The public nuisance and health concerns caused by odours from livestock facilities are among the key issues that affect neighbouring communities and the growth of the livestock industry across Canada. A setback distance is the common regulatory practice to reduce odour impact on the neighbouring areas. The air dispersion modeling method may be a more accurate tool for establishing setback distances since it considers site-specific airborne emissions, such as odour and gases from the animal production site as well as weather conditions and then estimates a concentration of the pollutant (odour, ammonia, etc.). Although various dispersion models have been studied to predict odour concentration from agricultural sources, limited field data exist to evaluate their applicability in agricultural odour dispersion. Thus, the purpose of this project was to evaluate the selected commercial air dispersion models with field plume measurements from swine operations. <p>Firstly, this thesis describes a sensitivity analysis of how the climatic parameters affect model simulations for four selected air dispersion models, ISCST3, AUSPLUME, CALPUFF, and CALPUFF. Under the steady state weather condition, mixing height had no effect on the livestock odour dispersion, while atmospheric stability, wind speed and wind direction had great effect on the livestock odour dispersion. Ambient temperature had a moderate effect compared with other parameters. Under variable weather conditions, the predicted odour concentrations were much lower than the results under steady state weather conditions. <p>A series of comparisons between model predictions of the same four models and field odour measurements were conducted. When using the livestock odour plume measurement data from University of Manitoba, three equations were used to convert the model predicted odour concentration to field measured odour intensity. The equations did not predict odour intensity very well. No model showed obvious better performance than the others. Scaling factors did not improve the results considerably. When using the odour plume measurement data from University of Minnesota, INPUFF2 performed better than CALPUFF. Scaling factors did improve the modeled results. When using the odour plume measurement data from University of Saskatchewan, INPUFF2 also performed better than CALPUFF. Scaling factors were still useful for the results improvements.<p>Finally, because CALPUFF is the US EPA preferred model and predicted the highest values under variable weather conditions in the sensitivity study, we used it to simulate odour plumes on selected three swine sites using hourly weather data from 1993 to 2002 in Yorkton, Saskatchewan. The maximum predicted distance were 2.9 km for 1 OU, which was lower than the recommended maximum setback distance of 3.2 km. <p>It is recommended that the variable weather conditions be used in the setback distance determination. CALPUFF is the preferred model and INPUFF2 is another option for field odour plume simulation, however scaling factors are needed to bring the model predictions close to the field measured results. Because the models evaluated were not developed for odour dispersion simulation, a model that can accurately predict livestock odour dispersion should be developed to take into account of the difference between odour and gas and wind direction shifts within the simulation time interval.

The public nuisance and health concerns caused by odours from livestock facilities are among the key issues that affect neighbouring communities and the growth of the livestock industry across Canada. A setback distance is the common regulatory practice to reduce odour impact on the neighbouring areas. The air dispersion modeling method may be a more accurate tool for establishing setback distances since it considers site-specific airborne emissions, such as odour and gases from the animal production site as well as weather conditions and then estimates a concentration of the pollutant (odour, ammonia, etc.). Although various dispersion models have been studied to predict odour concentration from agricultural sources, limited field data exist to evaluate their applicability in agricultural odour dispersion. Thus, the purpose of this project was to evaluate the selected commercial air dispersion models with field plume measurements from swine operations. <p>Firstly, this thesis describes a sensitivity analysis of how the climatic parameters affect model simulations for four selected air dispersion models, ISCST3, AUSPLUME, CALPUFF, and CALPUFF. Under the steady state weather condition, mixing height had no effect on the livestock odour dispersion, while atmospheric stability, wind speed and wind direction had great effect on the livestock odour dispersion. Ambient temperature had a moderate effect compared with other parameters. Under variable weather conditions, the predicted odour concentrations were much lower than the results under steady state weather conditions. <p>A series of comparisons between model predictions of the same four models and field odour measurements were conducted. When using the livestock odour plume measurement data from University of Manitoba, three equations were used to convert the model predicted odour concentration to field measured odour intensity. The equations did not predict odour intensity very well. No model showed obvious better performance than the others. Scaling factors did not improve the results considerably. When using the odour plume measurement data from University of Minnesota, INPUFF2 performed better than CALPUFF. Scaling factors did improve the modeled results. When using the odour plume measurement data from University of Saskatchewan, INPUFF2 also performed better than CALPUFF. Scaling factors were still useful for the results improvements.<p>Finally, because CALPUFF is the US EPA preferred model and predicted the highest values under variable weather conditions in the sensitivity study, we used it to simulate odour plumes on selected three swine sites using hourly weather data from 1993 to 2002 in Yorkton, Saskatchewan. The maximum predicted distance were 2.9 km for 1 OU, which was lower than the recommended maximum setback distance of 3.2 km. <p>It is recommended that the variable weather conditions be used in the setback distance determination. CALPUFF is the preferred model and INPUFF2 is another option for field odour plume simulation, however scaling factors are needed to bring the model predictions close to the field measured results. Because the models evaluated were not developed for odour dispersion simulation, a model that can accurately predict livestock odour dispersion should be developed to take into account of the difference between odour and gas and wind direction shifts within the simulation time interval.

The Swedish defence agency (FOI) has developed a particle model (called Pello) that simulates the dispersion of aerosols and gases. At the boundaries, such as the ground, the particles can either reflect back into the domain (the atmosphere) or be absorbed. Which of the events that occurs is decided by a certain probability, which in the present model depends on mere physical properties. In this thesis we have investigated a newly proposed boundary behaviour which also depends on the time step used in the numerical simulations. We verified the accuracy of the new model by using a dispersion model with an explicit solution. To gain a better understanding of how important parameters at the boundary influence each other, we performed a sensitivity analysis. Simulations showed an overall improving concentration profile as the time step became smaller and the new model working well. The convergence order of the simulations was found to be close to 0.5. In this thesis we have shown that there exist an upper limit for the time step, which depends on the specific model. The present used time step at FOI does not have this versatile property. But having this upper limit for the time step close to the boundary, and a uniform time step can be time demanding. This lead us to the conclusion that an adaptive time step should be implemented.

No description available.

Until present, many researchers relied on the conventional plug flow dispersion models to analyse the concentration profiles obtained from the tracer injection experiments to evaluate the dispersion coefficients in packed beds. The Fickian concept in the limit of long time duration is assumed to be applicable and it implies that the mean-square displacement of the tracer profile is constant with time and the concentration profile is Gaussian. There were very few studies on identifying the conditions under which this assumption is valid and delineate the range of applicability of the existing plug flow dispersion models. If the time scales of a tracer injection experiment are not sufficient for a tracer to traverse the bed radius and sample the velocity variations, this could give rise to persisting non-Fickian transients where the mean-square displacement of the tracer profile is not constant with time and the concentration profile deviates from the normal Gaussian distribution. These transients cannot be predicted by the conventional plug dispersion models. An extended axial non-Fickian macroscopic dispersion model is derived to describe the transient development of a solute tracer when injected into a fluid flowing through a cylindrical packed bed or empty tube and some non-Fickian effects in the dispersion process. The flow profile in beds packed with uniform particles exhibits radial non-uniformity due to the oscillatory variation in porosity because of the wall confinement (wall effect). Compared with the axial plug flow dispersion model, the extended model contains time-dependent coefficients such as the transient axial dispersion coefficient and higher order derivatives (higher than second order) of the cross-sectionally averaged concentration. Including them provides some insight on non-Fickian transport in the dispersion process. The model provides time criteria on the basis that the effelongitudinal dispersion coefficient in the packed bed reaches its asymptotic value and the non-Fickian transients will die out. Some experimental conditions in the literature were checked by these criteria and found to be either marginally satisfied, or not satisfied at all, which indicates that the Fickian concept is not valid. The model results for tracer dispersion in cylindrical packed beds show that the longitudinal dispersion coefficient converges to its asymptotic value on a time scale proportional to R2/(DT) where R is the column radius and (DT) is the area averaged lateral dispersion coefficient. The extended model encouraged study of the consequences of the additional dispersion terms in other applications such as the pulse spread in the field of capillary column inverse gas chromatography (CCIGC). CCIGC is used to evaluate the solute-polymer diffusion coefficient Dp and the partition coefficient K at infinite dilute conditions. The tube geometry in CCIGC is more complex than the conventional Taylor dispersion problem due to the polymer coating on the inside of the capillary wall. The extended CCIGC model presented in this study has advantages over the previous models by including the effects of Taylor dispersion and higher order derivatives of the pulse area-averaged concentration. Taylor dispersion effect causes more pulse spread in the longitudinal direction and by not including it in the CCIGC regression models may cause a significant error in the measured Dp values. The extended CCIGC model provides for the first time criteria on capillary dimensions for the transient coefficients (multiplying the second and higher order derivatives) to become constant and for the non-Fickian effects associated with the higher order derivatives to be neglected. Model results show that Taylor dispersion effect has a significant effect on the elution profiles at high values of Dp and/or low values of gas diffusion coefficients Dg and it can be used to increase the sensitivity range of the previous CCIGC models at extremely low and high Dp values.

Impact of odour emissions from livestock operation sites on the air quality of neighboring areas has raised public concerns. A practical means to solve this problem is to set adequate setback distance. Air dispersion modeling was proved to be a promising method in predicting proper odour setback distance. Although a lot of air dispersion models have been used to predict odour concentrations downwind agricultural odour sources, not so much information regarding the capability of these models in odour dispersion modeling simulation could be found because very limited field odour data are available to be applied to evaluate the modeling result. A main purpose of this project was evaluating AERMOD and CALPUFF air dispersion models for odour dispersion simulation using field odour data.<p> Before evaluating and calibrating AERMOD and CALPUFF, sensitivity analysis of these two models to five major climatic parameters, i.e., mixing height, ambient temperature, stability class, wind speed, and wind direction, was conducted under both steady-state and variable meteorological conditions. It was found under steady-state weather condition, stability class and wind speed had great impact on the odour dispersion; while, ambient temperature and wind direction had limited impact on it; and mixing height had no impact on the odour dispersion at all. Under variable weather condition, maximum odour travel distance with odour concentrations of 1, 2, 5 and 10 OU/m3 were examined using annual hourly meteorological data of year 2003 of the simulated area and the simulation result showed odour traveled longer distance under the prevailing wind direction.<p> Evaluation outcomes of these two models using field odour data from University of Minnesota and University of Alberta showed capability of these two models in odour dispersion simulation was close in terms of agreement of modeled and field measured odour occurrences. Using Minnesota odour plume data, the difference of overall agreement of all field odour measurements and model predictions was 3.6% applying conversion equation from University of Minnesota and 3.1% applying conversion equation from University of Alberta between two models. However, if field odour intensity 0 was not considered in Minnesota measured odour data, the difference of overall agreement of all field odour measurements and model predictions was 3.1% applying conversion equation from University of Minnesota and 1.6% applying conversion equation from University of Alberta between two models. Using Alberta odour plume data, the difference of overall agreement of all field odour measurements and model predictions was 0.7% applying conversion equation from University of Alberta and 1.2% applying conversion equation from University of Minnesota between two models. However, if field odour intensity 0 was not considered in Alberta measured odour data, the difference of overall agreement of all field odour measurements and model predictions was 0.4% applying conversion equation from University of Alberta and 0.7% applying conversion equation from University of Minnesota between two models. Application of scaling factors can improve agreement of modeled and measured odour intensities (including all field odour measurements and field odour measurements without intensity 0) when conversion equation from University of Minnesota was used.<p> Both models were used in determining odour setback distance based on their close performance in odour dispersion simulation. Application of two models in predicting odour setback distance using warm season (from May to October) historical annul hourly meteorological data (from 1999 to 2002) for a swine farm in Saskatchewan showed some differences existed between models predicted and Prairie Provinces odour control guidelines recommended setbacks. Accurately measured field odour data and development of an air dispersion model for agricultural odour dispersion simulation purpose as well as acceptable odour criteria could be considered in the future studies.

Impact of odour emissions from livestock operation sites on the air quality of neighboring areas has raised public concerns. A practical means to solve this problem is to set adequate setback distance. Air dispersion modeling was proved to be a promising method in predicting proper odour setback distance. Although a lot of air dispersion models have been used to predict odour concentrations downwind agricultural odour sources, not so much information regarding the capability of these models in odour dispersion modeling simulation could be found because very limited field odour data are available to be applied to evaluate the modeling result. A main purpose of this project was evaluating AERMOD and CALPUFF air dispersion models for odour dispersion simulation using field odour data.<p> Before evaluating and calibrating AERMOD and CALPUFF, sensitivity analysis of these two models to five major climatic parameters, i.e., mixing height, ambient temperature, stability class, wind speed, and wind direction, was conducted under both steady-state and variable meteorological conditions. It was found under steady-state weather condition, stability class and wind speed had great impact on the odour dispersion; while, ambient temperature and wind direction had limited impact on it; and mixing height had no impact on the odour dispersion at all. Under variable weather condition, maximum odour travel distance with odour concentrations of 1, 2, 5 and 10 OU/m3 were examined using annual hourly meteorological data of year 2003 of the simulated area and the simulation result showed odour traveled longer distance under the prevailing wind direction.<p> Evaluation outcomes of these two models using field odour data from University of Minnesota and University of Alberta showed capability of these two models in odour dispersion simulation was close in terms of agreement of modeled and field measured odour occurrences. Using Minnesota odour plume data, the difference of overall agreement of all field odour measurements and model predictions was 3.6% applying conversion equation from University of Minnesota and 3.1% applying conversion equation from University of Alberta between two models. However, if field odour intensity 0 was not considered in Minnesota measured odour data, the difference of overall agreement of all field odour measurements and model predictions was 3.1% applying conversion equation from University of Minnesota and 1.6% applying conversion equation from University of Alberta between two models. Using Alberta odour plume data, the difference of overall agreement of all field odour measurements and model predictions was 0.7% applying conversion equation from University of Alberta and 1.2% applying conversion equation from University of Minnesota between two models. However, if field odour intensity 0 was not considered in Alberta measured odour data, the difference of overall agreement of all field odour measurements and model predictions was 0.4% applying conversion equation from University of Alberta and 0.7% applying conversion equation from University of Minnesota between two models. Application of scaling factors can improve agreement of modeled and measured odour intensities (including all field odour measurements and field odour measurements without intensity 0) when conversion equation from University of Minnesota was used.<p> Both models were used in determining odour setback distance based on their close performance in odour dispersion simulation. Application of two models in predicting odour setback distance using warm season (from May to October) historical annul hourly meteorological data (from 1999 to 2002) for a swine farm in Saskatchewan showed some differences existed between models predicted and Prairie Provinces odour control guidelines recommended setbacks. Accurately measured field odour data and development of an air dispersion model for agricultural odour dispersion simulation purpose as well as acceptable odour criteria could be considered in the future studies.

In order to understand the impact from a point source of pollution a comprehensive air quality monitoring programme must be in place. Measurement stations are at fixed locations and can only measure the relevant concentration if the wind is blowing in a particular direction. With changing wind direction, a measurement station needs to be coupled with a dispersion model to predict the impact from a point source, such as the BCL Limited smelter. The smelter produces 55 500 tonnes per annum of nickel/copper granules; in the process emitting waste gases to the atmosphere through a 153 m stack. The EEGAIR-BCL1. dispersion model was developed for BCL Limited to complement the air quality monitoring programme in the Selebi-Phikwe area. The model was developed for the smelter specifically, using local meteorological data at Selebi-Phikwe and smelter specific emission data. Results from the model were tested against data obtained from existing measurements stations at Selebi-Phikwe for May 2002, using statistical analyses. The average index of agreement indicates that the model predictions were more accurate for Kopano (0,44) and SPSS (0,25) stations than for WUC (0,08) station. Yearly averaged emission data was used for model input and better correlation can be expected when using actual hourly and/or monthly emission data. Based on the EEGAIR-BCl1 dispersion model results, it was found that the highest impact from the smelter stack was at a distance of 4 km to 7 km west from the stack. The model results indicate that, on average, the impact from the smelter stack in residential areas of Selebi-Phikwe was between 0 and 50 µ/m3 for May 2002. / Dissertation (MEng(Environmental Engineering))--University of Pretoria, 2006. / Chemical Engineering / unrestricted

ADMS 4 is a relatively new generation air quality dispersion model program written by Cambridge Environmental Research Consultants in the UK and used to verify datasets in the USA on several types of plants, in various settings across the country. Typical results indicate an approximate 20% under prediction. It was decided to use this model to verify SO2 emissions typical of the Eastern Highveld, which has several varying industries emitting SO2 gases in the region. Two datasets were identified with complete data recorded, supplied by ESKOM as measured at their Elandsfontein site for the winter periods mid 2006 and 2007. The meteorological data recorded at Elandsfontein was used for the model and the SO2 predicted emissions by ADMS 4 then compared to the actual measured SO2 emissions at the monitoring site, to determine the efficiency level of ADMS 4. Efficiency levels of 42% for the 2006 dataset and 58% for the 2007 dataset were achieved. The high level of under prediction is ascribed to the influence of the local petrochemical refinery (SASOL) as well as the local steelworks plants also emitting SO2 gas, but was not entered into the model database for evaluation in conjunction with the power stations’ emissions data. / Dissertation (MSc)--University of Pretoria, 2013. / Geography, Geoinformatics and Meteorology / Unrestricted

The use of hazardous toxic substances is very common in the industrial sector. The substances are often stored in tanks in storage compartments or transported between industrial premises. In case of an accident involving these substances, severe harm can affect both population and the environment. This leaves a demand for an accurate prediction of the substance concentration distribution to mitigate the risks as much as possible and in advance create suitable safety measures. Toxic gases and vapors are often denser than air making it affected by negative buoyancy forces. This will make the gas descend and spread horizontally when reaching the ground. Swedish Defence Research Agency (FOI) carries today a model called LillPello for simulating the dispersion of gases, yet it does not account for the specific case of a dense gas. Therefore, this thesis aims to implement the necessary effects needed to accurately simulate the dispersion of a dense gas. These effects were implemented in Fortran 90 by solving five conservation equations for energy, momentum (vertical and horizontal) and mass. The model was compared against experimental data of a leak of ammonia (NH3). By analyzing the result of the simulations in this thesis, we can conclude that the overall result is satisfactory. We can notice a small concentration underestimation at all measurement points and the model produced a concentration power law coefficient which lands inside the expected range. Two out of the three statistical quantities Geometric Mean (MG), Geometric Variance (VG) and Factor of 2 (FA2) produced values within the ranges of acceptable values. The drawback of the model as it is implemented today is its efficiency, so the main priority for the future of this thesis is to improve this. The model should also be analyzed on more experiments to further validate its accuracy. / Användandet av giftiga ämnen är vanligt inom den industriella sektorn. Ämnena är oftast lagrade i behållare positionerade i lagringsutrymmen eller så transporteras ämnena mellan industrilokaler. I samband med en olycka innehållande dessa substanser kan stora skador drabba både befolkning och miljön. Detta leder till ett behov av att noggrant kunna förutspå koncentrationsfördelningen för att minska riskerna, samt i förväg kunna skapa lämpliga säkerhetsåtgärder. Giftiga gaser och ångor är oftast tyngre än luft vilket gör att gasen blir påverkad av negativ bärkraft. Detta gör att gasen sjunker och sprids horisontalt när den når marken. Totalförsvarets Forskningsinstitut (FOI) besitter idag en modell kallad LillPello som simulerar spridning av gaser, men den hanterar inte det specifika fallet av en tunggas. Därför siktar detta projekt på att, in i LillPello, implementera de nödvändiga effekterna som behövs för att korrekt kunna simulera spridningen av en tunggas. Dessa effekter är implementerad i Fortran 90 genom att lösa fem konserveringsekvationer för energi, momentum (vertikal och horisontell) samt massa. Modellen jämfördes mot data från ett fältexperiment där ammoniak (NH3) släpptes ut. Genom att analysera resultatet från simuleringar kan vi dra slutsatsen att det övergripande resultatet är tillfredsställande. Vi kan notera en underskattning för alla koncentrationsmätningar i simuleringarna och modellen producerade en potenslagsexponent vars värde hamnade innanför den accepterade gränsen. Två utav de tre beräknade statistiska kvantiteterna: Geometriskt medelvärde (MG), Geometrisk varians (VG) och Faktor av 2 (FA2) producerade värden inom de acceptabla gränserna. Största nackdelen med modellen är dess effektivitet och därför är största prioritet för det fortsatta arbetet inom detta projekt att effektivisera implementeringen. Modellen ska även bli vidare analyserad mot fler experiment för att validera dess noggrannhet.