Search

As air pollution is regarded as the single largest environmental health risk in Europe it is important that communication to the public is up to date and accurate and provides means to avoid exposure to high air pollution levels. Long- and short-term exposure to outdoor air pollution is associated with increased risks of mortality and morbidity. Up-to-date information on present and coming days' air quality helps people avoid exposure during episodes with high levels of air pollution. Air quality forecasts can be based on deterministic dispersion modelling, but to be accurate this requires detailed information on future emissions, meteorological conditions and process-oriented dispersion modelling. In this paper, we apply different machine learning (ML) algorithms – random forest (RF), extreme gradient boosting (XGB), and long short-term memory (LSTM) – to improve 1, 2, and 3 d deterministic forecasts of PM10, NOx, and O3 at different sites in Greater Stockholm, Sweden. It is shown that the deterministic forecasts can be significantly improved using the ML models but that the degree of improvement of the deterministic forecasts depends more on pollutant and site than on what ML algorithm is applied. Also, four feature importance methods, namely the mean decrease in impurity (MDI) method, permutation method, gradient-based method, and Shapley additive explanations (SHAP) method, are utilized to identify significant features that are common and robust across all models and methods for a pollutant. Deterministic forecasts of PM10 are improved by the ML models through the input of lagged measurements and Julian day partly reflecting seasonal variations not properly parameterized in the deterministic forecasts. A systematic discrepancy by the deterministic forecasts in the diurnal cycle of NOx is removed by the ML models considering lagged measurements and calendar data like hour and weekday, reflecting the influence of local traffic emissions. For O3 at the urban background site, the local photochemistry is not properly accounted for by the relatively coarse Copernicus Atmosphere Monitoring Service ensemble model (CAMS) used here for forecasting O3 but is compensated for using the ML models by taking lagged measurements into account. Through multiple repetitions of the training process, the resulting ML models achieved improvements for all sites and pollutants. For NOx at street canyon sites, mean squared error (MSE) decreased by up to 60 %, and seven metrics, such as R2 and mean absolute percentage error (MAPE), exhibited consistent results. The prediction of PM10 is improved significantly at the urban background site, whereas the ML models at street sites have difficulty capturing more information. The prediction accuracy of O3 also modestly increased, with differences between metrics. Further work is needed to reduce deviations between model results and measurements for short periods with relatively high concentrations (peaks) at the street canyon sites. Such peaks can be due to a combination of non-typical emissions and unfavourable meteorological conditions, which are rather difficult to forecast. Furthermore, we show that general models trained using data from selected street sites can improve the deterministic forecasts of NOx at the station not involved in model training. For PM10 this was only possible using more complex LSTM models. An important aspect to consider when choosing ML algorithms is the computational requirements for training the models in the deployment of the system. Tree-based models (RF and XGB) require fewer computational resources and yield comparable performance in comparison to LSTM. Therefore, tree-based models are now implemented operationally in the forecasts of air pollution and health risks in Stockholm. Nevertheless, there is big potential to develop generic models using advanced ML to take into account not only local temporal variation but also spatial variation at different stations.

A growing trend in developed countries is the use of wood as fuel for domestic heating due to measures taken to reduce the usage of fossil fuels. However, this imposed another issue with the environment and human health. That is, the emission from wood burning contributed to the increased level of atmospheric particulates and the wood smoke caused various respiratory diseases. The aim of this study was to investigate the impact of wood burning on the polycyclic aromatic hydrocarbons (PAHs) in air PM10 using known wood burning tracers, i.e. levoglucosan, mannosan and galactosan from the measurement at the urban background and residential areas in Sweden. A yearly measurement from three residential areas in Sweden showed a clear seasonal variation of PAHs during the cold season mainly from increased domestic heating and meteorology. Together, an increased sugar level assured the wood burning during the same period. The sugar ratio (levoglucosan/(mannosan+galactosan)) was a good marker for wood burning source such as the wood type used for domestic heating and garden waste burning. On the Walpurgis Night, the urban background measurement demonstrated a dramatic increase in levoglucosan, benzo[a]pyrene (B[a]P) and oxygenated PAHs (OPAHs) concentrations from the increased wood burning. A significant correlation between levoglucosan and OPAHs was observed suggesting OPAHs to be an indicator of wood burning together with levoglucosan. The levoglucosan tracer method and modelling used in predicting the B[a]P concentration could not fully explain the measured levels in the cold season. The model showed that the local wood source contributed to 98 % of B[a]P emissions in the Stockholm area and 2 % from the local traffic. However, non-local sources were dominating in the urban background (60 %). A further risk assessment estimated that the airborne particulate PAHs caused 13.4 cancer cases per 0.1 million inhabitants in Stockholm County.

Air pollution concentrations have been decreasing in many cities in the developed countries. We have estimated time trends and health effects associated with exposure to NOx, NO2, O3, and PM10 (particulate matter) in the Swedish cities Stockholm, Gothenburg, and Malmö from the 1990s to 2015. Trend analyses of concentrations have been performed by using the Mann–Kendall test and the Theil–Sen method. Measured concentrations are from central monitoring stations representing urban background levels, and they are assumed to indicate changes in long-term exposure to the population. However, corrections for population exposure have been performed for NOx, O3, and PM10 in Stockholm, and for NOx in Gothenburg. For NOx and PM10, the concentrations at the central monitoring stations are shown to overestimate exposure when compared to dispersion model calculations of spatially resolved, population-weighted exposure concentrations, while the reverse applies to O3. The trends are very different for the pollutants that are studied; NOx and NO2 have been decreasing in all cities, O3 exhibits an increasing trend in all cities, and for PM10, there is a slowly decreasing trend in Stockholm, a slowly increasing trend in Gothenburg, and no significant trend in Malmö. Trends associated with NOxand NO2 are mainly attributed to local emission reductions from traffic. Long-range transport and local emissions from road traffic (non-exhaust PM emissions) and residential wood combustion are the main sources of PM10. For O3, the trends are affected by long-range transport, and there is a net removal of O3 in the cities. The increasing trends are attributed to decreased net removal, as NOx emissions have been reduced.Health effects in terms of changes in life expectancy are calculated based on the trends in exposure to NOx, NO2, O3, and PM10 and the relative risks associated with exposure to these pollutants. The decreased levels of NOx are estimated to increase the life expectancy by up to 11 months for Stockholm and 12 months for Gothenburg. This corresponds to up to one-fifth of the total increase in life expectancy (54–70 months) in the cities during the period of 1990–2015. Since the increased concentrations in O3 have a relatively small impact on the changes in life expectancy, the overall net effect is increased life expectancies in the cities that have been studied.

Abstract Coral-eating Crown-of-Thorns Sea stars (Acanthaster spp.) are major contributors to coral reef loss in the Indo-Pacific region. A release from food limitation of their planktotrophic larvae through enhanced pelagic productivity is one of the main hypothesis explaining population outbreaks (‘nutrient limitation hypothesis’). To improve the understanding of these outbreaks we developed an automated flow- through larvae rearing system that maintained food (microalgae) at set levels over the course of four 15d experiments. This resulted in stable food concentrations in experimental tanks. Increased algae concentrations had a significant positive effect on larval development and size at 10 and 15 days post fertilization (dpf). Larvae densities had no effect at 10 dpf. At 15 dpf greater larvae densities were associated with declines in larvae size. Larval development was slowed under higher larvae densities. Thus, the effects of algae concentration and larvae density were additive at 15 dpf, with larvae under low densities at a given algae concentration being further developed than those under higher densities. The development of a flow-through system gives greater insight into the effect of algae and larvae concentrations on Acanthaster development, and the system can be applied to further test the nutrient-limitation hypothesis for present and future outbreaks.

Late Quaternary paleoceanographic changes at the Lomonosov Ridge, central Arctic Ocean, were reconstructed from a multicore and gravity core recovered during the 2014 SWERUS-C3 Expedition. Ostracode assemblages dated by accelerator mass spectrometry (AMS) indicate changing sea-ice conditions and warm Atlantic Water (AW) inflow to the Arctic Ocean from ∼ 50 ka to present. Key taxa used as environmental indicators include Acetabulastoma arcticum (perennial sea ice), Polycope spp. (variable sea-ice margins, high surface productivity), Krithe hunti (Arctic Ocean deep water), and Rabilimis mirabilis (water mass change/AW inflow). Results indicate periodic seasonally sea-ice-free conditions during Marine Isotope Stage (MIS) 3 ( ∼ 57–29 ka), rapid deglacial changes in water mass conditions (15–11 ka), seasonally sea-ice-free conditions during the early Holocene ( ∼ 10–7 ka) and perennial sea ice during the late Holocene. Comparisons with faunal records from other cores from the Mendeleev and Lomonosov ridges suggest generally similar patterns, although sea-ice cover during the Last Glacial Maximum may have been less extensive at the new Lomonosov Ridge core site ( ∼ 85.15° N, 152° E) than farther north and towards Greenland. The new data provide evidence for abrupt, large-scale shifts in ostracode species depth and geographical distributions during rapid climatic transitions.

Continental slopes north of the East Siberian Sea potentially hold large amounts of methane (CH4) in sediments as gas hydrate and free gas. Although release of this CH4 to the ocean and atmosphere has become a topic of discussion, the region remains sparingly explored. Here we present pore water chemistry results from 32 sediment cores taken during Leg 2 of the 2014 joint Swedish–Russian–US Arctic Ocean Investigation of Climate–Cryosphere–Carbon Interactions (SWERUS-C3) expedition. The cores come from depth transects across the slope and rise extending between the Mendeleev and the Lomonosov ridges, north of Wrangel Island and the New Siberian Islands, respectively. Upward CH4 flux towards the seafloor, as inferred from profiles of dissolved sulfate (SO42−), alkalinity, and the δ13C of dissolved inorganic carbon (DIC), is negligible at all stations east of 143° E longitude. In the upper 8 m of these cores, downward SO42− flux never exceeds 6.2 mol m−2 kyr−1, the upward alkalinity flux never exceeds 6.8 mol m−2 kyr−1, and δ13C composition of DIC (δ13C-DIC) only moderately decreases with depth (−3.6 ‰ m−1 on average). Moreover, upon addition of Zn acetate to pore water samples, ZnS did not precipitate, indicating a lack of dissolved H2S. Phosphate, ammonium, and metal profiles reveal that metal oxide reduction by organic carbon dominates the geochemical environment and supports very low organic carbon turnover rates. A single core on the Lomonosov Ridge differs, as diffusive fluxes for SO42− and alkalinity were 13.9 and 11.3 mol m−2 kyr−1, respectively, the δ13C-DIC gradient was 5.6 ‰ m−1, and Mn2+ reduction terminated within 1.3 m of the seafloor. These are among the first pore water results generated from this vast climatically sensitive region, and they imply that abundant CH4, including gas hydrates, do not characterize the East Siberian Sea slope or rise along the investigated depth transects. This contradicts previous modeling and discussions, which due to the lack of data are almost entirely based on assumption.

Semi-parameterized street canyon models, as e.g. the Operational Street Pollution Model (OSPM®), have been frequently applied for the last two decades to analyse levels and consequences of air pollution in streets. These models are popular due to their speed and low input requirements. One often-used simplification is the assumption that emissions are homogeneously distributed in the entire length and width of the street canyon. It is thus the aim of the present study to analyse the impact of this assumption by implementing an inhomogeneous emission geometry scheme in OSPM. The homogeneous and the inhomogeneous emission geometry schemes are validated against two real-world cases: Hornsgatan, Stockholm, a sloping street canyon; and Jagtvej, Copenhagen; where the morning rush hour has more traffic on one lane compared to the other. The two cases are supplemented with a theoretical calculation of the impact of street aspect (height / width) ratio and emission inhomogeneity on the concentrations resulting from inhomogeneous emissions. The results show an improved performance for the inhomogeneous emission geometry over the homogeneous emission geometry. Moreover, it is shown that the impact of inhomogeneous emissions is largest for near-parallel wind directions and for high aspect ratio canyons. The results from the real-world cases are however confounded by challenges estimating the emissions accurately.

Currently residential wood combustion (RWC) is increasing in Europe because of rising fossil fuel prices but also due to climate change mitigation policies. However, especially in small-scale applications, RWC may cause high emissions of particulate matter (PM). Recently we have developed a new high-resolution (7 × 7 km) anthropogenic carbonaceous aerosol emission inventory for Europe. The inventory indicated that about half of the total PM2.5 emission in Europe is carbonaceous aerosol and identified RWC as the largest organic aerosol source in Europe. The inventory was partly based on national reported PM emissions. Use of this organic aerosol inventory as input for two chemical transport models (CTMs), PMCAMx and EMEP MSC-W, revealed major underestimations of organic aerosol in winter time, especially for regions dominated by RWC. Interestingly, this was not universal but appeared to differ by country. In the present study we constructed a revised bottom-up emission inventory for RWC accounting for the semivolatile components of the emissions. The revised RWC emissions are higher than those in the previous inventory by a factor of 2–3 but with substantial inter-country variation. The new emission inventory served as input for the CTMs and a substantially improved agreement between measured and predicted organic aerosol was found. The revised RWC inventory improves the model-calculated organic aerosol significantly. Comparisons to Scandinavian source apportionment studies also indicate substantial improvements in the modelled wood-burning component of organic aerosol. This suggests that primary organic aerosol emission inventories need to be revised to include the semivolatile organic aerosol that is formed almost instantaneously due to dilution and cooling of the flue gas or exhaust. Since RWC is a key source of fine PM in Europe, a major revision of the emission estimates as proposed here is likely to influence source–receptor matrices and modelled source apportionment. Since usage of biofuels in small combustion units is a globally significant source, the findings presented here are also relevant for regions outside of Europe.

From May 2008 to March 2009 aerosol emissions were measured using the eddy covariance method covering the size range 0.25 to 2.5 μm diameter (Dp) from a 105 m tower, in central Stockholm, Sweden. Supporting chemical aerosol data were collected at roof and street level. Results show that the inorganic fraction of sulfate, nitrate, ammonium and sea salt accounts for approximately 15% of the total aerosol mass < 1 μm Dp (PM1) with water soluble soil contributing 11% and water insoluble soil 47%. Carbonaceous compounds were at the most 27% of PM1 mass. It was found that heating the air from the tower to 200 °C resulted in the loss of approximately 60% of the aerosol volume at 0.25 μm Dp whereas only 40% of the aerosol volume was removed at 0.6 μm Dp. Further heating to 300 °C caused very little additional losses Dp. The chemical analysis did not include carbonaceous compounds, but based on the difference between the total mass concentration and the sum of the analyzed non-carbonaceous materials, it can be assumed that the non-volatile particulate material (heated to 300 °C) consists mainly of carbonaceous compounds, including elemental carbon. Furthermore, it was found that the non-volatile particle fraction Dp correlated (r2 = 0.4) with the BC concentration at roof level in the city, supporting the assumption that the non-volatile material consists of carbonaceous compounds. The average diurnal cycles of the BC emissions from road traffic (as inferred from the ratio of the incremental concentrations of nitrogen oxides (NOx) and BC measured on a densely trafficked street) and the fluxes of non-volatile material at tower level are in close agreement, suggesting a traffic source of BC. We have estimated the emission factors (EFs) for non-volatile particles Dp to be 2.4 ± 1.4 mg veh−1 km−1 based on either CO2 fluxes or traffic activity data. Light (LDV) and heavy duty vehicle (HDV) EFs were estimated using multiple linear regression and reveal that for non-volatile particulate matter in the 0.25 to 0.6 μm Dp range, the EFHDV is approximately twice as high as the EFLDV, the difference not being statistically significant.

Carbon-containing particles have deleterious effects on both Earth's climate and human health. In Europe, the main sources of light-absorbing carbon (LAC) emissions are the transport (67%) and residential (25%) sectors. Information on the spatiotemporal variability of LAC particles in urban areas is relevant for air quality management and to better diagnose the population exposure to these particles. This study reports on results of an intensive field campaign conducted at four sites (two kerbside stations, one urban background site and a rural station) in Stockholm, Sweden, during the spring 2006. Light-absorbing carbon mass (MLAC) concentrations were measured with custom-built Particle Soot Absorption Photometers (PSAP). The spatiotemporal variability of MLAC concentrations was explored by examining correlation coefficients (R), coefficients of divergence (COD), and diurnal patterns at all sites. Simultaneous measurements of NOx, PM10, PM2.5, and meteorological variables were also carried out at the same locations to help characterize the LAC emission sources. Hourly mean (± standard deviation) MLAC concentrations ranged from 0.36±0.50 at the rural site to 5.39±3.60 μg m−3 at the street canyon site. Concentrations of LAC between urban sites were poorly correlated even for daily averages (R0.30) even at spatial scales of few kilometers. This high variability is connected to the distribution of emission sources and processes contributing to the LAC fraction at these sites. At urban sites, MLAC tracked NOx levels and traffic density well and mean MLAC/PM2.5 ratios were larger (26–38%) than at the background sites (4–10%). The results suggest that vehicle exhaust emissions are the main responsible for the high MLAC concentrations found at the urban locations whereas long-range transport (LRT) episodes of combustion-derived particles can generate a strong increase of levels at background sites. To decrease pollution levels at kerbside and urban background locations in Stockholm, we recommend abatement strategies that target reductions of vehicle exhaust emissions, which are the main contributors to MLAC and NOx concentrations.

Unlike exhaust emissions, non-exhaust traffic emissions are completely unregulated and in addition, there are large uncertainties in the non-exhaust emission factors required to estimate the emissions of these aerosols. This study provides the first published results of direct measurements of size resolved emission factors for particles in the size range 0.25–2.5 μm using a new approach to derive aerosol emission factors based on carbon dioxide (CO2) emission fluxes. Aerosol fluxes were measured over one year using the eddy covariance method at the top of a 105 m high communication tower in Stockholm, Sweden. Maximum CO2 and particle fluxes were found when the wind direction coincided with the area of densest traffic within the footprint area. Negative fluxes (uptake of CO2 and deposition of particles) coincided with periods of sampling from an urban forest area. The fluxes of CO2 were used to obtain emission factors for particles by assuming that the CO2 fluxes could be directly related to the amount of fuel burnt by vehicles in the footprint area. The estimated emission factor for the fleet mix in the measurement area was, in number 1.8 × 1011 particle veh−1 km−1 (for 0.25–2.5 μm size range). Assuming spherical particles of density 1600 kg m−3 this corresponds to 27.5 mg veh−1 km−1. For particles (0.8–2.5 μm) the emission factors were 5.1 × 109 veh−1 km−1 for number and 11.5 mg veh−1 km−1 for mass. But a wind speed dependence was noted for high wind speeds. Thus, for wind speeds larger than 9 m s−1, as measured in the tower at 105 m (U105), the emission factor for particle number and mass was parameterised as: Ef (Number, 0.8–2.5 μm) = (6.1 ± 1.7)109 U105 −50 ± 188 and Ef (Mass, 0.8–2.5 μm) = (20 ± 12) U105 − 171 ±122.

Natural abundance radiocarbon analysis facilitates distinct source apportionment between contemporary biomass/biofuel (14C "alive") versus fossil fuel (14C "dead") combustion. Here, the first compound-specific radiocarbon analysis (CSRA) of atmospheric polycyclic aromatic hydrocarbons (PAHs) was demonstrated for a set of samples collected in Lycksele, Sweden a small town with frequent episodes of severe atmospheric pollution in the winter. Renewed interest in using residential wood combustion (RWC) means that this type of seasonal pollution is of increasing concern in many areas. Five individual/paired PAH isolates from three pooled fortnight-long filter collections were analyzed by CSRA: phenanthrene, fluoranthene, pyrene, benzo[b+k]fluoranthene and indeno[cd]pyrene plus benzo[ghi]perylene; phenanthrene was the only compound also analyzed in the gas phase. The measured Δ14C for PAHs spanned from −138.3‰ to 58.0‰. A simple isotopic mass balance model was applied to estimate the fraction biomass (fbiomass) contribution, which was constrained to 71–87% for the individual PAHs. Indeno[cd]pyrene plus benzo[ghi]perylene had an fbiomass of 71%, while fluoranthene and phenanthrene (gas phase) had the highest biomass contribution at 87%. The total organic carbon (TOC, defined as carbon remaining after removal of inorganic carbon) fbiomass was estimated to be 77%, which falls within the range for PAHs. This CSRA data of atmospheric PAHs established that RWC is the dominating source of atmospheric PAHs to this region of the boreal zone with some variations among RWC contributions to specific PAHs.

The combined effect of residential wood combustion (RWC) emissions with stable atmospheric conditions, which frequently occurs in Northern Sweden during wintertime, can deteriorate the air quality even in small towns. To estimate the contribution of RWC to the total atmospheric aerosol loading, positive matrix factorization (PMF) was applied to hourly mean particle number size distributions measured in a residential area in Lycksele during winter 2005/2006. The sources were identified based on the particle number size distribution profiles of the PMF factors, the diurnal contributions patterns estimated by PMF for both weekends and weekdays, and correlation of the modeled particle number concentration per factor with measured aerosol mass concentrations (PM10, PM1, and light-absorbing carbon MLAC) Through these analyses, the factors were identified as local traffic (factor 1), local RWC (factor 2), and local RWC plus long-range transport (LRT) of aerosols (factor 3). In some occasions, the PMF model could not separate the contributions of local RWC from background concentrations since their particle number size distributions partially overlapped. As a consequence, we report the contribution of RWC as a range of values, being the minimum determined by factor 2 and the possible maximum as the contributions of both factors 2 and 3. A multiple linear regression (MLR) of observed PM10, PM1, total particle number, and MLAC concentrations is carried out to determine the source contribution to these aerosol variables. The results reveal RWC is an important source of atmospheric particles in the size range 25–606 nm (44–57%), PM10 (36–82%), PM1 (31–83%), and MLAC (40–76%) mass concentrations in the winter season. The contribution from RWC is especially large on weekends between 18:00 LT and midnight whereas local traffic emissions show similar contributions every day.

Urban aerosol sources are important due to the health effects of particles and their potential impact on climate. Our aim has been to quantify and parameterise the urban aerosol source number flux F (particles m−2 s−1), in order to help improve how this source is represented in air quality and climate models. We applied an aerosol eddy covariance flux system 118.0 m above the city of Stockholm. This allowed us to measure the aerosol number flux for particles with diameters >11 nm. Upward source fluxes dominated completely over deposition fluxes in the collected dataset. Therefore, the measured fluxes were regarded as a good approximation of the aerosol surface sources. Upward fluxes were parameterised using a traffic activity (TA) database, which is based on traffic intensity measurements. The footprint (area on the surface from which sources and sinks affect flux measurements, located at one point in space) of the eddy system covered road and building construction areas, forests and residential areas, as well as roads with high traffic density and smaller streets. We found pronounced diurnal cycles in the particle flux data, which were well correlated with the diurnal cycles in traffic activities, strongly supporting the conclusion that the major part of the aerosol fluxes was due to traffic emissions. The emission factor for the fleet mix in the measurement area EFfm=1.4±0.1×1014 veh−1 km−1 was deduced. This agrees fairly well with other studies, although this study has an advantage of representing the actual effective emission from a mixed vehicle fleet. Emission from other sources, not traffic related, account for a F0=15±18×106 m−2 s−1. The urban aerosol source flux can then be written as F=EFfmTA+F0. In a second attempt to find a parameterisation, the friction velocity U* normalised with the average friction velocity has been included, F=EF . This parameterisation results in a somewhat reduced emission factor, 1.3×1014 veh−1 km−1. When multiple linear regression have been used, two emission factors are found, one for light duty vehicles EFLDV=0.3±0.3×1014 veh−1 km−1 and one for heavy-duty vehicles, EFHDV=19.8±4.0×1014 veh−1 km−1, and F0=19±16×106 m−2 s−1. The results show that during weekdays ~70–80% of the emissions came from HDV.

Disturbances in central serotonergic systems have been hypothesized to be involved in seasonal affective disorder (SAD). Association between SAD and the shorter allele of the serotonin transporter promoter repeat length polymorphism (5-HTTLPR) has been reported in an American sample. We have genotyped 82 SAD patients and 82 healthy controls from Sweden, Finland, and Germany for this and five other polymorphisms in the genes coding for serotonin receptors 5-HT2A and 5-HT2C, tryptophan hydroxylase and white. No associations with SAD or seasonality (seasonal variations in mood and behavior) were detected. Although minor effects cannot be excluded, our results suggest that these polymorphisms do not play a major role in the pathogenesis of SAD in the northern European population.