Microbial indicators and pathogens: removal, relationships and predictive capabilities in water reclamation facilities

Four water reclamation facilities in north-eastern Spain were monitored over 2 years to determine the occurrence and concentrations of a set of microbial indicators (total coliforms, Escherichia coli, enterococci, spores of sulphite reducing clostridia, somatic coliphages, F-specific RNA phages, phages infecting Bacteroides fragilis strain RYC2056 and phages infecting Bacteroides tethaiotaomicron strain GA-17), and two selected pathogens (cytopathogenic enteroviruses and viable Cryptosporidium oocysts). The indicator (survival) and index (presence) functions of the various indicators tested were evaluated through the wastewater treatments. The inactivation pattern of all groups of bacteriophages tested was closer to the inactivation of enteroviruses than to the inactivation of the conventional bacterial indicators tested. The inactivation of sulfite reducing clostridia spores and bacteriophages more closely approximates the reduction of viable Cryptosporidium than do the conventional bacterial indicators. We observed neither index functions nor a predictive relationship between any of microbial indicators and viable Cryptosporidium oocysts. In contrast, several regression models (r > 0.6) and discriminant functions (67-88% well classified samples) based mostly on numbers of bacteriophages were able to predict both the presence and concentrations of enteroviruses. A combination of both bacterial and bacteriophage indicators seem to be the best choice for ensuring the microbial quality of reclaimed water.     

噬菌体作为水中病毒指示物的研究进展

噬菌体作为潜在的水中病毒指示生物，可用于污水再生利用过程中的病毒学安全评价、阐明病毒灭活机理以及改进病毒检测方法等领域的研究。介绍了常用的指示噬菌体——SC噬菌体、F—RNA噬菌体和Bacteroides fragilis噬菌体在环境中的分布、存活和去除特性及其作为水质评价和病毒指示生物的研究进展。     

Survival of infectious Poliovirus-1 in river water compared to the persistence of somatic coliphages, thermotolerant coliforms and Poliovirus-1 genome

The microbiological quality of water is currently assessed by search for fecal bacteria indicators. There is, however, a body of knowledge demonstrating that bacterial indicators are less resistant to environmental factors than human pathogenic viruses and therefore underestimate the viral risk. As river water is often used as a resource for drinking water production, it is particularly important to obtain a valid estimation of the health hazard, including specific viral risk. This work was conducted to compare the survival of infectious Poliovirus-1 used as a pathogenic virus model to the persistence of, on the one hand, thermotolerant coliforms commonly used as indicators and on the other hand, to somatic coliphages and Poliovirus-1 genome considered as potential indicators. We studied the behavior of infectious Poliovirus-1 and the three (potential) indicators of viral contamination in river water at three different temperatures (4 degrees C,18 degrees C and 25 degrees C). This experiment was performed twice with river water sampled at two different periods, once in winter and once in summer. Our results showed that the survival of thermotolerant coliforms can be 1.5-fold lower than infectious Poliovirus-1. In contrast, under all our experimental conditions, somatic coliphages and Poliovirus-1 genome persisted longer than infectious Poliovirus-1, surviving, respectively, 2-6-fold and about 2-fold longer than infectious Poliovirus-1. According to our results exclusively based on survival capacity, somatic coliphages and viral genome, unlike thermotolerant coliforms appear to be better indicators of viral contamination in river water. Moreover, the disappearance of viral genome is well-correlated to that one of infectious virus irrespective of the conditions tested.     

Coliphages as an indicator of faecal pollution in water. Its relationship with indicator and pathogenic microorganisms

The study was designed to test the proposal that Escherichia coli specific bacteriophages might serve as universal faecal pollution indicators in water. A highly specific, sensitive and rapid technique for the detection and quantification of these virus particles was developed. The numerical relationship between E. coli and its parasitic phages was investigated in three different aqueous ecosystems such as sea water in the vicinity of sewage outfalls, river water contaminated by domestic and industrial sewage discharges, and estuarine waters, and found to be very close. In addition, the results obtained indicate that the coliphages are good indicators of the presence of the pathogenic microorganisms studied. In nearly all the water samples tested, the results suggest that coliphages are better indicators of faecal pollution than the classical indicator systems currently employed.     

Occurrence and levels of indicator bacteriophages in bathing waters throughout Europe

Somatic coliphages, F-specific RNA bacteriophages, bacteriophages infecting Bacteroides fragilis, Escherichia coli and enterococci were counted in bathing waters in the late spring and summer. We tested fresh and marine bathing waters from North, South, East and West Europe expected to contain between 100 and 500 E. coli per 100 ml, although wider ranges were sometimes found. Bacteriophages were counted after concentration, since a preliminary study proved that this step was necessary to obtain positive counts. During monitoring, a first-line quality control with reference materials for bacteria and bacteriophages was performed by all the laboratories participating in the study. The same microbes were also counted in raw sewage samples from various areas in Europe, where the bacterial indicators and the three groups of bacteriophages were detected in roughly the same numbers. All groups of bacteriophages were detected in both fresh and marine bathing waters throughout Europe. Reliable and complete results from 147 samples showed that for log-transformed values, E. coli and bacteriophages were slightly correlated. However, the slope of the regression line changed according to E. coli concentration and the correlation diminished when this concentration was close to zero per 100 ml. The ratios between E. coli and phages in bathing waters differed significantly from those in sewage. The relative amounts of bacteriophages, mainly somatic coliphages and phages infecting Bact. fragilis RYC2056, increased in bathing waters with low E. coli concentration, especially in seawater samples containing <100 E. coli per 100 ml. The relationship of bacteriophages with respect to enterococci paralleled that of bacteriophages with respect to E. coli. Somatic coliphages and bacteriophages infecting Bact. fragilis are useful to predict the presence of some pathogens with the same origin as present bacterial indicators but with higher survival rates.     

Comparative disinfection of treated sewage with chlorine and ozone: Effect of nitrification

Chlorine and ozone were compared in pilot plants (capacity about 3.2 m³ h⁻¹), which were fed with the same activated sludge treated and filtered water. Together with physico-chemical analysis the water was analysed for different types of microorganisms, including vegetative bacteria (total and thermotolerant coliforms, faecal streptococci and Pseudomonas aeruginosa), bacterial spores (spores of aerobic bacteria at 37°C and sulphite reducing clostridia) and bacterial viruses (somatic coliphages and F-specific bacteriophages).The average chlorine and ozone dose were, respectively, 3.65 and 15.3 mg l⁻¹ of water, while after a contact time for both of about 25 min the average residual concentrations were 1.79 and 0.35 mg l⁻¹ of water. These residuals were measured with the DPD-method. The ammonia-N concentration varied greatly (0.06–72.0 mg l⁻¹) and was used to group the data into four classes: (1) non-nitrified water, defined as water in which nitrate-N was smaller than ammonia-N; (2) moderately nitrified water, in which nitrate-N was larger than ammonia-N and the ammonia-N was higher than 2 mg l⁻¹; (3) well nitrified water, defined as water in which ammonia-N was lower than 2 mg l⁻¹; (4) very well nitrified water, in which ammonia-N was smaller than 0.5 mg l⁻¹.This classification indicated that the concentrations of most other impurities decreased with a better nitrification. Statistical analysis of the data showed also that ozone was a better disinfectant than chlorine in the case where the disinfection is based upon their residual content. The degree of nitrification had a greater effect on chlorine disinfection than on ozone disinfection.During chlorination the total residual chlorine decreased, with better nitrification; the chlorine demand increased; the composition of the residual chlorine changed very much and the inactivation of bacterial viruses improved. The vegetative bacteria showed a varying pattern; most were inactivated in moderately nitrified water, when the dichloramine concentration was highest and false positive FAC concentration was lowest of the four classes. Reduction of bacterial spores was not observed.During ozonization other effects were indicated. Reductions of most organims increased slightly with better nitrification; only reductions of F-specific bacteriophages decreased. There was also a small decrease of bacterial spores. The treated effluent had a high ozone consumption and the inactivation of the organisms was low in relation to ozone dose and residual ozone.The bromide concentration (0.3–2.9 mg l⁻¹) effected the chemistry of chlorine and ozone and had a positive effect on chlorine and ozone disinfection of total coliforms.For most types of micoorganisms the disinfection coefficients of the Selleck model and the germicidal efficiencies could be determined.     

Characterization of Enterococcus faecalis-infecting phages (enterophages) as markers of human fecal pollution in recreational waters

Enterophages are a novel group of phages that specifically infect Enterococcus faecalis and have been recently isolated from environmental water samples. Although enterophages have not been conclusively linked to human fecal pollution, we are currently characterizing enterophages to propose them as viral indicators and possible surrogates of enteric viruses in recreational waters. Little is known about the morphological or genetic diversity which will have an impact on their potential as markers of human fecal contamination. In the present study we are determining if enterophages can be grouped by their ability to replicate at different temperatures, and if different groups are present in the feces of different animals. As one of the main objectives is to determine if these phages can be used as indicators of the presence of enteric viruses, the survival rate under different conditions was also determined as was their prevalence in sewage and a large watershed. Coliphages were used as a means of comparison in the prevalence and survival studies. Results indicated that the isolates are mainly DNA viruses. Their morphology as well as their ability to form viral plaques at different temperatures indicates that several groups of enterophages are present in the environment. Coliphage and enterophage concentrations throughout the watershed were lower than those of thermotolerant coliforms and enterococci. Enterophage concentrations were lower than coliphages at all sampling points. Enterophages showed diverse inactivation rates and T₉₀ values across different incubation temperatures in both fresh and marine waters and sand. Further molecular characterization of enterophages may allow us to develop probes for the real-time detection of these alternative indicators of human fecal pollution.     

Removal of native coliphages and coliform bacteria from municipal wastewater by various wastewater treatment processes: implications to water reuse

The efficacy of a conventional activated sludge wastewater treatment process and the membrane bioreactor technology in removing microbial pathogens was investigated. Total and fecal coliforms and somatic and F-specific coliphages were used as indicators of pathogenic bacteria and viruses. Up to 5.7 logs removal of coliforms and 5.5 logs of coliphages were observed in the conventional treatment process with advanced tertiary treatment. Addition of chemical coagulants seemed to improve the efficacy of primary and secondary treatment for microorganism removal. Complete removal of fecal coliforms and up to 5.8 logs removal of coliphages was observed in the MBR system. It was shown that the MBR system was capable of high removal of coliphages despite the variation in feed coliphage concentrations. The results of this study indicated that the MBR system can achieve better microbial removal in far fewer steps than the conventional activated sludge process with advanced tertiary treatment. The final effluent from either treatment processes can be potentially reused.     

Concentrations of coprostanol that correspond to existing bacterial indicator guideline limits

Coprostanol is a faecal sterol that has been proposed as an alternative measure of faecal pollution. While the technique has been used successfully to trace sewage-derived organic matter in a range of environments, it has not been embraced for use as a water quality indicator. This is mostly because of a lack of epidemiological evidence relating coprostanol abundance to any health risk. However, there is a valuable reason why the concentration of coprostanol should be related as quantitatively as possible to the abundance of bacterial indicators currently used to measure faecal pollution. The measurement of coprostanol (and concurrently other faecal sterols) offers many diagnostic and quantitative advantages over traditional techniques for detecting human sewage pollution versus faecal contamination from animal sources. Knowing the amount of coprostanol expected given a certain amount of human sewage pollution would provide a measure against which water managers could quantitatively assess faecal pollution as a whole and relate that assessment to variables with which they are more familiar. This study determines the relationships between coprostanol concentrations and indicator bacterial counts and synthesises the results from several environments to propose coprostanol concentrations broadly equivalent to existing bacterial standards. Our data suggest that 60 and 400 ng L⁻¹ of coprostanol correspond to currently defined primary and secondary contact limits for bacteria measured as thermotolerant coliforms (commonly referred to as faecal coliforms) or enterococci.     

Efficiency of removal of coliforms, faecal coliforms and coliphages in the tubli sewage treatment plant, bahrain

Total coliforms, faecal coliforms and coliphages in the raw sewage were detected in high numbers during the months averaging a maximum temperature of 20°C. The secondary and tertiary effluents show a 99–100% removal of the three indicators of faecal pollution. A direct correlation between the faecal coliforms and coliphages suggests that coliphages alone may be used as an indicator for the presence of pathogenic bacteria in contaminated waters.     

Differential persistence of F-specific RNA phage subgroups hinders their use as single tracers for faecal source tracking in surface water

The four subgroups of F-specific RNA bacteriophages (I-IV) have been proposed as potential tracers for faecal source tracking. Groups II and III predominate in human sources while groups I and IV are most abundant in animal sources. The four subgroups of naturally occurring F-specific RNA bacteriophages were identified in different samples by plaque hybridization with genotype-specific probes and the persistence of each subgroup was evaluated. The proportions of the F-specific RNA bacteriophage subgroups were measured in wastewaters, after inactivation in surface waters or after wastewater treatment and in mixtures of wastewater of human and animal origin. Our results indicate that phage groups differ in their persistence in the environment and to different disinfecting treatments. The greater survival of subgroups I and II in treated samples hinders the interpretation of results obtained with F-specific RNA bacteriophages. The phages of subgroups III and IV were the least resistant to all treatments. These results should be considered when using genotypes of F-specific RNA as sole tracers for faecal source tracking.     

Airborne enteric coliphages and bacteria in sewage treatment plants

The concentrations of airborne culturable microorganisms were determined in wastewater and sludge treatment processes of seven sewage treatment plants. Two types of coliphages, Salmonella and total viable bacteria were sampled by the BioSampler and the numbers of faecal coliforms and enterococci were obtained from the Andersen 6-stage impactor. The BioSampler recovered higher numbers of airborne coliphage viruses than has been measured with other liquid samplers in previous studies, suggesting that this sampler has improved efficiency for sampling airborne coliphages. Airborne coliphages were detected in many stages of the wastewater or sludge treatment process. The highest microbiological air contaminations were found in pre-treatment and aerated grit separation stages of the operation. This was attributed to aerosolisation of microorganisms by mechanical handling or forced aeration. Aeration and settling processes located outdoors caused low microbial concentrations, but the brush aerator released more microorganisms into the air. Our results emphasize the necessity for controlling the exposure of sewage workers to airborne microorganisms, especially in process areas that involve mechanical agitation or forced aeration of wastewater.     

Occurrence and densities of fungi from northern Greek coastal bathing waters and their relation with faecal pollution indicators

The objective of this study was to determine the frequency and densities of yeasts and filamentous fungi in coastal water samples as well as their correlation with the indicator bacteria of faecal pollution. The prevalence of fungi was investigated in parallel with the standard pollution indicator microorganisms in 197 marine water samples from six northern Greek prefectures during the bathing season May–October 1999. Filamentous fungi were isolated from all the examined samples and yeasts from 29 (14.7%) of them; among the positive samples, their mean counts were 90.9 and 38.4 cfu/100 ml, respectively. A total of 23 genera of filamentous fungi and four genera of yeasts were identified. Prevailing genera of filamentous fungi were Penicillium, Aspergillus and Alternaria spp., whereas Candida spp. was the most frequently isolated yeast. Counts of yeasts were significantly (p<0.01) correlated with those of total and faecal coliforms, whereas no correlation was found between filamentous fungi and the indicator bacteria of faecal pollution. Significantly higher counts of total and faecal coliforms (p<0.05), and enterococci (p<0.001), were found during the months with the higher water temperatures and bather numbers. In the six prefectures, significant differences were observed in the counts of filamentous fungi and yeasts as well as in the counts of all the faecal pollution indicators. The results of this study indicate that coastal water can be a path for contamination of swimmers with yeasts and filamentous fungi and that the pollution indicator microorganisms cannot always predict their presence.     

Application of real-time PCR assays to genotyping of F-specific phages in river water and sediments in Japan

Genotyping of F-specific RNA phages is currently one of the most promising approaches to differentiate between human and animal fecal contamination in aquatic environments. In this study, a total of 18 river water and sediment samples were collected from the Tonegawa River basin, Japan, in order to describe the genogroup distribution of F-specific RNA and DNA phages using genogroup-specific real-time PCR assays. F-specific phages were detected in nine (100%) river water and six (67%) sediment samples. Eighty-five phage plaques were isolated from these samples and subjected to real-time PCR assays specific for the phages. F-specific RNA phages of human genogroups (II and III) were detected in 32 (38%) plaques, whereas those of animal genogroups (I and IV) were detected in 17 (20%) plaques. No correlation was observed between the genogroup distribution of F-specific RNA phages and the occurrence of human adenovirus genomes, suggesting that genotyping of the phages alone is inadequate for the evaluation of the occurrence of viruses in aquatic environments. SYBR Green-based real-time PCR assay revealed the presence of F-specific DNA phages in four (5%) plaques, which were further classified into two genogroups (fd- and f1-like phages) by sequence analysis. Thirty-two (38%) plaques were not classified as the F-specific phage genogroups, indicating the limited applicability of these real-time PCR assays to a wide range of aquatic environmental samples worldwide.     

Microorganisms in bioaerosol emissions from wastewater treatment plants during summer at a Mediterranean site

Measurements were conducted at a Mediterranean site (latitude 35 degrees 31' north and longitude 24 degrees 03' east) during summer, to study the concentration of microorganisms emitted from a wastewater treatment plant under intensive solar radiation (520-840 W/m2) and at elevated air temperatures (25-31 degrees C). Air samples were taken with the Air Sampler MAS 100 (Merck) at each stage of an activated-sludge wastewater treatment (pretreatment, primary settling tanks, aeration tanks, secondary settling tanks, chlorination, and sludge processors). Cultivation methods based on the viable counts of mesophilic heterotrophic bacteria, as well as of indicator microorganisms of faecal contamination (total and faecal coliforms and enterococci), and fungi were performed. During air sampling, temperature, solar radiation, relative humidity and wind speed were measured. The highest concentrations of airborne microorganisms were observed at the aerated grit removal of wastewater at the pretreatment stage. A gradual decrease of bioaerosol emissions was observed during the advanced wastewater treatment from the pretreatment to the primary, secondary and tertiary treatment (97.4% decrease of mesophilic heterotrophic bacteria, and 100% decrease of total coliforms, faecal coliforms and enterococci), 95.8% decrease of fungi. The concentration of the airborne microorganisms at the secondary and tertiary treatment of the wastewater was lower than in the outdoor control. At the same time, the reduction of the microbial load at the waste sludge processors was 19.7% for the mesophilic heterotrophic bacteria, 99.4% for the total coliforms, and 100% for the faecal coliforms and the enterococci, 84.2% for the fungi. The current study concludes that the intensive solar radiation, together with high ambient temperatures, as well as optimal wastewater treatment are the most important factors for low numbers of microbes in the air.     

Microbiological quality of water from hand-dug wells used for domestic purposes in urban communities in Kumasi,Ghana

Assessment was done on the microbiological quality of water in hand-dug wells in urban communities in Kumasi, Ghana. A total of 256 water samples were taken from eight wells and examined for faecal coliforms, enterococci and helminths. High contamination levels were recorded in the wells, more so in the wet season, with faecal coliforms levels between 6.44 and 10.19 log units and faecal enterococci between 4.23 and 4.85 CFU per 100 ml. Influence on protection and lining of wells on water quality was not pronounced but mechanization reduced contamination significantly by about 3 log units. This study shows a stronger influence of poor sanitation and improper placement of wells on water quality compared to improvements made from lining and protection of wells. In the race to increase access to drinking water in poor urban settlements, quality of groundwater could be a major barrier, if provision of drinking water is not matched with improvements in sanitation and urban planning.     

Risk factors contributing to microbiological contamination of shallow groundwater in Kampala,Uganda

A study of water quality variation in shallow protected springs in Kampala was undertaken over a 12-month period to assess the causes of microbiological contamination. The microbiological quality of water was assessed using thermotolerant coliforms and faecal streptococci. Sanitary inspections and hazard assessments were undertaken to identify faecal sources (hazards), contaminant pathways and contributory factors. Data were collected on rainfall and population as additional factors potentially exerting an influence on microbiological quality. Initial analysis of the data showed a significant relationship between median level of contamination and rainfall, in particular to short-term rainfall events. Total sanitary risk score showed a significant relationship with median level of contamination, but population density may be a confounding factor. The raw microbiological data were transformed into five water quality targets: <1 and < or =10 cfu 100ml(-1) for faecal streptococci; and <1, < or =10 and < or=50 cfu 100 ml(-1) for thermotolerant coliforms. The presence of individual risk factors as well as variables for rainfall and population density were analysed with respect to failure to meet these water quality targets using contingency tables. Logistic regression models were developed for each of the five water quality targets. The analysis strongly suggested that there is rapid recharge of the springs after rainfall and this leads to microbiological contamination. On-site sanitation was less important than other sources of faecal matter, which was consistent with a low sanitation coverage in the study area. The study suggested that improving sanitary completion and local environmental hygiene was more important than controlling on-site sanitation in improving the quality of these springs.     

Occurrence, molecular characterization and antibiogram of water quality indicator bacteria in river water serving a water treatment plant

Water pollution by microorganisms of fecal origin is a current world-wide public health concern. Total coliforms, fecal coliforms (Escherichia coli) and enterococci are indicators commonly used to assess the microbiological safety of water resources. In this study, influent water samples and treated water were collected seasonally from a water treatment plant and two major water wells in a Black Belt county of Alabama and evaluated for water quality indicator bacteria. Influent river water samples serving the treatment plant were positive for total coliforms, fecal coliforms (E. coli), and enterococci. The highest number of total coliform most probable number (MPN) was observed in the winter (847.5 MPN/100 mL) and the lowest number in the summer (385.6 MPN/100 mL). Similarly E. coli MPN was substantially higher in the winter (62.25 MPN/100 mL). Seasonal variation of E. coli MPN in influent river water samples was strongly correlated with color (R² = 0.998) and turbidity (R² = 0.992). Neither E. coli nor other coliform type bacteria were detected in effluent potable water from the treatment plant. The MPN of enterococci was the highest in the fall and the lowest in the winter. Approximately 99.7 and 51.5 enterococci MPN/100 mL were recorded in fall and winter seasons respectively. One-way ANOVA tests revealed significant differences in seasonal variation of total coliforms (P < 0.05), fecal coliforms (P < 0.01) and enterococci (P < 0.01). Treated effluent river water samples and well water samples revealed no enterococci contamination. Representative coliform bacteria selected by differential screening on Coliscan Easygel were identified by 16S ribosomal RNA gene sequence analysis. E. coli isolates were sensitive to gentamicin, trimethoprim/sulfamethazole, ciprofloxacin, vancomycin, tetracycline, ampicillin, cefixime, and nitrofurantoin. Nonetheless, isolate BO-54 displayed decreased sensitivity compared to other E. coli isolates. Antibiotic sensitivity pattern can be employed in microbial source tracking.     

Monitoring source water for microbial contamination: evaluation of water quality measures

Watershed management programs often rely on monitoring for a large number of water quality parameters to define contaminant issues. While coliforms have traditionally been used to identify microbial contamination, these indicators cannot discriminate among potential contaminant sources. Microbial source tracking (MST) can provide the missing link that implicates the sources of contamination. The objective of this study was to use a weight-of-evidence approach (land use analysis using GIS, sanitary surveys, traditional water quality monitoring, and MST targets) to identify sources of pollution within a watershed that contains a raw drinking water source. For the study watersheds, statistical analyses demonstrated that one measure each of particulate matter (turbidity, particle counts), organic matter (total organic carbon, dissolved organic carbon, UV(254) absorbance), and indicator organisms (fecal coliforms, enterococci) were adequate for characterizing water quality. While these traditional parameters were useful for assessing overall water quality, they were not intended to differentiate between microbial sources at different locations. In contrast, the MST targets utilized (Rhodococcus coprophilus, sorbitol-fermenting Bifidobacteria, and male-specific coliphages) pinpointed specific sources of microbial pollution. However, these targets could not be used for routine monitoring due to a high percentage of non-detects.     

Monitoring of water quality from roof runoff: Interpretation using multivariate analysis

The quality of harvested rainwater used for toilet flushing in a private house in the south-west of France was assessed over a one-year period. Temperature, pH, conductivity, colour, turbidity, anions, cations, alkalinity, total hardness and total organic carbon were screened using standard analytical techniques. Total flora at 22 °C and 36 °C, total coliforms, Escherichia coli and enterococci were analysed. Overall, the collected rainwater had good physicochemical quality but did not meet the requirements for drinking water. The stored rainwater is characterised by low conductivity, hardness and alkalinity compared to mains water. Three widely used bacterial indicators - total coliforms, E. coli and enterococci - were detected in the majority of samples, indicating microbiological contamination of the water. To elucidate factors affecting the rainwater composition, principal component analysis and cluster analysis were applied to the complete data set of 50 observations. Chemical and microbiological parameters fluctuated during the course of the study, with the highest levels of microbiological contamination observed in roof runoffs collected during the summer. E. coli and enterococci occurred simultaneously, and their presence was linked to precipitation. Runoff quality is also unpredictable because it is sensitive to the weather. Cluster analysis differentiated three clusters: ionic composition, parameters linked with the microbiological load and indicators of faecal contamination. In future surveys, parameters from these three groups will be simultaneously monitored to more accurately characterise roof-collected rainwater.