A two-year monitoring program of Cryptosporidium parvum oocysts, Giardia duodenalis cysts, Escherichia coli, Clostridium perfringens spores and adenovirus was conducted in three large rivers in France used for recreational activities and as a resource for drinking water production. Fifty-liter river water and one thousand-liter tap water samples were concentrated using hollow-fiber ultrafiltration and analyzed by molecular biology or laser-scanning cytometry. In order to evaluate watershed land use influence on microorganism concentration changes, occurrence and seasonality of microorganisms were studied. The highest concentrations of protozoan parasites and C. perfringens were found for one of the three sites, showing a high proportion of agricultural territories, forests and semi-natural environments, which may be partly attributable to soil leaching due to rainfall events. On the contrary, the highest concentrations of adenoviruses were found at the two other sites, probably due to strong urban activities. Health risk assessment was evaluated for each waterborne pathogen regarding exposure during recreational activities (for a single or five bathing events during the summer). The calculated risk was lower than 0.5% for parasites and varied from 1% to 42% for adenovirus. A theoretical assessment of microorganism removal during the drinking water treatment process was also performed, and it showed that an absence of microorganisms could be expected in finished drinking water. This hypothesis was confirmed since all tested tap water samples were negative for each studied microorganism, resulting in a risk for drinking water consumption lower than 0.01% for parasites and lower than 0.5% for adenovirus.

The present study aimed to compare several methods for quantifying and discriminating between the different physiological states of a bacterial population present in drinking water. Flow cytometry (FCM), solid-phase cytometry (SPC), epifluorescence microscopy (MSP) and culture method performances were assessed by comparing the results obtained for different water samples. These samples, including chlorinated and non-chlorinated water, were collected in a drinking water treatment plant. Total bacteria were quantified by using SYBR Green II (for FCM) and 4',6'-diamino-2-phenylindole (DAPI) (for MSP), viable and non-viable bacteria were distinguished by using SYBR Green II and propidium iodide dual staining (for FCM), and active cells were distinguished by using CTC (for MSP) and Chemchrome V6 (for FCM and SPC). In our conditions, counts using microscopy and FCM were significantly correlated regarding total bacteria and active cells. Conversely, counts were not significantly similar using solid-phase and FCM for active bacteria. Moreover, the R2A medium showed that bacterial culturability could be recovered after chlorination. This study highlights that FCM appears to be a useful and powerful technique for drinking water production monitoring.

A 4-month sampling campaign has been conducted for the monitoring of three drinking water treatment plants using flow cytometry and culture-based methods to provide information related to changes in bacterial concentration according to treatments. Flow cytometry is a fast and user-friendly technique enabling bacteria quantification and viability assessment in less than 1 hour. Specific profiles regarding log-reduction of total bacteria were obtained for each treatment plant. Chlorination appeared to be the most effective by causing metabolism inactivation and nucleic acid damages. Ozonation showed a significant impact on cell activity in contrast with ultraviolet treatment which strongly affected bacterial DNA. In addition, the results showed that active bacteria quantified by flow cytometry were significantly correlated with culturable bacteria. This alternative approach appeared as gainful compared to culture methods as it greatly facilitates the diagnosis of treatment plant process for drinking water production monitoring.

Anaerobic sulfite-reducing and aerobic spore-forming bacteria are used as indicators to verify the reliability of water treatment plant efficiency. The objectives of this study aimed to highlight the more accurate endospore indicator for water treatment monitoring and to select an efficient filtration system in order to detect low amounts of endospores from large volumes. The performance of two different ultrafiltration modules was assessed with two filtration configurations (tangential and dead-end). The Hemoflow™ and the Dizzer® filters present recovery yields of 73% and 77% for 100-liter samples of drinking and sand filtered water, respectively. Both systems enabled detection of endospores at a concentration level of 100 spores in 100 liters, by analyzing the totality of the volume, whereas the standard method did not highlight the presence of spores, analyzing only 100 mL subsamples. In addition, experiments on a sand filtration pilot confirmed that the concentration calculation using the standard method is not reliable, due to extrapolation from the small volumes analyzed. This study highlights that the collection of large volumes using a fieldable dead-end hollow-fiber ultrafiltration system in association with the detection of anaerobic sulfite-reducing spore-forming bacteria is an effective way to monitor drinking water production.