Project uses geospatial, time of day, and weather data as part of E. coli risk model

"Can we predict when surface water, such as streams or canals, is more likely to carry pathogens and exceed certain guideline levels for E. coli? That is our goal." said project investigator Martin Wiedmann, Gellert Family Professor in Food Safety at Cornell University's Department of Food Science in Ithaca, New York (picture). His co-investigator is Channah Rock, University of Arizona water quality specialist and associate professor in the department of Soil, Water and Environmental Science housed at the Maricopa Agricultural Center, Arizona. Wiedmann said he wanted to work with Rock after learning of her research and hearing her presentations at past Center for Produce Safety Symposia. "I think it really illustrates the power of CPS," he said. Their research project, titled "Remotely sensed and field-collected hydrological, landscape and weather data can predict the quality of surface water used for produce production",- is designed to provide growers with a model to identify times and locations that may carry a higher potential risk of microbial contamination. Using those risk assessment results, Wiedmann said growers could then take a more targeted approach to water sampling, water treatment, and water use. "If it's rained within the last 24 hours and you're within 5 miles of an old housing development, you might want to consider treating or another water source," he said. The project's eventual goal is to develop a Web-based model where producers in regions where the researchers have developed strong data could assess their risks under certain conditions. One year into the two-year project, Wiedmann and his colleagues have examined remote sensing imagery for a chosen watershed within Upstate New York. The images, taken by satellites orbiting the Earth, have become readily available from different agencies in recent years. The researchers were mostly interested in land uses upstream from produce production fields. "If you have a stream where you have houses that are not on a municipal septic system and they're older than 25 or 30 years, it might make a risk factor from the septic systems more likely," he said. The analysis also looked at upstream agricultural uses that involved livestock, including pastures. Wiedmann and his team collected water samples from various locations within the watershed and analyzed them using two different sampling methods. They wanted to determine whether there were variations in results between the two methods. They also sought to correlate geography to levels of generic E. coli and presence of three pathogens: Shiga-toxin producing E. coli (STEC), Salmonella and Listeria monocytogenes. "We're particularly going to look at E. coli numbers and how they correlate to presence of pathogens," he said. "We're going to a lot of different sites so we can predict across different areas." Part of the project also is examining whether time of day or weather events play a role in pathogen levels. As part of data collection, the researchers pulled samples during the early morning, mid-morning, afternoon and evening; during different days of the week; and before and after rain storms. Rock and her team have been collecting water samples using the same methods and parameters as the Cornell University colleagues, and have analyzed more than 640 samples to date. Because of the Arizona vegetable production calendar, growers in the Southwest desert plant in early fall for harvest during late winter and spring. Upstate New York growers, on the other hand, plant in the spring and harvest during the summer. The Arizona team is currently launching into their next sampling campaign to complete the year. Additionally, Rock will focus on surface water supplies, including canals, as Arizona producers typically don't rely on streams or rivers because of the arid climate and proximity to water sources. Wiedmann said having data from two different production regions should yield a more robust model that should readily translate to other areas. More informationRemotely sensed and field-collected hydrological, landscape and weather data can predict the quality of surface water used for produce production? ?