Patricia Turgeon1, Pascal Michel2,1, Patrick Levallois3,4, André Ravel2,1, Marie
Archambault1, Martin-Pierre Lavigne2,1, Serge Olivier Kotchi2,1, Stéphanie Brazeau2,1
1. Groupe de recherche en épidémiologie des zoonoses et santé publique, Faculté de médecine vétérinaire, Université de Montréal, Saint-Hyacinthe, QC, Canada
2. Laboratoire de lutte contre les zoonoses d’origine alimentaire, Agence de la santé publique du Canada, Saint-Hyacinthe, QC, Canada
3. Centre Hospitalier Universitaire de Québec, Université Laval, Québec, QC, Canada 4. Institut national de santé publique du Québec, Québec, QC, Canada
Abstract
Large amounts of non-pathogenic and pathogenic microorganisms may flow into natural recreational waters through various sources of fecal contamination originating from agriculture practices, human activities and wildlife animals.
Monitoring of natural recreational water microbial quality is most often based mainly on testing a set of microbiological indicators. The cost and labour involved in testing numerous water samples may be significant when a large number of sites must be monitored repetitively over time. In addition to water testing, ongoing monitoring of key environmental factors known to influence microbial contamination may also be carried out as an additional component. Monitoring of environmental factors can now be performed using remote sensing technology which represents an
increasingly recognized source of rigorous and recurrent data, especially when monitoring over a large or difficult to access territory is needed. To determine
whether this technology could be useful in the context of recreational water monitoring, we evaluated a set of agroenvironmental determinants associated with fecal contamination of recreational waters through a multivariable logistic regression model built with data extracted from satellite imagery. We found that variables describing the proportions of land being agricultural and impervious surfaces, as derived from remote sensing observations, were statistically associated (OR=11 and 5.2, respectively) with a higher level of fecal coliforms in lake waters in the
southwestern region of Quebec, Canada. From a technical perspective, remote sensing may provide important added-value in the monitoring of microbial risk from recreational waters and further applications of this technology should be investigated to support public health risk assessments and environmental monitoring programs relating to water quality.
Keywords: recreational waters, remote-sensing, public health, fecal contamination, agroenvironment.
Introduction
In the last decade, the popularity of natural recreational water activities has grown in many countries and is expected to further increase with demographic changes and global warming (Pond 2005). Recreational water quality can be impaired by different types of contamination including fecal contamination, algae, chemical contamination and contamination with other free-living organisms like amibes and Leptospira sp. (WHO 2003). People exposed to recreational water impacted by fecal contamination can develop health problems, especially gastro-intestinal diseases, which can lead to severe sequelae (Craun et al. 2004). Pathogens found in these waters can come from different sources, including agricultural activities, human activities and wild animal populations. Currently, monitoring of recreational water quality is mainly based on the detection of microbiological fecal indicators, which can provide reliable
information on water quality relatively quickly (WHO 2003). This information is crucial in order to take necessary public health actions such as closing beaches when needed. In addition to direct testing of water, the World Health Organization also recommends the assessment and monitoring of all the sources of fecal
contamination and environmental characteristics that may influence beach water quality, like land use and topography. Combination of these two monitoring
components provides a basis for a beach classification which can be used to assist in the identification of effective management intervention (WHO 2003). In addition to a better characterization of the hazard, monitoring of proximal environmental factors also provide insights into the possible sources on which contamination may be attributed. There are various known data that can be used to describe
environmental factors, including census and field records. However, in the context of a large territory like the Province of Quebec in Canada, where there are hundreds of natural recreational bodies of water and a small population density, there is a need for new techniques to characterize the environment with greater efficiency and consistency.
Tele-epidemiology consists of monitoring and studying the distribution of human and animal diseases strongly linked to climatic and environmental variations through the application of space-based technologies, including remote-sensing. (Marechal et al. 2008). Remote sensing data have been previously used for the surveillance of a number of public health problems, including vector-borne diseases and algal bloom detection (Boone et al. 2000; Kutser et al. 2006; Herbreteau et al. 2007; Leblond et al. 2007). This type of data could also be useful in the surveillance of fecal contamination of recreational waters, but scientific data on the application of such technology in this domain is limited. Agricultural lands, including croplands and pastures, have been known to influence water quality downstream since they can be a source of fecal microorganisms originating from spread manure or grazing animals (Rodgers et al. 2003; Ramos et al. 2006). Urban areas have also been linked with fecal water pollution through wastewater treatment plant discharge, urban runoff and direct spillage of impacted water (Basnyat et al. 2000; Marsalek and
Rochfort 2004). Conversely, some land covers like forests and wetlands have been associated with better water quality as they can act as a filter of microorganisms (Kao and Wu 2001; Matteo et al. 2006); Matteo et al. 2006; (Knox et al. 2008). Since remote sensing data can effectively provide information on land use and land cover known to influence recreational water quality, this information could be a useful addition to recreational water surveillance programs focusing on environmental determinants associated with the risk of fecal contamination (Beck et al. 2000; Tong and chen 2002; Campbell 2007). This paper had two objectives. The first was to evaluate the use of remote sensing, and more specifically Landsat-5 satellite imagery, in identifying agroenvironmental determinants associated with fecal contamination of recreational freshwaters, and the second was to discuss the strengths and the limitations of this approach in the context of recreational water quality monitoring programs.