BO[D.sub.5] values that differ significantly from those of the other two types.

To further investigate how much difference there was among the ATUs, Multi-Flo ATUs were chosen as the reference for the models examining ATU variance across the different seasons because samples collected from the Multi-Flo ATUs tended to show the least variation by season for the three sewage parameters (Figure 1, Figure 2, Figure 3). When only the records for samples collected from Multi-Flo ATUs were considered, the regression models for spring, summer, and fall found no significant difference among the seasons. Samples from the Whitewater and Norweco ATUs did, however, significantly differ from Multi-Flo samples for the season winter.

[FIGURES 1-3 OMITTED]

Comparison with Standards

Overall, the performance of the ATUs in this study failed to consistently meet either state or national effluent standards (Table 4).

Discussion

This population-based study of residential onsite ATUs in one Pacific Northwest county found that over a third of the ATUs failed to meet NSF certification standards for BO[D.sub.5] and TSS in effluent (<30 mg/L). In addition, over two-thirds of the ATUs failed to meet the Washington State Board of Health Treatment Standard 2 criteria for BO[D.sub.5] and TSS (<10 mg/L), and a majority of the systems (59 percent) failed to meet state standards for fecal coliform (<800 fecal coliforms per 100 mL). The high prevalence of substandard performance is particularly remarkable when it is recognized that most (77 percent) of the systems were not even one year old at the time of sampling.

The results of this study generally agree with those of previous studies and demonstrate that many systems actually in use do not produce an effluent that meets local or state standards. This study found only 36 percent of the systems tested produced an effluent that met the BO[D.sub.5] Treatment Standard 2 of less than 10 mg/L. While many studies also found generally high BO[D.sub.5] effluent levels (Brewer, Lucas, & Prascak, 1978; Hutzler, Fancy, & Waldorf, 1978; Kellam, Hagedorn, & Reneau, 1993), some studies have found that the systems perform well (Island County Health Department, 1999; Mancl & Vollmer, 2001). Similarly, TSS levels were high in many cases, with only 39 percent meeting the Treatment Standard 2 of less than 10 mg/L. Again, this result is consistent with those of other studies, which also found generally high TSS levels in effluent (Brewer, Lucas, & Prascak, 1978; Hutzler, Fancy, & Waldorf, 1978; Kellam, Hagedorn, & Reneau, 1993; Mancl & Vollmer, 2001), but it is in contrast with the findings of one study that found only 8 percent BO[D.sub.5] and 15 percent of TSS sampling results not in compliance with Treatment Standard 2 requirements (Island County Health Department, 1999). The systems tested in this study may have performed worse because of influent waste strength or hydraulic overloading by homeowners. This deficiency in performance is particularly remarkable given that the ATUs were examined early in their life span.

This study also found that only 38 percent of the tested systems produced an effluent that met the Treatment Standard 2 requirement of less than 800 fecal coliforms per 100 mL. Three other studies found that 35 percent (Kellam, Hagedorn, & Reneau, 1993), 36 percent (Mancl & Vollmer, 2001), and 55 percent (Island County Health Department, 1999) of the fecal-coliform results met their respective state requirements.

Overall, effluent quality varied significantly by type of system. Multi-Flo units generally had better effluent quality for BO[D.sub.5], TSS, and fecal coliform than did units of the other two types, while the Whitewater and Norweco units generally were comparable. By contrast, the only other study that evaluated the same brands of ATUs found a statistical difference in BO[D.sub.5] levels between Whitewater and Multi-Flo units and no statistical significance difference in TSS between the two types of unit (Island County Health Department, 1999).

There were substantial seasonal variations in effluent quality, especially for BO[D.sub.5]. Generally, higher levels of BO[D.sub.5] were found in spring, and lower fecal-coliform counts were found in winter. Effluent quality also varied by day of week on which samples were analyzed; analysis that occurred later in the week was significant for BO[D.sub.5] and fecal coliforms, with the majority of the samples analyzed either on a Tuesday or Thursday. Performance differed somewhat by season, with winter having the best effluent quality and fall having the worst. The only other study that could be found that evaluated the effect of seasonal variation found no overall trends (Kellam, Hagedorn, & Reneau, 1993).

Grab samples have more limited accuracy, characterizing ATU performance at one point in time, than do samples collected over a longer period of time (e.g., 24 hours). The accuracy of grab samples, however, is sufficiently acceptable for the method to be used routinely by regulatory agencies to monitor and evaluate ATU performance in the field. This study used the data available in the existing database of a regulatory agency, allowing the authors to evaluate the performance of a large number of ATUs of several types in a large geographic region. Future studies should consider using longer sample collection periods.

Although ATUs are certified through NSF according to a rigorous testing protocol, the results obtained during such testing do not appear to be reflective of in-the-field results found by this and other studies. Currently no active programs require periodic monitoring of effluent quality from ATUs in the state. The findings of this study indicate that performance monitoring should be ongoing to evaluate not only the initial performance, typically called the design life (Jantrania, 1998), but also the long-term performance of the units. One would have expected to see better results from newer ATUs than from older units, but that was not the case here. If one accepts that the state treatment standard is reasonable, then the findings of this study bring into question both the reliability of ATUs for routine residential use and the reduction of the horizontal separation requirement to within 100 feet of drinking-water supplies. Reduction of regulatory horizontal separation criteria based on treatment of effluent is not advisable until effluent standards are enforceable.

The use of ATUs will continue as the availability of better-drained lots diminishes, leaving poorly drained soils for development. This study could not evaluate many factors that influence effluent quality; however, it is significant that effluent from ATUs analyzed in this research did not meet state or national requirements more than 60 percent of the time and that most of the systems were no more than 12 months old at the time of sampling.

As ATUs become more prominent in the state of Washington, efforts are needed to improve their performance. The authors recommend that the following measures be considered:

* more in-the-field evaluations of influent waste strength and how this factor affects the performance of the ATUs,

* improvement of homeowners' knowledge about their systems,

* operation and maintenance agreements that would require periodic effluent sampling and testing (as well as allow maintenance and operating personnel to troubleshoot performance problems), and

* more frequent (i.e., more frequent than semiannual) operation and maintenance inspections to minimize malfunctions and assist in homeowner education.

TABLE 1

ATU Type and Effluent Characterization

Non-chlorinated         Log BO[D.sub.5]          Log TSS
Samples

ATU type                    n = 275              n = 267
  Multi-Flo            0.8 [+ or -] 0.7 *    0.9 [+ or -] 0.7
  Whitewater           1.4 [+ or -] .5       1.3 [+ or -] 0.6
  Norweco              1.5 [+ or -] .4       1.3 [+ or -] 0.6
  Significance **           p < .001             p < .001

Season                      n = 281              n = 273
  Winter               1.0 [+ or -] 0.7 *    1.2 [+ or -] 0.7
  Spring               1.5 [+ or -] 0.6      1.2 [+ or -] 0.7
  Summer               1.3 [+ or -] 0.6      1.2 [+ or -] 0.7
  Fall                 1.3 [+ or -] 0.6      1.4 [+ or -] 0.6
  Significance **           p <.001               p <.26

Day of week                 n = 273              n = 265
  MTuW                 1.4 [+ or -] 0.7 *    1.3 [+ or -] 0.8
  ThF                  1.2 [+ or -] 0.6      1.2 [+ or -] 0.6
  Significance **           p < .02               p <.09

Time of day                  n = 85               n = 83
  Morning              1.1 [+ or -] 0.6 *    1.2 [+ or -] 0.7
  Afternoon            0.8 [+ or -] 0.8      0.8 [+ or -] 0.8
  Significance **           p < .12              p < .03

Non-chlorinated              Log FC
Samples

ATU type                    n = 169
  Multi-Flo            2.1 [+ or -] 1.1
  Whitewater           3.2 [+ or -] 1.3
  Norweco              2.8 [+ or -] 0.7
  Significance **           p < .001

Season                      n = 173
  Winter               2.6 [+ or -] 1.2
  Spring               2.7 [+ or -] 1.3
  Summer               2.9 [+ or -] 1.3
  Fall                 3.4 [+ or -] 1.1
  Significance **            p <.04

Day of week                 n = 168
  MTuW                 3.2 [+ or -] 1.6
  ThF                  2.7 [+ or -] 0.9
  Significance **            p <.01

Time of day                  n = 39
  Morning              2.5 [+ or -] 1.0
  Afternoon            2.1 [+ or -] 0.7
  Significance **           p < .20

* Mean and ([+ or -]) standard deviation.

** One-way ANOVA.

TABLE 2

Differences in ATU Effluent Quality, Adjusted for When Samples
Were Collected

Main Effects          Log BO[D.sub.5]             Log TSS

                   B          95% CI           B          95% CI
ATU
  Norweco         .19      (-0.00, 0.38)     -.03      (-0.19, 0.25)
  Multi-Flo      -.47     (-0.65, -0.30)     -.31     (-0.51, -0.18)
  Whitewater        0                           0

Season
  Winter         -.23     (-0.44, -0.002)    -.07      (-0.12, 0.31)
  Summer         -.14      (-.36, 0.08)      -.01      (-0.26, 0.24)
  Fall           -.13      (-0.35, 0.01)      .16      (-0.09, 0.42)
  Spring            0                           0

Day of Week
  M-W             .11      (-0.06, 0.27)      .01      (-0.09, 0.28)
  Th-F              0                           0

Main Effects          Log FC

                   B          95% CI
ATU
  Norweco        -.29      (-0.81, 0.23)
  Multi-Flo      -1.0      (-1.5, -0.60)
  Whitewater        0

Season
  Winter          .36      (-0.24, 0.95)
  Summer          .75       (0.13, 1.4)
  Fall            .74       (0.01, 1.4)
  Spring

Day of Week
  M-W             .44      (-0.06, 0.82)
  Th-F              0

0 = reference category.

95% CI = 95% confidence interval.

TABLE 3

Seasonal Variation in ATU Performance

Interaction        Log BO[D.sub.5]            Log TSS

                B          95% CI          B          95% CI
Multi-Flo
  Spring       .42      (0.21, 0.64)      .06      (-0.18,0.29)
  Summer       .24      (0.05, 0.43)      .02     (-0.18, 0.23)
  Fall         .32      (0.11, 0.52)      .21     (-0.01, 0.43)
  Winter         0                          0

Norweco
  Spring       .47      (0.06, 0.88)     -.71     (-1.2, -0.20)
  Summer      -.07     (-0.37, 0.23)     -.40     (-0.78, -0.03)
  Fall         .23     (-0.08, 0.54)      .02     (-0.37, 0.40)
  Winter         0                          0

Whitewater
  Spring       .12     (-0.15, 0.39)      .03     (-0.29, 0.35)
  Summer      -.02     (-0.27, 0.24)     -.09     (-0.40, 0.22)
  Fall        -.05     (-0.30, 0.21)     -.03     (-0.34, 0.27)
  Winter         0                          0

Interaction        Log FC

                B          95% CI
Multi-Flo
  Spring       .10     (-0.50, 0.69)
  Summer       .28     (-0.17, 0.73)
  Fall         .74      (0.24, 1.3)
  Winter         0

Norweco
  Spring      -1.8     (-2.1, -0.88)
  Summer      -.49      (-1.1, 0.10)
  Fall        -.19     (-0.81, 0.43)
  Winter         0

Whitewater
  Spring       .01     (-0.77, 0.78)
  Summer       .14     (-0.47, 0.95)
  Fall         .39     (-0.28, 1.1)
  Winter         0

0 = reference category.

95% CI = 95% confidence interval.

TABLE 4

Performance of ATUs Compared with State and National Standards

                           Percentage of Samples Exceeding Limits

Standard                                BO[D.sub.5]

NSF Standard                 [greater than or equal to]30 mg/L
  Multi-Flo                                 12%
  Whitewater                                36%
  Norweco                                   34%

Treatment Standard 2         [greater than or equal to]10 mg/L
  Multi-Flo                                 28%
  Whitewater                                74%
  Norweco                                   89%

                           Percentage of Samples Exceeding Limits

                                            TSS

NSF Standard                 [greater than or equal to]30 mg/L
  Multi-Flo                                 12%
  Whitewater                                34%
  Norweco                                   28%

Treatment Standard 2         [greater than or equal to]10 mg/L
  Multi-Flo                                 34%
  Whitewater                                71%
  Norweco                                   75%

                           Percentage of Samples Exceeding Limits

                                      Fecal Coliforms
                                         per 100 mL

NSF Standard                            No standard
  Multi-Flo                                 N/A
  Whitewater                                N/A
  Norweco                                   N/A

Treatment Standard 2        [greater than or equal to]800 Fecal
  Multi-Flo                                 35%
  Whitewater                                74%
  Norweco                                   69%

N/A = not applicable.

REFERENCES

Brewer, W.S., Lucas, J., & Prascak, G. (1978). An evaluation of the performance of household aerobic sewage treatment units. Journal of Environmental Health, 41 (2), 82-85.

Hutzler, N.J., Fancy, J., & Waldorf, L.E. (1978). Performance of aerobic treatment units. In Proceedings of the 2nd National Home Sewage Treatment Symposium (pp. 149-163). St. Joseph, MI: American Society of Agricultural Engineers.

Island County Health Department (1999). Wastewater sampling assessment, Island County Special On-Site Demonstration Grant Program. Coupeville, WA: Author.

Jantrania, A. (Virginia Department of Health, Division of On-site Sewage and Water Services.) (1998). Monitoring protocol for onsite systems, http://plymouth.ces.state.nc.us/septic/98jantra3 .html (14 Sept. 2000).

Kellam, J., Boardman, G., Hagedorn, C., & Reneau, R. (1993). Evaluation of the performance of five aerated package treatment systems. (Bulletin 178). Blacksburg, VA: Virginia Water Resources Research Center, Virginia Polytechnic Institute and State University.

Lenore, C., Eaton, A., & Greenberg, A. (1998). Standard methods for the examination of water and wastewater (20th ed.). Washington, DC: American Public Health Association.

Mancl, K. & Vollmer, M. (2001). Management of individual mechanical sewage-treatment systems: How much is needed? Journal of Environmental Health, 63(9), 22-25.

Washington State Board of Health (1995). Chapter 246-272 WAC: On-site sewage systems. In Rules and Regulations of the State Board of Health (pp. 11-13). Olympia, WA: Author.

Corresponding Author: Meliss Maxfield, Water Quality Manager, Division of Drinking Water, Washington State Department of Health, 7171 Cleanwater Lane, Building 3, Olympia, WA 98504. E-mail: [e-mail]