Local Variations Research Articles

7 Tesla MRI Liver Fat Quantification in Mice: Data Quality Assessment.

The objective of this study was to evaluate the robustness of proton density fat fraction (PDFF) data determined by magnetic resonance imaging (MRI) and spectroscopy (MRS) via spatially resolved error estimation. Using standard T2* relaxation time measurement protocols, in-vivo and ex-vivo MRI data with water and fat nominally in phase or out of phase relative to each other were acquired on a 7 T small animal scanner. Based on a total of 24 different echo times, PDFF maps were calculated in a magnitude-based approach. After identification of the decisive error-prone variables, pixel-wise error estimation was performed by simple propagation of uncertainty. The method was then used to evaluate PDFF data acquired for an explanted mouse liver and an in vivo mouse liver measurement. The determined error maps helped excluding measurement errors as cause of unexpected local PDFF variations in the explanted liver. For in vivo measurements, severe error maps gave rise to doubts in the acquired PDFF maps and triggered an in-depth analysis of possible causes, yielding abdominal movement or bladder filling as in vivo occurring reasons for the increased errors. The combination of pixel-wise acquisition of PDFF data and the corresponding error maps allows for a more specific, spatially resolved evaluation of the PDFF value reliability.

Coupling MAR (Modèle Atmosphérique Régional) with PISM (Parallel Ice Sheet Model) mitigates the positive melt–elevation feedback

Abstract. The Greenland Ice Sheet is a key contributor to sea level rise. By melting, the ice sheet thins, inducing higher surface melt due to lower surface elevations, accelerating the melt coming from global warming. This process is called the melt–elevation feedback and can be considered by using two types of models: either (1) atmospheric models, which can represent the surface mass balance (SMB), or SMB estimates resulting from simpler models such as positive degree day models or (2) ice sheet models representing the surface elevation evolution. The latter ones do not represent the surface mass balance explicitly as well as polar-oriented climate models. A new coupling between the MAR (Modèle Atmosphérique Régional) regional climate model and the PISM (Parallel Ice Sheet Model) ice sheet model is presented here following the CESM2 (Community Earth System Model; SSP5-8.5, Shared Socioeconomic Pathway) scenario until 2100 at the MAR lateral boundaries. The coupling is extended to 2200 with a stabilised climate (+7 ∘C compared to 1961–1990) by randomly sampling the last 10 years of CESM2 to force MAR and reaches a sea level rise contribution of 64 cm. The fully coupled simulation is compared to a one-way experiment where surface topography remains fixed in MAR. However, the surface mass balance is corrected for the melt–elevation feedback when interpolated on the PISM grid by using surface mass balance vertical gradients as a function of local elevation variations (offline correction). This method is often used to represent the melt–elevation feedback and prevents a coupling which is too expensive in computation time. In the fully coupled MAR simulation, the ice sheet morphology evolution (changing slope and reducing the orographic barrier) induces changes in local atmospheric patterns. More specifically, wind regimes are modified, as well as temperature lapse rates, influencing the melt rate through modification of sensible heat fluxes at the ice sheet margins. We highlight mitigation of the melt lapse rate on the margins by modifying the surface morphology. The lapse rates considered by the offline correction are no longer valid at the ice sheet margins. If used (one-way simulation), this correction implies an overestimation of the sea level rise contribution of 2.5 %. The mitigation of the melt lapse rate on the margins can only be corrected by using a full coupling between an ice sheet model and an atmospheric model.

Observation of rapid decay and optical properties of micro-pyramidically grown molybdenum diselenide MoSe2 crystal

In two-dimensional transition metal dichalcogenides (2D-TMDs), the existence of excitons and trions promises their fascinating optical properties for applications in nanophotonic. The more exotic the nanostructure geometry, the less explored its optical responses. The chemical vapor deposition (CVD) technique enables the synthesis of nanostructures with different geometries ranging from monolayers to pyramids driven by a screw-dislocation mechanism. Here we report the CVD synthesis of micro-pyramids with a curious optical response exotic geometry like pyramid. The photoluminescence of as-grown micro-pyramids is significantly diminished compared to their flat-layered structures. In contrast, nanostructured micro-pyramids exhibit a large number of higher-order phononic resonances when excited resonantly by the Raman signal, allowing a more detailed exploration of exciton dynamics. The spectral peak position, intensity, and ratio of the spectral peak reveal local variations in the atomic arrangement of the center and sides of the micro-pyramid. We have investigated the optical response of MoSe2 micro-pyramids and correlated them with their growth-induced nano geometry. In this study, we have taken into account the first step towards producing tunable nanophotonic devices with applications ranging from optoelectronics to nonlinear optics.

Opposing effects of thyroid hormones on hypothalamic subunits and limbic structures in hyperthyroidism patients: A comprehensive volumetric study.

Thyroid hormones play a critical role in brain development, but paradoxically, patients with hyperthyroidism often exhibit cognitive decline and irritability. This study aims to explore the pattern of atrophy in hyperthyroid patients, changes in specific areas of the brain, including hypothalamic subfields and limbic structures, and their relationships with hormonal levels and psychometric tests. This prospective cross-sectional study involves 19 newly diagnosed, untreated hyperthyroid patients, and 15 age and gender-matched control subjects. The participants underwent psychometric and cognitive tests and volumetric MRI. The hypothalamic subfield (anterior-inferior, anterior-superior, superior-tubular, inferior-tubular, and posterior hypothalamus) and limbic structures (fornix, basal forebrain, nucleus accumbens, and septal nucleus) were segmented using voxel-based morphometry, surface-based morphometry, and deep learning algorithms. The groups were compared using the t-test, and correlation analyses were performed between clinical parameters and volumetric measurements. The correlation between hormonal parameters and volumetric measurements in patient and control groups was assessed with the Meng test. Hyperthyroid patients displayed widespread grey matter loss and sulcal shallowing in the left hemisphere. However, no local gyrification index changes or cortical thickness variations were detected. The limbic structures and hypothalamic subunits did not show any volume discrepancies. Free thyroxine in the patient group negatively correlated with bilateral anterior-inferior and right septal nucleus, but positively correlated with left anterior-inferior in the control group. Thyroid stimulating hormone in the patient group showed a positive correlation with bilateral fornix volume, a correlation absent in the control group. Disease duration negatively correlated with right anterior-inferior, right tubular inferior, and right septal nucleus. Changes in cognitive and psychometric test scores in the patient group correlated with the bilateral septal nucleus volume. Hyperthyroidism primarily leads to a reduction in grey matter volume and sulcal shallowing. Thyroid hormones have differing volumetric effects in limbic structures and hypothalamic subunits under physiological and hyperthyroid conditions.

Investigations of Water Flow Behaviors Induced by Local Temperature Variations through a Single Rough Fracture for the Enhanced Geothermal Systems

Investigations of Water Flow Behaviors Induced by Local Temperature Variations through a Single Rough Fracture for the Enhanced Geothermal Systems

Quality Control of DAS VSP Data in Desert Environment Using Simulations and Matching Filters.

The unconsolidated near surface and large, daily temperature variations in the desert environment degrade the vertical seismic profiling (VSP) data, posing the need for rigorous quality control. Distributed acoustic sensing (DAS) VSP data are often benchmarked using geophone surveys as a gold standard. This study showcases a new simulation-based way to assess the quality of DAS VSP acquired in the desert without geophone data. The depth uncertainty of the DAS channels in the wellbore is assessed by calibrating against formation depth based on the concept of conservation of the energy flux. Using the 1D velocity model derived from checkshot data, we simulate both DAS and geophone VSP data via an elastic pseudo-spectral finite difference method, and estimate the source and receiver signatures using matching filters. These field geophone data show high amplitude variations between channels that cannot be replicated in the simulation. In contrast, the DAS simulation shows a high visual similarity with the field DAS first arrival waveforms. The simulated source and receiver signatures are visually indistinguishable from the field DAS data in this study. Since under perfect conditions, the receiver signatures should be invariant with depth, we propose a new DAS data quality control metric based on local variations of the receiver signatures which does not require geophone measurements.

Precursory Signals to Injection Induced Fault Reactivation in the Laboratory Using Active Ultrasonic Monitoring Methods

AbstractInduced earthquakes are still highly unpredictable, and often caused by variations in pore fluid pressure. Monitoring and understanding the mechanisms of fluid‐induced fault slip is essential for seismic risk mitigation and seismicity forecasting. Fluid‐induced slip experiments were performed on critically stressed faulted sandstone samples, and the evolution of the actively sent ultrasonic waves throughout the experiment was measured. Two different fault types were used: smooth saw‐cut fault samples at a 35° angle, and a rough fault created by in situ faulting of the samples. Variations in the seismic slip velocity and friction along the fault plane were identified by the coda of the ultrasonic waves. Additionally, ultrasonic amplitudes show precursory signals to laboratory fault reactivation. Our results show that small and local variations in stress before fault failure can be inferred using coda wave interferometry for time‐lapse monitoring, as coda waves are more sensitive to small perturbations in a medium than direct waves. Hence, these signals can be used as precursors to laboratory fault slip and to give insight into reactivation mechanisms. Our results show that time‐lapse monitoring of coda waves can be used to monitor local stress changes associated with fault reactivation in this laboratory setting of fluid‐induced fault reactivation. This is a critical first step toward a method for continuous monitoring of natural fault zones, contributing to seismic risk mitigation of induced and natural earthquakes.

Land development and frictions to housing supply over the business cycle

Using a novel data set of U.S. residential land developments, we document that the average time to develop residential properties—which includes both the time spent preparing land infrastructures and construction—is about three years, consistent with sizable lags in housing investment projects. We show that the time to develop is highly dispersed across locations, a finding that helps quantify the housing supply elasticity that is relevant for assessing local housing variations over the business cycle. We also show that incorporating long and dispersed time to develop into an otherwise standard housing investment model helps rationalize some empirical facts on the housing market. Our model implies that policies to boost housing supply are less effective in immediately stabilizing house prices for regions where land development takes a long time.

Exploring the dynamics of finite-energy Airy beams: a trajectory analysis perspective.

In practice, Airy beams can only be reproduced in an approximate manner, with a limited spatial extension and hence a finite energy content. To this end, different procedures have been reported in the literature, based on a convenient tuning of the transmission properties of aperture functions. In order to investigate the effects generated by the truncation and hence the propagation properties displayed by the designed beams, here we resort to a new perspective based on a trajectory methodology, complementary to the density plots more commonly used to study the intensity distribution propagation. We consider three different aperture functions, which are convoluted with an ideal Airy beam. As it is shown, the corresponding trajectories reveals a deeper physical insight about the propagation dynamics exhibited by the beams analyzed due to their direct connection with the local phase variations undergone by the beams, which is in contrast with the global information provided by the usual standard tools. Furthermore, we introduce a new parameter, namely, the escape rate, which allow us to perform piecewise analyses of the intensity distribution without producing any change on it, e.g., determining unambiguously how much energy flux contributes to the leading maximum at each stage of the propagation, or for how long self-accelerating transverse propagation survives. The analysis presented in this work thus provides an insight into the behavior of finite-energy Airy beams, and therefore is expected to contribute to the design and applications exploiting this singular type of beams.

Retrieval of Rotational Temperatures From the Arecibo Observatory Ebert‐Fastie Spectrometer and Their Inter‐Comparison With Co‐Located K‐Lidar and SABER Measurements

AbstractRotational temperatures in the Mesosphere‐Lower Thermosphere region are estimated by utilizing the OH(6,2) Meinel band nightglow data obtained with an Ebert‐Fastie spectrometer (EFS) operated at Arecibo Observatory (AO), Puerto Rico (18.35°N, 66.75°W) during February‐April 2005. To validate the estimated rotational temperatures, a comparison with temperatures obtained from a co‐located Potassium Temperature Lidar (K‐Lidar) and overhead passes of the Sounding of the Atmosphere by Broadband Emission Radiometry (SABER) instrument onboard NASA's Thermosphere Ionosphere Mesosphere Energetics and Dynamics (TIMED) satellite is performed. Two types of weighting functions are applied to the K‐Lidar temperature profiles to compare them with EFS temperatures. The first type has a fixed peak altitude and a fixed full width at half maximum (FWHM) for the whole night. In the second type, the peak altitude and FWHM vary with the local time. SABER measurements are utilized to estimate the OH(6,2) band peak altitudes and FWHMs as a function of local time and considerable temporal variations are observed in both the parameters. The average temperature differences between the EFS and K‐Lidar obtained with both types of weighting functions are comparable with previously published results from different latitude‐longitude sectors. We found that the temperature comparison improves when the time‐varying weighting functions are considered. Comparison between EFS, K‐Lidar, and SABER temperatures reveal that on average, SABER temperatures are lower than the other two instruments and K‐Lidar temperatures compare better with SABER in comparison to EFS. Such a detailed study using the AO EFS data has not been carried out previously.

Impacts of Local Green's Functions on Modeling Atmospheric Loading Effects for GNSS Reference Stations

AbstractThe Green's function approach is well‐established and widely used for modeling the surface mass loading displacements. Global mean Green's functions (MGFs) are commonly applied without considering local variations of the crustal structure. Derived from the modified layered Earth structure, the local Green's functions (LGFs) are theoretically beneficial to generate more accurate deformation, since they consider interior information of the local crust. This paper analyzed the differences among MGFs from Gutenberg–Bullen A model and two sets of LGFs derived from modified PREM Earth models consolidating with two crust models TEA12 and CRUST1.0, hereafter called PREMTEA and PREMCRU, respectively. Utilizing MGFs and two sets of LGFs, we modeled the corresponding 3D atmospheric loading displacements for 984 ITRF2014 stations and compared them with the ITRF2014 residuals. The results show that LGFs from PREMTEA and PREMCRU perform well in further promoting scatter reduction for ∼72%, ∼56%, and ∼85% of stations for the Up, East and North components, respectively. The improvements for the North components are significant (up to 3.6%). In particular, stations in the east coastal areas of North America and the west edge of Greenland exhibit further promoting scatter reduction for the East components (up to ∼2.5%), while those located in west coastline of North America show better performance for the North components. Nevertheless, there are significant anomalies in northern Europe for PREMCRU, the mutation margin of which should be carefully considered when using a resolution higher than 1°. In the area around station MORP (358.31°W, 55.21°N, sited at coastline of Britain), we suggest using PREMTEA model.

Community-level variation in TB testing history in Blantyre, Malawi.

SETTING: Equitable access to TB testing is vital for achieving global diagnosis and treatment targets, but access to diagnostic services is often worse in poorer communities. The SCALE (Sustainable Community-wide Active case-finding for Lung hEalth) survey estimated TB prevalence in Blantyre City, Malawi, and recorded previous engagement with TB services.OBJECTIVE: To explore local variation in the prevalence of ever-testing for TB in Blantyre and investigate potential socio-economic drivers.DESIGN: We fit a mixed-effects model to self-reported prior TB testing from survey participants across 72 neighbourhood clusters, adjusted for sex, age and HIV status and with cluster-level random intercepts. We then evaluated to what extent cluster-level variation was explained by two alternate poverty indicators.RESULTS: We observed substantial variation between clusters in previous TB testing, with little correlation between neighbouring clusters. Individuals residing in less affluent households, on average, had lower odds of having undergone prior testing. However, adjusting for poverty did not explain the cluster-level variations observed.CONCLUSION: Despite a decade of increased active case-finding efforts, access to TB testing is inconsistent across the population of Blantyre. This likely reflects health inequities that also apply to TB testing in many other settings, and motivates collection and analysis of TB testing data to identify the drivers behind these inequities.

An ICME framework for short fiber reinforced ceramic matrix composites via direct ink writing

A manufacturing-driven ICME framework is proposed to model short fiber reinforced ceramic matrix composites (CMCs) via direct ink writing. Currently, there lacks efforts to investigate the effects of properties of short fiber reinforced CMCs due to fiber alignment variance. A multi-scale modeling approach is presented to use representative volume elements to capture the homogenized mechanical behavior at various fiber aspect ratios and volume ratios. The orthotropic material properties are mapped to model the printing process. A series of tensile tests simulations show that with 20∘ standard deviation in fiber alignment, the fracture plane has the maximum local tensile stress range at a printing angle of 30∘ and has the minimum local tensile stress range at 90 ∘ . When the standard deviation increases from 20∘ to 40∘, the average tensile strength across the fracture plane decreases by 2%, but the stress variations increase 27.6%.

A dynamic voltage scaling circuit design based on critical path replica and time warning techniques

Critical path replica (CPR) is a widely used technique in synchronous digital circuit design. However, the existing CPR technique cannot accurately reflect the timing of the circuit due to local process variations (LPV). An improved CPR technique based on load capacitance matching (LCM) is proposed in this paper, which can track critical path delay across wide voltage range. The impact of LPV is simulated under wide voltage range, and a configurable delay line is designed to eliminate the effect of LPV. Furthermore, a low overhead mixed-threshold transition detector (TD) circuit is also proposed to monitor timing violations of the replica path, which generate an ‘error’ signal used to dynamically regulate the chip’s operating voltage. The proposed techniques are implemented on a CORDIC chip using the 55-nm CMOS process. Simulation results show that in the near-threshold voltage (NTV) region, the supply voltage can be reduced from 0.8 to 0.6 V, enabling a maximum of 42.6% power saving at the TT corner, 25°C with lesser than 1% area overhead as compared to the baseline design.

Effects of Viscosity Law on High-Temperature Supersonic Turbulent Channel Flow for Chemical Equilibrium

Direct numerical simulations of temporally evolving high-temperature supersonic turbulent channel flow for chemical equilibrium were conducted with a Mach number of 3.0, a Reynolds number of 4880, and a wall temperature of 1733.2 K to investigate the influence of the viscosity law. The mean and fluctuating viscosity for the mixture rule is higher than that for Sutherland’s law, whereas an opposite trend is observed in the mean temperature, mean pressure, and dissociation degree. The Trettel and Larsson transformed mean velocity, the Reynolds shear stress, the turbulent kinetic energy budget, and the turbulent Prandtl number are insensitive to the viscosity law. The semilocal scaling that take into account local variation of fluid characteristics better collapses the turbulent kinetic energy budget. The modified strong Reynolds analogies provide reasonably good results for the mixture rule, which are better than those for Sutherland’s law. The streamwise and spanwise coherencies for the mixture rule are stronger and weaker than those for Sutherland’s law, respectively. The relationship between viscosity and species components can help to identify the traveling wave packet.

Physico-chemical properties of illite in the organic matter (OM)-rich claystone of Nankai Trough contrast with coeval OM-poor claystone: IODP Expedition 348

Naturally occurring microbially induced smectite-to-illite (Sm-I) reaction is of great interest to understand the control for the conventional Sm-I transformation, the source of bioavailable iron (Fe) in sediments, and the evolution of the microbial community in the deep biosphere. Direct comparisons of physico-chemical properties of illite-like phases including collapsed high-charge smectite and transient phase, and mature illite were performed for the organic matter (OM)-rich claystone in contrast with those of coeval OM-poor claystone (2163 m–2217 m below sea floor) buried at Nankai Trough (Integrated Ocean Discovery Program Site C0002). These clay layers are interbedded with 70°–90° dipping angle, such that they have nearly identical burial conditions of temperature, time, and fluid geochemistry except for OM content, one of the significant controls that can fuel heterotrophic microbial Fe (III) reduction. Unlike the abiotic Sm-I reaction, the local variations in the crystal structure, elemental composition, and oxidation states of Fe for the illite and illite-like phases in the OM-rich claystone suggest that illitization is promoted through the reduction of structural Fe (III) in smectite by microbial activity, however the effects of OM cannot be ruled out.

The impact of local cost-of-living differences on relative poverty incidence: an application using retail scanner data and small area estimation models

AbstractEstimating economic poverty indicators at the local level is essential for well-targeted data-driven welfare policies. However, Italy is a country characterized by strong geographical heterogeneity represented by unequal price levels among different areas, and computing poverty indicators with a national monetary poverty threshold can be misleading. This work proposes a novel approach to estimate monetary poverty incidence at the provincial level in Italy considering the different price levels within national boundaries. To account for local price variation, Spatial Price Indices (SPIs) are computed using scanner data on retail prices. The SPIs are estimated in two ways, referring to the mean local prices and using the 20th percentile of the prices. These two kinds of SPIs are used to adjust the national poverty line when computing the poverty incidence at the provincial level using Small Area Estimation (SAE) models. Our findings suggest that adjusting the national poverty line using the SPIs to compute a monetary poverty index can modify the poverty mapping results from the map produced with the traditional national poverty line that ignores price differences.

A multi-channel neural network model for multi-focus image fusion

The objective of multi-focus image fusion (MFIF) is to generate a fully focused image through integrating multiple partially focused source images. Most of the existing methods do not fully consider the local gradient variation rate of the source image, which makes it difficult to accurately distinguish the small defocused (focused) region covered by the large focused (defocused) region. In addition, these methods also cause edge blurring because they do not take into account misregistration of the source images. To address these issues, in this paper, we propose a simple and effective multi-focus image fusion framework based on multi-channel Rybak neural network (MCRYNN) model. Specifically, the proposed MCRYNN model is highly sensitive to local gradient changes of images based on input receptive fields, which can process multiple source images in parallel and extract the features of focused regions. Moreover, the decision maps can accurately be generate in proposed method based on the information interaction effect of parallel network structure for multi-focus image fusion task. Finally, we conduct qualitative and quantitative experiments on public datasets, and the results show that the performance of the proposed method outperforms the state-of-the-art methods.

Mechanical properties and acoustic emission characteristics of two dissimilar layers of rock-like specimens with prefabricated parallel fissures

AbstractThrough the uniaxial compression test of double parallel fissured layered rock mass, the mechanical properties of layered rock mass with different fissure dip angle, and the characteristics of acoustic emission (AE) parameters in the process of fracture are studied. The influence of fissure dip angle on the progressive damage, and macroscopic fracture of layered rock mass is explored. The fracture mode, local stress variation characteristics, and stress field evolution law of fissured layered rock mass are analyzed from a mesoscopic point of view. The results show that with the increase of the fissure dip angle α, the peak strength and the elastic modulus of the layered rock mass decrease first and then increase. The low frequency-high amplitude (LF-HA) signals of AE all appear in the crack propagation stage. With the increase of fissure dip angle α, the LF-HA signal ratio increases first, then decreases and then increases, and shows significant stage characteristics. The cracks are mainly generated around the relatively low strength A rock and prefabricated fissures, and all pass through the interface between A rock and B rock. Eight types mesoscopic displacement field models are found, and the final failure mode of the model is tensile-shear mixed failure. The upper and lower regions of the fissure are tensile stress areas, while the left and right regions are compressive shear stress areas, which are distributed in a “butterfly” type. The stress difference at the fissure tip is negatively correlated with the mechanical parameters of the layered rock mass.

Improving 3-day deterministic air pollution forecasts using machine learning algorithms

Abstract. 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.