Research Project: Design and Implementation of Monitoring and Modeling Methods to Evaluate Microbial Quality of Surface Water Sources Used for Irrigation

Location: Environmental Microbial & Food Safety Laboratory

2020 Annual Report

Objectives
Objective 1 - Elucidate spatial variability of indicator bacteria concentrations in surface waters (e.g., streams, ponds, reservoirs), and describe factors responsible for this variability. Sub-objective 1.A. Research and quantify lateral patterns of indicator bacteria concentrations in ponds and reservoirs, and evaluate the effect of algal populations, flow patterns and water quality parameters on these patterns. Sub-objective 1.B. Research and quantify patterns of vertical indicator bacteria distributions in water column in ponds and reservoirs. Sub-objective 1.C. Develop a model to estimate indicator bacteria concentrations at the intake of irrigation water based on vertical and lateral indicator bacteria distributions in the water of pond or reservoir. Objective 2 - Elucidate temporal variability of indicator bacteria concentrations in watersheds as a function of land use and meteorological conditions, and develop/validate predictive models. Sub-objective 2.A. Develop a model to evaluate stream bottom sediment as an indicator bacteria source between rainfall events. Sub-objective 2.B. Research survival of manure-borne indicator bacteria in soil to predict contribution of soil E. coli reservoir to runoff leaving fields and pastures. Sub-objective 2.C. Develop a modeling-based method for site-specific optimization of stream water sampling scheduling to provide the most representative indicator bacteria concentrations in irrigation water for a given annual number of samples.

Approach
Taken as a whole, this project strives to acquire, package and disseminate the knowledge about microbial quality of irrigation water in the way that offers wide applicability of results. No resources can currently be made available to monitor a large enough number of sites across the country to build a reliable statistical model that would relate microbial water quality to a multitude of environmental variables that vary based on prevailing local conditions at specific sites. This project relies on mechanistic rather than statistical models. It is designed on the premise that processes affecting microbial water quality stay the same whereas rates of those processes vary as they reflect local conditions. The project will develop observation methods that will improve data collection to fine-tune the model to a specific site by finding the site-specific rates. Models will be tested to make sure that simulation results are quantitatively and qualitatively similar to results of measurements. Data for such testing will be collected at field sites that reflect represent major contrasting combinations of environmental and management factors affecting water quality in irrigation water sources. The satisfactory performance of the models will provide confidence that the models and the corresponding data collection will be applicable at sites other than observed. As a disclaimer, it is realized that the current knowledge about microbial water quality controls still is far from being exhaustive, and some sites may exhibit microbial water quality features that are not understood and modeled well. The project is designed to efficiently utilize the best current knowledge about the processes controlling the microbial water quality of surface water. The integrated monitoring and modeling approach of this project can be re-applied as new knowledge will become available about the processes and factors controlling the microbial quality of surface water used for irrigation.

Progress Report
1. For Sub-Objective 1A, progress was made in observing and quantifying the diurnal changes of spatial patterns of E. coli concentrations in irrigation ponds in Maryland. Substantial decrease of concentrations during the morning hours was observed that indicated the need of tailoring sampling times to the time of water use for irrigation. Water quality parameters seemed to affect the spatial patterns change, but more data are needed to quantify this effect. 2. For Sub-Objective 1A, progress was made in the analysis of the spatiotemporal patterns of the pigment phycocyanin which is used to indicate cyanobacteria presence in waters. Toxins from these harmful cyanobacteria blooms can be carried from irrigation waters to nearby crop fields, and ultimately fresh produce. We used data from intensive monitoring of two irrigation ponds in Maryland combined with artificial intelligence-based -data processing and determined that the most influential predictors of phycocyanin are chlorophyll A, colored dissolved organic matter and turbidity; these parameters can be measured with existing sensors in situ and estimated with remote sensing methods. That opens the opportunity of the expedited characterization of cyanobacteria populations in irrigation ponds. 3. For Sub-Objective 1A, progress was made in the research of the temporal stability of water quality parameters in irrigation ponds. Three major methods of characterization of finding spatial patterns persisting in time were applied to the three years of monitoring performed at irrigation ponds in Maryland. We observe several types of temporal stability and preliminary conclude that the temporal stability of water quality parameters can be used to guide monitoring. 4. For Sub-Objective 1A, progress has been made in understanding the effect of the available controls of the artificial intelligence methods on their performance in recognizing the spatial s=distribution of E, coli in irrigation ponds from drone-based video imagery. We found that the recognitions of E. coli spatial patterns can be substantially improved if controls will be changed from defaults. The consequences of this change for the statistics of the recognition efficiency needs to be investigated. 5. For Sub-objective 2A, data analysis methods for the hyperspectral imagery were tested with images from freshwater ponds and reservoirs. Advanced artificial intelligence algorithms showed promise with data obtained in South Korea and in Maryland’ Preliminary results were submitted to journals; the work is in progress. 6. For Sub-objective 1.B, the unique dataset was collected biweekly that included concentrations of E. coli, major phytoplankton species, and twelve water quality parameters at three depths in 10 locations across the irrigation ponds coupling; the data analysis is pending. 7. For Sub-objective 2B, the reanalysis was performed with the data on the effect of the manure consistency and weathering on indicator bacteria removal in the runoff. We found that the percentage of manure-borne E. coli and enterococci removed in the runoff in one hour was exponentially dependent on the weathering duration, and did not depend on the manure consistency. These results can be useful for professionals involved in developing guidance on manure application practices to prevent or mitigate microbiological impairment of irrigation and recreation water sources. 8. For Sub-objectives 1.A, 1.C, 2.A, and 2C, substantial conceptual contributions were made to the collaborative projects on modeling Low Impact Development (LID) in Urban watersheds.UNIST, Korea, prediction of infiltration parameters of soils (Cankiri Karatekin University, Turkey, and Environmental Microbial and Food Safety Laboratory), recognition of patterns in yield maps (Hydrology and Remote Sensing Laboratory), changes in soil hydrology after amendment applications (Alicante University, Spain), and temporal patterns of Listeria monocytogenes in water and sediment of the Conococheague creek, Pennsylvania (Food and Drug Administration and Wilson College).

Accomplishments
1. Estimating microbial water quality. Understanding the microbial quality of irrigation water sources can prevent human consumption of microbes that can cause disease. The microbial quality of irrigation water is based on the concentrations of the indicator E. coli. ARS scientists from Beltsville, Maryland, tested the method of using the water quality sensors from a boat and using artificial intelligence to estimate E. coli concentrations across irrigation ponds. The results of this work provide the knowledge base for better monitoring of microbial water quality to improve food safety.

Review Publications
Stocker, M., Pachepsky, Y.A., Hill, R., Sellner, K., Staver, K. 2019. Intra-seasonal variation of E. coli and environmental covariates in two irrigation ponds in Maryland, USA. Science of the Total Environment. 670:732-740.
Park, S., Baek, S., Chun, J., Park, Y., Pachepsky, Y.A., Park, J., Cho, K. 2019. Deep learning model for membrane biofouling prediction using optical coherence tomography imaging. Journal Membrane Science. 587:117164. https://doi.org/10.1016/j.memsci.2019.06.004.
Garcia-Gutierrez, C., Pachepsky, Y.A. 2019. On the information content of coarse data with respect to the particle size distribution of complex granular media: rationale approach and testing. Entropy Journal. 21(6):601. https://doi.org/10.3390/e21060601.
Smith, J.E., Kiefer, L.A., Stocker, M., Blaustein, R., Ingram, S., Pachepsky, Y.A. 2019. Depth-dependent response of fecal indicator bacteria in sediments to changes in water column nutrient levels. Environmental Quality. 48:1074–1081.
Smith, J.E., Stocker, M.D., Pachepsky, Y.A. 2019. The effect of temperature oscillations and sediment texture on fecal indicator bacteria survival in sediments. Water, Air, and Soil Pollution. 230(11):270. https://doi.org/10.1007/s11270-019-4278-7.
Stocker, M.D., Smith, J.E., Hernandez, C., Macarisin, D., Pachepsky, Y.A. 2019. Seasonality of E. coli and enterococci concentrations in creek water, sediment, and periphyton. Water, Air, and Soil Pollution. 230(9):223. https://doi.org/10.1007/s11270-019-4263-1.
Baek, S., Pachepsky, Y.A., Chun, J., Yoon, K., Park, Y., Cho, K. 2020. Assessment of a low impact development (LID) practice using the coupled SWMM and HYDRUS models. Journal of Environmental Modeling and Software. 261:109920. https://doi.org/10.1016/j.jenvman.2019.109920.
Morgan, B.J., Stocker, M.D., Valdes-Avellan, J., Kim, M.S., Pachepsky, Y.A. 2019. Drone-based imaging to assess the microbial water quality in an irrigation pond: a pilot study. Science of the Total Environment. 716:135757. https://doi.org/10.1016/j.scitotenv.2019.135757.
Pachepsky, Y.A., Kierzewski, R.A., Stocker, M., Mulbry Iii, W.W., Sellner, K. 2018. Temporal stability of E. coli concentrations in waters of two irrigation ponds in Maryland. Applied and Environmental Microbiology. 84:e01876-17. https://doi.org/10.1007/s11356-019-07030-9.
Jeon, D., Pachepsky, Y.A., Harriger, D., Picard, R., Coppock, C.R. 2019. Effect of the time scale on the uncertainty of geometric mean concentrations of fecal indicators in creek water under baseflow conditions. Environmental Research Letters. 10:1-5. https://doi.org/10.1038/s41598-020-58603-5.
Pyo, J., Pachepsky, Y.A., Kim, M., Baek, S., Lee, H., Cha, Y., Park, Y., Cho, K. 2017. Simulating seasonal variability of phytoplankton in stream water using the modified SWAT model. Environmental Modelling & Software. 122:104073. https://doi.org/10.1016/j.envsoft.2017.11.005.
Baek, S., Ligaray, M., Pyo, J., Park, J., Kang, J., Pachepsky, Y.A., Chun, J., Cho, K. 2020. A novel water quality module of the SWMM model for assessing Low Impact Development (LID) in Urban watersheds. Journal of Hydrology. 586:124886. https://doi.org/10.1016/j.jhydrol.2020.124886.