Publications

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, accurate and provides means to avoid exposure to high air pollution levels. Long- as well as 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 help 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-day 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 MLs but that the degree of improvement of the deterministic forecasts depends more on pollutant and site than on what machine learning (ML) algorithm is applied. Deterministic forecasts of PM10 is improved by the MLs through the input of lagged measurements and Julian day partly reflecting seasonal variations not properly parameterised in the deterministic forecasts. A systematic discrepancy by the deterministic forecasts in the diurnal cycle of NOx is removed by the MLs considering lagged measurements and calendar data like hour of the day and weekday reflecting the influence of local traffic emissions. For O3 at the urban background site the local photochemistry not properly accounted for by the relatively coarse Copernicus Atmosphere Monitoring Service ensemble model (CAMS) used here for forecasting O3, but compensated using the MLs by taking lagged measurements into account. The machine learning models performed similarly well for the sites and pollutants. Performance measures like Pearson correlation, root mean square error (RMSE), mean absolute percentage error (MAPE) and mean absolute error (MAE), typically differed less than 30% between ML models. At the urban background site, the deviations between modelled and measured concentrations (RMSE errors) are smaller than uncertainties in the measurements estimated according to recommendations by the Forum for Air Quality Modeling (FAIRMODE) in the context of the air quality directives. At the street canyon sites modelled errors are higher, and similar to measurement uncertainties. Further work is needed to reduce deviations between model results and measurements for short periods with relatively high concentrations (peaks). Such peaks can be due to a combination of non-typical emissions and unfavourable meteorological conditions and may be difficult to forecast. We have also shown that deterministic forecasts of NOx at street canyon sites can be improved using MLs even if they are trained at other sites. For PM10 this was only possible using LSTM.
An important aspect to consider when choosing ML is that the decision tree based models (RF and XGB) can provide useful output on the importance of features that is not possible using neural network models like LSTM, and also that training and optimisation is more complex with LSTM, which could be important to consider when selecting ML algorithm in an operational forecast system. A random forest model is now implemented operationally in the forecasts of air pollution and health risks in Stockholm. Development of the tuning process and identification of more efficient predictors may make forecast more accurate.

Three secondary flows, namely the inward radial flow along the cyclone lid, the downward axial flow along the external surface of the vortex finder, and the radial inward flow below the vortex finder (lip flow) have been studied at a wide range of flow rate 0.22–7.54 LPM using the LES simulations. To evaluate these flows the corresponding methods were originally proposed. The highly significant effect of the Reynolds number on these secondary flows has been described by equations. The main finding is that the magnitude of all secondary flows decrease with increasing Reynolds number. The secondary inward radial flow along the cyclone lid is not constant and reaches its maximum value at the central radial position between the vortex finder external wall and the cyclone wall. The secondary downward axial flow along the external surface of the vortex finder significantly increases at the lowest part of the vortex finder and it is much larger than the secondary flow along the cyclone lid. The lip flow is much larger than the secondary inward radial flow along the cyclone lid, which was assumed in cyclone models to be equal to the lip flow, and the ratio of these two secondary flows is practically independent of the Reynolds number.

Plants combine both chemical and structural means to appear colorful. We now have an extensive understanding of the metabolic pathways used by flowering plants to synthesize pigments, but the mechanisms remain obscure whereby cells produce microscopic structures sufficiently regular to interfere with light and create an optical effect. Here, we combine transgenic approaches in a novel model system, Hibiscus trionum, with chemical analyses of the cuticle, both in transgenic lines and in different species of Hibiscus, to investigate the formation of a semi-ordered diffraction grating on the petal surface. We show that regulating both cuticle production and epidermal cell growth is insufficient to determine the type of cuticular pattern
produced. Instead, the chemical composition of the cuticle plays a crucial role in restricting the formation of diffraction gratings to the pigmented region of the petal. This suggests that buckling, driven by spatiotemporal regulation of cuticle chemistry, could pattern the petal surface at the nanoscale.

Aerosol optical thickness (AOT) has decreased substantially in Europe in the summer half year (April–September) since 1980, with almost a 50% reduction in Central and Eastern Europe, according to Modern-Era Retrospective analysis for Research and Applications, Version 2 (MERRA-2) reanalysis. At the same time, strong positive trends in ERA5 reanalysis surface solar radiation downward for all-sky and clear-sky conditions (SSRD and SSRDc, respectively) and temperature at 2 m are found for Europe in summer during the period 1979–2020. The GEBA observations show as well strong increases in SSRD during the latest four decades. Estimations of changes in SSRDc, using the Santa Barbara DISORT Atmospheric Radiative Transfer (SBDART) model, show similarly strong increases when fed by MERRA-2 AOT. The estimates of warming in this study, caused by increases in SSRD and SSRDc, are based on energy budget approximations and the Stefan Boltzmann law. The increases in near surface temperature, estimated both for clear-sky and all-sky conditions, are up to about 1°C for Central and Eastern Europe. The total warming over large parts of this region for clear-sky conditions is however nearly double the global mean temperature increase of 1.1°C, while somewhat less for all-sky conditions. The effects of aerosols on warming over the southerly Iberian Peninsula are weaker compared to countries further north. The rapid total warming over the Iberian Peninsula is probably caused by greenhouse warming, drier surface conditions, and to some degree decline in aerosols. Reduced cloud cover is found for large parts of Europe in summer during the latest four decades.

The formation of sea ice in polar regions is possible because a salinity gradient or halocline keeps the water column stable despite intense cooling. Here, we demonstrate that a unique water property is central to the maintenance of the polar halocline, namely, that the thermal expansion coefficient (TEC) of seawater increases by one order of magnitude between polar and tropical regions. Using a fully coupled climate model, it is shown that, even with excess precipitations, sea ice would not form at all if the near-freezing temperature TEC was not well below its ocean average value. The leading order dependence of the TEC on temperature is essential to the coexistence of the mid/low-latitude thermally stratified and the high-latitude sea ice–covered oceans that characterize our planet. A key implication is that nonlinearities of water properties have a first-order impact on the global climate of Earth and possibly exoplanets.

The Arctic is warming at more than twice the rate of the global average. This warming is influenced by clouds, which modulate the solar and terrestrial radiative fluxes and, thus, determine the surface energy budget. However, the interactions among clouds, aerosols, and radiative fluxes in the Arctic are still poorly understood. To address these uncertainties, the Ny-Ålesund Aerosol Cloud Experiment (NASCENT) study was conducted from September 2019 to August 2020 in Ny-Ålesund, Svalbard. The campaign’s primary goal was to elucidate the life cycle of aerosols in the Arctic and to determine how they modulate cloud properties throughout the year. In situ and remote sensing observations were taken on the ground at sea level, at a mountaintop station, and with a tethered balloon system. An overview of the meteorological and the main aerosol seasonality encountered during the NASCENT year is introduced, followed by a presentation of first scientific highlights. In particular, we present new findings on aerosol physicochemical and molecular properties. Further, the role of cloud droplet activation and ice crystal nucleation in the formation and persistence of mixed-phase clouds, and the occurrence of secondary ice processes, are discussed and compared to the representation of cloud processes within the regional Weather Research and Forecasting Model. The paper concludes with research questions that are to be addressed in upcoming NASCENT publications.

The Arctic region is sensitive to climate change and is warming faster than the global average. Aerosol particles change cloud properties by acting as cloud condensation nuclei and ice-nucleating particles, thus influencing the Arctic climate system. Therefore, understanding the aerosol particle properties in the Arctic is needed to interpret and simulate their influences on climate. In this study, we collected ambient aerosol particles using whole-air and PM10 inlets and residual particles of cloud droplets and ice crystals from Arctic low-level clouds (typically, all-liquid or mixed-phase clouds) using a counterflow virtual impactor inlet at the Zeppelin Observatory near Ny-Ålesund, Svalbard, within a time frame of 4 years. We measured the composition and mixing state of individual fine-mode particles in 239 samples using transmission electron microscopy. On the basis of their composition, the aerosol and cloud residual particles were classified as mineral dust, sea salt, K-bearing, sulfate, and carbonaceous particles. The number fraction of aerosol particles showed seasonal changes, with sulfate dominating in summer and sea salt increasing in winter. There was no measurable difference in the fractions between ambient aerosol and cloud residual particles collected at ambient temperatures above 0 ∘C. On the other hand, cloud residual samples collected at ambient temperatures below 0 ∘C had several times more sea salt and mineral dust particles and fewer sulfates than ambient aerosol samples, suggesting that sea spray and mineral dust particles may influence the formation of cloud particles in Arctic mixed-phase clouds. We also found that 43 % of mineral dust particles from cloud residual samples were mixed with sea salt, whereas only 18 % of mineral dust particles in ambient aerosol samples were mixed with sea salt. This study highlights the variety in aerosol compositions and mixing states that influence or are influenced by aerosol–cloud interactions in Arctic low-level clouds.