Based on the ground-level observed NO₂ concentration, satellite-observed NO₂ column concentration from the Ozone Monitoring Instrument (OMI) and meteorological parameters, we comprehensively consider the seasonal and regional differences in the relationship between NO₂ column concentration and measured NO₂ concentration and establish a two-stage combined ground NO₂ concentration estimation (TSCE-NO₂) model using a support vector machine for regression (SVR) and a genetic algorithm optimized back propagation neural network (GABP). On this basis, the spatial-temporal variation in the modelled ground-level NO₂ concentration over China during the period of 2005–2019 was analysed. The results show that the TSCE-NO₂ model proposed in this study provides a reliable estimation of the modelled ground-level NO₂ concentration over China, effectively filling the spatial and temporal gaps in China's air quality ground monitoring network (the model's correlation coefficient, R, is 0.92, the mean absolute error, MAE, is 3.62 μg/m³, the mean square percentage error, MSPE, is 0.72%, and the root-mean-square error, RMSE, is 5.93 μg/m³). The analysis results of the spatial and temporal variation indicate that (1) the perennial ground-level NO₂ concentration over China is high in the eastern area and low in the western area, and the high values are mainly distributed along the northern coast, the eastern coast, the middle reaches of the Yangtze River, the middle reaches of the Yellow River, the Pearl River Delta and the Sichuan Basin. (2) The modelled ground-level NO₂ concentrations over China are highest in winter, followed by those in autumn and spring, and they are lowest in summer. Before 2011, the ground-level NO₂ concentration over China increased at a rate of 0.348 ± 0.132 μg/(m³∙a) but decreased at a rate of 0.312 ± 0.188 μg/(m³∙a) after 2011. (3) From 2011 to 2019, measures such as energy savings and emission reductions alleviated NO₂ pollution on the premise of ensuring sustained China's GDP growth.

Land-use and land-cover changes due to urban expansion is recognized as one of the crucial factors affecting carbon dioxide emissions. In this study, the land conversion effects on soil CO₂ fluxes associated with temperate re-established grasslands within the Forest Botanical Garden found on an anthropogenic landform were investigated. The present work analyses the capabilities and requirements of the CALPUFF Lagrangian puff air quality modelling system to simulate the spatial distribution of ecosystem respired CO₂ in the urban domain. The results are validated against the available observations of CO₂ fluxes in the urban environment using the closed-chamber method with the measurements of the stable carbon isotope ratio (δ¹³C) of daytime soil-respired CO₂. The isotope mass balance partitioning approach was applied to distinguish biogenic portions of CO₂ from the admixture of atmospheric air.The spatial and temporal amplitude of the simulated CO₂ concentrations from the CALPUFF model showed considerable agreement with the tracer measurements of the biogenic CO₂ component in the near-ground air (0.25 m). In most cases, however, the CALPUFF predictions of ecosystem-derived CO₂ showed a general tendency toward considerable underestimation of real concentration levels. Such discrepancies are related to the difficulties associated with the optimization of biospheric CO₂ flux and uptake from ecosystems by means of local-scale modelling. The modelled results implied that the CALPUFF performance in the dispersion simulation of CO₂ concentrations within the urban ecosystems is very sensitive to the initial meteorological conditions, grid resolution, measurement timescale, and the calculated gas flux rate from soils. A significant negative correlation was found between hourly values of the average modelled CO₂ and observed wind speed during the entire study campaign (r = −0.58 and ρ = −0.82 for Pearson and Spearman statistics, respectively, p < 0.05). Moreover, analysis of the impact of the deposition parameters on changes in the atmospheric concentration of carbon dioxide indicated significant dependency of the temporal CO₂ distribution patterns on the precipitation-based events. According to the obtained estimates, the wet deposition rate during rain events was approximately two orders higher than the average dry deposition flux.Overall, the present case study indicates that the CALPUFF model has a rather acceptable predictive ability. A better agreement of model predictions and all field measurements, however, require further studies of CO₂ exchange between the ecosystem and atmosphere and understanding where they need to be improved.

Chronic atmospheric nitrogen (N) deposition could influence the functioning of ecosystems as well as their biodiversity. However, N deposition in urban forest ecosystems, especially natural evergreen broad-leaved forests, is not well known. In this study, the concentrations and fluxes of dissolved inorganic N (DIN) and dissolved organic N (DON) in bulk deposition, throughfall, and stemflow were assessed in an urban evergreen broad-leaved forest site over three years, in order to clarify the characteristics of N deposition. At the study site, bulk DIN deposition was 3.7 kg N ha⁻¹ year⁻¹ (1.5 kg N ha⁻¹ year⁻¹ for NH₄–N and 2.2 kg N ha⁻¹ year⁻¹ for NO₃ + NO₂–N), which is the same level as that found in rural areas. In contrast, 6.5 kg N ha⁻¹ year⁻¹ for bulk DON deposition contributed to 66% of the bulk N deposition, which suggests the importance of bulk DON deposition in Japanese forest ecosystems. Passing through the tree canopy, DIN was enriched by 8.8 kg N ha⁻¹ year⁻¹ (3.7 kg N ha⁻¹ year⁻¹ for NH₄–N and 5.1 kg N ha⁻¹ year⁻¹ for NO₃ + NO₂–N) and DON was enriched by 1.5 kg N ha⁻¹ year⁻¹ as net throughfall in the evergreen broad-leaved forest. This reveals that dry deposition of DIN dominates the total DIN deposition onto the urban forest floor, compared to that found in the rural areas, due to the non-negligible N emissions from outside and possibly because of the evergreen broad-leaved forest's greater ability to capture N.

Well integrity failure resulting in migration of natural gas outside of the surface casing can cause atmospheric greenhouse gas emissions and groundwater quality impacts from existing and historic energy wells. Spatial and temporal variability in gas migration can result in errors in detection (i.e., presence/absence) and efflux estimations. This field-based case study used automated dynamic closed chambers to record repeated (~every 18 min) CO₂ and CH₄ efflux measurements over a two-week period around a single petroleum production well in Alberta, Canada. Long-term efflux measurements supplemented soil gas compositional and isotopic characterization, along with surface concentration measurements. Effluxes were spatially concentrated around the wellhead and only occasionally detectable more than a few meters away. Estimated total emissions attributable to gas migration ranged from 48 to 466 g CH₄ d⁻¹ (or 0.07–0.7 m³ CH₄ d⁻¹). Methane effluxes and concentrations were temporally variable on second-to-hourly and diel scales. Multivariate stepwise regression analysis indicates that multiple meteorological factors, particularly wind speed and air temperature, were related to the temporal variability. Despite temporal variability, elevated concentrations and effluxes were consistently detectable around the well. Major soil gas composition suggests that gas migration near the wellhead causes advective displacement of soil gas, while more distal measurements are indicative of episodic and diffusion-dominated transport. Values of ¹³C–CO₂ and ¹³C–CH₄ samples were consistent with CH₄ oxidation within the unsaturated zone. Although these results reflect a single well, the findings are salient to gas migration detection and emission estimation efforts.

Accuracy prediction of air quality is of crucial importance for people to take precautions and improve environmental conditions. By introducing adjacency matrix in Long Short-Term Memory (LSTM) cell, we propose in this research a Graph-based Long Short-Term Memory (GLSTM) model to predict PM2.5 concentration in Gansu Province of Northwest China. We regard all air quality monitoring stations as a graph, and construct a parameterized adjacency matrix on the basis of the adjacency matrix of the graph. Through the combination of parameterized adjacency matrix and LSTM, we introduce spatiotemporal information to achieve PM2.5 prediction. The advantage of GLSTM is that it can realize synchronous operation of all stations, making it unnecessary to train different model for each monitoring station to obtain the overall PM2.5 variation of a certain area. The parameterized adjacency matrix also enhances the interpretability of the model. By visualizing the parameterized adjacency matrix obtained from the end-to-end PM2.5 prediction task in training, the importance of introducing spatial information, i.e. the distribution importance of surrounding stations to a specific station is clearly demonstrated. We compared our model with several newly reported methods, and found that it achieved the best results on PM2.5 prediction tasks at almost all stations, which proved the effectiveness of the GLSTM model.

A source apportionment study was carried out at four sites in Emilia-Romagna region, southern Po Valley, one of the most critical regions in Europe in terms of atmospheric pollution. PM₂.₅ daily samples were collected during 4 years from April 2013 to October 2017 at one rural site (San Pietro Capofiume) and three urban background locations in the cities of Bologna, Rimini, Parma which show different features and are located in the central, coastal and inner part of the investigated region. Samples were analyzed to achieve a complete chemical characterization (carbon fractions, ions, and elements). A source apportionment analysis by Positive Matrix Factorization (PMF) was performed and 6 PM₂.₅ factors were identified at all sites but the rural one (where 5 out of 6 of them were detected); the factors were associated to traffic with dust resuspension, biomass burning, oil combustion/ship emission, mix anthropogenic (not found at the rural site), ammonium nitrate and ammonium sulfate with organics. Chemical profiles of factors were very similar among all the 4 sites, indicating that main pollution sources are basically the same at the 4 sites, while some differences emerged with regard to source contributions. Factors related to secondary components seem to explain almost 50% or even more of PM₂.₅ mass concentration in all seasons. Traffic and biomass burning are the most relevant contributors to PM₂.₅ in terms of primary components. A not negligible contribution of biomass burning results in Rimini during the summer, suggesting other possible sources of wood combustion, such as cooking or open burning of agricultural pruning bonfires. Agriculture is not singled out as a PMF factor, but a rough estimate based on ammonium concentrations and ammonia data from emission inventory indicates a contribution from this source of about 10% of PM₂.₅ mass, thus resulting the single productive activity with the highest impact on PM₂.₅ at the investigated sites. Back trajectory analysis points out the relevant extra-regional contributions of two factors; indeed, oil combustion/ship emission is related to long-range transport of air masses overpassing the Mediterranean sea and secondary sulfate from Eastern Europe countries occasionally impacts on the Po Valley.

In Belgium, 16 shredding facilities manage annually tens of thousands tons of wastes from different origins (end-of-life vehicles, electronic waste, electrical transformers, …). These materials contain hazardous persistent organic pollutants such as polychlorinated dibenzo-dioxins/furans (PCDD/Fs), polychlorinated biphenyls (PCBs) and polybrominated diphenyl ethers (PBDEs). The shredding process promotes the production and the emission of dust contaminated by these compounds. The objective of this study is to measure the concentrations of PCDD/Fs, PCBs and PBDEs in fallout dust collected in the vicinity of 3 shredding plants located in Wallonia (French speaking part of Belgium). Samples were collected by using Owen gauges and pollutant levels were measured by GC-MS. The median deposition levels measured for ∑PCDD/Fs, ∑dioxin-like PCBs, 5 × ∑6 DIN PCBs and PBDE 209 were 1.9 pg TEQ/m².day, 4.4 pg TEQ/m².day, 246.5 ng/m².day and 253.8 ng/m².day, respectively. These levels represent high concentrations compared to those observed in most of the remote, rural and urban areas studied around the world and were similar to those measured in other heavily industrialized districts. Consequently, the health effects of this high exposure to pollutants among workers and residents in the vicinity of these shredding facilities are of concern.

Cities located on Eastern Mediterranean is exposed to both local and distant anthropogenic and natural sources. In this study, to determine the effect of these sources on Particulate Matter (PM) concentrations in a coastal city, particulate matter with diameters less than 2.5 μm (PM2.5) and with diameters between 2.5 and 10 μm (PM2.5−10) were collected once in a two-day period for 24 h between July 2014 and July 2015 in downtown Antalya which is located on the Mediterranean coast of Turkey. Antalya is one of the fast growing city of Turkey with a population of 2.3 million. Samples were analyzed using an energy-dispersive X-ray fluorescence for 15 elements (Na, Mg, Al, Si, S, K, Ca, Ti, V, Mn, Fe, Cu, Zn, As, Pb). Statistical parameters were calculated for all measured elements in fine (PM2.5) and coarse (PM2.5−10) fraction. Crustal and marine elements, such as Ti, Ca, Al and Na were abundant in the course fraction. Only S was found in higher concentration in the fine fraction. Monthly variation of Crustal Enrichment Factor (EFC) results of Si showed that the area was under influence of non-local crustal dust especially during spring and late summer. EFC also indicated that during winter season, fine fraction K was due to local wood combustion. Source regions of S was determined using Potential Source Contribution Function (PSCF) and compared with previous studies conducted at a rural site of Antalya approximately twenty years ago. Most of the source regions affecting S concentrations at the Eastern Mediterranean region were found out to be same: western Anatolia, Marmara region, the Aegean Sea coasts of Greece and some parts of Bulgaria and Romania. However, due to decrease in SO2 emissions over the northeast coast of Black Sea and between Caspian Sea and Ukraine, the region was not turned up to be a source region.

The inappropriate combustion of furniture-industry waste can be a source of serious environmental problems. We proposed a method which is capable to distinguish the emission resulting from the combustion of wood-based materials, the essential component of such waste. The originality of the approach consists in the classification of gas mixtures instead of focussing on the individual pollutants emitted to the atmosphere. The classification of emission was based on the measurements applying differential ion mobility spectrometry (DMS) and Fourier transform infrared (FTIR) spectrometry, for comparison. There were successfully distinguished emissions associated with combustion of wood-based materials: OSB board, MDF board and plywood (≥95% correct classifications in the class (ccc)) and wood: pellet and kindling wood (≥92% ccc). Results of classification based on DMS and FTIR measurements were similar. Emissions from the combustion of individual materials were best distinguished using DMS (100% ccc), as compared with FTIR, which offered lower performance, mostly >90% ccc. Regarding pairwise classification, the most distinctive were emissions from the combustion of plywood (DMS: 99% ccc; FTIR: 99% ccc), MDF board (DMS: 99% ccc; FTIR: 99% ccc) and OSB board (DMS: 99% ccc; FTIR: 98% ccc). Emissions form kindling wood (DMS: 100% ccc; FTIR: 95% ccc) and pellet (DMS: 97% ccc; FTIR: 98% ccc) caused a bit more confusion. In most cases, results of classification based on DMS and FTIR measurements were comparable. The success of classification based on DMS measurements proved that it is possible to detect the harmful emission without determining the chemical composition of the flue gas. This solution represents a new approach to air quality monitoring, which recently attracts increasing attention.

Solar eclipse (with maximum obscuration of 85.3% and magnitude of 0.893) occurred on 26 December 2019 during morning hours (08:10 to 11:15 LT with a peak at 09:33 LT) over Gadanki (13.5ᵒN, 79.2ᵒE) has provided a unique opportunity to test the hypothesis of ‘Extended Fumigation Effect’ or ‘Second Fumigation’ on the surface and boundary layer pollutants. To capture this event, a campaign using multi-instrument (AWS, Aethalometer, PM sensors, ceilometer, radiosonde) on multi-platform (surface, surface based remote sensing, drone, tethered balloon, in-situ balloon) was conducted. Eclipse obscuration caused decrease in surface temperature by 4.3 °C around 10:00 LT. Boundary layer remained shallow until 09:00 LT (between 500 m and 900 m) but near the termination of the eclipse and soon after the termination a convective boundary layer showed a rapid increase to above 1 km within a short time (1 h). A Fumigation peak (common phenomenon in normal days) in black carbon occurred with a sharp peak concentration of 9.4 μg/m³ at around 07:00 LT and then started decreasing. However, concentration started to increase unusually again at around 08:20 LT and remained at the range of 4–6 μg/m³instead of a normal decreasing trend, which is about 2–3 times of the mean concentration at this period of time. Similar variation in PM₁, PM₂.₅, and PM₁₀are also observed. Background instability estimated using radiosonde measurements suggests Fumigation, Fumigation/Lofting and Trapping before, during and after the eclipse, respectively.