In order to obtain supporting evidence for biological denitrification in anaerobic lagoons degenerate ?Ò-Proteobacterial AOB primers were used to create and sequence clone libraries to detect the presence of ammonia-oxidizing bacteria at three field sites. Although there were PCR products from almost all samples, the clone libraries that were created show that not all PCR reactions produce only PCR products from ammonia-oxidizing bacteria. However, these primers did verify the presence of ammonia-oxidizers at one site, although their presence was not verified at the other sites. The presence of ammonia-oxidizers at the Battelle site implies that aerobic ammonia-oxidation is occurring. Clones were created and sequenced that were significantly different from other known sequences and tended to form very closely related phylogenetic groups. These phylogenetic groups were not isolated to one field site, and often more than one site had representatives in a closely related group. Future research in this field includes the design of new primer sets based on the sequences of the nitrifying bacteria clones reported in this research, creation of enrichment cultures, and use of new primers for fluorescent in situ hybridization.

GPS and stormwater models are two extremely powerful technologies that can effectively predict the movement of rainfall runoff and soluble pollutants via surface water, when applicable correctly. Using the Cherry Point Marine Corps Air Station, North Carolina, topographic information, historical and observed rainfall datasets, streamflow measurements, subsurface conduits attributes and GPS acquired data was inserted into XPSWMM to model the downstream movement of user-defined tracer elements. The overall objective of the research was to develop a hydrologic/hydraulic model to predict pollutant movement from a spill site to subcatchment outlets on the Marine Corps Air Station at Cherry Point. Other related research objectives were to: 1) to use spatial information gathered from the GIS to construction drainage areas in efforts to estimate catchment characteristics, 2) to evaluate the results of peak outflow rates gathered from several event-based hydrologic models and to explain the evolution from lumped parameter models to process-based, rainfall-runoff simulations and 3) to generate continuous simulations for rainfall-runoff processes using a calibrated/validated version of XPSWMM and 4) to introduce the concept of using pulse tracers to estimate travel times via surface water to understand associated reaction times.

Ammonia is a very important atmospheric pollutant. Agricultural activities, livestock production in particular, have been reported to be the largest contributor of ammonia emissions into the atmosphere. Accurate estimation of ammonia emission rate from individual operations or sources is important and yet a challenging task for both regulatory agencies and animal producers. The overall research objective of this study was to develop an emission model which can be used to estimate ammonia emission from broiler litter. In the reported model, the ammonia flux is essentially a function of the litter's total ammoniacal nitrogen (TAN) content, moisture content, pH, and temperature, as well as the Freundlich partition coefficient (Kf), mass transfer coefficient (KG), ventilation rate (Q), and emission surface area (A). A dynamic flow-through chamber system and a wind tunnel were designed to measure ammonia fluxes from broiler litter. The dynamic flow-through chamber experiments evaluated the reported model with various litter samples under a constant temperature and wind profile. The wind tunnel experiments evaluated the reported model under various temperatures and wind profiles. Model parameters such as Kf and KG were estimated. The results from the two experiments were consistent with each other. The estimated KG ranged from 1.11 to 27.64 m h-1, and the estimated Kf ranged from 0.56 to 4.48 L kg-1. Regression sub-models were developed to estimate Kf as a function of litter pH and temperature and to estimate KG as a function of air velocity and temperature. Sensitivity analysis of the model showed that ammonia flux is very sensitive to litter pH and to a lesser extent temperature. A validation metric based on the mean and covariance in the measurement and in the model parameters were used to validate the ammonia emission model in the presence of measurement and model parameter uncertainties.

<p>Ventricular Fibrillation (VF) is a severe cardiac arrhythmia that canresult in sudden death, a leading cause of death in the United States. DuringVF, the electrical activity of the heart becomes disordered, the ventriclescontract erratically, and an insufficient supply of blood is pumped to the body.Identification of the electrical activation sites during VF is important for theunderstanding and improved treatment of the disorder. Unipolar electrogramsof four pigs were recorded following the induction of VF. The data from theVF recordings was preprocessed using a Rule-Based Method (RBM) , aCurrent Source Density (CSD) method, and a Transmembrane CurrentMethod (TCM) to separate local activations from distant activity. RBM usesthe magnitude of the derivative of the voltage to identify activations. CSD is ascalar quantity that represents the magnitude of the current source or sink.TCM estimates a value proportional to the transmembrane current. Afeedforward artificial neural network (ANN) using backpropagation wastrained to identify the local activations in the electrograms of VF based on theRBM and CSD calculations. Another feedforward ANN usingbackpropagation was trained using data preprocessed with not only RBM andCSD, but also TCM. In order to improve the ability of the ANNs to detect localactivations, a new training method, called staged training, was utilized. Instaged training, the ANNs were trained in stages using different sets oftraining examples. Examples were included in a particular training set basedon the minimum magnitude of the voltage derivative. When training was donein stages, both the ANN with RBM and CSD data only and the ANN with RBM,CSD, and TCM data were able to more accurately distinguish activations.Overall, the ANN, which used only RBM and CSD data, produced betterresults than the ANN that also included TCM data. <P>

<p>Eighteen lots of shelled corn were tested for aflatoxin contamination. The variability and distributional characteristics associated with the aflatoxin testing procedure were investigated. The total variance associated with testing shelled corn was estimated and partitioned into sampling, sample preparation, and analytical variances. All variances were found to increase with an increase in aflatoxin concentration. Using regression analysis, mathematical expressions were developed to model the relationship between aflatoxin concentration and the total, sampling, sample preparation, and analytical variances. The expressions for these relationships were used to estimate the variance for any sample size, subsample size, and number of analyses for a specific aflatoxin concentration. For example, testing a lot with 20 parts per billion (ppb) aflatoxin using a 2.5 lb sample, Romer mill and 50 g subsample, and HPLC analysis, the total, sampling, sample preparation, and analytical variances are 274.9 (CV=82.9%), 214.0 (CV=73.1%), 56.3 (CV=37.5%), and 4.6 (CV=10.7%), respectively. The percentage of the total variance for sampling, sample preparation, and analytical is 77.8, 20.5, and 1.7 %, respectively. Next, fifteen positively skewed distributions were each fitted to 18 empirical distributions of aflatoxin test results for shelled corn. The compound gamma distribution was selected to model the sample aflatoxin test results for shelled corn. The method of moments technique was chosen to estimate the parameters of the compound gamma distribution. Mathematical expressions were developed to calculate the parameters of the compound gamma distribution for any lot aflatoxin concentration and test procedure. Observed acceptance probabilities were compared to operating characteristic curves predicted from the compound gamma distribution and all 18 distributions of sample aflatoxin test results were found to lie within a 95% confidence band. Using the mean and variance relationships to compute the parameters of the compound gamma distribution, 16 sampling plans, based on four sample sizes and four sample acceptance levels were created and analyzed. For a given sample size, decreasing the sample acceptance level, using a sample acceptance level equal to the regulatory guideline: (a) decreases the percentage of lots accepted while increasing the percentage of lots rejected at all aflatoxin concentrations; (b) increases misclassification of lots (both false positives and false negatives) while decreasing the percentage of correct decisions; and (c) decreases the average aflatoxin concentration in the lots accepted and lots rejected. For a given sample size where the sample acceptance level is less than the regulatory guideline, the number of false positives increases and the number of false negatives decreases when compared to the situation where the sample acceptance level equals the regulatory guideline. For a given sample size, where the sample acceptance level is greater than the regulatory guideline, the number of false positives decreases and the number of false negatives increases when compared to the situation where the sample acceptance level equals the regulatory guideline. Increasing the sample size for a given sample acceptance level, where the legal limit equals the sample acceptance level: (a) increases the percentage of lots accepted at lower concentrations while increasing the percentage of lots rejected at higher concentrations; (b) decreases misclassification of lots (both false positives and false negatives) while increasing the percentage of correct decisions; and (c) decreases the average aflatoxin concentration in the lots accepted while increasing the average aflatoxin concentration in the rejected lots.<P>

<p><br>The use of in-stream wetlands is a growing practice being used to mitigatethe impacts of non-point source (NPS) pollution. Wetlands promote physical,chemical, and biological processes that attenuate and convert nutrientswhich can lead to improved water quality. Wetland performance is sensitiveto site conditions, making it difficult to precisely quantify their possibleimpact. Factors such as site specific soil, hydrologic, and vegetativecharacteristics influence wetland effectiveness. Typical design criteriainclude surface area/depth, retention time, plant coverage, and other hydrodynamicrecommendations. One recommendation made by Scheuler (1992) for stormwaterwetlands is that the design surface area be at least 1% of the contributingwatershed size. Most urban areas are severely limited with regard to availableland areas for wetland creation, thus, to be practical, smaller wetlandswill be necessary in many locations.<p>This research involved a two year study to quantify impacts of an in-streamconstructed wetland on water quality. The one hectare (2.4 ac) in-streamwetland was built to intercept drainage waters from approximately 240 hectares(600 ac) of agricultural and urban watershed, which resulted in a wetland:watershedarea ratio of 0.004:1. The wetland was instrumented to monitor hydrologyand water chemistry.<p>Water level recorders were used to measure stage at the wetland inletsand at the outlet. Weir equations and discharge curves combined with statisticalmodeling and calibration techniques were used to determine the flows throughthe wetland. A water balance was computed using inflow, outflow, precipitation,and potential evapotranspiration. A watershed scale balance showed thatthe total volume of flow leaving the wetland was comparable to the estimatedvolume of drainage and runoff from various land uses.<p>Water quality samples were analyzed for total Kjeldahl nitrogen (TKN),ammonium-nitrogen (NH-N),total phosphorus (TP), and ortho-phosphorus (OP). Measurements of watertemperature and dissolved oxygen levels were also made within the wetland.Background data acquisition began in early 1996. The evaluation periodbegan in August of 1997 and continued through December of 1999.<p>Over the evaluation period, NO-N concentrations were reducedthrough the wetland by 60%, NH-N concentrations by 30%, andTKN levels by 9.5%. This resulted in a 20% drop in total nitrogen concentration.Phosphorus levels increased 55% between the wetland inlets and outlet.Actual reduction of NH-N and TKN concentrations may be slightlyunderestimated due to unaccounted for inputs. This may also contributeto the increases in phosphorus concentrations observed between the inletsand the outlet. In the first full year, NO-Nlevels were reduced 70% and 33%, respectively. A significant decrease inNO-N concentrations through the wetland was detected duringthe first winter, and in other nitrogen forms in the first full growingseason. NO-N levels were 60% lower at the wetland outlet throughoutthe year. Ammonium nitrogen concentrations dropped 30% through the wetlandduring the growing season and 20% during the dormant season. TKN levelswere reduced (15%) in winter months, but not during the growing season.Phosphorus concentrations were higher at the wetland outlet than at theinlet throughout the year, but showed larger increases during the growingseason.<p>Monthly nutrient reductions were generally associated with temperaturechanges. Higher temperatures resulted in greater reduction of NO-Nand NH-N concentrations. Larger increases in TKN and phosphorusconcentrations were also associated with higher temperatures.<br><P>

<p>BIRGAND, FRANÇOIS. Quantification and Modeling of In-Stream Processes in Agricultural Canals of the Lower Coastal Plain. (Under the direction of Dr. R. Wayne Skaggs).Excess nutrient loads have been recognized to be the major cause of serious water quality problems recently encountered in the North Carolina estuaries and coastal waters. There has been a particular concern in coastal watersheds because agricultural and forested lands are located in close proximity to recreational and environmentally sensitive waters. The key to nutrient management at the watershed scale is the understanding and quantification of the fate of nutrients at the field scale and after they enter the aquatic environment. There is no accepted method to describe and predict fate of nutrients in canals and streams. The purpose of this research was to investigate the magnitude of the effects of in-stream processes in agricultural canals of the lower coastal plain and to propose a modeling approach for quantifying nitrogen transformations in such canals. This was accomplished in four steps.The first step was an extensive review of the literature on nitrogen retention in agricultural streams. Nitrogen removal rates in most agricultural canals and streams vary between 50 and 800 mg N/m²/d, with mass transfer coefficient varying between 0.01 and 0.10 m/d. The magnitude of nitrogen retention in streams and canals of agricultural watersheds has been reported to vary between less than 5% to more the 60% of the gross load. In the second step, the effects of biogeochemical processes on chemical and nutrient loads was evaluated in a 1125-m long agricultural canal reach of the lower coastal plain near the town of Plymouth, NC. Chemical and nutrient loads at both ends of the reach were measured by continuous measurement of flow and concentrations. Flow measurements were made using trapezoidal flumes in which flow velocity and depth was continuously measured and recorded with velocity meters. Nutrient concentrations were measured on water samples taken both manually and automatically at strategic times along the hydrographs so that linear interpolation between two consecutive samples could be made. Nutrient addition due to seepage along the reach was estimated. After corrections for lateral contribution, it was estimated that, over the 14-month measuring campaign, 3% of the total nitrogen load entering the upstream end was retained within the reach. This was mostly due to the combination of nitrate retention and release of organic nitrogen (ON) within the reach. Up to 10.2 % of the total phosphorus load measured at the upstream station was retained while 10% of the total suspended solids was also retained. There was a release of inorganic carbon equal to 18.7% more that the load measured at the upstream end.Measurements of algae and macrophyte biomass within the reach, and, measurements of nitrogen and carbon concentration profiles at the sediment-water interface revealed that most of nitrate retention was likely due to denitrification after diffusion from the water-column to the sediment. Release of organic nitrogen was attributed to flux of refractory organic nitrogen from the sediment into the water-column. Assimilation by algae and macrophytes may have accounted for as much as 20% of the total retention of inorganic nitrogen. Rates of nitrate removal and release of organic nitrogen were estimated using the model DUFLOW. Nitrate removal rates varied between 200 and 800 mg NO3-N/m²/d, while release rates of organic nitrogen varied between 100 and 400 mg ON/m²/d. A mass transfer coefficient of 0.3 m/d was obtained for nitrate at two distinct periods of the year.A simple approach was proposed for modeling nitrogen transformations in canals of the lower coastal plain. Transformations are simplified as the combination of downward diffusion of water-column nitrate into the sediment and an upward diffusion of organic nitrogen from the sediment.<P>

<p> Degradation of water quality in the streams and estuaries of North Carolina in recent years has resulted in regulations to reduce the introduction of numerous types of contaminants to this system. In the Neuse and Tar-Pamlico River Basins, excessive amounts of nitrogen have been identified as causing increased algal growth, low dissolved oxygen concentrations, and have been linked to increased growth of toxic microorganisms such as Pfiesteria piscicida. There are numerous sources of nitrogen to the basins; however, agricultural nonpoint sources have been identified as the largest contributor of nitrogen. Riparian buffers, controlled drainage, and nutrient management have been identified as effective BMPs for reducing nitrogen transport to streams under many landscape conditions. As a result, a combination of nutrient management, controlled drainage, and riparian buffer best management practices have been mandated in the Neuse River Basin to reduce the loss of agricultural nonpoint source pollution. A large portion of the agricultural nonpoint source nitrogen losses to surface waters in the Neuse River Basin originate in the Middle Coastal Plain. These lands are drained by irregularly spaced first and second order streams that have often been channelized (i.e. deepened) to enhance drainage. The riparian vegetation has often been removed from these channelized streams. The effectiveness of riparian buffers and controlled drainage are not well documented under these landscape conditions that are common in the Middle Coastal Plain region. Controlled drainage may not be economical in this region because multiple control structures would be required to maintain a suitable water table elevation in this gently sloping landscape. Implementation of riparian buffers has met strong resistance from the agricultural community due to the potential loss of land. A few studies have also found that nitrogen rich groundwater may enter deeply incised or channelized streams below the active treatment zone of the buffer, rendering the buffer ineffective. A study to evaluate the effect of riparian buffer vegetation type and width on shallow groundwater quality was implemented at the Center for Environmental Farming Systems near Goldsboro, North Carolina. The effect of controlled drainage, riparian buffers, and a combination of both was studied. The hydrologic portion of the riparian ecosystem management model (REMM) was evaluated and tested against field measurements.Five riparian buffer vegetation types were established as follows: cool season grass (fescue), deep-rooted grass (switch grass), forest (pine trees), native vegetation, and no buffer (no-till corn and rye rotation). These vegetation types were established at two buffer widths perpendicular to the channelized streams, 8 m (25 ft) and 15 m (50 ft). In addition, a continuous native vegetation buffer under free drainage and a continuous no buffer treatment under controlled drainage was established. For about 50% of the time monitored, the 15 m riparian buffer plots resulted in a statistically lower NO3-N concentration in the mid depth ditch wells (screen depth 1.5-2.1 m below ground surface) compared to the 8 m plots. Width was not a statistically significant variable at the deep well depth (2.1-3.5 m screen depth). Vegetation type had no statistically significant effect on NO3-N concentration. Nitrate concentration decreased 69 and 28% as groundwater flowed beneath the 8 m wide riparian buffer plots toward the channelized streams and 84 and 43% in the 15 m plots, at the deep and mid depth, respectively. The wider buffers were approximately 15% more effective at removing nitrate, but the improvement was not linearly correlated to the width increase. The primary reason vegetation differences were not observed was likely due to the limited time for vegetation establishment and development during the relatively short 2.5 year study period. Five years or longer may be required for some types of vegetation to mature to the point of impacting the nitrogen in the shallow groundwater. Furthermore, differences in localized groundwater flow paths and soil physical and chemical properties may indefinitely over shadow vegetation effects at this site. Controlled drainage did not raise the water table elevation near the ditch as compared to the free drainage treatment. Over seventeen storm events, the riparian buffer (free drainage) treatment had an average groundwater table depth of 0.92 m, compared to 0.96 and 1.45 m for the combination and controlled drainage treatments, respectively. Again, the lack of hydrologic treatment effect may be due to localized differences in soil properties and groundwater flow paths. Percent NO3-N concentration decrease for those treatments was 22 and 35%, 75 and 51%, and 77 and 69%, for the deep and mid depth wells, for each respective treatment. Although more nitrate was apparently removed from the groundwater on the controlled drainage treatments, this effect could not be correlated to water table depth.Daily predicted water table depth from the riparian ecosystem management model (REMM) was compared to observed depths over a simulation period of two years. The model performed well during some periods but poorly during large storm events. Average absolute errors ranged from 150 to 650 mm. Model instability during large storm events and anomalies in evapotranspiration calculations must be addressed before this model can be a reliable planning tool for regions such as the Middle Coastal Plain of North Carolina.Based on this research, several recommendations for further study are presented. Monitoring of riparian buffer vegetation plots should continue with the expectation that vegetation may have a significant impact over time as the vegetation types become established. Eventually the vegetation in the buffer will reach a steady state with respect to nitrogen in the buffer; however, this may take many years. Quantification of the relative proportion of dilution and denitrification for a given nitrate concentration decrease beneath the buffers should be investigated. One approach would be installation of redox probes at the deep well depth to give an indication if conditions are favorable (i.e. reducing) for denitrification. Also, the deep groundwater (i.e. below the impermeable layer) chemistry should be compared to the shallow groundwater chemistry to determine the relative proportions of constituents such as calcium and magnesium. This analysis would give an indication if dilution of the shallow groundwater were occurring as a result of deep groundwater upwelling. The REMM simulations may be improved by measuring groundwater velocity into the riparian buffer, improving the estimates of surface water runoff into the riparian buffers, and by modifying the model to simulate a single buffer zone rather than three.<P>

<p>ABSTRACTMarbis, Juan M. CO2 Enrichment and Hot Water Heat in a Greenhouse as a Mean of Recovering Bioresources From Swine Waste. (Under the direction of Daniel H. Willits) Predictions of heating and cooling requirements of a greenhouse located at the Barham Farm, Zebulon, N.C. were made via computer simulation. Kimball?s Modular Energy Balance Model (MEB) was used to simulate thermal behavior of the greenhouse. The weather inputs to the model were provided by data collected at the greenhouse and a Typical Meteorological Year (TMY2) data file for Raleigh, NC. Greenhouse air temperature (Tai), inside CO2 concentration (CO2) and inside relative humidity (RH) levels were used to validate the accuracy of the model. Absolute percentage differences ranging from 5.92% to 10.67% for Tai were observed. CO2 levels showed the biggest differences between observed and predicted data, from 14.93% to 42.33%, and RH showed a difference of 9.79% to 19.41%. Heating times were under-predicted, showing percentage difference between observed and predicted periods from -3.01% to -34.87%. On the other hand, cooling periods were over-predicted. With the exception of the month of February were cooling periods were under-predicted. Percentage difference for cooling periods ranged from -3.59% to 27.80%. The use of supplemental heat using a 10,000 gallon hot water tank serving as a waste heat collector was simulated. No data was available on its operation. Based on specific operating assumptions, it is expected that heat from the hot water will supply approximately 35% of the total energy demand of the greenhouse in a typical meteorological year. It was observed that the use of supplemental heat is most sensitive to its initial water temperature and the cutoff temperature. Outside weather conditions to which the use of hot water is most sensitive are solar radiation and wind speed.<P>

<p>Gurjer, Yeshwant Ramesh. Use of Heat Pumps for Heating and Night Cooling of Greenhouses. (under the direction of Daniel H. Willits). The use of heat pumps for heating and night cooling of greenhouses was investigated using a computer simulation program and weather data from Typical Meteorological Year (TMY2) datasets for Raleigh, NC, and Wilmington, NC. A greenhouse computer model taken from the literature, along with the heat pump subroutines (HPHEAT and HPCOOL) developed in this study, were used for the simulations. The use of heat pumps for heating only, and for heating plus night cooling, were examined separately using both standard and time-of-use electricity rates for two North Carolina utility providers. When heat pumps were used only for heating, standard electricity rates were predicted to provide greater savings in utility costs compared to time-of-use rates. When heat pumps were used for both heating and night cooling, the predicted savings in utility costs was greater for time-of-use electricity rates. Night cooling was predicted to decrease the average daily temperature (0.10C to 1.10C) and average nighttime temperature (0.10C to 2.00C) inside the greenhouse providing the potential for increased yield. Although a heat pump of capacity 36.5 W/m2 of floor area provided less savings in total costs compared to a capacity of 76 W/m2, the heat pump with the smaller capacity (36.5 W/m2) may be a better proposition because of the lower purchase cost. <P>