Abstract. Interpreting atmospheric CO2 observations over cities from space requires transport models that accurately link concentration patterns to surface fluxes, making realistic urban boundary-layer representation critical. This study examines how urban physics parameterizations influence boundary-layer dynamics and near-surface CO2 mixing ratios over the Paris metropolitan area under winter and summer conditions. Using the Weather Research and Forecasting (WRF) model, four configurations are evaluated: no-urban representation (No_URB), a single-layer urban canopy model (SLUCM), and two multi-layer schemes (BEP: Building Effect Parameterization and BEM: Building Energy Model). Model outputs are assessed against surface energy flux observations, turbulence measurements, planetary boundary layer height (PBLH), and near-surface CO2 mixing ratios from dense urban and suburban monitoring networks, alongside wind, temperature and humidity. Urban physics exert strong control on wintertime boundary-layer structure and CO2 variability, with scheme differences driven primarily by sensible heat flux, friction velocity, and turbulent kinetic energy, producing large contrasts in PBLH and CO2 accumulation. In summer, PBLH diurnal patterns converge across schemes, with a characteristic plateau during active convection, and CO2 variability becomes dominated by convective mixing. BEM provides the most physically consistent representation across both seasons. Sensitivity tests with three planetary boundary layer schemes show that Mellor–Yamada–Janjic coupled with BEM best reproduces wintertime CO2, capturing realistic nighttime accumulation and daytime mixing, while Yonsei University and BouLac exhibit systematic biases. These results demonstrate that realistic urban physics combined with an appropriate turbulence scheme are essential for physically consistent urban CO2 simulations, particularly in winter.

Urban nighttime economic activities (UNEA) have grown rapidly in recent decades but are increasingly affected by intensifying heatwaves, especially compound heatwaves. However, the nonlinear effects of compound heatwaves on UNEA and their complex interactions with local conditions remain underexplored. Here, using 500-m resolution daily nighttime light data, we develop a machine learning with Shapley additive explanation (SHAP) approach to quantify these effects in Shanghai, China. Our results first show that the random forest model consistently outperforms alternative models across five evaluation metrics. A high cumulative heatwave index (CHWI, representing excess heat during compound heatwaves) strongly increases UNEA, with larger CHWI values producing greater marginal effects. This nonlinear effect is observed in both central and non-central areas, and the trend in central areas is more stable. In addition, meteorological factors (e.g., precipitation, wind speed) show the strongest interactive association with CHWI in both central and non-central areas. Beyond meteorological factors, point-of-interest (POI) density emerges as the dominant positive interaction in central areas, whereas gross domestic product (GDP) and aging rates prevail in non-central areas. Our findings reveal the complex ways in which compound heatwaves influence UNEA and provide essential evidence to guide adaptation and mitigation policies aimed at safeguarding urban nighttime economies.

Effective land management and climate adaptation call for tools that are reliable and easy to use. In this paper, we present the development and evaluation of machine learning emulators trained to reproduce ecosystem model outputs of urban vegetation carbon fluxes: photosynthesis (GPP) and net ecosystem exchange (NEE), on two different time scales, daily and monthly. By training the emulators on a set of JSBACH simulations, they learn to map meteorological drivers to the target fluxes. We demonstrate that our daily emulators achieve a mean R $^ {2}$ of 0.93, and the monthly emulators a mean R $^ {2}$ of 0.91, averaged across plant functional types and targets, relative to JSBACH. Thus, simulating monthly mean values provides a computationally efficient alternative to daily outputs, enabling informed decision-making with a negligible reduction in precision. Additionally, we compare the NEE estimates with eddy covariance measurements at an urban site in Helsinki and the GPP estimates with satellite-based solar-induced fluorescence data across Helsinki. Our emulators lay the groundwork for further development of lightweight tools that avoid the setup and computational requirements of the underlying process-based model and can inform planning and monitoring of land use and climate adaptation.

The structure and behavior of the urban boundary layer are critical for a wide range of applications ranging from urban air quality to urban heat. There still however exists large uncertainties on how the urban boundary layer responds to surface forcing, how this influences ventilation at various spatial scales, and how these processes should optimally be simulated. Recent advancements in computational fluid dynamics have made it possible to examine the urban boundary layer and its structure in detail. This highlights how large eddy simulation modelling can bring comprehensive insights into the urban boundary layer at both local and city scales, demonstrating also its practical application in urban planning or implementation of urban atmospheric measurements.

Abstract. Carbonyl sulfide (COS) is a useful tracer of gross primary productivity (GPP), but the magnitude and variability of anthropogenic COS emissions in urban environments remain poorly constrained. COS fluxes were investigated in Zurich, Switzerland, using tall-tower eddy covariance (EC) measurements as part of the ICOS Cities project from August 2022 to April 2023. Source-area variability was analysed using the Flux Footprint Prediction (FFP) model, and compared with biogenic COS fluxes simulated by the Simple Biosphere 4 (SiB4) model. Zurich acted as a net sink of COS throughout most of the campaign (median −4.4 pmol m-2 s-1). The most vegetated south-western sector exhibited approximately three times the COS uptake of the south-eastern sector (−6.7 and −1.9 pmol m-2 s-1, respectively), while the south-eastern sector exhibited the weakest uptake, consistent with stronger anthropogenic fluxes. Footprint-weighted land-cover analysis showed that spatial variability in COS exchange reflected the balance between vegetation uptake and anthropogenic fluxes, with the latter masking the biospheric signal in densely built sectors. SiB4 reproduced the seasonal evolution of COS exchange but underestimated the sink following the preceding compound drought event and during an exceptionally warm winter, highlighting limitations in representing stomatal conductance and urban vegetation. These results demonstrate both the potential and the challenges of using COS as a tracer of urban GPP. Improved anthropogenic COS emission inventories, ecosystem models, and urban flux observations are needed to improve COS-based estimates of urban carbon uptake.

While climate-induced population migration has received rising attention, the role played by human climate endeavors remains underexplored. Here, we combine machine learning with attribution mapping to analyze the impacts of 4,713 heat-related policies (HPs) on 11,177 migration flows between U.S. counties. We find that heat adaptation policies (APs) and heat mitigation policies (MPs) have significant and opposing impacts on internal migration: APs reduce out-migration, while MPs increase it. These policies have heterogeneous effects on migration among policy types. Behavioral and cultural MPs at origins lead to a 0.24

Abstract. Urban air quality strategies increasingly rely on transitioning to battery electric vehicles (BEVs), yet their impact on non-exhaust aerosol emissions remains uncertain. This study uses high-resolution Large Eddy Simulation (LES) to investigate aerosol concentrations in a planned Helsinki neighborhood where a highway corridor is being converted into a residential boulevard. We consider three scenarios with varying BEVs shares: a baseline year 2022 (10 % of BEVs), and projected years 2035 (60 %) and 2040 (100 %). The findings reveal a "dual impact" of vehicle electrification. Increased BEV shares significantly reduce particle number concentrations (PN2.5), with a projected 60 % decrease by 2040 compared to the baseline. Conversely, fine particle mass (PM2.5) is projected to increase by approximately 15 % by 2040. This divergence occurs because while BEVs eliminate tailpipe exhaust, their greater weight increases unregulated non-exhaust emissions (NEE) from tire and brake wear. Results show high spatial variation, with pollutants concentrated along boulevard roads and limited penetration into residential blocks. This study underscores the challenges of BEV adoption in realistic urban environments and provides vital insights for sustainable urban planning and pollution mitigation strategies.

While China's urban nighttime economic activities (UNEA) are booming as urbanization deepens, the vitality of these activities is being affected by increasingly frequent and severe heatwaves. However, the direction and magnitude of this impact remain poorly understood. Guided by a novel behavioral adaptation framework that conceptualizes nighttime light (NTL) dynamics as signals of UNEA climate resilience, we quantify the effects of daytime heatwaves (DHW), nighttime heatwaves (NHW), and compound heatwaves (CHW) on UNEA using daily 500-meter resolution nighttime light (NTL) data from 2013 to 2020 in Beijing, Shanghai, and Shenzhen. The results show that DHW has a consistent and significant positive impact on UNEA. In Shanghai, each 1 °C increase in the cumulative heatwave index (CHWI, representing excess heat accumulation during heatwaves) leads to a cumulative increase of 4.6% (±0.29%, 95% CI, P = 0.000) in NTL intensity, though inter-annual analysis reveals that this magnitude exhibits temporal variability. In comparison, NHW suppresses UNEA in Beijing and Shanghai, and the effects of CHW exhibit variations in selected cities. Moreover, spatial econometric analysis reveals that the heatwave-NTL relationship in Beijing has moderate spatial spillovers, while that in Shanghai and Shenzhen is minimal, reflecting the role of urban spatial structure in mediating impact propagation. We finally explore the heterogeneous effects of each heatwave-city combination across central and non-central areas, workdays and non-workdays, days of the week, and calendar years. Notably, DHW consistently exerts its strongest positive influence on Wednesdays across all three cities, highlighting how social rhythms interact with climatic shocks. Our findings uncover the diverse pathways through which heatwaves affect UNEA and offer a nuanced, theoretical foundation for targeted resilience policies.

Urban ecosystems have the potential to provide carbon (C) sequestration; however, the specifics of urban C sequestration and its behaviour under future climate change are poorly quantified. Here, we explore the C sequestration behaviour of five urban ecosystem types (lawn, mesic meadow, dry meadow, park and forest) and two irrigated types (irrigated park and lawn) in Finnish cities under climate change. Our focus was on the role of meteorological drivers and we wanted to understand how weather conditions influence regional differences in the C sequestration of the urban ecosystem types. We utilized the JSBACH model forced with simulated meteorological data between 2006 and 2100 from the CORDEX initiative, downscaled data from the CanESM2, MIROC5 and CNRM-CM5 climate models. The findings indicate that, overall, urban ecosystems with trees exhibit a higher net ecosystem production (NEP) and demonstrate more stable C sequestration in the face of future climate change. In contrast, different grassland systems are smaller C sinks and can become C sources under different climate projections. Irrigated versions of park and lawn had lower C sequestration than their non-irrigated counterparts through the simulation period. We found that changes in annual precipitation and incoming radiation were the predominant climatic drivers of NEP, while changes in temperature had a mixed or weak effect. These findings highlight the importance of long-term planning, management of urban ecosystems, and understanding of the trade-offs between various ecosystem types and services associated with these urban green spaces in future climates.

Understanding new particle formation (NPF) and the fate of nanoparticles is crucial because of their close links to air quality, cloud formation, and climate. These effects vary spatially and temporally owing to diverse aerosol sources and their relatively short atmospheric lifetime. Here, we present a comprehensive analysis of long-term trends in NPF-associated nucleation-mode particles and cloud condensation nuclei (CCN) concentrations across diverse observation environments using quality-controlled particle number size distribution (PNSD) and CCN data from 37 sites, primarily from Global Atmosphere Watch (GAW) stations. We identify declining decadal trends in both NPF occurrences and nucleated particle concentrations across most site types, with the strongest declines in urban areas. We observe simultaneous reductions in both CCN concentrations and nucleation-mode particles, suggesting that newly formed particles are a potential source of CCN. This, in turn, suggests that cloud microphysical properties and radiative effects can be indirectly influenced through aerosol-cloud interactions that modify cloud droplet formation. These findings indicate that decreasing anthropogenic emissions could influence the climate forcing potential of aerosol-cloud interactions, with important implications for future climate projections.

Urban boundary layer (UBL) processes are important to understand the complex interactions between the surface and atmosphere. Increased surface roughness and heat storage are typically attributed to the larger fraction of built surfaces in urban areas. Heatwaves lead to stronger heat storage and to extreme convective UBLs, which is one of many applications where large-eddy simulations (LES) offer an accurate modelling methodology of complex turbulent characteristics in urban flow studies. This study investigates the structure, turbulence characteristics and measurement implications of UBL in Paris, France, during a heatwave in July 2022. High-resolution LES simulations (4–16 m) produced with PALM model system are evaluated against observations from an obervation campaign, including eddy-covariance fluxes (EC), Doppler wind lidar (DWL) profiles, and near-surface measurements from the Météo-France observation network. In addition, wavelet analysis is used to examine resolved turbulent scales and vertical coupling at typical urban measurement heights. PALM reproduces the temporal variability of sensible heat fluxes at the SIRTA observatory with a root mean squared error (RMSE) of 73.4 W m^-2 and a bias of 45.0 W m^-2 , while momentum-related quantities show larger deviations, particularly during weak nighttime stratification. Near-surface temperatures are well captured across the Météo-France network, with an RMSE of 1.18 ^∘C and mean bias of - 0.63 ^∘C . DWL comparisons show that PALM captures the growth and decay of the UBL at urban and suburban sites, with maximum UBL heights reaching 2.4 km in the city centre, approximately 10

The size distribution of particles emitted from road traffic substantially influences urban air quality and human health. There is however a limited understanding on the importance of the emission number size distributions, especially for particles smaller than 50 nm, in neighbourhood-scale air quality modelling. We used the large-eddy simulation PALM-SALSA model to investigate the influence of traffic emissions and their number size distributions on pedestrian-level particle concentrations in Beijing. The simulations were driven by particle size distributions from emission model GAINS, roadside field measurements, and engine laboratory experiments. With constant total traffic emissions and only modifying the shape of the emission number size distribution to better represent smaller particles, the modelled total particle number concentrations (Ntot) decreased less than 4% compared to the simulation using GAINS. When experimental-based nucleation mode vehicle emissions were introduced, the total particle number emissions increased to 12-fold when compared to GAINS, leading to a pronounced rise in Ntot (up to 560%), particularly in the number concentration of smallest particles. Owing to aerosol dynamics, the increase in Ntot within the street canyon was lower than the relative increase in vehicle emissions. PM1 (Particulate Matter with a diameter smaller than 1 mu m) remained largely unchanged. These results highlight the importance of nucleation-mode traffic emissions for modelling urban No, whereas their influence on PM1 and LDSA (Lung-Deposited Surface Area) appears comparatively minor at the neighbourhood scale.

Quantifying urban CO2 emissions from space can be approached using different methodologies, including direct plume-based analyses, but combining satellite observations with atmospheric transport models requires the ability to realistically reproduce fine-scale spatial gradients over cities. Using the Grand Paris area as a testbed, we investigate the sensitivity of simulated near-surface CO2 concentrations to urban physics parameterization and horizontal resolution within the WRF-Chem modeling framework coupled to a high-resolution fossil fuel emission inventory. At mesoscale resolution (900 m), a hierarchy of urban representations ranging from simulations without urban physics to multi-layer urban canopy models is evaluated, showing that the Building Energy Model (BEM) provides the most physically consistent simulation of surface energy fluxes, boundary-layer development, and near-surface CO2 variability. Building on this configuration, we compare mesoscale simulations with Large-Eddy Simulation (LES) runs at 300 m and 100 m resolution. Model results are evaluated against dense urban CO2 observations from the high-precision Picarro network, a complementary mid-cost sensor network from ICOS-Cities, and surface sensible and latent heat flux observations from the ICOS ETC Level-2 fluxes data product. An extensive urban observation network including wind lidars and ceilometers from Urbisphere project provides an exceptional constraint for the evaluation of boundary-layer structure and vertical mixing at fine scales. The LES simulations substantially enhance the representation of spatial heterogeneity and localized CO2 enhancements associated with major emission sources, which are smoothed or underestimated at mesoscale resolution. However, increased resolution also amplifies sensitivity to local wind fields and emission inventory uncertainties. These results highlight that both urban physics and model resolution critically shape the ability of transport models to reproduce observed urban CO2 gradients.

Abstract. Quantifying urban CO2 emissions using spaceborne total column observations requires atmospheric transport models capable of resolving fine-scale spatial and temporal variability across metropolitan areas. Using the Grand Paris region as a testbed, we evaluated the sensitivity of simulated CO2 fields to spatial resolution and model physics within the Weather Research and Forecasting model coupled to a high-resolution fossil fuel emissions inventory. Simulations were conducted at mesoscale (900 m) and Large-Eddy Simulation (LES) resolutions (300 m and 100 m), and evaluated against dense urban CO2 observations from the Paris Picarro and rooftop mid-cost sensors network. Horizontal resolution strongly affects plume morphology: high-resolution simulations produce sharper gradients, higher peak mixing ratios and more realistic temporal variability. LES simulations capture localized enhancements and intra-urban variability more accurately than the 900 m mesoscale run, which systematically underestimates spatial contrasts by a factor of 2–3. Aggregation experiments indicate that averaging LES simulations to 900 m significantly modifies plume magnitudes and spatial gradients. Domain-mean relative errors for near-surface CO2 reach 29–34 % in winter and 59–61 % in summer, while column-integrated XCO2 shows 21–22 % errors in winter and 58–59 % in summer, with maxima exceeding 70 %. While LES improves the representation of urban CO2 variability, it increases sensitivity to local wind and inventory uncertainties, meaning misplaced point sources can cause large local biases. These findings emphasize the need for inversion frameworks able to incorporate scale-dependent transport and inventory uncertainties to ensure unbiased top-down emission estimates and accurate spatial attribution within city domains.

This article presents the results of long-term air quality measurements (2015–2024) from the Helsinki Traffic Supersite, including gaseous compounds (NO, NO2, CO, CO2, O3), particulate matter (PM2.5, PM10, particle number (PN) and size distribution), chemical composition of PM (black carbon (BC), organic aerosol (OA), inorganic species), lung-deposited surface area (LDSA), polycyclic aromatic hydrocarbons, volatile organic compounds (VOC), as well as road surface conditions and meteorology. In addition to the long-term observations, large number of targeted short-term measurement campaigns were conducted to investigate emerging phenomena, to further develop supersite measurements and gain new information about sources impacting air quality.The observed air quality improvements at the Traffic Supersite were driven by declining traffic exhaust concentrations (NOx (-8.0%/yr), BC (-7.1%/yr), PN (-3.7%/yr), and anthropogenic VOCs (-3.1 to -8.9%/yr)). The observed emission factors (g/kgfuel) also decreased for NOx (-7.6%/yr), BC (-7.5%/yr), and PN (-4.1%/yr), reflecting fleet renewal. The observed rate of decrease in PN concentration was lower than that of other parameters, particularly in the smallest size classes (< 30 nm). Organic aerosol analysis showed that traffic hydrocarbon tracers (m/z 57) declined faster (-7.9%/yr) than total OA and oxidized OA (m/z 44), which are linked to secondary formation and long-range transport. Long-term results indicate that compliance with the upcoming EU Air Quality Directive limit values for 2030 is already largely achievable, with the most challenging aspect being the exceedances of the PM10 daily limit due to road dust events. However, achieving the EU Zero Pollution target for 2050 remains challenging, and the WHO guideline values for PM10, PM2.5 and NO2 continue to be exceeded. These unique findings highlight the importance of comprehensive, long-term supersite measurements for understanding pollutant sources and trends, and for supporting urban planning and policy development.

With several cities worldwide pursuing carbon neutrality in the upcoming decades, there is an increasing interest in quantifying cities’ anthropogenic carbon emissions using atmospheric observations. The challenge with both in-situ and remote sensing methods is however that the observations include both anthropogenic and biogenic signals. To reduce uncertainties in anthropogenic emission estimations, it is critical to partition biogenic fluxes of carbon dioxide (CO2) from the observed data. In this study, we for the first time examine the suitability of carbonyl sulfide (COS), a proxy for photosynthesis, on partitioning biogenic CO2 uptake from the ecosystem exchange measured with the eddy covariance (EC) technique over an urban area in Helsinki, Finland. The urban vegetation acts as a clear sink for COS whereas anthropogenic processes show minimal COS emissions within the source area of the measured net carbon flux. We show that two different COS flux based methods are able to produce the dynamics of photosynthesis by an independent light-response curve based estimation. Together with commonly used soil and vegetation respiration proxy, we removed biogenic signals from the urban net CO2 exchange and demonstrated that together with CO2 fluxes, COS flux can successfully be used to get realistic estimations of anthropogenic carbon emissions using the EC method.

To inform tree species selection for vegetation restoration, we assessed the water balance of three plantations — exotic broadleaf Acacia mangium, native broadleaf Schima wallichii, and native conifer Cunninghamia lanceolata — based on 2017–2018 field measurements in hilly areas of South China. Hydrological components were quantified using rain gauges (canopy interception), litter devices (litter interception), sapflow sensors (stand-scale transpiration), soil moisture sensors (soil water storage), and weirs (surface runoff). No clear evidence was found that any tree species exhibited excessive deep water source extraction throughout the two whole study years, but drainage decreased in most months of 2018 despite 114.5 mm more rainfall than in 2017. A. mangium consistently showed evapotranspiration-to-precipitation ratios above one and dry-season drainage deficits in both years, whereas the two native species exceeded this threshold only in 2018 and did not show drainage deficits in either dry season. This suggests the exotic A. mangium was more water-limited and may pose greater risks to catchment water balance in the dry season. Among the three species, native S. wallichii demonstrated the greatest canopy interception during extreme rainfall in 2018, the lowest wet-season surface runoff, and the highest drainage in 2018 as well as in the wet season of 2017 (excluding April), indicating its relative hydrological advantage for afforestation especially under increasingly extreme rainfall. All species had low stand-scale transpiration-to-precipitation ratios, with C. lanceolata showing significantly lower transpiration, significantly higher soil evaporation, and the highest surface runoff—suggesting that, after 30 years of growth, this native conifer may pose risks to water balance sustainability due to inefficient water use and increased flood risks. Longer-term and larger-scale research is needed to confirm these patterns.

In response to the urgent environmental crises, the European Union has intensified legislative actions, including the Nature Restoration Law (NRL). The NRL provides a framework aimed at halting and reversing the degradation of ecosystems and biodiversity loss across the EU. A pivotal element of the NRL is Article 8, dedicated to the restoration of urban ecosystems. While the outlined land use measures limit environmental degradation, they also raise concerns about how cities will manage these requirements with a simultaneously growing population. We examine the potential implications of the NRL on carbon sequestration and biodiversity in Helsinki, Finland, assessing the impacts under various urban development scenarios. Our results indicate that whilst the law presents an improvement compared to business-as-usual growth, this is contrasted with increasing density as cities need to accommodate the growing population. The research, therefore, advocates for an approach merging top-down quantitative targets with qualitative, locally tailored actions.

Understanding how urban forest carbon exchange responds to short-term heatwaves and droughts is critical for its effective management, especially given the increase in frequency and compounded extreme events. The impacts of individual heatwaves or droughts on urban forest carbon fluxes have been studied, however, their combined effects and the roles of key environmental factors remain unclear. This study employs daily-scale eddy covariance data collected over 102 days in the summer of 2022 to examine urban forest carbon fluxes-gross primary productivity (GPP), ecosystem respiration (Re), and net ecosystem exchange (NEE)-responses to extreme heatwave-drought events. A random forest model combined with SHAP analysis was employed to evaluate carbon flux differences and identify key environmental drivers. The results show that concurrent heatwave-drought conditions led to the largest carbon sink capacity reduction, decreasing NEE by 66 % (95 % CI: 50 %-82 %, p < 0.001) compared to normal conditions observed within the same period, driven by extreme temperatures and soil moisture deficits. Heatwaves with sufficient water availability resulted in the strongest physiological activity, with GPP and Re increasing by approximately 17 % (95 % CI: 2 %-33 %, p < 0.01) and 47 % (95 % CI: 23 %-74 %, p < 0.01), respectively, relative to normal periods; however, the trade-off between the two processes reduced the overall carbon sink compared to normal periods. Short-term droughts during non-heatwave periods had negligible effects, with carbon sequestration remaining stable. Solar radiation and soil moisture were the most influential environmental factors, varying in importance across conditions. These findings demonstrate the high sensitivity of urban forest carbon dynamics to the combined effects of heat and drought. Even though limited by preliminary evidence from a single growing season, this still highlights the need for targeted management interventions to protect the functionality of ecosystems in the face of increasing climate extremes.