A large-eddy simulation of a line source in a convective atmospheric boundary layer—I. Dispersion characteristics

The dispersion of a passive line source in the convective atmospheric boundary layer is calculated with a large-eddy simulation. The results of the calculation are presented in the form of dispersion parameters. The total dispersion parameter is made up of two contributions: meandering and relative diffusion. Both contributions are calculated as a function of diffusion time and of source height. The results for vertical dispersion show that initially the contribution by meandering is large. Furthermore, the meandering component becomes larger as source height increases. These results can be interpreted in terms of the well-known looping plume in the convective boundary layer. The horizontal dispersion parameter is also influenced by meandering. However, in this case the lowest source has the largest meandering component. Finally, we consider the velocity statistics and demonstrate that they are consistent with the dispersion parameters.     

Water tank experiment of gas diffusion from a stack in stably and unstably stratified layers under calm conditions

Water tank experiments were conducted to investigate plume rrise and diffusion of gases discharged from a stack under calm conditions with stable and unstable thermal stratifications.First, tracer liquid with fluorescent dye and salt was emitted from a model of a stack in a stably stratified layer and the behavior of a plume was recorded by video camera. Using these results. we obtained formulae for plume rise height, and horizontal and vertical width of a plume as a function of thedensimetric Froude number, stability ratio and time.After the completion of these experiments under stable conditions, an unstably stratified layer was developed from a ground surface by heating a floor and the resulting concentration distribution was measured with electric conductivity probes. Using these results, we analyzed the height of the convection layer and concentration distribution as a function of the overall Richardson number and time.     

Concentration fluctuation profiles from a water channel simulation of a ground-level release

Measurements of concentration fluctuation profiles were made in a plume of saline tracer from a ground level source in a water channel simulation of a neutrally stable rpugh surface atmospheric boundary layer. Vertical profiles of mean concentration showed good agreement with gradient diffusion theory relations between the mean velocity power law and concentration profile exponents. The structure and development of vertical and crosswind profiles of mean, variance, intermittency and conditional intensity of non-zero fluctuations were satisfactorily represented using an isotropic meandering plume model combined with a clipped normal concentration probability distribution. Near the edges of the plume, the clipped normal pdf overestimated the fraction of non zero concentration periods.     

Particle model simulation of diffusion in low wind speed stable conditions

A Lagrangian particle model (LAMDA), previously developed and applied to the simulation of atmospheric dispersion in neutral and convective windy conditions, was modified to deal with stable low wind speed conditions. These last are among the most difficult to be treated. In fact, on the one hand, nearly calm situations, associated to strong stability and air stagnation, make the lower layers of the atmosphere poorly diffusive, and, on the other hand, the large fluctuations in the wind direction (meandering), spread the airborne pollutants over wide angular sectors. An ad hoc algorithm to simulate the effect of meadering on the dispersion is proposed. The model is validated by comparing its simulation results to three tracer experiments held in stable low wind speed conditions by the Idaho National Engineering Laboratory (U.S.A.) in 1974. These experiments present plume spread of different width (48, 138 and 360°, respectively, at an arc located 200 m downwind from the source) and are comprehensive of a wide set of conditions, ranging from strong to weak stability and from low wind speed to calm. The results of the comparison are discussed. The ability of the model to simulate the g.l.c. distributions with a good degree of confidence is illustrated.     

Velocity oscillations and plume dispersion in a residential neighborhood during wintertime nights

Measurements of velocity and tracer plume concentrations during stable atmospheric conditions were obtained in the Boise River Valley as part of the EPA Integrated Air Cancer Project during December, 1986. Wind speed, temperature, and wind direction were measured at two levels on a 30 m tower. Spectral and autocorrelation analyses of the velocity component data clearly indicate the occurrence of wave-like oscillations in the flow and the almost complete lack of turbulent energy. The predominate wave-like motion had an oscillation period of about 1000 s. Halogenated atmospheric tracers were released from as many as four houses during the night-time drainage conditions. Hourly averaged horizontal dispersion coefficients were very large compared to the Pasquill-Gifford curves and the urban McElroy-Pooler dispersion curves. The time-averaged dispersion coefficients formed an upper bound on very short-term dispersion coefficients obtained from mobile traverses of the tracer plume with a continuous SF₆ analyzer. These results agree with the concept of a narrow instantaneous plume with a broad meander driven by the observed 1000 s oscillation. Vertical dispersion rates were slightly smaller than the Pasquill-Gifford class F curve. Results from a single tracer release from a side canyon near the neighborhood showed that drainage flow from the tributary impacted the main residential sampling site at Elm Grove Park and represented a significant fraction of the upstream air flow at Elm Grove Park. For sources with equal emission rates, a source in the tributary adds about 10% to the mean of the concentration caused by a neighborhood source.     

寒区多孔介质中LNAPL迁移的多场耦合模拟

建立了包含水、气、NAPL、固(冰)四相的温度场、流场及化学场多场耦合模型,适用于定量模拟LNAPL泄漏过程中发生冻结后的迁移分布.通过与非冻条件下LNAPL迁移模型对比,模拟结果显示,冻结条件下因冰的形成LNAPL会发生与非冻条件下不同的迁移行为,冰的形成阻碍LNAPL下渗,并产生冻结势使LNAPL向地表方向迁移.相...     

An approximate model of atmospheric plumes produced by large area fires

We develop a simple analytical model to estimate the thickness of a smoke layer formed by a plume of a large area fire and to account for crosswinds. We take advantage of the dominant flow features in the upper part of the rising plume and in the smoke layer far from the plume to model the transition of a vertical flow to a horizontally spread thin smoke (or dust) layer. A one-dimensional approximation is used to model the buoyancy decay and plume overshoot. That solution is matched and combined with a solution based on the work required to establish our asymptotic horizontal smoke layer. The solution estimates the thickness of the spreading smoke layer in a quiescent atmosphere or in an ambient wind field.     

A synoptic analysis of the meteorological conditions affecting dispersion of pollutants emitted from tall stacks in the coastal plain of Israel

The meteorological conditions affecting dispersion of pollutants emitted from tall stacks in the coastal plain of Israel were analysed according to a synoptic approach. The SO₂ highest concentrations (episodes) measured around the largest power plant were first screened in order to characterize the associated synoptic configurations. Almost all the episodes were found to fall into three synoptic categories, all occurring during the warm season. According to the synoptic classification, the main elements influencing dispersion of pollutants were then analysed: the wind and temperature profiles and estimations of the plume rise and its partial penetration into the stable layer aloft. The following conclusions were reached: (a) horizontal wind shear is absent in the appropriate atmospheric layer; (b) for the shallow Persian trough synoptic category, in which the SO₂ episodes are the most frequent (55% of the total number), the CBL is shallow (450 m) and over two-thirds of the plume volume is expected to penetrate into the stable layer; (c) for the deep Persian trough synoptic category, which is the second most frequent among the episodes (22% of the total number), plume rise would be unobstructed in the deep CBL (900 m). A prediction index defined as the ratio of the frequency of each synoptic category among the SO₂ episodes to its annual frequency was developed. The marked differences obtained in the index values were tentatively explained by the different thermal stratification in the coastal CBL for each synoptic category. The synoptic classification was thus shown to be an efficient tool for predicting spells of SO₂ high concentrations.     

Time scales of physical and chemical processes in chemically reactive plumes

The rate of non-linear chemical reactions in a dispersing plume, such as the formation of NO₂, is discussed on the basis of various models and measurements of the oxidation rate of NO. Results of calculations with four different reactive plume models are compared with a set of plume measuremeent and discussed in terms of the ration of the time scales of physical (T_p) and chemical processes (T_c), the Damköhler number:T_p/T_c. In modelling a fast non-linear chemical reaction an interaction between physical processes of dispersion and mixing and chemical reactions can be expected. Models may differ in that they assume either chemical equilibrium (T_p/T_c⪢1) or no chemical equilibbrium (T_p/T_c⪡1).From our study it is concluded that for the first 5 km from a stack concentrations in the plume change more rapidly by dilution due to dispersion of the plume than by chemical reactions. Farther off it is the other way around and the chemistry is fast enough to achieve chemical equilibrum in the plume parcels.Conversion of NO is then inhibited by the mixing rate of the plume with its surrounding air. To incorporate the mixing rate of the plume correctly into a model, in principle the concentration distribution in a momentary plume has to be used. If a non-linear chemical reaction is modelled it is therefore important to establish the averaging times on which the dispersion parameters in the model are based. The contribution of plume meandering to dispersion, which is especially important in the daytime, can be eliminated by taking shorter averaging times. This leads to a marked improvement of modelling results. Further research is required to investigate how the contribution of fluctuations in the concentrations around the mean has to be incorporated into a model of a chemically reactive plume.     

Physico-chemical study by two aircraft of a plume from a coal-fired power plant

During the experimental field program that was conducted in October 1985 ton study the physics and chemistry of the Le Havre power plant plume, joints flights were performed with two aircraft from EDF and KEMA. The two episodes with very different meteorological conditions were investigated. The first episode was characterized by fairly constant wind speed and direction with altitude, low solar irradiation and neutral vertical dispersion conditions. The second episode was characterized by low wind speed and changing wind direction, high solar irradiation and stable vertical dispersion conditions. During the latter episode, the concentrations of emitted gases, SO₂ and NO_x remained high, even at several tens of km downwind of the stock. Increased SO₄^2- and NO₃^- concentrations were detected in the power plant plume. In addition, important photochemical activity took p-lace in the plume of a nearby industrial area. The former episode showed more significantly vertical dilution and oxidation of NO but did not reveal any contribution of the power plant plume to aerosol formation.     

An intercomparison of semiempirical diffusion models under low wind speed,stable conditions

The performances of some diffusion models are analysed using concentration data measured at ground level up to 400 m from the emission point, in a series of diffusion tests conducted by U.S. National Oceanic and Atmospheric Administration (NOAA) under inversion conditions with light winds. All tested models are simple semiempirical formulae based on the Gaussian formulation, with different assumptions concerning dispersion parameters; each model utilizes a minimum set of information, i.e. vertical stability category, mean wind speed and standard deviation of the horizontal wind direction over the considered averaging time, σ_θ.Results show that for cases with very low wind speed and large plume spread, explicit consideration of diffusion along the mean wind direction, which is neglected in the standard plume model, significantly improves model results; moreover, when σ_θ is very large (greater than 50–60°), the analysis suggests that standard deviations of the horizontal wind speed may significantly differ from the estimates commonly found in the literature.     

Diurnal wind-structure variations and dispersion of pollutants in the boundary layer

A coupled boundary-layer model and Lagrangian particle model are used to investigate the role of boundary-layer shear especially that produced by inertial oscillations in affecting the horizontal dispersion of pollutants on time-scales of 24–36 h. The coupled models show that the amplitude and the effective periods of the inertial oscillations are the main cause of nocturnal accelerating dispersion. The effective width of the plume in the morning is determined by whether the morning daytime mixing coincides with the phase of the inertial oscillation being at a maximum or minimum value. The phase of the oscillation is determined by latitude. Thus, latitude is shown to be an extremely important parameter in determining horizontal dispersion. An analytical model is introduced to investigate the role of external parameters such as latitude in influencing the horizontal dispersion. The analytical model is based on a simple Ekman-type model for the daytime and nighttime boundary layer. The Ekman model is used to provide initial conditions to an inertial oscillation regime between the nighttime boundary layer and the old daytime boundary layer. The analytical model was able to reproduce the magnitude and phase of the inertial oscillations reasonably well. However, the Ekman model overestimates the shear in the boundary layer causing the inertial oscillation to be too large. A semi-empirical method was used to provide more reasonable estimates of the daytime boundary-layer structure. This semi-empirical approach gave rates of the horizontal dispersion which were in general agreement with the numerical results.     

A model of instrument smoothing and thresholding in measurements of turbulent dispersion

Since instruments used to measure the concentration of contaminants dispersing in turbulent flows cannot fully resolve all of the scales present, they inevitably smooth out the concentration fluctuations. Measurement and data processing systems may also have associated thresholds, so it is of interest to consider the effect of any smoothing and thresholding on such concentration fluctuations. A one-dimensional model of the effect of smoothing and thresholding on statistics of the concentration fluctuations, in particular the variance, is presented. The major effect in the model is the reduction of variance through averaging. For threshold values occurring in many experiments the effect of thresholding is small, but in some cases the threshold is large enough to give a significant effect. This effect can either be to decrease or to increase the perceived variance, depending on the values of the intermittency and of the smoothing scale. The variation of the effect of smoothing throughout a plume is considered, including a modification to the model for the case of a meandering plume. It is demonstrated that, for a given instrument resolution, the most important model parameter to estimate is the representative absolute width of regions of non-zero concentration. The model shows the same qualitative dependences as are found from experiment, and gives good quantitative agreement with experiment in many of the cases considered. The results suggest, in agreement with results derived elsewhere using signal processing theory, that in the experiments considered in this paper the instruments have relatively little effect on the variance.     

人工注气及生物产气对孔隙介质中DNAPL运移影响

重非水相液体(DNAPLs)的污染和治理问题是国内外研究热点,由地下水曝气技术注入的人工气体或生物作用等产生的气体必然对DNAPL在地下水系统中的运移及修复产生影响.采用CCD实时监测系统,以TCE为目标污染物,通过3个二维砂箱试验,以饱水条件下DNAPL的运移为对照,研究了人工注气及生物产气对DNAPL在孔隙介质中运移的影响.结果表明:(1)基于CCD相机的透射光监测系统是一种研究二维孔隙介质中多相流迁移规律的有效非侵入式监测方法,应用于实时监测水/NAPL,水/气,水/NAPL/气多相系统中流体渗流过程.(2)人工注气和生物产气两种条件下,气体在多孔介质中的分布特征有差异,前一条件下气体连续分布,后一条件下则以非连续分布的气泡为主.(3)在水/气两相系统中,气体的存在使得DNAPL污染羽的整体形状更加不规则;同时缩短了TCE污染羽前缘整体平均垂向下迁距离;也导致迁移路径上孔隙中TCE的截留量变小.DNAPL的入渗过程受重力影响以垂向渗流为主,垂向渗流时易于驱替孔隙中的水分,然而水平渗流时优先驱替孔隙中的空气.     

Simulation of dispersion in moderately complex terrain—Part C. A dispersion model for operational use

An operational dispersion model for use in areas with complex terrain is presented. The model uses mean and turbulence quantities simulated with the fluid dynamic model presented in Part A. A large number of wind and turbulence fields are simulated with the fluid dynamic model. These simulations are put into a database and can be used in the calculations of dispersion with the operational model. To get relevant meteorological data for the model a Doppler sodar and a 10 m high mast with a temperature profile and wind and wind direction at one level are used. The model calculates a trajectory for the plume centerline from the simulated wind field, and approximates the concentration field with a bi-Gaussian distribution. For convective conditions the mixing height and the surface heat flux, used as input for the model, are being determined from the sodar measurements through relations related to the temperature structure parameter C_T² and the standard deviation of the vertical velocity. The horizontal and vertical standard deviations for the plume are determined by using the simulated turbulence quantities from the dynamic model and Eulerian velocity spectra. Simulations with the model is compared with dispersion measurements performed in an area in the southern Sweden, the Vänersborg-Trollhättan region. The geographical area is characterized by topographical features on the meso-γ-scale, i.e. 2–20 km. Thus there are forested hills, a relatively flat agricultural area and an extended lake area within the model domain. The terrain height relief is typically 80 m. The simulations show, in general, good agreement with the measured data both for unstable and stable stratifications.     

Heavily-contaminated snowfalls in the remote Scottish Highlands: A consequence of regional-scale mixing and transport

Heavily-contaminated (black) snowfalls frequently occur in the Cairngorm Mountains (Scotland), in spite of their relative remoteness from major source regions. The frequency of such pronounced events has been characterized throughout one complete snow season. Daily samples of snow were collected at 1100 m under sub-arctic conditions, and the reported results from the analyses of major ions and a large number of trace elements (dissolved and particulate phases) represent a novel precipitation chemistry data set. This paper focuses on the major ion, black carbon and particulate concentrations, but formic acid concentrations are also reported. The heavily-contaminated snowfalls are associated with back-trajectories which originate in the east (Eastern Europe, European U.S.S.R. and the Baltic) and, in spite of providing a small proportion (30%) of the total snowfall, produce the bulk of the pollutant wet deposition (e.g. 70% of non-marine SO₄²⁻; 80% of black carbon). The covariances of the constituent concentrations indicate that airborne pollution is relatively aged and well-mixed before removal by the snowflakes. Back-trajectory, synoptic and upper-air analyses have been combined to produce a qualitative model of pollutant transport from source to receptor region. Once air is advected out of a (typically) long-lived anticyclone over Eastern Europe (which produces the regional-scale mixing), transport westwards of the ‘regional plume’ is usually associated with an elevated stable layer which can effectively decouple the plume from the surface and inhibit dispersion. Over the North-Sea-NE Scotland region, horizontal convergence, convection and topographic uplift can all contribute to precipitation formation and pollutant wet deposition. Although there are indications that the NE Scottish Highlands are particularly susceptible to these ‘black’ snowfalls, comparison with data from monitoring stations within 35 km of the study site, and 600–800 m lower elevation, does not indicate enhancement of ion concentration in snowfall with increasing altitude.     

An improved method for predicting seasonal and annual shadowing from cooling tower plumes

An improved model developed at Argonne National Laboratory and the University of Illinois (ANL/UI) for predicting long-term shadowing due to cooling tower plumes is presented, and its assumption are compared with those used in previous models. The model is based on a method for the selection of representative categories of similar plumes developed by Dunn and Policastro (Dunn, 1980, Proc. IAHR Cooling Tower Workshop, San Francisco; Policastro et al., 1984, Report EPRI CS-3403-CCM). At a given site this method reduces the large number of meteorological data cases in a season or year to a much smaller number (≈35) of representative cases, each of which will have a predicted plume substantially different from the others. Plume predictions for the reduced set of category representative cases are then made with the validated ANL/UI plume model. With category representative plume shape, wind speed, wind direction and sun angles available for each hour, full effects of sun angles for the latitude and longitude of the site to be studied are included. The ANL/UI model yields seasonal and annual isopleths of hours of additional shadowing or of percentage reduction in total and direct solar energy arriving at the ground on a horizontal surface. Results for two hypothetical sites with 500 MWe generating capacity are presented and contrasted, one at Syracuse, NY, and the other at Spokane, WA.     

A meso-scale plume dispersion model. Preliminary evaluation in a heterogeneous area

A plume dispersion model for meso-γ-scale applications is developed. The model is formulated with the aim to study plume spreading in moderately complex terrain. An eddy diffusivity analogue is used in order to facilitate future inclusion of chemical transformations. The full diffusivity tensor, obtained from a second-order closure turbulence model, is used to described the dispersion process. The model is driven with simulated meteorological fields, obtained by means of a second-order turbulence closure model. An initial model evaluation has been undertaken using model predictions for an area featuring horizontally inhomogeneous dispersion conditions. Comparison with measurements, as well as with predictions of several other model types for distances out to 40 km, shows that the model is able to describe realistically plume widths and concentration in areas with strongly varying dispersion conditions. Non-negligible errors in plume position were however found for all models compared.     

确定大气环境承载力的烟云足迹法

本文讨论了大气承载力的定义,同时也给出了计算大气承载力的烟云足迹法的原理和计算公式,此外还给出了根据区域气象条件计算烟云足迹函数的具体算式和大气环境承载力的计算实例,烟云足迹分析法可为区域间的排污权交易提供技术基础.     

高斯烟团模型在石化类项目大气环境风险评价中的应用

利用高斯烟团模式对某石化项目进行风险识别与分析,并对主要计算参数进行选取和设定,重点介绍了源强确定的思路,进而对项目中120万t/a延迟焦化装置的脱硫装置发生特大灾害性事故状态下的硫化氢泄漏进行模拟,最后确定安全距离为下方向250 m。介绍了高斯烟团模型适用性以及其解决问题的总体思路和步骤,对今后确定事故状态下装置的安全距离以及高斯烟团模型的应用有一定的参考价值。