Climate change can affect air pollution both physically, by modifying transport and mixing of pollutants as well as emissions of natural origin, and chemically by affecting chemical reactions. Such changes can lead to a “climate change penalty” which can be defined as the deterioration of air quality due to a warming climate, in the absence of changes in anthropogenic polluting activities (Fu and Tian, 2019). Changes in climate also influence the frequency and intensity of natural wildland fires (Lozano et al., 2017). Vice versa, pollutants in the atmosphere and their precursor emissions can influence climate by modifying clouds properties, precipitation, and radiation (Myhre et al., 2013). Feedbacks between climate change and air pollution should thus be taken into account in developing mitigation and adaptation strategies. However, the existing knowledge gaps on the involved mechanisms result in large uncertainties in the understanding and quantification of these feedbacks. For instance, although the main cause of wildfires currently is anthropogenic, i.e., caused by agriculture and forest management practices (Sofiev et al., 2013), projections suggest that the risk of wildland fires will increase due to higher temperatures and droughts (IPCC, 2022; McCarty et al., 2021; Dupuy et al., 2020). Recent findings from the latest Coupled Model Intercomparison Project Phase 6 (CMIP6: e.g., Turnock et al., 2020), the sixth assessment report from the IPCC (AR6), as well las the latest assessment on the short lived climate forcers (SLCFs) from the Arctic Monitoring and Assessment Programme (AMAP 2021) shows that future anthropogenic pollutant source strengths and levels decrease under all scenarios over Europe, but also in some other regions globally. This suggests that natural sources such as mineral dust, sea-salt, and wildland fires can be important drivers of air pollution and climate over Europe, where the feedback mechanisms and climate adaptation/mitigation scenarios will be particularly important.

Due to the above reasons, the aim of this study is to review the latest literature on the interactions between climate change and air pollution and assess the effects of climate change on surface ozone (O₃) and fine particulate matter (PM_2.5) levels over Europe, and how they will change in the future. Section 1 introduces the major mechanisms and pathways of meteorology-air pollution interactions and summarizes the current knowledge and findings regarding the different physical and chemical processes in relation to climate change and air pollution; Section 2 summarizes the recent findings on the specific impacts of climate change on ozone and particulate matter concentrations, respectively; Section 3 compares the relative impact of future emissions on air quality versus the impacts on air quality caused by climate change alone; Section 4 summarizes recent findings on the impacts of atmospheric composition on climate change; Section 5 summarizes the impacts of climate change on wildland fires and consequently on air pollution and climate change; Section 6 summarizes the co-benefits of climate adaptation and mitigation on air pollution; and finally, Section 7 offers some recommendations regarding needs for future research. Therefore, the current review paper attempts to give a more holistic picture of future air pollution and climate change, as well as feedbacks, impacts of emerging sources such as wildland fires, and climate adaptation and mitigation, compared to the existing reviews.

There are several approaches that are used in the literature in order to understand how O₃ and particulate matter (PM) are influenced by changes in climate or emissions (Fiore et al., 2019). These include 1) sensitivity studies in which individual meteorological parameters are perturbed, 2) statistical downscaling of future changes in meteorology, using correlations between observed changes in air quality and meteorological variables from climate models, 3) the direct calculation of air quality by using fully coupled global or regional chemistry–climate models (CCM), global-to-urban CCMs, or off-line global or regional chemical transport models (CTMs) forced by either projections of meteorological fields from separate atmosphere–ocean general circulation models, or dynamically downscaled meteorology. In addition, there is also large literature available on the statistical relationships between observed meteorological parameters and air pollution concentrations (Im et al., 2006, Im et al. 2008, Jacob and Winner, 2009). Such studies aid in understanding the short-term changes in air pollution concentrations with changes in meteorological factors, which can be used to extrapolate for future conditions (e.g., Cheng et al., 2007; Lin et al., 2007).

Several of the model studies included in this review have also analyzed the combined effect of future changes in both climate and air pollutant and precursor emissions such as NOx, CO, SO₂, NMVOCs, and primary particles. For Europe, the anthropogenic emissions of the main air quality components (e.g., SOx, NOx, and PM_2.5) are projected to decrease in the future posing the central question whether the benefits of these emission reductions are counteracted by a possible climate penalty.

The latest CMIP6 models under different future scenarios, shared socio-economic pathways (SSPs) showed that O₃ concentrations over Europe can increase by up to 4 ppb compared to the 2005–2014 mean under low mitigation scenarios (e.g., SSP3-7.0 and SSP5-8.5), while high mitigation scenarios lead to decreases by up to 16 ppb (in SSP1-2.6) (Turnock et al., 2020). Surface PM_2.5 concentrations over Europe, on the other hand, are projected to decrease consistently by all models by 1–4 ug m⁻³. The recent studies conclude that expected emission changes in comparison to impacts of climate will have the largest impact on European O₃ levels (e.g., Colette et al., 2015; Westervelt et al., 2016). This conclusion is, however, very dependent on the scenarios used for Europe, but also for the developments outside Europe. In the study by Fortems-Cheiney et al. (2019), the high emission scenario RCP8.5 and the related higher European O₃ concentrations, would to a large degree be caused by inflow from outside Europe. In the more detailed study by Lacressonniere et al. (2014), it is also shown that long-range transport from outside Europe can increase the O₃ levels in Northern Europe in spite of decreasing European emissions. The impact of the European emissions was studied directly by Fortems-Cheiney et al. (2019) as they were running a CTM with both current day (2005) and future (2050) European emissions and future climate data. The applied Current Legislation emissions for 2050 projects a significant decrease (more than 40%) in most components across Europe (except for the Turkish area). When emissions are fixed to 2005 under RCP8.5 for 2050, the mean summer O₃ levels increase by up to +12% (+4.30 ppb), except in urban areas with high NOx emissions. Studies also showed that the anthropogenic emissions within the European region have the largest impact on the European PM concentrations towards the end of the century, compared to developments in both climate and long-range transport (Cholakian et al., 2019).

The future developments in health impacts related to air pollution have also been assessed in a number of studies (see e.g., the review by Doherty et al., 2017 and the global study by Silva et al., 2017). Geels et al. (2015) combined two CTMs with a health assessment model and focused on the development in the number of premature deaths related to PM_2.5 and O₃ in Europe in the 2050s and 2080s. Climate change alone in the RCP4.5 scenario, where emissions were kept constant, had only minor impacts (±2%–4%) on the PM_2.5 related number of cases, while the number of cases related to O₃ increased by up to 10% (relative to the 2000s). Including the emissions from the same RCP4.5 scenario leads, however, to significant decreases in the health effects especially for PM_2.5 with a significant decline in the number of premature deaths (∼65% in 2050s and ∼80% in 2080s). But, as pointed out by Doherty et al. (2017), the health assessments for future conditions are related to additional uncertainties as e.g., the temperature effects on applied pollutant-response relationships are not well known.

Air pollutants like O₃ and different types of aerosols [e.g., black carbon (BC), SO₄ ^2-, NO₃ ⁻ and organic aerosols (OA)] have important radiative effects on climate, while other pollutants such as carbon monoxide (CO), VOCs, nitrogen oxides (NOx) and SO₂, although being negligible greenhouse compounds themselves, have an important indirect effect on climate by altering the abundances of radiatively active gasses such as O₃ and CH₄, a well as by acting as precursors for secondary inorganic and organic particles (Stohl et al., 2015).

Tropospheric climate–chemistry interactions are to a large extent related to chemical and physical processes and to compounds that show large variability in global distribution and trends, mainly driven by precursor emissions that also show large spatial variability. Atmospheric composition influences climate by regulating the radiation budget and cloud properties (Boucher et al., 2013; Stohl et al., 2015; Bellouin et al., 2020), through the presence of water vapor (H₂O), CO₂, CH₄, O₃, and aerosols. The main radiative effect of the gasses is through the greenhouse effect (i.e., through interference with long-wave radiation), thus heating the atmosphere, while aerosols may either heat or cool the surface (through interference with short-wave radiation), depending on their optical properties (Table 2: Bond et al., 2013; Stohl et al., 2015).

Aerosol–cloud interactions represent an area with potential for strong interactions in the climate system (Myhre et al., 2013). The largest uncertainty in the climate projections caused by the interaction of aerosols with clouds (Lee et al., 2016; Mülmenstädt and Feingold, 2018). The representation of aerosol-cloud interactions in ESMs remains a challenge in the Sixth Assessment Report (AR6) from IPCC, due to the limited representation of important sub-grid scale processes, from the emissions of aerosols and their precursors to precipitation formation and lack of certain processes such as contribution from anthropogenic aerosols acting as ice nucleating particles (INP) and aerosol effects on deep convective clouds. Aerosols determine the cloud radiative properties and thus the radiative heating/cooling and participate in the precipitation process. Although considerable progress has been made in recent years to include parameterization of aerosol–cloud droplet interaction and explicit microphysics for cloud water/ice content in climate models, inadequate understanding of the processes contributes to significant uncertainties in model simulated future climate changes (Table 2: Malavelle et al., 2017; Toll et al., 2019). According to IPCC AR6, effective radiative forcing due to aerosol-cloud interactions (ERF_aci) based on observational evidence alone (–1.0 ± 0.7 W m⁻²) is very similar to the one based on model-evidence alone (–1.0 ± 0.8 W m⁻²), with a medium confidence (Table 2), compared to the low confidence in AR5.

Regarding the trends in emissions of air pollutants affecting radiative forcing, large differences are observed in the temporal and geographical distributions of human-related (anthropogenic) emissions and natural emissions (Shindell and Faluvegi, 2009; Flanner, 2013). Anthropogenic emissions respond to regulations and to different growth rates in energy use, transportation, and industry. For example, emissions from Europe generally show reductions thanks to the regulatory actions taken over the last decades (EEA, 2019b), while emissions in regions like Southeast Asia, and in other developing countries show large increases during recent years (IEA, 2020). In addition to geographical differences, there are also sectoral differences, such as the large increases observed in aircraft and shipping (Isaksen et al., 2009). On the other hand, natural emissions of key climate precursors (NOx, CO, biogenic VOCs and sulfur compounds) are difficult to regulate and exhibit large year-to-year variations (e.g., in the case of biomass burning).

Climate change causes an increasing number of wildfires in some regions (Lozano et al., 2017) because it increases wildfire risk by raising the average temperature of the atmosphere, intensifying droughts, and prolonging the dry season (Figure 4). Driving mechanisms behind the fires have varied in different historical periods. Pechony et al. (2010) argued that in the preindustrial period, fires were strongly driven by lightning, rather than temperature. With the Industrial Revolution, this changed to an anthropogenic-driven regime (Sofiev et al., 2013), and there are now indications that we are moving towards a temperature-driven fire regime during this century. However, currently the main cause of wildfires is anthropogenic activities related to agriculture and forest management practices, as well as from recreational habits (Sofiev, 2013; Dijkstra et al., 2022). In Southern Europe, more than 95% of the fires with a known cause have an anthropogenic origin (Ganteaume et al., 2013; Parente et al., 2018), while in central Europe, more than 99% of fires are caused by humans (Ganteaume et al., 2013).

Modelling the impacts of climate change on wildland fires is very challenging due to the complex dynamics of fuel accumulation, vegetation dynamics and their interactions with changing climate, as well as the difficulty in representing the fire behavior patterns due to impacts of land management and human ignitions (Hantson et al., 2020; Sanderson and Fisher, 2020). Thus, projection of wildland fires over decadal to century timescales requires more mechanistic approaches, capable of capturing the numerous interacting system components that affect the evolution of fire risk. Process-based global fire models based on these principles have progressed rapidly over the past decade, although their use in fully coupled earth system models is still not standard practice (Sanderson et al., 2020), leading to omitting a potentially important component of the global carbon–climate feedback.

Interaction between the anthropogenic activity, land cover, meteorological conditions and intensity of wildfires (expressed as the amount of biomass burned or pollutants released) basically resides to the human activities playing the fire ignition and suppression roles whereas vegetation and weather modulate the natural fire lifetime and potential for spreading (Prestemon et al., 2013; Fusco et al., 2016; Balch et al., 2017). Most studies that have estimated future fires (risk or area burned) have relied on statistical regressions. However, it is highly regionalized. For instance, in Indonesia (one of the largest sources of fire-induced pollution worldwide) the use of fire for clearing and preparing land on degraded peat is a routine repeated every year—with fires routinely getting out of control. However, it is the El-Nino-induced drought that can make them devastating (Field et al., 2015). In the Western US, worsening climatic conditions will lead the fires to become the main source of PM_2.5 if human behavior and practices remain the same (Yue et al., 2013; Liu et al., 2016). Similarly, Ford et al. (2018) found that, despite the reduction of anthropogenic PM_2.5, raising fire contribution will lead to a net increase of PM_2.5 in several US regions.

Wildfires are known to be the largest contributor to global biomass burning. They are a large source of atmospheric trace gasses and aerosols, in addition to major vegetation disturbance (Knorr et al., 2016). They emit into the atmosphere primarily CO₂ among other greenhouse gasses and carbonaceous particulate matter among other aerosols (Saarikoski et al., 2007). However, our knowledge of wildfires as an integrated part of the Earth system is still rather weak (Bowman et al., 2009; Langmann et al., 2009; Keywood et al., 2013).

In Europe, climate change is expected to increase the probability of forest fires in areas where increase in summertime temperature is accompanied by decrease in precipitation, as in the Mediterranean area (Mouillot et al., 2002; Bedia et al., 2014; Kalabokidis et al., 2014). In more northern regions the estimations of precipitation amounts are less conclusive (Jylhä et al., 2009; Ylhäisi et al., 2010; Venäläinen et al., 2020; Ruosteenoja and Jylhä 2021), which makes predicting the fire sensitivity much more difficult. A global modelling study of Knorr et al. (2017), unlike the research quoted above, gives a calmer picture. The authors conclude that at the global scale anthropogenic emission reduction will exceed increases in wildfire emissions and the WHO target of 10 μg m⁻³ will be in reach in many regions. However, the authors also admitted high uncertainty connected with the fire abatement measures that were accounted for.

Air pollution due to PM_2.5 released from vegetation fires is a notable risk factor for public health everywhere in the world, also in Europe (Johnson et al., 2020). This should be taken into consideration when evaluating the overall health and socio-economic impacts of the fires. Kollanus et al. (2017) evaluated that in 27 European countries, over a thousand premature deaths were attributable to the vegetation-fire originated PM_2.5 concentrations in 2005 and 2008. Estimated impacts were highest in southern and Eastern Europe, but all countries were affected by fire-originated PM_2.5. Saarnio et al. (2010) analyzed smoke plumes originating from wild-land fires, which were detected in Helsinki, Finland, during a 1-month period in August 2006. The measurements showed that the major growth in PM concentration was caused by submicron particles consisting mainly of particulate organic matter. They reported an increase of a factor of 1.5–3.5 in PM_2.5 concentrations. Saarikoski et al. (2007) calculated that the wildland fires increased the PM_2.5 concentrations by a factor of 3.6–4.3. Sofiev et al. (2008) considered the influence of wild-land fires on air quality in Europe, using spring and summer seasons of 2006–2007 as prominent examples. High concentrations of nearly all pollutants were detected in Central, Eastern and Northern Europe in late spring of 2006, causing widespread allergic symptoms and other illnesses associated with poor air quality.

The aerosols produced by fires seem to be different from those emitted from e.g., anthropogenic industrial or transport sources: intriguing observations of Dusek et al. (2005) showed that peat-fire smoke has much weaker interaction with clouds than one would expect: less than 50% of smoke particles of suitable size were activated to CCN, whereas this number often approaches 100%. Therefore, the fire aerosols may not be very efficient shielding agents compared to other types of particles. The other factor is darkening of the surface after the fires: albeit small (at 1 km scale the albedo change is ∼0.002 in sub-Saharan Africa (Gatebe et al., 2014)) these changes occur over wide areas and constitute noticeable additional radiative forcing.

Generally, there is some agreement that carbonaceous aerosol emissions from anthropogenic sources may decrease in the future (Lamarque et al., 2011; Bond et al., 2013; Shindell et al., 2013). Global total carbonaceous emissions are expected to be reduced due to new, less polluting technologies along with a shift away from the burning of coal and wood in the residential sector, while this reduction will vary regionally and not be reduced everywhere (Streets et al., 2004). However, the future trends of wildfires and their aerosol emissions is more uncertain due to complicated interactions between fire, land, atmosphere, and, most importantly, anthropogenic activities. Even now the state-of-the-art Earth system models have very limited capacity to simulate wildfire—climate interactions (Carslaw et al., 2010).

Estimating the fire emission heights is of particular importance when evaluating the coupling between terrestrial and atmospheric processes. Emission heights affect the aerosol long-range transport, aerosol-cloud interaction, and radiation (Luderer et al., 2006; Samset et al., 2013). One large problem is that emission heights are strongly affected by fire intensity and atmospheric stability. As illustrated by Joshi et al. (2007) and Luo et al. (2013), both of these are expected to change in the future. The emission height and its effect on atmospheric black carbon and its transport has been evaluated using present day climate conditions by Veira et al. (2015a,b), Sofiev et al. (2012, 2013), Kukkonen et al. (2014), and Daskalakis et al. (2015). Attempts to apply that same approach to future climate have more recently been done by Veira et al. (2016).

Radiative forcing due to aerosols originating from wildfires has mostly been investigated using present-day climatology and prescribed wildfire emission inventories (satellite based). Based on the literature review by Veira et al. (2016), the most recent climate-aerosol models find that the total wildfire top of the atmosphere radiative forcing range between +0.18 W m⁻² (Tosca et al., 2013) and −0.29 W m⁻² (Jones et al., 2007).

There is a general consensus among studies that there are additional (co) benefits from mitigation of green house gases (GHG) emissions in terms of improved air quality (West et al., 2013; Bollen 2015; de Oliveira and Doll 2016; Klausbruckner et al., 2016; Rao et al., 2017), but these are rarely taken into account in policy developments and assessments at the European and national scale (Sokhi et al., 2022). The estimated improvements in air quality under different climate mitigation scenarios depend on socioeconomic and technological development and the assumptions on the development of air pollution in the next few decades. Rao et al. (2017) examined the impacts of different air pollution narratives under the global Shared Socio-economic Pathway (SSP) framework, concluding that scenarios with limited climate change mitigation present higher air pollution impacts than scenarios with more stringent mitigation trajectories. An overview paper (von Schneidemesser et al., 2015) concludes that any policy actions intended to mitigate one of these two issues must necessarily take into account the feedback with the other, to avoid that benefits to one sector will worsen the situation in another.

The effect of air pollution policies on climate have been investigated in some studies. Shindell et al. (2012) identified 14 emission control measures targeting short-lived climate pollutants (SLCPs) and relative to their reference scenario, found that all CH4 and BC control measures avoid a global mean surface temperature increase of ∼0.5°C by 2050, although significant uncertainties exist regarding the extent to which reductions of SLCFs impact climate change (Myhre et al., 2013). In light of the estimated large adverse health impacts of air pollution around the globe, emission controls on SO₂, BC, and O₃ precursors may continue to be imposed irrespective of climate policy (Anenberg et al., 2012; Silva et al., 2013).

Reduction of non-SLCF pollutants, such as NO₂ and SO₂, generally leads to exacerbation of climate change because these species are the key aerosol precursors and responsible for almost 1 W m⁻² or shielding from the solar radiation roughly-equally split between direct and indirect effects (IPCC AR5, Physical science basis). The secondary inorganic aerosols mainly consisting of nitrates and sulfates are highly reflective and highly soluble, thus facilitating both direct (reflection of sunlight back to space) and indirect (increasing the cloud albedo and modifying the cloud cover and rain patterns) aerosol effects. Recent studies showed that, e.g., The International Convention for the Prevention of Pollution from Ships—Annex VI (MARPOL-VI) required reduction of sulfur content in ship fuels brings substantial health benefits (global avoided mortality 74–200 thousand cases annually, 95% confidence interval) but adds about 0.07 W m⁻² due to lost cooling effects of sulfates (Sofiev et al., 2018). Estimates of other studies of the shipping cooling effect showed similar results suggesting the range from ∼0.04 W m⁻² up to 0.11 W m⁻² (Eyring et al., 2007; Fuglestvedt et al., 2008; Skeie et al., 2009; Lund et al., 2012).

Several studies show potential air quality co-benefits in Europe, including in the United Kingdom and Ireland (Karlsson et al., 2020; Alam et al., 2017, 2018; Wilkinson et al., 2009) and the Nordic countries (Åström et al., 2013; Williams, 2007). Rafaj et al. (2014) associate large potential co-benefits in Europe with renewable energy, due to reduced Mercury emissions to air. Collette et al. (2013) looked at combined projections of air quality and climate impact at the regional scale over Europe under the CMIP5 climate scenarios and conclude that exposure to air pollution will decrease substantially by 2050 according to a mitigation pathway that aims at keeping global warming below 2°C by the end of the century where exposure weighted SOMO35 and PM_2.5 are reduced by 80% and 78%, respectively in Europe. Large potentials for reaping air quality co-benefits are also shown for several European cities, e.g., Rotterdam (Tobollik et al., 2016) and Barcelona, Malmö, Sofia and Freiburg (Creutzig et al., 2012). Markandya et al. (2018) estimated that the co-benefits would make a valuable contribution towards covering the mitigation costs, from 7% to 84% in the EU-27 countries.

The impacts of climate adaptation policies on greenhouse gas emissions and air pollution have also been discussed. Classic adaptation strategies to increased temperatures like air conditioning have been shown to result in increased GHG emissions as well as increased air pollution (Abel et al., 2018; Watts et al., 2019). In a study in Eastern US, Abel et al. (2018) found that in a warmer climate 3.8% of the total increase in PM_2.5 and 6.7% of the total increase in O₃ are attributable to extra air conditioning use. Other climate adaptation actions like improving urban design to reduce heat stress can have significant co-benefits for air pollution (Cheng and Berry, 2013).

The atmospheric concentrations of secondary pollutants like ozone or organic aerosols are expected to increase with the projected global warming, mainly due to an increase in the oxidation capacity of the atmosphere and the increase in natural emissions. On the other hand, the projected increases in precipitation are expected to enhance removal of pollutants from the atmosphere, resulting in a weak positive or negative change in the pollutants concentrations, in particular in the first half of the century. This strongly suggests that feedback between pollutants and climate change play a crucial role in the direction of the policy response on future climate change and air pollution. The net changes caused by the feedback vary geographically and seasonally, and, importantly, all studies reviewed here showed that changes in anthropogenic emissions had much larger impacts on air pollutant concentrations in the future, compared to the corresponding impacts caused by the projected changes in climate.

Studies showed that in most cases, there may be immediate air quality co-benefits from mitigation of GHGs due to reduced emissions of co-emitted air pollutants, such as aerosols, from the same sources such as energy production and industrial combustion. However, for instance increasing the residential combustion of wood due to climate policies may result in a substantial deterioration of air quality (e.g., Kukkonen et al., 2020). Noteworthy, the reverse relation does not necessarily hold: improving air quality often raises the challenges to climate change mitigation, since it may diminish the amount of solar radiation scattered back into space.

Socioeconomic and technological developments may influence the potential for air quality co-benefits. Recent findings suggest that limiting climate change mitigation today would entail higher air pollution levels in the future. Due to the adverse human health impacts of these pollutants, policies will continue to be implemented to reduce the emissions, in order to follow for example, the recent air quality recommendations from WHO. This implies that stricter climate policies are needed to counteract the increased global warming resulting from the pollutant removals.

As this review shows, the projections of how climate and health relevant air pollution components (e.g., PM_2.5 and O₃) will develop in the future across Europe, are still associated with major uncertainties. The differences in the modeled projections can to a large degree be attributed to the differences in the design in these studies e.g., in terms of the applied emission and climate scenarios. However, the chemistry-transport models also show differences in meteorology-air pollution relationships and so far, these differences have not been analyzed in detail (e.g., Kukkonen et al., 2012).

More advanced ecosystem models in recent studies have shown that the complex feedbacks between ecosystems and changes in climate, CO₂ levels, land-use etc. will modify the natural emissions of O₃ precursors to a larger degree than has been previously known. A robust quantification of the direction of this change is not possible with the current understanding, which is one of the main knowledge-gaps related to projections of future surface concentrations of O₃.

More research is needed for further development of climate and air pollution models, in terms of their capabilities in representing different chemical/physical processes and the response to climate change (Sokhi et al., 2022). There is a growing need to have integrated air pollution and related observations from both ground-based and remote sensing instruments, including in particular those on satellites. Use of upcoming new satellite observations such as the Plankton, Aerosol, Cloud, ocean Ecosystem (PACE) observations in general and SPEXone observations in particular, which can provide novel aerosol products related to size, composition, and absorption with high accuracy, while other instruments on PACE can provide accurate and novel cloud retrievals (droplet number concentration, droplet size, cloud phase), which can help improve the climate models in better quantifying aerosol-radiation and aerosol-cloud-precipitation interactions. Another emerging and growing field in observation science is the use of low-cost sensors, which can provide widespread information of air pollutant concentrations, but yet associated with large uncertainties when compared to existing monitoring networks. In order to get more robust projections with current generation models, multi-model studies will be needed, where scenarios and projection timelines are harmonized, as well as large ensemble studies using harmonized scenarios. For health impact studies, it is also important to increase the spatial resolution of the driving climate data in order to perform spatially high-resolution simulations of the evolution of the main air pollutants in the future. Higher accuracy in the air pollution simulations will also lead to more robust forecasting on the impacts of extreme weather and natural hazards such as wildland fires.