Faecal bacterial loads during flood events in Northwestern Mediterranean coastal rivers

SummaryIn Mediterranean coastal rivers, floods last often less than a few hours but supply large amounts of contaminants to transitional and coastal waters. Estimating flood loads requires appropriate sampling strategies. We applied flood-scale sampling for the survey of two rivers flowing into the Thau lagoon (France). Two bacterial indicators were considered, thermotolerant coliforms (TTC) and faecal streptococci (FC). During floods, concentrations of indicator bacteria associated with non-mineral suspended solids increased quickly with the rising flow, their decrease during the recession period was slow and erratic. Statistical analysis was performed on total bacterial flood loads measured during 20 floods, versus hydrological variables and land-use characteristics. The analysis highlighted the significant impacts of human pollution sources together with the magnitude of the flood. Regarding the results, the best linear regression models linked total bacterial flood loads to peak discharge for both TTC and FS, reinforcing the assumption that in-stream bacterial stores play an important role in the level of bacterial flood loads in Mediterranean coastal rivers. At an annual scale, between 13.9 and 16.6 log10cfu of TTC could be supplied depending on the hydrological conditions during the year. Over the 12 year period, from 1994 to 2006 it was shown that the flood loads were responsible for at least 98% of the TTC total annual load and in 8 of 12 years the floods contributed to at least 99.9% of the annual loads. Over the same period on average the single major flood represents 74% of the total annual load. The contribution of in-stream bacterial stores was demonstrated but spatial variations in total flood loads showed that this contribution is difficult to evaluate. Bacteria from land stores appeared to be negligible in both catchments.

► We analyze temporal dynamics and spatial evolution of bacterial loads during Mediterranean flood events.
► The peak discharge and flood volume are of primary importance to characterize bacterial flood loads.
► The best linear regression models linked total bacterial flood loads to peak discharge.
► Bacterial load of the major flood of the year could represent 99% of the total annual load.
► Within intermittent river, bacterial in-stream stores and overflows of sewage systems contribute largely to bacterial load.