Cyberbiosecurity: A New Perspective on Protecting U.S. Food and Agricultural System.

Our national data and infrastructure security issues affecting the "bioeconomy" are evolving rapidly. Simultaneously, the conversation about cyber security of the U.S. food and agricultural system (cyber biosecurity) is incomplete and disjointed. The food and agricultural production sectors influence over 20% of the nation's economy ($6.7T) and 15% of U.S. employment (43.3M jobs). The food and agricultural sectors are immensely diverse and they require advanced technologies and efficiencies that rely on computer technologies, big data, cloud-based data storage, and internet accessibility. There is a critical need to safeguard the cyber biosecurity of our bio economy, but currently protections are minimal and do not broadly exist across the food and agricultural system. Using the food safety management Hazard Analysis Critical Control Point system concept as an introductory point of reference, we identify important features in broad food and agricultural production and food systems: dairy, food animals, row crops, fruits and vegetables, and environmental resources (water). This analysis explores the relevant concepts of cyber biosecurity from food production to the end product user (such as the consumer) and considers the integration of diverse transportation, supplier, and retailer networks. We describe common challenges and unique barriers across these systems and recommend solutions to advance the role of cyber biosecurity in the food and agricultural sectors.

A comparison of nutrient losses from two simulated pastureland management scenarios.

Fecal deposits by grazing animals on pasturelands have the potential to leach nutrients to runoff during rainfall events. Unlike croplands, grazing systems such as pasturelands or rangelands have little opportunity to ameliorate nutrient runoff through in-field or edge-of-field management practices. Thus, we investigated the amounts and concentrations of nutrients in overland flow from simulated grazing lands. Two grazing management scenarios were simulated: continuous grazing represented by two sparsely vegetated (SV) plots and rotational grazing represented by two densely vegetated (DV) plots. In addition, there were two control plots. The plots were treated with standard cowpats and rainfall was simulated until overland flow at the edge of the plots reached steady-state. Higher runoff was observed from DV plots (9.97 mm) than SV plots (7.05 mm), but the average total suspended solids concentration in runoff from SV plots was approximately 17 times the concentration observed in runoff from the DV plots. The average total phosphorus (TP) concentrations were highest in plots simulating continuous grazing (5.91 mg L(-1)). In both DV and SV plots at least 83% of the TP was found to be in the dissolved form. The average total Kjeldhal nitrogen (TKN) and total nitrogen concentrations observed in runoff samples from SV plots were 1.25 and 1.46 mg L(-1), respectively. Organic nitrogen comprised 95% of the TKN observed in runoff samples from SV plots. The SV plots have relatively higher loads for those nutrients in the particle associated form compared to DV plots, whereas DV plots had higher loads for those nutrients in the dissolved form. Grazing lands without any additional manure applications were found to release nutrients in high levels and vegetation did not show any effect on removing dissolved nutrients from runoff. These results are useful to inform selection of appropriate management practices to reduce nutrient transport to surface waters in watersheds dominated by grazed lands.

Attachment of Escherichia coli and enterococci to particles in runoff.

Association of Escherichia coli and enterococci with particulates present in runoff from erodible soils has important implications for modeling the fate and transport of bacteria from agricultural sources and in the selection of management practices to reduce bacterial movement to surface waters. Three soils with different textures were collected from the Ap horizon (silty loam, silty clay loam, and loamy fine sand), placed in portable box plots, treated with standard cowpats, and placed under a rainfall simulator. Rainfall was applied to the plots until saturation-excess flow occurred for 30 min, and samples were collected 10, 20, and 30 min after initiation of the runoff event. The attachment of E. coli and enterococci to particles present in runoff was determined by a screen filtration and centrifugation procedure. Percentage of E. coli and enterococci attached to particulates in runoff ranged from 28 to 49%, with few statistically significant differences in attachment among the three soils. Similar partitioning release patterns were observed between E. coli and enterococci from the silty loam (r = 0.57) and silty clay loam soils (r = 0.60). At least 60% of all attached E. coli and enterococci were associated particles within an 8- to 62-microm particle size category. The results indicate that the majority of fecal bacteria attach to and are transported with manure colloids in sediment-laden flow regardless of the soil texture.

Evaluation of a filtration/dispersion method for enumeration of particle-associated Escherichia coli.

The transport processes of microorganisms in storm-generated flows have not been clearly elucidated, limiting the ability of computational models to effectively design and evaluate watershed remediation plans. Although several studies have identified association with particulates as a critical factor in predicting microbial transport and fate, no generally accepted method exists for the identification of the planktonic and particle-associated fractions of microorganisms in water samples. In this study, a filtration/dispersion method proposed for quantification of these fractions was verified using laboratory-composed samples of bovine Escherichia coli and sterile topsoil. A statistical experimental design allowed quantification of potential sources of experimental error (e.g., filter retention, die-off), although no significant sources of methodological error were identified. On average, 78% of E. coli cells were particle associated after 1 h of contact time. Further application of the method was illustrated by an isotherm experiment examining the association of a bovine strain of E. coli with sterile topsoil. Further examination of this method in laboratory or field-based studies of microbial partitioning between the planktonic and particulate phases in surface runoff appears justified.

Comparison of HSPF and SWAT models performance for runoff and sediment yield prediction.

A watershed model can be used to better understand the relationship between land use activities and hydrologic/water quality processes that occur within a watershed. The physically based, distributed parameter model (SWAT) and a conceptual, lumped parameter model (HSPF), were selected and their performance were compared in simulating runoff and sediment yields from the Polecat Creek watershed in Virginia, which is 12,048 ha in size. A monitoring project was conducted in Polecat Creek watershed during the period of October 1994 to June 2000. The observed data (stream flow and sediment yield) from the monitoring project was used in the calibration/validations of the models. The period of September 1996 to June 2000 was used for the calibration and October 1994 to December 1995 was used for the validation of the models. The outputs from the models were compared to the observed data at several sub-watershed outlets and at the watershed outlet of the Polecat Creek watershed. The results indicated that both models were generally able to simulate stream flow and sediment yields well during both the calibration/validation periods. For annual and monthly loads, HSPF simulated hydrologic and sediment yield more accurately than SWAT at all monitoring sites within the watershed. The results of this study indicate that both the SWAT and HSPF watershed models performed sufficiently well in the simulation of stream flow and sediment yield with HSPF performing moderately better than SWAT for simulation time-steps greater than a month.

Development of bacteria and benthic total maximum daily loads: a case study, Linville Creek, Virginia.

Two total maximum daily load (TMDL) studies were performed for Linville Creek in Rockingham County, Virginia, to address bacterial and benthic impairments. The TMDL program is an integrated watershed management approach required by the Clean Water Act. This paper describes the procedures used by the Center for TMDL and Watershed Studies at Virginia Tech to develop the Linville Creek TMDLs and discusses the key lessons learned from and the ramifications of the procedures used in these and other similar TMDL studies. The bacterial impairment TMDL was developed using the Hydrological Simulation Program-Fortran (HSPF). Fecal coliform loads were estimated through an intensive source characterization process. The benthic impairment TMDL was developed using the Generalized Watershed Loading Function (GWLF) model and the reference watershed approach. The bacterial TMDL allocation scenario requires a 100% reduction in cattle manure direct-deposits to the stream, a 96% reduction in nonpoint-source loadings to the land surface, and a 95% reduction in wildlife direct-deposits to the stream. Sediment was identified as the primary benthic stressor. The TMDL allocation scenario for the benthic impairment requires an overall reduction of 12.3% of the existing sediment loads. Despite the many drawbacks associated with using watershed-scale models like HSPF and GWLF to develop TMDLs, the detailed watershed and pollutant-source characterization required to use these and similar models creates information that stakeholders need to select appropriate corrective measures to address the cause of the water quality impairment when implementing the TMDL.

Variation in root density along stream banks.

While it is recognized that vegetation plays a significant role in stream bank stabilization, the effects are not fully quantified. The study goal was to determine the type and density of vegetation that provides the greatest protection against stream bank erosion by determining the density of roots in stream banks. To quantify the density of roots along alluvial stream banks, 25 field sites in the Appalachian Mountains were sampled. The riparian buffers varied from short turfgrass to mature riparian forests, representing a range of vegetation types. Root length density (RLD) with depth and aboveground vegetation density were measured. The sites were divided into forested and herbaceous groups and differences in root density were evaluated. At the herbaceous sites, very fine roots (diameter < 0.5 mm) were most common and more than 75% of all roots were concentrated in the upper 30 cm of the stream bank. Under forested vegetation, fine roots (0.5 mm < diameter < 2.0 mm) were more common throughout the bank profile, with 55% of all roots in the top 30 cm. In the top 30 cm of the bank, herbaceous sites had significantly greater overall RLD than forested sites (alpha = 0.01). While there were no significant differences in total RLD below 30 cm, forested sites had significantly greater concentrations of fine roots, as compared with herbaceous sites (alpha = 0.01). As research has shown that erosion resistance has a direct relationship with fine root density, forested vegetation may provide better protection against stream bank erosion.

Simulating fecal coliform bacteria loading from an urbanizing watershed.

The fate and transport of fecal coliform bacteria in the urbanizing Polecat Creek watershed, located in Virginia, was simulated using the Hydrological Simulation Program-FORTRAN (HSPF). Both point and nonpoint sources of fecal coliform were included in the simulation. Hydrologic and water quality parameters of HSPF were calibrated and validated using observed data collected from October 1994 to June 2000 at three monitoring stations. The percent errors in total runoff volumes between observed and simulated values ranged from 0.4 to 4.2% for the calibration period, and 0.4 to 6.7% for the validation period. The geometric mean of simulated fecal coliform concentrations at the outlet of the watershed was 10% lower than that of observed values for the calibration period. HSPF moderately under-predicted the geometric mean concentration by 16.4% for one sub-watershed and slightly over-predicted by 7.3% for another. Observed fecal coliform concentrations were compared with the range defined by the minimum and maximum simulated concentrations occurring within a 3-day window centered on the day the water sample was collected. Over 42% of grab sample data collected at the three monitoring sites in the watershed fell within the max min range of simulated concentrations over the 3-days window for the calibration period. For all monitoring sites, 39.5% of the total samples taken during the validation period fell in the range of simulated concentrations over the 3-day window period. Results presented in this study demonstrate that HSPF reasonably represents the hydrology and water quality of an urbanizin watershed and that it could be utilized as a planning tool for future assessment of land use impacts on fecal coliform on in-stream concentrations.