RESIDENTIAL GRAYWATER REUSE STUDY

WATER QUALITY- ANALYSIS

The summary of the study findings is reflected in both our brochure Can I Use my Graywater? and the new rules from ADEQ titled Reclaimed Water General Permit: Graywater Irrigation effective January 2001.

* On a continuum from highest to lowest risk by source of graywater:

• Kitchen sink
• Washing machines
• Tub/showers
• Bath sinks

* Risk factors include but are not limited to:

• washing diapers
• household pets
• feral animals and birds
• wet dry irrigation cycles
• organic matter in the irrigated soil

Description of Project Methodology & Selection Criteria

The criteria for site selection included:

• Storage of graywater
• Septic system
• Graywater filtration
• Graywater disinfectant
• Water sources:

1. Clothes washer
2. Kitchen sink
3. Bath sink
4. Bath tub/shower
5. Other

Vectors:

• Children
• Pets
• Washing diapers

Graywater application:

• Surface (spray, drip, flood, furrow)
• Subsurface
• Food crops
• Fruit trees
• Turf

Site selection was made after development of the site and criteria matrix. This enabled the group to select the most representative set of sites. The grant request called for a total of ten sites to be selected. Ultimately eleven sites were chosen, based upon a consensus that chances were high that one site would be lost during the course of the study.

Data collection was done in accordance with ADEQ's Quality Assurance Project Plan (QAPP) and the Field Manual for Water Quality Sampling, published by ADEQ and the UA Water Resources Research Center. The only variance from the guidelines, which has been agreed to by ADEQ, PCDEQ and UA, was the measurement of electrical conductivity and pH in the UA laboratory rather than in the field.

All eleven study sites were visited and detailed information about each household and their systems were gathered and compiled.

Graywater is defined as all wastewater generated in the household, excluding toilet wastes (Gerba et al., 1995). It can come from the sinks, showers, tubs, and washing machine of a home. It has been reused for purposes such as landscape irrigation and toilet flushing. But little data is available about the chemical and microbial quality of this water. Studies of graywater from a single family home have shown the presence of total and fecal coliforms and heterotrophic plate count (HPC) bacteria (Karpiscak et al., 1987; Gerba et al., 1995). If graywater reuse is truly to be a viable option for residential water conservation, then concerns about its safety need to be addressed, especially those related to the potential for transmission of disease. Graywater may be used by homeowners to irrigate both ornamental and food plants, and there is epidemiological evidence that the use of wastewater, particularly for the irrigation of food crops, has resulted in disease transmission when undisinfected effluent was used (Crook, 1985).

In order to assess the risks involved in the reuse of graywater, more information about the quality of this water and the factors that influence it is needed. Therefore, a yearlong study of eleven Tucson, Arizona households that recycle graywater was undertaken. Graywater, graywater-irrigated soil, and potable water irrigated soil were analyzed for fecal coliforms, fecal streptococci, and Escherichia coli. Fecal coliform bacteria were included because they are indicator organisms commonly used in water quality monitoring (APHA, 1999) and reuse standards (Maier et. al, 2000). Escherichia coli was selected because it is thought to be the most specific indicator of true fecal contamination (Gleeson and Gray, 1997). Fecal streptococci and coliphages have been selected because they have been suggested as indicators for some of the more environmentally resistant pathogens, such as enteric viruses (Maier et. al, 2000). Participating homes were chosen for the presence and absence of children and animals, methods of graywater storage, and household sources of captured graywater. Together, these site characteristics and the bacterial data gathered were to provide information about which factors in and around a home influence the quality of recycled household graywater before and after application to soil.

Materials and Methods

Water samples were collected from available graywater in 1L sterile plastic bottles. Soil samples were collected from yard sites irrigated by graywater or potable water. Samples were placed in sterile plastic tubes or bags using an ethanol-disinfected spatula. Samples were transported on ice to the laboratory, where they were held at 4°C until processing. Samples were processed within 8 hours of receipt in the laboratory.

Soil moisture was measured by drying 10g portions of soil at 100°C for 24 hours. Moisture content was calculated as described in the Environmental Microbiology Laboratory Manual (Pepper, Gerba, and Brendecke, 1995).

Fecal coliforms

Fecal coliforms in water were quantitated using spread plate and membrane filtration techniques on mFC agar (Difco, Detroit, MI). Volumes of up to 10 mL were assayed. Fecal coliforms in soil were processed according to a modified protocol for the elution of bacteria from soils (Zuberer, 1994). 10g portions of soil were mixed with glass beads and 95 mL of 0.1% peptone. Bottles were shaken on a rotary shaker for 20 minutes to elute the organisms from the soil particles. The eluent was analyzed for fecal coliforms using the spread plate technique. Plates were incubated inverted at 44.5±5°C for 24 hours. Blue or blue-white colonies were counted under a magnifier with light source. Random presumptive colonies were selected and aseptically transferred to EC broth with MUG (Difco, Detroit, MI). Broth tubes were incubated at 44.5±5°C for 24 hours. Tubes were examined for growth and fluorescence. Growth indicated the presence of fecal coliforms. Fluorescence under ultraviolet light indicated the presence of E. coli. EC with MUG tubes were compared to a positive control using E. coli ATCC# 15597.

Escherichia coli

E. coli was quantitated using the Simplateâ system for total coliforms and E. coli (IDEXX, Westbrook, ME). Water and soil samples of up to 1 mL were processed according to the manufacturer's instructions. After incubation at 37°C for 24 hours, plate wells showing a purple color were positive for total coliforms, and purple wells fluorescing under UV light were positive for E. coli. Numbers of E. coli were calculated according to the most probable number method using a table provided by the manufacturer, and results were expressed as MPN per 100 mL of water or gram of dry soil.

Fecal Streptococci

Fecal streptococci in water were analyzed using the spread plate and membrane filtration techniques on KF streptococcus agar (Difco, Detroit, MI). Volumes of up to 10 mL were analyzed. Eluent from soil was obtained as described above and analyzed using the spread plate technique. Plates were incubated inverted at 41.5°C for 48 hours. Light to dark pink colonies were counted under a magnifier with light source. Random presumptive colonies were selected and transferred to brain heart infusion agar (Difco, Detroit, MI). They were confirmed according to the protocol outlined in the Standard Methods for the Examination of Water and Wastewater (APHA, 1999), using brain heart infusion and bile esculin azide media (Difco, Detroit, MI). All confirmation tests were compared to positive controls using the fecal streptococcus Enterococcus faecalis and negative controls using Pseudomonas aeruginosa.

Coliphages

Coliphages in soil and water were analyzed using the double layer agar technique (Adams, 1959). Samples were held at 4°C for 24 hours before processing. TSA bottom agar was used. MEndo (Difco, Detroit, MI) top agar was used to suppress overgrowth of Gram positive organisms in the top agar layer. The host used was E. coli ATCC# 15597. 3 mL volumes of water and soil eluent were assayed. Plates were incubated inverted at 37°C for 24 hours. Plaques were counted under a magnifier with light source.

Protozoan parasites

Water samples were examined for Giardia and Cryptosporidium according to a modified Information Collection Rule protocol (See appendix A). Volumes of 1 L were examined. Parasites were detected using indirect immunofluorescent antibody staining and examination under an ultraviolet microscope.

Statistical Analysis

Statistical analysis was performed using analysis of variance (ANOVA) (Sokal and Rohlf, 1995) with SYSTAT Version 9 software (SPSS Inc., 1999). The level of significance was defined as 95% (a=0.05). Therefore, p values less than 0.05 generated by ANOVA are considered to be statistically significant.

Discussion Graywater Quality

Fecal coliforms

Fecal coliforms were consistently detected in all samples from all sampling sites (Figures 1&2). Seasonal variation can be seen in fecal coliform levels (Figure 1). Levels seem to peak in April, then decline in May-June, and then rise again in August-September. When the year is separated into quarters, statistical analysis (Table 16) shows significant difference in levels over time (µ=0.05, p=4.13´10-12). Highest overall fecal coliform levels were found at Site 14 (Figure 2), a site with graywater coming exclusively from the kitchen sink. This site had no pets, no children, and no storage, suggesting that the kitchen sink may represent a significant source of contamination, with levels surpassing that of graywater from all other combined household sources, as shown by the lower levels of fecal coliform bacteria present in graywater from other sites. Sites 1, 6, 17, and 18 had the next highest fecal coliform levels, with levels in graywater from these houses being roughly equal. Sites 1, 6, and 17 all included kitchen sink water in their graywater, again indicating the kitchen sink may represent a significant contamination source. Sites 1, 17 and 18 used in ground storage, possibly providing an environment conducive to bacterial growth. The lowest levels of fecal coliforms in graywater were found at site 19, a site utilizing washing machine water exclusively.

Levels of fecal coliforms were roughly equal in houses with and without children under 12 (Figure 9, Table 9). Again, a higher level is seen in house 14, a site with graywater coming exclusively from the kitchen sink. However, statistical analysis (Table 16) indicated that there is a significant difference in fecal coliform levels in houses with and without children (µ=0.05, p=6.88´10-12). Therefore, the presence of children may make a small difference in graywater fecal coliform load.

Fecal coliform levels were higher in households including the kitchen sink in their graywater than in houses excluding the kitchen sink (Figure 7, Table 10). However, Site 18 stands out (Figure 7). With no children and only one household pet (cat), the reasons for this higher level of fecal coliform contamination are unclear. Statistical analysis (Table 16) shows a significant difference in fecal coliform levels with the presence or absence of kitchen sink water (µ=0.05, p=8.89´10-12). The higher levels of contamination in graywater including the kitchen sink again points to the kitchen sink water as a contamination source, possibly due to the introduction of large amounts of organic matter, providing nutrient sources for organisms present. Washing of meat and poultry products in the sink may also introduce organisms into the graywater supply.

Fecal coliform levels in graywater also are higher in households using in ground storage tanks than in households using above ground tanks (Figure 8, Table 11). Statistical analysis (Table 16) indicates that storage does make a significant difference in fecal coliform levels (µ=0.05, p=4.82´10-12). The higher levels in sites with in ground storage may indicate that storage tanks may provide a favorable environment for bacterial growth while shielding organisms from sunlight, which can inactivate them.

Fecal coliform levels were slightly higher in houses without animals than in those with animals (Figure 10, Table 12). House 14, a house without any animals, had higher fecal coliform levels than some houses with animals. Statistical analysis (Table 16) indicates that animals do make a significant difference in fecal coliform levels (µ=0.05, p=6.23´10-12). The impact of the presence of animals on fecal coliform levels, though significant, may be small.

One of the interesting points that statistical analysis brings to light (Figure 8) is interaction between factors that influence graywater quality. Based on the site characteristics and samples collected, it is possible to analyze two way interactions- that is, to see if two factors interact to produce a significant difference in fecal coliform levels. In this study, it is possible to do this for quarter (time) and animals, source and animals, quarter and source, and quarter and children. Analysis of variance showed that all of these two-way interactions produce a significant difference in fecal coliform levels. Therefore, presence or absence of animals, presence or absence of kitchen sink water, presence or absence of children, and storage method all impact the way that fecal coliform levels vary over time. This suggests that what may be happening here with the factors that influence graywater quality is an additive or synergistic effect. Rather then being viewed in isolation, the interaction of these factors and their impact on graywater quality must be considered.

Escherichia coli

E. coli in water was quantitated using the Simplateâ system, a most probable number method for total coliform bacteria and E. coli. If all wells in a Simplateâ are positive for E. coli, then the MPN of organisms is greater than the maximum number that can be detected by the plate- a limitation of the method.

E. coli was sampled on a total of 10 dates in August through December. E. coli was detected on all dates except one in mid-October (Figure 5). E. coli was highest in September, with levels declining sharply until mid October and then beginning to rise again in November and December. E. coli levels reached another peak in mid December.

E. coli was detected in graywater from 6 of 10 sites (Figure 6). Sites 10, 13, 14, and 19 had no detectable E. coli, while site 17 had the highest levels.

These results disagree with the results for E. coli using ECMUG broth, a presence/absence test for E. coli. Samples positive for E. coli as a percentage of all samples taken can be seen in Figures 3 and 4. These samples, taken from March through December, show a seasonal variation in number of samples testing positive for the presence of E. coli (Figure 3). Here, samples with E. coli seem to decline from March to August, then rise again until September, when there is a sharp decline. Samples with E. coli then rise sharply again in mid September, then decline until mid November, then rise again. The midyear decline may be due to inactivation by the typically high summer temperatures.

E. coli was detected in samples from all sites. Site 17 was the only site with 100% of samples positive for E. coli (Figure 4). This is a site with kitchen sink water, a child under the age of 5, and in ground storage. Sites 7 and 14 had the next highest percentage of positive samples, 60%. Site 14 used exclusively kitchen sink water, while site 7 used washing machine and bath water. Contrary to the Simplateâ results, which detected no E. coli in graywater from sites 10, 13, 14, and 19 (Figure 4), ECMUG testing showed the presence of E. coli in graywater from these sites, ranging from 25 to 60% of total samples (Figure 6). The Simplates also indicated that no E. coli was present in any samples on 10/18 (Figure 3), whereas approximately 80% of samples taken on that date were positive for E. coli using ECMUG (Figure 4). These differences are probably due to variation between two methods for the detection of E. coli.

From results of quantitation of E. coli by the Simplate method, it is difficult to tell how the presence or absence of kitchen sink water influences the levels of E. coli in graywater (Figure 11, Table 10). Levels of E. coli when all sites are averaged differ by one order of magnitude. While the highest E. coli level was found in house 17, which included kitchen sink water, the method detected no E. coli in house 14, a site using exclusively kitchen sink graywater. E. coli at site 7, without kitchen sink water, was higher than house 1, using kitchen sink water. It may be that while the kitchen sink introduces fecal bacteria into graywater, it does not necessarily introduce higher levels of E. coli than would be there in the absence of kitchen sink water. This is reinforced by the ECMUG results (Figure 4). The highest percentage of samples positive for E. coli was found at site 17. However, the other sites using kitchen sink graywater, 1, 6, and 14, did not have consistently higher percentages of samples positive for E. coli than sites that did not use kitchen sink graywater.

Levels of E. coli were higher in houses using in ground storage (Figure 13, Table 11). The somewhat higher levels in other sites with in ground storage may indicate that storage tanks may provide a favorable environment for bacterial growth while shielding organisms from sunlight, which can inactivate them.

The impact of children in a household on E. coli levels in graywater is questionable (Figure 13, Table 10). Averaged across all sites, levels of E. coli in houses with and without children were roughly equal (Table 9). The highest level of E. coli was found in house 17, a house with children. However, E. coli in houses 1, 2, and 7, houses without children, were higher than in house 5, a house with children. House 5 E. coli levels were also equal to house 18, a house with children. There were houses both with and without children in which no E. coli was detected. Therefore, the presence of children younger than 12 in a household may not increase the load of E. coli in graywater. This is confirmed by the ECMUG results. Again, house 17 had the highest number of samples positive for E. coli (Figure 4). However, households without children did not have consistently lower numbers of samples positive for E. coli than did houses 5 and 10, the other households with children. In some cases, houses without children had equal or greater numbers of samples positive for E. coli when compared with houses with children.

As with children, the impact of animals on the levels of E. coli in household graywater is questionable (Figure 14, Table 12). Averaged across all sites, levels of E. coli in houses with and without animals were roughly equal (Table 12). The highest levels of E. coli were found in houses 17 and 7, houses with animals. However, houses with and without animals had undetectable levels of E. coli. Houses 5 and 18, with animals, had E. coli equal to house 2, without animals. House 1, without animals, had higher E. coli than houses 5 and 18. Therefore, the presence of animals in a household may not increase the load of E. coli in graywater.

Fecal Streptococci

Fecal streptococci were detected in graywater from all sites (Figure 27), with sites 14, 17, and 19 having the highest levels. Like the fecal coliforms, the fecal streptococci indicate that fecal contamination is making its way into household graywater.

Protozoan Parasites

Of those samples that could be examined by IFA staining, all were negative for protozoan parasites (Table 13). This is consistent with the fact that there was no evidence (based on self-reported illnesses of residents) that anyone in the households might be shedding protozoan parasites.

Coliphages

There was only one occurrence of coliphages in graywater, occurring in August at site 5. This indicates that it was only a random occurrence, and coliphages are not usually present in this graywater.

Irrigated Soil Quality

Fecal coliforms

Fecal coliforms were detected in most samples of graywater irrigated soil (Figure 15). Seasonal variation can be seen in fecal coliform levels in graywater irrigated soil (Figure 15). Levels are highly variable from month to month. Peaks seem to occur in June and August-September. When the year is separated into quarters, statistical analysis (Table 17) shows significant difference in levels over time (µ=0.05, p=4.17´10-12). Fecal coliforms were detected in potable water irrigated (background) soil in fewer months of the year. In most samples, fecal coliform levels in potable water irrigated soil were lower than in graywater irrigated soil, although a few times they were slightly higher. Fecal coliforms in potable water irrigated soil seem to have a differing pattern of seasonal variation, with peaks in January and August and sharp reductions in April-June.

Fecal coliforms were detected at various times in graywater irrigated soil from all sites (Figure 16). Highest overall fecal coliform levels were found at sites 6K, 6W, 10 and 14. Sites 6K and 14, with the highest levels of soil contamination, are irrigated with graywater coming exclusively from the kitchen sink. Site 14 has no pets, no children, and no storage, suggesting that the kitchen sink may represent a significant source of contamination, with contamination from the water being introduced into the soil, as shown by the lower levels of fecal coliform bacteria present in graywater from other sites. However, houses 1 and 17, also using kitchen sink graywater, have lower levels of fecal coliforms in soil than 6K and 14. Therefore, the contamination at sites 6K and 14 is probably not entirely due to the kitchen sink, but may have other contributing factors. 6W and 10 are irrigated with washing machine water, suggesting that the washing machine can still serve as a source of fecal contamination.

Levels of fecal coliforms in background soil were lower than levels in graywater irrigated soil for most sites (Figure 16). For some sites, fecal coliforms were at undetectable levels in background soil. Statistical analysis (Table 17) shows that type of water used for irrigation makes a significant difference in soil fecal coliform levels (µ=0.05, p=8.82´10-12). This shows that graywater irrigation does introduce fecal coliform contamination into the soil at levels above what is normally present.

Levels of fecal coliforms in graywater irrigated soil were higher in houses with children under 12 than in houses without (Figure 25, Table 9). Statistical analysis (Table 17) indicated a significant difference in soil fecal coliform levels because of the presence or absence of children (µ=0.05, p=7.28´10-12). Analysis showed that children had a statistically significant impact on fecal coliform levels in graywater (Table 16), and this effect appears to carry over into the soil.

Across all sites, fecal coliform levels were highest in graywater irrigated soil at sites including the kitchen sink in their graywater (Table 10). Fecal coliform levels were highest in graywater irrigated soil at two sites including the kitchen sink in their graywater, 6K and 14 (Figure 21). Statistical analysis (Table 17) showed a significant difference in fecal coliform levels in graywater with and without the kitchen sink (µ=0.05, p=5.84´10-12). The higher levels if contamination in graywater irrigated soil including the kitchen sink again points to the kitchen sink water as a contamination source. However, this may not always be the case, since sites 1 and 17, also users of kitchen sink water, have lower levels of fecal coliforms in graywater irrigated soil than some households without kitchen sink water.

Fecal coliform levels in graywater irrigated soil are higher in households using above ground storage tanks than in households using in ground tanks (Figure 22, Table 11). Statistical analysis (Table 17) shows that storage makes a significant difference in fecal coliform levels (µ=0.05, p=1.20´10-11). This differs from the trend seen in the graywater, where levels in water sampled from in ground tanks were slightly higher (Figure 8). It may be that over time, the use of an above ground surge tank creates an environment similar to a continuous laboratory culture, with nutrients always being introduced to support growth of bacteria which are then washed out onto the soil. Bacteria in this state are not subject to the growth and die off kinetics that would probably be seen in a closed environment such as a storage tank.

Fecal coliform levels in graywater irrigated soil were similar in houses with animals (Figure 23, Table 12). This trend also occurs in the background soil (Figure 24). However, statistical analysis shows a significant difference in soil with and without animals (µ=0.05, p=1.12´10-11), so it is possible that the impact of animals, thought significant, is still small.

As with the graywater data, statistical methods can be used to analyze two way interactions. That is, to see if two factors interact to produce a significant difference in fecal coliform levels. With the soil data, it is possible to do this for quarter (time) and animals, storage and quarter, quarter and source, quarter and children, and storage and animals. Analysis of variance showed that all of these two-way interactions produce a significant difference in fecal coliform levels in soil (Table 17). Again, this suggests that what may be happening here with the factors that influence graywater quality is an additive or synergistic effect. Rather then being viewed in isolation, the interaction of these factors and their impact on graywater quality must be considered.

Escherichia coli

The presence or absence of E. coli in soil was analyzed using ECMUG. E. coli was sampled on a total of 10 dates in August through December . E. coli was not detected in graywater irrigated soil on dates in April, May, June, November, and December (Figure 17). The number of samples testing positive for E. coli was highest in august through November. Background soil E. coli did not follow the same trend (Figure 18). The number of samples positive for E. coli peaked in March, but no E. coli was detected in background soil from April through June. This indicates that irrigation with graywater does introduce E. coli into the soil that would not otherwise be present.

E. coli was detected in 12 of 13 graywater irrigated soil sites (Figure 19). Site 1 had no detectable E. coli, while sites 6T/S, 6K, 6W and 14 had the highest percentages of samples positive for E. coli.

E. coli was detected in 7 of 12 background soil sites (Figure 20). Sites 1, 5, 6, 18, and 30 had no detectable E. coli, while sites 5, 13, and 14 had the highest percentages of samples positive for E. coli. Again, this indicates that irrigation with graywater does introduce E. coli into the soil that would not otherwise be present.

Fecal Streptococci

Fecal streptococci were detected in most samples of graywater irrigated and background soil (Figure 28). For most sites, fecal streptococci levels were higher in graywater irrigated than in background soil. Fecal streptococci do seem to be present even in potable water irrigated soil. However, the differences in levels between the two soils suggest that graywater irrigation introduces additional fecal streptococci into the soil.

Coliphages

There were two occurrences of coliphages in soil, one in graywater irrigated and one in background soil. One was at background site 2 in December, and one was at Site 6K in September. These occurrences were not correlated with any coliphages in the graywater at these sites, and thus appear to be random.

Soil Controls

Eight soil samples were collected in January from Tucson, Arizona residential yard sites irrigated with potable water (Table 6). They were analyzed for fecal coliforms and fecal streptococci. No fecal coliforms were detected in any of the soils (Table 7). There was fecal streptococci detected in soil at one of the sites. These controls indicate that low or undetectable levels of these organisms that are normally present in potable water irrigated soil.

Go to Main page