College of Agriculture & Natural Resources
Permanent URI for this communityhttp://hdl.handle.net/1903/1598
The collections in this community comprise faculty research works, as well as graduate theses and dissertations.
Browse
3 results
Search Results
Item A Machine Learning Model for Food Source Attribution of Listeria monocytogenes(MDPI, 2022-06-16) Tanui, Collins K.; Benefo, Edmund O.; Karanth, Shraddha; Pradhan, Abani K.Despite its low morbidity, listeriosis has a high mortality rate due to the severity of its clinical manifestations. The source of human listeriosis is often unclear. In this study, we investigate the ability of machine learning to predict the food source from which clinical Listeria monocytogenes isolates originated. Four machine learning classification algorithms were trained on core genome multilocus sequence typing data of 1212 L. monocytogenes isolates from various food sources. The average accuracies of random forest, support vector machine radial kernel, stochastic gradient boosting, and logit boost were found to be 0.72, 0.61, 0.7, and 0.73, respectively. Logit boost showed the best performance and was used in model testing on 154 L. monocytogenes clinical isolates. The model attributed 17.5 % of human clinical cases to dairy, 32.5% to fruits, 14.3% to leafy greens, 9.7% to meat, 4.6% to poultry, and 18.8% to vegetables. The final model also provided us with genetic features that were predictive of specific sources. Thus, this combination of genomic data and machine learning-based models can greatly enhance our ability to track L. monocytogenes from different food sources.Item Monitoring and Predicting the Microbial Water Quality in Irrigation Ponds(2022) Stocker, Matthew Daniel; Hill, Robert L; Environmental Science and Technology; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Small- to medium-sized farm ponds are a popular source of irrigation water and provide a substantial volume of water for crop growth in the United States. The microbial quality of irrigation waters is assessed by measuring concentrations of the fecal indicator bacteria Escherichia coli (E. coli). Minimal guidance currently exists on the use of surface irrigation waters to minimize consumer health risks. The overall objective of this work was to provide science-based guidance for microbial water quality monitoring of irrigation ponds. Spatial and temporal patterns of E. coli were evaluated in two Maryland irrigation ponds over three years of observations. Patterns of E. coli were stable over the three years and found to be significantly correlated to patterns of water parameters such as temperature, dissolved oxygen, turbidity, and pH. The EPA Environmental Fluid Dynamics Code model was used to evaluate the spatial 3D heterogeneity of E. coli concentrations within the ponds. Significant differences in E. coli concentrations by sampling depth were found. Spatial heterogeneity of E. coli within the pond also resulted in substantial temporal variation at the irrigation pump, which was dependent on the intake location. Diurnal variation of E. coli concentrations was assessed for three farm ponds. E. coli concentrations declined from 9:00 to 15:00 for each pond, but statistically significant declines were only observed in two of the three ponds. Dissolved oxygen, pH, and electrical conductance were found to be the most influential environmental variables affecting E. coli concentrations. To better describe the relationships between E. coli and the environmental variables, four machine learning algorithms were used to estimate E. coli concentrations using water quality parameters as predictors. The random forest algorithm provided the highest predictive accuracy with R2 = 0.750 and R2 = 0.745 for Ponds 1 and 2, respectively, in the multi-year dataset containing 12 predictors. Temperature, electrical conductance, and organic matter content were identified as the most influential predictors. It is anticipated that the recommendations contained in this dissertation will be used to improve microbial monitoring strategies and protect public health.Item Novel Applications in Wetland Soils Mapping on the Delmarva Coastal Plain(2018) Goldman, Margaret Anne; Needelman, Brian A; Environmental Science and Technology; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)On the Delmarva Peninsula, depressional wetlands provide a range of ecosystem services, including water purification, groundwater recharge, provision of critical habitat, and carbon storage. Concern for the health of the Chesapeake Bay and the establishment of the Bay Total Maximum Daily Load have led to growing interest in restoring depressional and other wetland types to mitigate agricultural nitrogen inputs. The ability of natural resource managers to implement wetland restoration to address nonpoint source pollution is constrained by limited spatial information on hydrogeologic and soil conditions favoring nitrogen removal. The goal of this study was to explore the potential of new digital soil mapping techniques to improve identification of wetland soils and map soil properties to improve assessment of wetland ecosystem services, including removing excess nitrogen, and inform natural resource decision making. Previous research on digital soil mapping has focused largely on the development of medium to low-resolution general purpose soil maps in areas of heterogeneous topography and geomorphology. This study was unique in its focus on mapping wetland soils to support wetland restoration decisions in a low relief landscape. A digital soil mapping approach involving the spatial disaggregation of soil data map units was used to create maps of natural soil drainage and texture class. The study was conducted in the upper part of the Choptank River Watershed on central Delmarva, where depressional wetlands occur in high densities and historical loss of wetlands is estimated to be high compared to similar Maryland watersheds. The soil disaggregation techniques developed in this study were successful in creating a more refined representation of natural soil drainage and texture class in forested depressional wetlands. Comparison of the disaggregated soils map with recently developed time-series inundation maps of the region demonstrate the need for further research to understand how indicators of historic and current hydrologic conditions can guide operational soils and wetland mapping and inform wetland restoration decisions.