Risks of soil chemical degradation from atmospheric gypsum plumes around selected extraction and processing enterprises, Northeast Brazil.

Gypsum plays a prominent role in agriculture, being considered an effective alternative to alleviate subsurface acidity due to its higher solubility and containing sulfur. However, another significant aspect is which pose long-term risks of groundwater contamination due to excessive applications of salts, pesticides, and other chemicals that will be leached, or even soil chemical depletion. So far, no study has focused on understanding the impacts of the atmospheric gypsum plume originating from gibbsite mining and processing on the leaching of soil bases and chemical degradation surrounding these sites. In this study, we evaluated the behavior of chemical characteristics in soil profiles distributed along the dispersion of the atmospheric plume and in areas without interference from the industrial sector in the state of Maranhão, Northeast Brazil. Fifty-three sampling points were collected at 7 locations based on the dispersion of the dust plume through wind drift. Each sampling point was represented by three composite soil samples at depths of 0.0-0.20, 0.40-0.60, and 1.00-1.20 m, where the chemical soil characteristics were evaluated. The average levels of Ca, Mg, and K in the studied layers are classified as low, with minimum values below the method's detection limit, and they also show imbalance due to higher concentrations of Ca in the surface layer in areas affected by atmospheric dispersion. The sum and saturation of bases at all depths are classified as low. Higher aluminum saturation values were observed in the deeper soil layers. The gypsum dust altered the soil's chemical characteristics at the evaluated depths; therefore, it is necessary to seek means to mitigate gypsum dust release during gibbsite extraction and processing and ensure that the soils in areas near these enterprises maintain their natural characteristics.

A study of particle dry deposition parameterizations in an atmospheric radioactive preparedness model: Application to the Chernobyl case.

Parameterization of dry deposition is key for modelling of atmospheric transport and deposition of radioactive particles. Still, very simple parameterizations are often encountered in radioactive preparedness models such as the SNAP model (SNAP=Severe Nuclear Accident Program) of the Norwegian Meteorological Institute. In SNAP a constant dry deposition velocity (=0.2 cm/s) neglecting aerodynamic and surface resistances, is presently used. Therefore, two new dry depositions schemes (the Emerson scheme and the EMEP (European Monitoring and Evaluation Programme) scheme) have been implemented in SNAP to evaluate the benefits of including aerodynamic and surface resistances codes with respect to model prediction skills. The three dry deposition schemes are evaluated using 137Cs total deposition from soil sample data (n = 540) for a 60 km radial zone out from the Chernobyl Nuclear Power Plant (ChNPP) collected during the months after the accident. The present study capitalizes on high resolution meteorological data (2.5 km horizontal resolution), a detailed land-use data set with 273 sub-classes and the hitherto most comprehensive source term description for the Chernobyl accident. Based on our findings it is recommended to replace the present simple SNAP scheme with the Emerson or EMEP dry deposition scheme.

Modelling the dynamic gas/particle partitioning process of semi-volatile organic compounds emitted from point sources: Quantitative analysis and impact assessment.

The deleterious impact of pollution point sources on the surrounding environment and human has long been a focal point of environmental research. When considering the local atmospheric dispersion of semi-volatile organic compounds (SVOCs) around the emission sites, it is essential to account the dynamic process for the gas/particle (G/P) partitioning, which involves the transition from an initial state to a steady state. In this study, we have developed a model that enables the prediction of the dynamic process for G/P partitioning of SVOCs, particularly considering the influence from emission. It is important to note that the dynamic processes of the concentrations of SVOCs in particle phase (CP) and in gas phase (CG) differ significantly. These differences arise due to the influence of two critical factors: particulate proportion of SVOCs in the emissions (ϕ0) and octanol-air partitioning coefficient (KOA). The validity of our model was assessed by comparing its predictions of the extremum value of the G/P partitioning quotient (KP) with the results obtained from the steady-state model. Remarkably, the characteristic time (tC), used to evaluate the timescale required for SVOCs to reach steady state, demonstrated different variations with KOA for CP and CG. Additionally, the values of tC were quite different for CP and CG, which were markedly influenced by ϕ0. For some SVOCs with high KOA values, it took approximately 35 h to reach steady state. Furthermore, it was found that the time to achieve 95 % of steady state (t95 ≈ 3tC) could reach approximately 105 h. This duration is sufficient for chemicals to disperse from their emission site to the surrounding areas. Therefore, it is crucial to consider the dynamic process of G/P partitioning in local atmospheric transport studies. Moreover, the influence of ϕ0 should be incorporated into future investigations examining the dynamic process of G/P partitioning.

Source term estimation for continuous plume dispersion in Fusion Field Trial-07: Bayesian inference probability adjoint inverse method.

In scenarios involving sudden releases of unidentified gases or concealed pollution emergencies, source control emerges as a critical procedure to safeguard residential air quality. Appropriate inverse source tracking methodology depending on diverse measurement data could be utilized to promptly identify pollutant source parameters. In this study, source term estimation (STE) method, i.e., jointly combining probability adjoint method with the Bayesian inference method, has been proposed. General form of the pollutant inverse transport equation was firstly established. Subsequently, the pollution source information, assumed from single continuous point releases during Fusion Field Trials 2007 under an unsteady wind field, was identified using the Bayesian inference probability adjoint inverse method. Metropolis-Hastings Markov Chain Monte Carlo (MH-MCMC) and Differential Evolution Markov Chain Monte Carlo (DE-MCMC) were then compared as sampling methods for Bayesian inference. Results indicated that the DE-MCMC algorithm has superior convergence and could present higher accuracy of pollutant source information than that of MH-MCMC algorithm, particularly for highly nonlinear and multi-modal distribution systems. Furthermore, the integration of Union standard Adjoint Location Probability (UALP) as prior information into the Bayesian inference probability adjoint inverse method effectively narrowed the sampling range, enhancing both the accuracy and robustness of the proposed approach. Finally, the impact of the covariance matrix on the inverse identification accuracy was explored. Overall, this research has provided insights into the future applicability of this Bayesian inference inversion technique for point source identification.

Multi-scenario validation of the robust inversion method with biased plume range and values.

Release rate estimation is a vital means of revealing the emission process of radionuclides and assessing the environmental consequences in an emergency. Inverse modeling is widely used in emergency cases, but is vulnerable to plume biases in atmospheric dispersion modeling. One promising solution is a model called "Simultaneously Estimates the Release rate And Corrects both the plume range and Transport pattern" (SERACT). This study investigates the feasibility and behavior of SERACT based on four wind tunnel experiments replicating complex dispersion scenarios with both dense buildings and heterogeneous topography. SERACT's performance is compared with that of Tikhonov-regularized inversion and its predecessor. The results demonstrate that SERACT successfully corrects the modeled plume biases and simultaneously improves the release rate estimations in all four complex local-scale scenarios. The release rates retrieved by SERACT provide better agreement with the true release rates than those given by the other methods for all scenarios, with an average deviation of only 5.83%. After correction, the simulated plume reproduces the concentrations at all sites and achieves the best Pearson correlation coefficient (1.00) and fraction of simulations within a factor of 2 of the measurements (1.00); these values are 7.33 and 2.09 times higher, respectively, than those of simulations using release rates obtained using Tikhonov-regularized inversion.

Spatial and temporal distribution of endotoxins, antibiotic resistance genes and mobile genetic elements in the air of a dairy farm in Germany.

Antimicrobial resistance (AMR) is a serious issue that is continuously growing and spreading, leading to a dwindling number of effective treatments for infections that were easily treatable with antibiotics in the past. Animal farms are a major hotspot for AMR, where antimicrobials are often overused, misused, and abused, in addition to overcrowding of animals. In this study, we investigated the risk of AMR transmission from a farm to nearby residential areas by examining the overall occurrence of endotoxins, antibiotic resistance genes (ARGs), and mobile genetic elements (MGEs) in the air of a cattle farm. We assessed various factors, including the season and year, day and nighttime, and different locations within the farm building and its vicinity. The most abundant ARGs detected were tetW, aadA1, and sul2, genes that encode for resistances towards antibiotics commonly used in veterinary medicine. While there was a clear concentration gradient for endotoxin from the middle of the farm building to the outside areas, the abundance of ARGs and MGEs was relatively uniform among all locations within the farm and its vicinity. This suggests that endotoxins preferentially accumulated in the coarse particle fraction, which deposited quickly, as opposed to the ARGs and MGEs, which might concentrate in the fine particle fraction and remain longer in the aerosol phase. The occurrence of the same genes found in the air samples and in the manure indicated that ARGs and MGEs in the air mostly originated from the cows, continuously being released from the manure to the air. Although our atmospheric dispersion model indicated a relatively low risk for nearby residential areas, farm workers might be at greater risk of getting infected with resistant bacteria and experiencing overall respiratory tract issues due to continuous exposure to elevated concentrations of endotoxins, ARGs and MGEs in the air of the farm.

Effects of Shape and Size on Microplastic Atmospheric Settling Velocity.

Microplastics (MPs) have been found in all terrestrial, marine, and riparian environments, including remote regions. This implies that atmospheric transport is an important pathway when considering MP sources and global budgets. However, limited empirical data exist to aid in effective development and parameterization of MP atmospheric transport models. This study measured the atmospheric settling and horizontal drift velocities of various sizes and shapes of MPs in two specially designed settling columns using a laser Doppler anemometer. The settling velocities were generally lower than modeled values, while shape exerted the most significant influence on the rate of settling. Rather than conforming to well-established, power-law models, each class of MP exhibited a linear but different relationship between MP size and settling velocity, with markedly higher slopes for the spheres and cylinders as compared to the films and fibers. Shape also had a substantial influence on particle drift, with the fibers and films exhibiting the greatest horizontal motion, as suggestive of their changing orientation in response to particle interactions and fluid drag. As a consequence, microplastic particles identified within atmospheric deposition samples collected at a single point may derive from entirely different sources representing a wide range in transport distance.

Improved Gaussian plume model for atmospheric dispersion considering buoyancy and gravitational deposition: The case of multi-form tritium.

Various types of radionuclides have different atmospheric dispersion characteristics, such as buoyancy and gravitational deposition phenomenon of light gas and heavy particles, respectively. Gaussian plume model was widely used to describe atmospheric dispersion behaviors of radioactive effluents, particularly for the purpose of engineering environmental impact assessment or nuclear emergency support. Nonetheless, buoyancy and gravitational deposition were rarely reported in previous work for tritium in particular, which might cause a deviation in evaluating near-surface concentration distribution and radiation dose to the public. Based on the multi-form tritium case, we made a quantitative description for the buoyancy and gravitational deposition phenomenon and discussed the feasibility of developing an improved Gaussian plume model to predict near-surface concentration distribution. Firstly, tritium concentration distribution near to the surface was predicted by using computational fluid dynamics method (CFD) and standard Gaussian plume model to reach consistency without consideration of buoyancy and gravitational deposition effects. Secondly, effects of buoyancy and gravitational deposition were identified by species transport model for gaseous tritium and discrete phase model for droplet tritium with integrating the buoyancy force caused by density variation of gaseous tritium and gravitational force of droplet tritium with enough size. Thirdly, buoyancy and gravitational deposition correction factors were obtained to modify the standard Gaussian plume model. Lastly, predictive results by improved Gaussian plume model were compared with CFD method. It was proved the improved correction method enables higher accuracy in predicting the atmospheric concentration distribution of gaseous pollutants with density variation or particles with gravitational deposition properties.

Accidental release of ammonia during ammonia bunkering: Dispersion behaviour under the influence of operational and weather conditions in Singapore.

Ammonia is identified as a potential marine fuel, and ammonia bunkering will take place in major bunkering ports, including Singapore. Due to its toxic nature, any accidental release of ammonia into the environment during bunkering operation has a risk of spreading rapidly and causing injury to the personnel in the vicinity and damage to the marine ecosystem. This safety study simulates how key operational parameters affect ammonia dispersion and the consequences. The results show that bunkering ammonia stored in fully-refrigerated tanks as an atmospheric pressure saturated liquid is the safest, and the severity of the consequence increases significantly with a release height of more than 5 m. A vertical release of ammonia results in the most severe consequence and shall be avoided at all times. Reducing the release duration and transfer flow rate can reduce the severity significantly. Based on the scenario used in this study, ammonia cloud disperses over a longer distance over water due to the high vaporisation rate driven by the large amount of heat generated from the dissolution of ammonia in seawater. The dispersion of ammonia over the sea spreads over a larger area during the day than at night.

Localisation of atmospheric release of radioisotopes using inverse methods and footprints of receptors as sources.

Releases of radionuclides to the atmosphere occasionally occur with no warning and with first observation at radioactivity monitoring stations. The Chernobyl accident of 1986 was first detected at Forsmark, Sweden, long before the official announcement by the Soviet Union, and the release of Ruthenium 106 detected across Europe in 2017 still has no official release location. The current study details a method based on footprint analysis of an atmospheric dispersion model to locate the source of an atmospheric release. The method was applied to the European Tracer EXperiment of 1994 to validate the method and to the Ruthenium observations of autumn 2017 to determine likely release locations and time characteristics of this release. The method can readily utilise an ensemble of numerical weather prediction data which improves the localisation results by taking into account meteorological uncertainties compared to only using deterministic weather data. In applying the method to the ETEX scenario, the most likely release location improved from a distance of 113 km from the true release location when using deterministic meteorology, to a distance of 63 km when using ensemble meteorology data, although such improvements may be scenario dependent. The method was constructed to be robust with respect to the choices of model parameters and measurement uncertainties. The localisation method can be useful for decision makers to enact countermeasures to protect the environment against the effects of radioactivity when observations are available from environmental radioactivity monitoring networks.

In-silico analysis of atmospheric diffusion, crop planting degrading scheme, and health risk of dioxins from a domestic waste incineration plant.

Based on a domestic waste incineration power generation project, the dioxin emission from the waste incineration plant (WIP), phytoextraction and microbial degradation of dioxins, and dioxins human health risks reduction were investigated through in-silico methods. Based on the dioxins concentrations in soil (9.97 × 10-9-7.00 × 10-5ng/g) predicted by atmospheric dispersion model system and the Level-III fugacity model, planting schemes under different wind directions were designed considering the dioxin absorption capacity and the economic benefits for crops (i.e., barley, peanut, pea, maize and wheat). The dioxins in soils can be further degraded by five crops' rhizosphere microorganisms and fertilizers, simulated through molecular dynamic simulations. The enhanced degradation rates of dioxin by rhizosphere microorganisms of five crops reached 15.70 %-28.66 %. Finally, healthy dietary plans were developed to reduce the risk of dioxin exposure to the sensitive populations living around WIP. Results showed that the consumption of maize, fungus, mushroom and bamboo fungus could effectively reduce dioxins toxicity to humans by 58.13 %. The systematic approach developed in this study provided theoretical support for soil remediation and human health risk control of dioxins-contaminated sites.

Bayesian inference and wind field statistical modeling applied to multiple source estimation.

We present a methodology to identify multiple pollutant sources in the atmosphere that combines a data-driven dispersion model with Bayesian inference and uncertainty quantification. The dispersion model accounts for a realistic wind field based on the output of a multivariate dynamic linear model (DLM), estimated from measured wind components time series. The forward problem solution, described by an adjoint transient advection-diffusion partial differential equation, is then obtained using an appropriately stabilized finite element formulation. The Bayesian inference tool accounts for uncertainty in the concentration data and automatically states the balance between the prior and the likelihood. The source parameters are estimated by a Metropolis in Gibbs Monte Carlo Markov chain (MCMC) algorithm with adaptive steps. The MCMC algorithm is initialized with a maximum a posteriori estimator obtained with particle swarm optimization to accelerate convergence. Finally, the proposed methodology seems to outperform inversion techniques from previous works.

Assessing Detection Efficiencies for Continuous Methane Emission Monitoring Systems at Oil and Gas Production Sites.

Continuous monitoring systems, consisting of multiple fixed sensors, are increasingly being deployed at oil and gas production sites to detect methane emissions. While these monitoring systems operate continuously, their efficiency in detecting emissions will depend on meteorological conditions, sensor detection limits, the number of sensors deployed, and sensor placement strategies. This work demonstrates an approach to assess the effectiveness of continuous sensor networks in detecting infinite-duration and fixed-duration emission events. The case studies examine a single idealized source and a group of nine different sources at varying heights and locations on a single pad. Using site-specific meteorological data and dispersion modeling, the emission detection performance is characterized. For these case studies, infinite-duration emission events are detected within 1 h to multiple days, depending on the number of sensors deployed. The percentage of fixed-duration emission events that are detected ranged from less than 10% to more than 90%, depending on the number of sources, emission release height, emission event duration, and the number of sensors deployed. While these results are specific to these case studies, the analysis framework described in this work can be broadly applied in the evaluation of continuous emission monitoring network designs.

Inversion of 137Cs emissions following the fukushima accident with adaptive release recovery for temporal absences of observations.

Temporal absences in observation records lead to release losses during the source term inversions of atmospheric radionuclide emissions. Consequently, objectively-estimated source terms for the Fukushima accident contain fewer release details and present large discrepancies when compared with the expert-judged one. This paper describes an objective method that can adaptively recover the missing releases caused by the temporal absences of observations. The proposed method assumes that the accident releases of radionuclides are piecewise-constant and comprise both peaks and constant releases. The missing releases are adaptively recovered as either peaks or constant releases by minimizing the total variation of the estimated source term. The proposed method is applied to the Fukushima accident and evaluated against regional airborne and deposited 137Cs observations. The results demonstrate that this method effectively recovers the missing releases, producing a source term that matches the timing of both on-site gamma dose rate peaks and accident events. The retrieved source term improves the simulation of air concentrations and reproduces most of the deposition patterns. This is the first time that an objective method has independently reproduced the details in the expert-judged one for the Fukushima accident.

Combined effects of vehicles and waste incineration on urban air halogenated and parent polycyclic aromatic hydrocarbons.

Traffic emissions and waste incineration are the main sources of PAHs in urban atmosphere, but their spatially superimposed effects are currently unclear. This study assessed the spatial distribution of PAHs and HPAHs concentrations in the atmosphere of Shenzhen by simulating the spatial and temporal dispersion of PAHs and HPAHs emissions from on-road vehicles and municipal solid waste incinerators (MSWIs). Generally, the concentrations of PAHs and HPAHs were higher on workdays than on weekends due to higher traffic volumes, while the prevailing wind direction of the northeast could cause more widespread dispersion of PAHs and HPAHs within Shenzhen's atmosphere. After superimposing the spatial distribution of pollutants emitted by vehicles and MSWIs, PAHs within 1000 m downwind of MSWIs are mainly contributed by MSWIs and beyond 1000 m by vehicles. The cancer risk values induced by exposure to PAHs and HPAHs via inhalation in Shenzhen were below the acceptable risk level for males and females in each age group, while adults faced the highest cancer risk, followed by adolescents and children. However, spatially, the cancer risk values were above the priority risk level for adult males in localized high-traffic areas in Futian and Luohu districts.

Flux of tritium from the sea to the atmosphere around a nuclear reprocessing plant: Experimental measurements and modelling for the Western English channel.

Tritium is released to the environment by nuclear industries in various forms, mainly HTO. In impact studies leading to estimated doses for the population, atmospheric discharges are mainly taken into consideration because they generally lead to values higher than those related to liquid discharges. However, the tritium released in liquid environments can be transferred to the atmosphere by evaporation and then be transported to terrestrial ecosystems by wind. This study was carried out in France near a fuel reprocessing plant (RP) which discharges tritium into the western English Channel. We highlighted the influence of a mass of water enriched with tritium on the HTO levels in atmospheric water vapour downwind through 18 field campaigns. A hydrodynamic model able to simulate tritium activity in the water was coupled with an evaporation an atmospheric transport model. It allows to reconstitute variations in atmospheric tritium on the coast, depending on liquid discharges of tritium from the reprocessing plant. On this basis, when seawater containing 20-100 Bq.L-1 of tritium flows between 0 and 10 km off the coast, variations in atmospheric activity onshore can increase of 2-15 Bq.L-1. Mean tritium quantities released by the sea into the atmosphere in the Western English Channel reached 130 TBq.y-1 over the 2017-2020 period. Emissions were estimated at 0.9-11.3 GBq km-2.y-1 and depends principally on the distance from the liquid discharge point. If we compare the "marine" source term, in HTO form, with the direct source term for gaseous discharges, the marine source term is one order of magnitude greater for the marine region affected by liquid discharges. Finally, we estimate that approximately 1.1% of tritium stock discharged at sea (regulated and controlled) return to the atmosphere each year at the scale of the Western English Channel.

Insights into near-surface distribution characteristics of multi-form tritium with consideration of atmospheric buoyancy and gravitational deposition.

Tritium contributes majority to the total airborne radioactive effluents from the nuclear facility because of its considerable production and difficulty in separation. Tritium inventory in the fusion reactor would reach an unprecedented magnitude which brings new safety concern. After being released into the atmosphere, inconsistent atmospheric dispersion behaviors might appear regarding different physicochemical forms such as gaseous state HT, gaseous-aerosol-droplet state HTO. In this study, atmospheric dispersion characteristics of multi-form tritium were investigated based on the computational fluid dynamics method validated by multi-fan type wind tunnel experiments. Species transport model and discrete phase model were used to describe atmospheric dispersion of gaseous and aerosol-droplet state tritium, respectively. Deposition velocity was predicted for gaseous and aerosol-droplet state tritium with different particle sizes. Conditions for describing the changes of particle diameter and its influencing on near-surface tritium distribution due to condensation were provided. The results show that buoyancy effect would strengthen along with the increasing gaseous tritium mass fraction in the airborne effluents. We also indicated that obvious gravitational deposition would appear once gaseous HTO was transformed into droplet state HTO with the particle diameter larger than 20 µm. Both the atmospheric buoyancy and deposition phenomenon would result in a quite different near-surface tritium distribution.

Dispersion of arsenic species from highly explosive historical volcanic eruptions in Patagonia.

Andean volcanic rocks typically have low to moderate arsenic (As) concentrations. However, elevated levels of As in groundwaters of southern South America have been reported as a consequence of weathering of volcanic glass. This study discusses the abundance, speciation and dispersion of As species in fresh volcanic ash from highly explosive (Volcanic Explosivity Index: 4-5) Patagonian eruptions, as well as the potential of As release to aqueous reservoirs. Synchrotron-based X-ray absorption and micro-focused X-ray photoelectron spectroscopies were used to evaluate As solid speciation. Batch experiments at different pH conditions were performed with the aim of understanding the controls on As release to aqueous reservoirs. Bulk chemical and mineralogical characterizations were performed by inductively coupled plasma optical emission spectroscopy, X-ray diffraction and scanning electron microscopy/energy dispersive spectroscopy. Finally, to understand how As-bearing phases are spatially distributed after eruptions, simulations of volcanic ash emission, transport and deposition were performed. Results indicate that the concentration, speciation, and mobility of As in fresh Patagonian volcanic ash depend on the silica content of source magmas. Although the main As host in volcanic ash is Al-silicate glass, this phase is stable at neutral pH characteristic of most aqueous reservoirs. Higher contributions of As to water are associated with the more mobile As species that concentrate onto the surface of Al-silicate glass. Atmospheric dispersion simulations revealed that primary fallout of As-bearing ash has affected large areas in Patagonia, but also reached the Chaco-Pampean plain, where the presence of As-rich groundwater has been widely documented.

Near-range atmospheric dispersion of an anomalous selenium-75 emission.

On May 15th of 2019, an anomalous emission of selenium-75 was detected at the stack of the Belgian Reactor 2 (BR2) in Mol. Although the release exceeded the prescribed limits for BR2, there was no harm to the population or food chain and so the event was classified as INES 1. However, it was very interesting from the perspective of near-range atmospheric dispersion. This publication brings together all available information at the near range - stack-monitored source term data, on-site meteorological data, ambient dose equivalent rates, deposition and concentration measurements - in one dataset that covers the initial puff release on May 15th, 2019 as well as a residual release through early November 2019. To demonstrate the consistency of this dataset, supporting atmospheric transport and dispersion calculations are performed using a Gaussian plume model. These calculations, in turn, are supported by near-range Flexpart calculations in the context of the residual release.

Air exchange rate and pollutant dispersion inside compact urban street canyons with combined wind and thermal driven natural ventilations: Effects of non-uniform building heights and unstable thermal stratifications.

In the present work, a delicate CFD research of a multi-street canyon model with varying thermal stratifications and non-uniformities of buildings was conducted to investigate the street ventilation and pollutant dispersion between the compact urban blocks. Non-isothermal turbulent wind flow, temperature field and pollutant dispersion in a two-dimensional computational domain were solved by the Renormalization Group (RNG) k-ε turbulence model along with the enhanced wall treatment. Present numerical results indicated that the variation of ground heating intensity has a significant influence on the airflow pattern in the step-down case, and the distribution of pollutants in the street canyons mainly depends on the variation of the upper clockwise vortex. The canyon ventilation performance became better as the unstable thermal stratification strengthened. Similarly, the increase of ground heating intensity could reduce ADF (atmospheric dispersion factor) in the step-down case and ADF became the lowest when Ri = -3.92 was maintained. Additionally, the increase of building unevenness further complicated the canyon airflow structure, which aggravated the pollution of the canyon. In the step-down configuration, as the standard deviation of adjacent building height gradually increases, canyon ventilation could be further enhanced. For the step-up configuration, the best ventilation performance was found at σH = 16.7 %. ADF of adjacent canyons also varied greatly. When σH = 33.3 % was maintained, the peak and bottom values of ADF were discovered in the step-up and step-down cases, respectively. Present research has provided a theoretical reference for guiding urban design and improve living environment in modern compact cities.