In the future, many Arctic–boreal processes will respond sensitively to climate change, affecting ecosystem productivity and functions. These changes may lead to unprecedented consequences, e.g. in the magnitude of the ecosystem carbon sinks, production of aerosol precursor gases, and surface albedo. We need first to develop methods for identifying the land regions and processes that are especially sensitive to climate change. Only after that are we able to analyse their responses.

Boreal forests are one of the largest terrestrial biomes, and account for around one-third of the Earth's forested area (Global Forest Watch, 2002; http://www.globalforestwatch.org/). Nearly 70 % of all boreal forests are located in the Siberian region. The forest biomass, soils, and peatlands in the boreal forest zone together constitute one of the world's largest carbon reservoirs (Bolin et al., 2000; Kasischke, 2000; Schepaschenko et al., 2013). Due to their large forest surface areas and huge stocks of carbon (∼ 320 gigatonnes of carbon; GtC), the boreal and Arctic ecosystems are significant players in the global carbon budget. Furthermore, permafrost, a dominant feature of Siberian landscapes, stores around 1700 GtC (Tarnocai et al., 2009). Boreal forests form the main vegetation zone in the catchment areas of large river systems, so they are an important part of the global water–energy–carbon feedbacks.

The forest biomass forms a climate feedback via the anticipated changes in nutrient availability and temperatures, affecting carbon sequestered into both the aboveground biomass and soil compartment. The Siberian forests are currently assumed to be a carbon sink, although with a large uncertainty range of 0–1 PgC yr-1 (Gurney et al., 2002). However, these ecosystems are vulnerable to global climate change in many ways, and the effects on ecosystem properties and functioning are complicated. While higher ambient CO2 concentrations and longer growing seasons may increase plant growth and productivity, as well as the storage of carbon to soil organic matter (e.g. Ciais et al., 2005; Menzel et al., 2006), warming affects respiration and ecosystem water relations in the opposite way (Bauerle et al., 2012; Parmentier et al., 2011). Expected acceleration of fire regimes might also substantially impact the carbon balance in Arctic and boreal regions (Shvidenko and Schepaschenko, 2013).

One example of the potentially large feedbacks is the critical role that permafrost plays in supporting the larch forest ecotone in northern Siberia. The boreal forests in the high latitudes of Siberia are a vast, rather homogenous ecosystem dominated by larch. The total area of larch forests is around 260 million ha, or almost one-third of all forests in Russia. Larch forests survive in the semi-arid climate because of the unique symbiotic relationship they have with permafrost. The permafrost provides enough water to support larch domination, and the larch in turn blocks radiation, protecting the permafrost from intensive thawing during the summer season. The anticipated thawing of permafrost could decouple this relationship, and may cause a strong positive feedback, intensifying the warming substantially.

The ambient temperature, radiation intensity, vegetation type, and foliar area are the main constraints for the emission of biogenic volatile organic compounds (BVOCs) (Laothawornkitkul et al., 2009). This makes BVOC emissions sensitive to both climate and land use changes, via, e.g., increased ecosystem productivity or the expansion of forests into tundra regions. Although the inhibitory effect of CO2 on the process level may be important, Arctic greening may strongly enhance the production of BVOCs in northern ecosystems (Arneth et al., 2007; Sun et al., 2013). Open tundra may also act as a significant source for BVOCs, especially if the snow cover period changes (Aaltonen et al., 2012; Faubert et al., 2012). This would lead to negative climate feedbacks involving either aerosol–cloud or aerosol–carbon cycle interactions (M. Kulmala et al., 2013, 2014; Paasonen et al., 2013). Linear trends in the annual maximum normalized difference vegetation index (NDVI) over 15 years in the northern areas of the Yamalo-Nenets Autonomous Okrug region in Russia, provide supporting evidence of the increasing biological activity and greening, and the potential for enhanced BVOC emissions (Fig. 2).

In summary, even small proportional changes in ecosystem carbon uptake can switch terrestrial ecosystems from a net carbon sink to a carbon source, with consequent impacts on atmospheric CO2 concentrations and global temperatures (e.g. Bala et al., 2013; Bodman et al., 2013, Mukhortova et al., 2015). This process has already been observed, particularly in disturbed forests of northern Asia (Shvidenko and Schepaschenko, 2014). Currently, we do not fully understand all the factors influencing carbon storage, or the links between biogeochemical cycles of carbon, water, and nutrients in a changing climate. However, the changes in these processes may be large, and their impacts may either amplify or decrease climate change, especially in the high northern latitudes.

The ecosystem structural changes are tightly connected to adaptation needs, and to the development of effective mitigation and adaptation strategies. Predictions concerning the shifting of vegetation zones are important for estimating the impacts of the region on future global GHG, BVOC, and aerosol budgets. Furthermore, natural and anthropogenic stresses, such as land use changes and biotic and abiotic disturbances, are shaping ecosystems in the Arctic and boreal regions and have many important feedbacks to climate (see, e.g., the review by Gauthier et al., 2015). In a warmer climate, northern ecosystems may become susceptible to insect outbreaks, drought, devastating forest fires, and other natural disasters. In addition, human impacts may cause sudden or gradual changes in ecosystem functioning. The ecosystem resilience is dependent on both the rate and magnitude of these changes. Recent studies have concluded that current estimates very likely overestimate the resilience of global forests and particularly boreal forests (Allen et al., 2015). In some cases, the changes may lead to system imbalance and possibly reaching a tipping point, after which the effects are irreversible. One of the most relevant research topics for the land system are to determine the structural changes and tipping points of the ecosystem changes in the northern pan-Eurasian region.

Part of the expected ecosystem structural changes is related to the lengthening of the growing season, which is taking place the Arctic–boreal regions due to climate change. This phenomenon, called “Arctic greening”, is due to increased plant biomass growth and advancing tree lines, turning previously open tundra into shrubland or forest (Myneni et al., 1997; Xu et al., 2015). However, “browning” as a proxy of decreased productivity has also been observed during recent decades in many boreal regions (Lloyd and Bunn, 2007), including vast territories of central Siberia, together with a general downward trend in basal area increment after the mid-20th century (Berner et al., 2013) and the overall decline in green from 2011 to 2014 in Arctic regions (Phoenix and Bjerke, 2016). Current predictions on the extent and magnitude of these processes vary significantly (Tchebakova et al., 2009; Hickler et al., 2012; Shvidenko et al., 2013a, b). It has been estimated that the northward shift of bioclimatic zones in Siberia will be as large as 600 km by the end of this century (Tchebakova et al., 2009). By taking into account that the natural migration rate of boreal tree species cannot exceed 200–500 m per year, such a forecast implies major vegetation changes in huge areas. In addition, we need to have a deeper understanding of the future role of the browning process and to re-analyse the previous model predictions of arctic greening: to what extent are they wrong, and why (Phoenix and Bjerke, 2016)? This has important biophysical consequences and climatic feedbacks. Changes in vegetation cover can, e.g., lead to albedo changes and therefore higher net absorption of radiation in regions covered by forests compared to open vegetation (Jeong et al., 2011). This modifies the local heat and vapour fluxes, and affects boundary layer conditions as well as both local and larger-scale climate (Sellers et al., 1997).

Northern peatlands contain a significant part of the global soil organic matter reservoirs (45 % of the world's soil carbon; Post et al., 1982), and comprise one of the world's largest GHG sources (in particular CH4) (IPCC, 2013). The hydrological conditions are a major factor in determining the functioning of peatlands as carbon source or sink, and the carbon balance of the vast northern peatlands is extremely sensitive to human influence, be it through either management or climate change. For example, thawing of permafrost peatlands in tundra regions might change tundra ecosystems from a stable state into a dynamically changing and alternating land–water mosaic, with dramatic impacts on their GHG production (Heikkinen et al., 2004; Repo et al., 2009). Today, peatland management activities range from drainage and peat harvesting to establishing crop plantations and forests. A complete understanding of the climatic effects of peatland management remains a challenging question (Maljanen et al., 2010).

Northern ecosystems are frequently suffering from increased stresses and deterioration. There is seldom a single and clear cause for forest dieback, but rather the ecosystems are suffering from multiple stresses simultaneously (e.g. Kurz et al., 2008a, b; Allen et al., 2010). This implies that a single stress factor may not be very dramatic for the resilience of the system, but when occurring simultaneously in combination with others, the system may cross a threshold (i.e. tipping point), and this may have dramatic consequences. Such perturbations and disturbances can include not only long-term pollutant exposures, but also stochastic events such as fires, flooding, windstorms, or insect population outbreaks, and human activities such as deforestation or the introduction of exotic plant or animal species. Disturbances of sufficient magnitude or duration can profoundly affect an ecosystem, and may force an ecosystem to reach a threshold beyond which a different regime of processes and structures predominates. Climate warming, precipitation changes during growth periods, and permafrost changes will substantially increase water stress, and consequently increase the risk of mortality for trees. This process has already clearly intensified over the entire circumpolar boreal belt (Allen et al., 2010). As a consequence, ecosystems may turn into carbon sources rather than sinks (Parmentier et al., 2011).

In the future, boreal forest diebacks may occur due to mass infections of invasive pathogens or herbivores, such as the autumnal moth (Epirrita autumnata) or mountain bark beetle (Dendroctonus ponderosae), which have previously been climatically controlled by harsh winter conditions. The growth and life cycles of herbivores or their habitat conditions may change in such a way that the outbreak frequencies and intensities of previously relatively harmless herbivore populations increase (Hunter et al., 2014). At the same time as climate is changing, boreal vegetation is also exposed to increased anthropogenic influences by pollutant deposition and land use changes (Dentener et al., 2006; Bobbink et al., 2010; Savva and Berninger, 2010). Large industrial complexes may lead to local forest diebacks, as has been observed in the Kola region (e.g. Nöjd and Kauppi, 1995; Tikkanen, 1995; Kukkola et al., 1997) and in some regions of Siberia (Baklanov et al., 2013). Societal transformations may lead to abandonment of agricultural land or deterioration of previously managed forests.

The major part of the northern Eurasian geographical region is covered by continuous permafrost. The fate of permafrost soils in high latitudes is important for global climate with regard to all greenhouse gases. Thawing of permafrost will also substantially alter the hydrological regimes, particularly in northern Asia, which will lead to increasing water stress in forests and explosive enlargement of fire extent and severity as well as post fire successions (Shvidenko et al., 2013b). These scenarios underline the urgent need for systematic permafrost monitoring, together with GHG measurements in various ecosystems. The treatment of permafrost conditions in climate models is still not fully developed (Bala et al., 2013). The major question is, how fast will the permafrost thaw proceed and how will it affect ecosystem processes and ecosystem–atmosphere feedbacks, including hydrology and greenhouse gas cycling.

Understanding of the feedbacks between carbon and water cycling, ecosystem functioning, and atmospheric composition related to permafrost thawing is one of the important topics of the land system study (Heimann and Reichstein, 2008; Schuur et al., 2009; Arneth et al., 2010). In high-latitude ecosystems with large, immobile carbon pools in peat and soil, the future net CO2 and CH4 exchange will depend on the extent of near-surface permafrost thawing, local thermal and hydrological regimes, and interactions with the nitrogen cycle (Tarnocai et al., 2009). The extra heat produced during microbial decomposition could accelerate the rate of change in active-layer depth, potentially triggering a sudden and rapid loss of carbon stored in carbon-rich Siberian Pleistocene loess (yedoma) soils (Khvorostyanov et al., 2008).

The connection between the climate and the thermal conditions in the subsurface layers (soil and bedrock) is an important aspect. The warming of the atmosphere will inevitably result in the warming of the permafrost layer, and is easily observed in deep borehole temperature data. However, the changes depend on the soil and rock type as well as on the pore filling fluids. As long as the pore fill is still ice, the climatic changes are reflected mainly in the thickness of the active layer, and in slow diffusive temperature changes of the permafrost layer itself. In areas where the ground is dominated by low ground temperatures and thick layers of porous soil types (e.g. sand, silt, peat), the latent heat of the pore-filling ice will efficiently “buffer” and retard the final thawing. This is one of the reasons why relatively old permafrost exists at shallow depths in high-porosity soils. On the other hand, quite different conditions prevail in low-porosity areas, e.g. in crystalline rock areas.

The permafrost dynamics affect methane fluxes in many ways. Hot spots such as mud ponds emitting large amounts of CH4 may form when permafrost mires thaw. In contrast, lakes have occasionally disappeared as a result of the intensification of soil water percolation (Smith et al., 2005). The rapid loss of summer ice, together with increasing temperature and melting ice deposits, results in coastal erosion, physical destruction of the surface in hilly areas, activation of old carbon and elevated CO2 and CH4 emissions from sea-bottom sediments (Vonk et al., 2012). High methane emissions have been observed from the East Siberian Arctic Shelf (Shakhova et al., 2010).

Atmospheric composition plays a central role in the northern Eurasian climate system. In addition to greenhouse gases and their biogeochemical cycling discussed in more detail in Sect. 3.2, key compounds in this regard are ozone and other oxidants, carbon monoxide, numerous organic compounds as well as different types of aerosols and their precursors (SO2 will be discussed in Sect. 3.5). At the moment, there is a serious gap in our knowledge on tropospheric composition and chemistry over Russia and China, with particularly few observation programs being active over Siberia (Crutzen et al., 1998; Ramonet et al., 2002; Paris et al., 2008; Kozlova et al., 2008; Uttal et al., 2015, Paris et al., 2010a, b; Sasakawa et al., 2010; Chi et al., 2013; Saeki et al., 2013; Ding et al., 2013a, b; Berchet et al., 2015; Heimann et al., 2014).

There is thus an urgent need for harmonized, coordinated and comprehensive greenhouse gas, trace gas, and aerosol in situ observations over northern Eurasia and China (long-term transport aspect) comparable to European and circumpolar data observations. In Fig. 3 we illustrate the geographical coverage of the ground stations that will be part of the coordinated, coherent, and hierarchic observation network in the northern Eurasian region and in China.

Little is known about whether and how the regional ozone budget in northern pan-Eurasia differs from that in the rest of the Northern Hemisphere (Ding et al., 2008; Berchet et al., 2013). Arctic tropospheric ozone is significantly influenced by long-range import of ozone and precursors from mid-latitude sources as well as by boreal wildfires (Ding et al., 2009; Wespes et al., 2012; Paris et al., 2010b; Vivchar et al., 2009). The role of biomass burning emissions in the ozone budget in high latitudes remains controversial. While most studies suggest significant ozone production in boreal smoke plumes (e.g. Paris et al., 2010b; Parrington et al., 2013; Jolleys et al., 2015), some observations from individual plumes suggest that O3 production in boreal wildfire plumes may be weaker, or even turn into net destruction, compared to fire plumes at lower latitudes (Liang et al., 2011; Jaffe and Wigder, 2012). Recent modelling work has suggested that boreal fires produce a substantial large-scale enhancement in summertime ozone at high latitudes, which appears to be highly sensitive to differences in partitioning of reactive nitrogen among models (Arnold et al., 2015). The boreal biosphere, on the other hand, provides a large sink for tropospheric ozone (Paris et al., 2010b; Parrington et al., 2013). Given their importance for air quality and global greenhouse gas budget, more atmospheric measurements of O3, its precursors and other pollutants over Siberia are needed (see Elansky et al., 2012). This is particularly the case in light of increasing local Arctic sources of ozone precursors (NOx, VOCs, volatile organic compounds) from, e.g., shipping and fossil fuel resource extraction (Roiger et al., 2015). Such data sets would be particularly useful for the evaluation of atmospheric chemistry models and satellite products.

The changes in the abundance of anthropogenic aerosols and their precursors in northern Eurasia have been extensive during the last decades (Granier et al., 2011), and this has almost certainly contributed to the very different regional warming patterns over these areas (e.g. Shindell and Faluvegi, 2009). The main anthropogenic aerosols in this context are primary carbonaceous particles, consisting of organic and black carbon as well as secondary sulfate particles produced during the atmospheric transport of sulfur dioxide. These species, as well as nitrate, have also been found to dominate the aerosol composition at the Zotino Tall Tower Observation Facility (ZOTTO) site in central Siberia (Mikhailov et al., 2015a, b; Ryshkevich et al., 2015). These aerosols cause large perturbations to the regional radiation budget downwind of major source areas in the northern Eurasian region, and the resulting changes in cloud properties and atmospheric circulation patterns may be important even far away from these sources (Koch and Del Genio, 2010; Persad et al., 2012). In the snow-covered parts of Eurasia, long-range transported aerosols containing black carbon and deposited onto snow tend to enhance the spring and early-summer melting of the snow, with concomitant warming over this region (Flanner et al., 2009; Goldenson et al., 2012; Meinander et al., 2013; Atlaskina et al., 2015).

The most important natural aerosol type over large parts of Eurasia is secondary organic aerosol originating from atmospheric oxidation of BVOCs emitted by boreal forests and possibly other ecosystems. Studies conducted in the Scandinavian part of the boreal zone indicate that new particle formation associated with BVOC emissions is the dominant source of aerosol particles and cloud condensation nuclei during summer time (Mäkelä et al., 1997; Kulmala et al., 2001; Tunved et al., 2006; Asmi et al., 2011; Hirsikko et al., 2011). The production of secondary organic aerosols associated with BVOC emissions has been estimated to induce large direct and indirect radiative effects over the boreal forest zone (Spracklen et al., 2008; Tunved et al., 2006; Lihavainen et al., 2009, 2015; Scott et al., 2014). The few continuous measurement data sets from Siberia suggest similarities in the frequency and seasonal pattern of new particle formation events between Siberia and Nordic stations (Dal Maso et al., 2007; Arshinov et al., 2012; Asmi et al., 2016). Measurements conducted at the ZOTTO site in central Siberia have shown that biogenic secondary organic aerosols reach high concentrations in summer and dominate the aerosol composition during this season (Mikhailov et al., 2015a, b; Ryshkevich et al., 2015). At this site, however, new particle formation events are seen much less frequently than at the Nordic stations (Heintzenberg et al., 2011). At present, relatively little is known about the overall contribution of biogenic emissions to aerosol number or mass concentrations, or to the cloud condensation nuclei budget, in northern Eurasia.

Other important natural aerosol types in northern Eurasia are sea spray, mineral dust, and primary biogenic aerosol particles. Sea spray aerosol makes an important contribution to the atmospheric aerosol over the Arctic Ocean and its coastal areas (Zábori et al., 2012, 2013), and influences cloud properties over these regions (Tjernström et al., 2014). The climatic effects of sea spray are expected to change in the future as a result of changes in the sea ice cover and ocean temperatures (Struthers et al., 2011). Mineral dust particles affect regional climate and air quality over large regions in Asia, especially during periods of high winds and moderate precipitation. Mineral dust and primary biological aerosol particles (PBAPs) particles are also effective ice nuclei (Hoose and Möhler, 2012), and have the potential to influence radiative and other properties of mixed-phase cold clouds in the Arctic–boreal regions. Over northern Eurasia, PBAPs typically contribute more than 20 % of PM2.5 (fine particulate matter with a diameter of 2.5 µm or less) organic aerosol mass concentrations (Heald and Spracklen, 2009) and 25 % of supermicron aerosol number concentrations (Spracklen and Heald, 2014). Ice nucleation, in general, is one of the key microphysical processes in the atmosphere that remain ill understood. However, a novel theoretical approach (Laaksonen, 2015; Laaksonen and Malila, 2016) has been shown to be superior to older theories in the case of water nucleation on solid surfaces, and it may open a completely new avenue in the studies of atmospheric ice formation.

Satellites provide information about spatial distributions of the column-integrated concentrations of aerosols (Andreae, 2009) and various trace gases including ozone and its precursors (Burrows et al., 2011). These atmospheric constituents are generally retrieved using passive instruments, which have good sensitivity near the surface. However, retrieving information on the near-surface concentrations of pollutants requires assumptions on their vertical distributions. For instance, the retrieval of tropospheric ozone from satellite observations requires corrections for the high concentrations in the upper troposphere and lower stratosphere. For aerosols, which can only be retrieved in clear-sky conditions, the situation may be complicated when disconnected layers are present with different types of aerosols. A solution may be the retrieval of aerosol vertical variation or the height of the aerosol layer using, e.g., active instruments (lidars), or retrieval using spectrally resolved observations in the oxygen A-band (e.g. Hollstein and Fisher, 2014), or instruments providing multiple viewing algorithms such as MISR (Nelson et al., 2013) or AATSR (Virtanen et al., 2014). Another complication for aerosols may be the vertical variation of the physical and chemical properties, which renders it difficult to obtain closure between column and ground-based in situ measurements (Zieger et al., 2015, and references cited therein). Nevertheless, good progress has been made in aerosol retrieval, and column-integrated aerosol measurements (aerosol optical depth, AOD) from satellites and ground-based observations compare favourably (e.g. de Leeuw et al., 2015; Kolmonen et al., 2015). Measurements of trace gases from space using wavelengths in the thermal infrared suffer from low sensitivity in the lower troposphere (Pommier et al., 2010). All these factors may render the comparison against local ground-based in situ observations difficult, although a possible way out could be the use of chemical transport models constrained by the satellite column measurements (e.g. de Laat et al., 2009; Stavrakou et al., 2012, 2014), possibly together with sub-orbital airborne measurements of relevant species. Satellite-measured AOD has been successfully applied to obtain information on ground-based aerosol mass concentrations (PM2.5) (Xu et al., 2015; van Donkelaar et al., 2015). In addition, the use of multiple satellite instruments, with different characteristics, is proposed to obtain more accurate information on the transport of aerosols and trace gases and their vertical distribution (e.g. Naeger et al., 2016). Recently, a technique has been demonstrated that makes it possible to derive CCN (cloud condensation nuclei) concentrations at cloud base using remote sensing of cloud properties (Rosenfeld et al., 2016).

Of particular interest is the pollutant transport to Arctic areas, where they can influence the radiation budget and climate in various ways (Stohl, 2006; Warneke et al., 2009; Meinander et al., 2013; Eckhardt et al., 2015). Model simulations suggest that European emissions dominate Arctic pollutant burdens near the surface, with sources from North America and Asia more important in the mid- and upper troposphere (Monks et al., 2015). The impact and influence of China and its polluted megacities on Arctic and boreal areas is a topic of key importance, given recent and rapid Chinese industrialization. Inter-continental pollution transport has also become of increased concern due to its potential influence on regional air quality. The pollutant export from North America and Asia has been characterized by intensive field campaigns (Fehsenfeld et al., 2006; Singh et al., 2006), but long-term research approaches are lacking.

Emissions from forest fires (van der Werf et al., 2006; Sofiev et al., 2013) and from agricultural fires in southern Siberia, Kazakhstan, and Ukraine (Korontzi et al., 2006) in spring and summer are large sources of trace gases such as carbon monoxide (Nédélec et al., 2005; Konovalov et al., 2014) as well as aerosol particles (Konovalov et al., 2015). Aerosols emitted by forest fires are of particular interest, since the strength of this source type depends on both climate change and human behaviour (Pechony and Shindell, 2010), and since particles emitted by these fires have potentially large radiative effects over Eurasia (Randerson et al., 2006). We need comprehensive top-down emissions estimates, using inverse modelling constrained by satellite observations, in order to provide quantitative information on the source strength of aerosols and trace gases emitted by open fires.

Air pollution in monsoon Asia has two main characteristics. First, the total pollutant emission rate from fossil fuel combustion sources is very high, leading to a high concentration of primary and secondary pollutants in Asia, especially in eastern China and northern India. Observations show that Asia is the only region where the concentrations of key pollutants, such as nitrogen oxides (Richter et al., 2005; Mijling et al., 2013) and their end-product ozone (Ding et al., 2008; Wang et al., 2009; Verstraeten et al., 2015), are still increasing. Second, in addition to the anthropogenic fossil fuel combustion pollutants, monsoon Asia is also influenced by intensive pollution from seasonal biomass burning and dust storms. For example, intensive forest burning activities often take place in south Asia during spring and in Siberia during summer, whereas intensive anthropogenic burning of agricultural straw takes place in the north and east China plains. Dust storms frequently occur in the Taklimakan and Gobi deserts in north-west China, and this dust is often transported over eastern China, southern China, the Pacific Ocean and even the entire globe (Nie et al., 2014). After mixing with other anthropogenic pollutants, biomass burning and mineral dust aerosols have been found to cause complex interactions in the climate system (Ding et al., 2013a; Nie et al., 2014).

The northern Eurasian urban environments are characterized by cities with strong anthropogenic emissions from local industry, traffic, and housing in Russia and China, and by megacity regions with alarming air quality levels like those of Moscow and Beijing. Bad air quality has serious health effects and damages ecosystems. In Beijing, for example, concentrations of atmospheric fine particles have been found to be more than 10 times higher than the safe level recommended by the World Health Organization (WHO) (Zheng et al., 2015). Furthermore, atmospheric pollutants and oxidants play a central role in climate change dynamics via their direct and indirect effects on global albedo and radiative transfer. A deeper understanding of the unpredicted chemical reactions between pollutants and identification of the most relevant feedbacks between air quality and climate at northern high latitudes and in China is the most urgent task helping us to find practical solutions for more healthy air (Kulmala, 2015).

In Siberian cities, the air quality is strongly linked to climatic conditions typical for Siberia. Stable atmospheric stratification and temperature inversions are predominant weather patterns for more than half of the year. This contributes to the accumulation of different pollutants in the lowest layers of the atmosphere, thus increasing their impact on ecosystems and humans. In addition to the severe climatic conditions, human impacts on the environment in industrial areas and large cities continue to increase. In winter time, shallow and stably stratified planetary boundary layers (PBL) typical for northern Scandinavia and Siberia are especially sensitive to even weak impacts and, therefore, deserve particular attention, especially in the conditions of environmental and climate change (Zilitinkevich and Esau, 2009; Esau et al., 2012; Davy and Esau, 2014; Wolf et al., 2014; Wolf and Esau, 2014). Unstably stratified PBLs interact with the free atmosphere mainly through turbulent ventilation at the PBL upper boundary (Zilitinkevich, 2012). This mechanism, still insufficiently understood and poorly modelled, controls the development of convective clouds, as well as dispersion and deposition of aerosols and gases, which are essential features of heat waves and other extreme weather events.

The worst air pollution episodes are usually associated with stagnant weather conditions with a shallow PBL, which promotes the accumulation of intensively emitted pollutants near the surface. The lower PBL is also influenced by the heavy pollution itself through its direct or indirect effects on solar radiation and hence the surface sensible heat flux (e.g. Ding et al., 2013b). The boundary layer–air pollution feedback will decrease the height of the PBL and result in an even more polluted PBL (Ding et al., 2013b; Wang et al., 2014; Petäjä et al., 2016). Therefore, considering the complex land-surface types (city clusters surrounded by agricultural areas) and pollution sources, improving our understanding of the associated feedbacks is very important for forecasting extreme air pollution episodes and for long-term policymaking. In order to understand this topic, more vertical measurements using aircraft, balloons, and remote sensing techniques, as well as advanced numerical models including all relevant processes and their couplings, are needed.

Planetary boundary layers are subject to diurnal variations, absorb surface emissions, control microclimate, air pollution, extreme colds, and heat waves, and are sensitive to human impacts. Very stable stratification in the atmosphere above the PBL prevents the compounds produced by the surface fluxes or surface emissions from efficiently penetrating from the PBL into the free atmosphere. This means that the PBL height and turbulent fluxes through the PBL upper boundary control local features of climate and extreme weather events, such as the heat waves associated with convection, or the strongly stable stratification events triggering the air pollution (Zilitinkevich et al., 2015). This concept (equally relevant to the hydrosphere) illustrates the importance of modelling and monitoring the atmospheric PBL height, which varies from dozens to thousands of metres (Zilitinkevich, 1991; Zilitinkevich et al., 2007; Zilitinkevich and Esau, 2009). To carry out a comprehensive inventory of the PBL height over northern Eurasia is urgently needed.

The ongoing environmental change and its amplification in northern Eurasia pose special challenges to the prediction of weather-related hazards, and also to long-term impacts. A key question is how the atmospheric dynamics (synoptic scale weather, boundary layer characteristics) will change in Arctic and boreal regions. The recent changes in the Arctic sea ice have been much more rapid than models and scientists anticipated about 10 years ago. The role of the Arctic Ocean in the climate system and sea ice changes have affected mid-latitude weather and climate, with central and eastern Eurasia among the regions with strongest effects (Vihma, 2014; Overland et al., 2015) (see Sect. 2.3.1).

The reliability of weather forecasts, and the extension of the time range of useful forecasts is needed for minimizing economic and human losses from extreme weather and extreme weather-related natural hazards. In Europe, this range is currently on average about 8–9 days (Bauer et al., 2015), which allows reliable early warnings to be issued for weather-related hazards, such as windstorms and extreme precipitation events with flash floods. The time range of useful forecasts has typically increased by a day per decade over the past three decades (Uppala et al., 2005). In the northern Eurasian region, improved predictions can be used, for instance, for better prediction of thermal comfort conditions in northern cities (Konstantinov et al., 2014). A strong urban heat island effect has already been observed in urban areas of the Arctic with complex spatial and temporal structures (Konstantinov et al., 2015).

Understanding of PBL processes is particularly important for improving the weather predictions. The representation of boundary layer clouds, and their further coupling to convection in stable conditions is not currently well understood. Quantification of the behaviour of the PBL over the northern Eurasian region is needed in analyses of spatial and temporal distribution of the surface fluxes, in predictions of microclimate and extreme weather events, and in modelling clouds and air quality.

The development of diagnostic and modelling methods for aero-electric structures is important for a study of both convective and electric processes in the lower troposphere (Shatalina et al., 2005, 2007). Convection in the PBL leads to the formation of aero-electric structures, manifested in ground-based measurements as short-period electric-field pulsations with periods from several seconds to several hundreds of seconds (Anisimov et al., 1999, 2002). The sizes of such structures are determined by the characteristic variation scales of aerodynamic and electrodynamics parameters of the atmosphere, including the PBL and surface-layer height as well as by the inhomogeneities in the ground (water) surface. Formed as a result of convective processes and the capture of positive and negative charged particles (both ions and aerosols) by convective elements (cells), aero-electric structures move with the airflow along the Earth's surface. The further evolution of convective cells results, in particular, in cloud formation.

The global electric circuit (GEC) is an important factor connecting the solar activity and upper atmospheric processes with the Earth's environment, including the biosphere and climate (Dolezalek et al., 1976; Singh et al., 2004). Thunderstorm activity maintains this circuit, whose appearance is dependent on atmospheric conductance variations over a wide altitude range. The anthropogenic impact on the GEC through aviation, forest fires, and electromagnetic pollution has been noted with great concern, and the importance of lightning activity in climate processes has been recognized. The GEC forms for two reasons: the continuous operation of ionization sources, which provides an exponential growth of the conductivity in the lower atmosphere, and the continuous operation of thunderstorm generators, providing a high rate of electrical energy generation and dissipation in the troposphere. Therefore, the GEC is influenced by both geophysical and meteorological factors, and can serve as a convenient framework for the analysis of possible inter-connections between atmospheric electrical phenomena and climate processes. Further exploration of the GEC as part of the climate system studies, specifically its effect on the balance between the Earth's ionosphere and global circuit, requires accurate modelling of the GEC stationary state and its dynamics (Mareev, 2010). Special attention should be paid to the observations and modelling of generators (thunderstorms, electrified shower clouds, mesoscale convective systems) in the global circuit.

The essential processes related to the interaction between the Arctic ocean and other components of the Earth system include the air–sea exchange of momentum, heat, and matter (e.g. moisture, aerosol, trace gases, CO2, and CH4) as well as the dynamics and thermodynamics of sea ice. The most dramatic change in the Arctic Ocean has been the rapid decline of the sea ice cover. Since the early the 1980s, the Arctic sea ice extent has decreased by roughly 50 % in summer and autumn (Cavalieri and Parkinson, 2012), while the winter sea ice thickness in the central Arctic has decreased by approximately 50 % (Kwok and Rothrock, 2009). Arctic sea ice changes have serious teleconnections. Despite the warming climate, wintertime cold spells in East Asia have become more frequent, stronger and longer lasting in this century compared with the 1990s (Kim et al., 2014). It also seems that the strong decline of the Arctic sea ice has favoured atmospheric pressure patterns that generate cold-air outbreaks from the Arctic to East Asia (Mori et al., 2014; Kug et al., 2015; Overland et al., 2015). The reasons for and the future evolution of the sea ice decline, as well as its effects on the ocean, atmosphere and surrounding continents are among the current topics of study on the Arctic climate system. Other major issues include the role of the ocean in the Arctic amplification of climate change, greenhouse gas exchange between the ocean, sea ice, and atmosphere as well as aerosol budgets in the marine Arctic (Smedsrud et al., 2013). The key question here is related to the changes of sea ice extent and thickness, and to the terrestrial snow cover change.

Many of the processes considered to be responsible for the Arctic amplification of climate warming are related to the ocean and sea ice (Döscher et al., 2014). Among these, the snow-/ice-albedo feedback has received the most attention (e.g. Flanner et al., 2011). This feedback is strongest when sea ice is replaced by open water, but it starts to play a significant role already in spring when the snowmelt on top of sea ice begins. This is because of the large albedo difference between dry snow (albedo about 0.85) and wet, melting, bare ice (albedo about 0.40). More work is needed to understand quantitatively the reduction of snow/ice albedo during the melting season, including the effects of melt ponds and pollutants in the snow. Other amplification mechanisms related to the ocean include increased heat transports from lower latitudes to the Arctic (Polyakov et al., 2010; Döscher et al., 2014) and fall–winter energy loss from the ocean (Screen and Simmonds, 2010). Furthermore, the melting of sea ice strongly affects evaporation, and hence the water vapour and cloud radiative feedbacks (Sedlar et al., 2011), and the PBL thickness, which controls the sensitivity of the air temperature to heat input into the PBL (Esau et al., 2012; Davy and Esau, 2016). The relative importance of the mechanisms affecting the Arctic amplification of climate warming are not yet well known (see also Pithan and Mauritsen, 2014; Cohen et al., 2014).

The rapid decline of the Arctic sea ice cover has tremendous effects on navigation and exploration of natural resources. To be able to predict the future evolution of the sea ice cover, the first priority is to better understand the reasons, including the role of black carbon (see Bond et al., 2013), behind the past and ongoing sea ice evolution. Several processes have contributed to the decline of Arctic sea ice cover, but the role of these processes needs better quantification (Smedsrud et al., 2013; Vihma et al., 2014). Further studies are needed on the impacts of changes in cloud cover and radiative forcing (Kay et al., 2008), atmospheric heat transport (Kapsch et al., 2013) and oceanic heat transport (Döscher et al., 2014). In addition, as the ice thickness has decreased, the sea ice cover becomes increasingly sensitive to the ice-albedo feedback (Perovich et al., 2008). Other issues calling for more attention include the reasons for the earlier onset of the spring melt (Maksimovich and Vihma, 2012), changes in the phase of precipitation (Screen and Simmonds, 2012), and large-scale interaction between the sea ice extent, sea surface temperature distribution, and atmospheric dynamics (cyclogenesis, cyclolysis, and cyclone tracks) as discussed, e.g. by Outten et al. (2013).

In addition to thermodynamic processes, another factor affecting the sea ice cover in the Arctic is the drift of sea ice. The momentum flux from the atmosphere to the ice is the main driver of sea-ice drift, which is poorly represented in climate models (Rampal et al., 2011). This currently hinders a realistic representation of sea-ice drift patterns in large-scale climate models. Furthermore, the progressively thinning ice pack is becoming increasingly sensitive to wind forcing (Vihma et al., 2012). In the future, research has to address the main processes that determine the momentum transfer from the atmosphere to the sea ice, including the effects of atmospheric stratification and sea ice roughness.

To understand better the links between the Arctic Ocean and terrestrial Eurasia, there is a particular need to study the effects of Arctic sea ice decline on Eurasian weather and climate (Sect. 2.2.3) Another poorly studied problem related to the Arctic Ocean is the role of sea ice as a source of aerosol precursors, and in the gas exchange between the ocean and atmosphere (Parmentier et al., 2013). Preliminary results of field studies at the drifting stations North Pole 35 and 36 (Makshtas et al., 2011) showed that the shrinking sea ice cover could be the reason for increasing CO2 uptake from the atmosphere over the annual cycle, and for the growth of the seasonal amplitude of CO2 concentrations in the Arctic.

Climate models project that air temperatures and precipitation will increase over the Arctic Ocean, and that this may have important effects on the structure of sea ice. Increased snow load on a thinner ice may in the future cause flooding of seawater on ice in the Arctic, which results in the formation of snow ice. Increased snowmelt and rain, on the other hand, results in increased percolation of water to the snow–ice interface, where it re-freezes, forming super-imposed ice (Cheng et al., 2008). Snow ice and super-imposed ice have granular structures, and differ thermodynamically and mechanically from the sea ice that currently prevails in the Arctic.

The changes in the Arctic Ocean have opened some, albeit limited, possibilities for seasonal prediction. These are mostly related to the large heat capacity of the ocean: if there is little sea ice in the late summer and early autumn, this tends to cause large heat and moisture fluxes to the atmosphere, favouring warm, cloudy weather in late autumn and early winter (Liu et al., 2011; Stroeve et al., 2012). On the other hand, the reduction of the sea ice thickness may decrease the possibilities for seasonal forecasting of ice conditions in the most favourable navigation season in late summer–early autumn. This is because a thin ice is very sensitive to unpredictable anomalies in the atmospheric forcing. For example, in August 2012 a single storm caused a reduction of the sea ice extent by approximately 1 million km2. The reduced sea ice extent in the winter months has significant impacts on convective clouds. Observations revealed a gradually increasing frequency of the convective cloud fields over Norwegian and Barents seas (Chernokulsky and Mokhov, 2012; Esau and Chernokulsky, 2015). The unusually strong atmospheric convection and weaker virtual potential temperature inversions create favourable conditions for the extreme Arctic cold outbreaks and meso-scale cyclones known as polar lows (Kolstad et al., 2009).

It is vital to enhance routine observations, data assimilation techniques and prediction models in order to properly monitor the physical state of the environment. Longer-term impacts of the reduced ice cover are largely unknown, because the scientific community has had only little time to create new knowledge on essential climate variables across the domain (see Sect. 2.3.1). To improve preparedness, new observational evidence is therefore needed to reduce uncertainties in the system dynamics both on short and longer time-scales.

The ice cover of the Arctic Ocean is undergoing fast changes, including a decline of summer ice extent and ice thickness (see Sect. 2.3.1). This results in a significant increase of the ice-free sea surface in the vegetation season, and an increase in the duration of the growing season itself. The key topic of future research is the joint effect of Arctic warming, ocean freshening, pollution load, and acidification on the Arctic marine ecosystem, primary production, and carbon cycle.

New ice-free areas of the Arctic Ocean could result in a pronounced growth of the annual gross primary production (GPP), increased phytoplankton biomass, and a loss of ice-rich algae communities associated with the low ice sheet surface (Bluhm et al., 2011). Progressive increase of oil and natural gas drilling and transportation over the shelf areas will be escalating the environmental changes of the Arctic marine ecosystems. Furthermore, there is a risk of irreversible changes in marine Arctic productivity and key biogeochemical cycles, and the potential for CO2 absorption by marine ecosystem. Processes involving the Arctic may also affect adjacent boreal areas.

We do not know how the climatically induced increase in GPP and phytoplankton biomass will influence the productivity of higher trophic levels of the Arctic ecosystem. In typical Arctic ecosystems, the most important consumers are large-sized herbivorous copepods, which have life cycles synchronized with the temperature as well as the seasonal algae dynamics (Kosobokova, 2012). Another important consumer community are the small-sized herbivorous copepods, which are important especially in shelf ecosystems. An increase in the phytoplankton production in fall, together with an increase in the sea temperature, may influence the populations of small-sized copepods, and increase their role in mass and energy flow in the ecosystems. Our current understanding of the role of small copepods in the Arctic ecosystems is limited (Arashkevich et al., 2010). An increase in surface water temperature may “open the Arctic doors” for new species, and change the Arctic pelagic food webs, energy flows, and biodiversity.

Increases in the Arctic sea temperature may lead to populations from neighbouring regions penetrating the Arctic ecosystem, changing the structure and functioning of native ecosystems. For example, a 1.5 ∘C water temperature increase in the Bering Sea during the mid-1970s allowed the Alaskan Pollock to penetrate the Arctic ecosystem, and occupy a place as a keystone species for several years, supporting one of the world's largest regional fish harvests (Shuntov et al., 2007). The Bering Sea ecosystem is very rich compared to the Arctic ecosystems. Currently, we are not aware of food sources sufficient for supporting massive invader populations even in case of climate-induced changes in ecosystems. However, the appearance of aggressive new species even in low numbers may dramatically impact the sensitive Arctic ecosystems and have effects on the future regulation of international fisheries in the Arctic.

We have only recently begun to understand the processes that regulate freshwater–marine ecosystem interactions in estuarine zones (Flint, 2010). The mechanisms determining the impact of riverine waters over the Arctic shelves and the central deep basin, and their dependence on specific climatic forces, are still poorly understood. In order to determine the impact of riverine waters, it is important to locate new flagship stations or permanent observation points in the estuaries of large Siberian rivers. The changing riverine discharge to the Arctic shelves may amplify the impact of climate warming on the Arctic marine ecosystems. Degradation of permafrost, soil erosion, changes in snow cover and summer precipitation may all lead to changes in flood timing, and also to an increase in the amount of fresh water and materials of terrestrial origin, including organic matter and nutrients, annually delivered to the Arctic shelves, and further to the Arctic basin (Gustafsson et al., 2011). Human-driven land use changes to drainage basins and associated river systems have the potential to increase the speed of delivery of pollutants to the Arctic sea.

In the last decade, the combined effects of air pollution and climate warming on fresh-water systems have received increasing attention (Skjelkvåle and Wright, 1998; Schindler, 2001; Alcamo et al., 2002; Sanderson et al., 2006; Feuchtmayr et al., 2009; Sereda et al., 2011). It is important to understand the future role of Arctic–boreal lakes, wetlands, and large river systems, including thermokarst lakes and running waters of all size, in biogeochemical cycles, and how these changes affect livelihoods, agriculture, forestry, and industry. The water chemistry of lakes without any direct pollution sources in the catchment area can be expected to reflect regional characteristics of water chemistry, as well as global anthropogenic processes, such as climate change and long-range air pollution (Müller et al., 1998; Moiseenko et al., 2001; Battarbee et al., 2005). The current ground-based streamflow-gauging network over the northern Eurasian region does not provide adequate spatial coverage for many scientific and water management applications, including the verification of the land-surface run-off contribution to the recipients of intra-continental run-off. Special field laboratories, with joint observation and modelling capabilities in hydrometeorology, sedimentology, and geochemistry are needed to understand the spreading of tracers and pollutants as part of current and future global environmental fluxes.

The gradient in water chemistry from the tundra to the steppe zones in Siberia can provide insight into the potential effects of climate change on water chemistry. In the last century, long-range trans-boundary air pollution led to changes in the geochemical cycles of sulfur, nitrogen, metals, and other compounds in many parts of the world (Schlesinger, 1997; Vitousek et al., 1997a, b; Kvaeven et al., 2001; Skjelkvåle et al., 2001). Environmental pollution problems also include the waterborne spreading of nutrients and pesticides from local agricultural areas, heavy metals often originating from mining areas, and other elements and chemicals, such as persistent organic pollutants from urban and industrial areas. Shifts in downstream loads cause changes in river and delta dynamics. One example of important study area is the Selenga River basin, which is located in the centre of Eurasia, extends from northern Mongolia into southern Siberia (Russia), and has its outlet at Lake Baikal. The Selenga River basin and Lake Baikal are located in the upstream part of the Yenisei River system, which discharges into the Arctic Ocean. Lake Baikal has the largest lake volume in the world at about 23 000 km3 (comprising 20 % of all unfrozen freshwater in the world), hosts a unique ecosystem (Granina, 1997), and is an important regional water resource (Garmaev and Khristoforov, 2010; Brunello et al., 2006). There are numerous industries and agricultural activities within the Selenga River basin, which affect the water quality of the lake and its tributaries. Mining is well-developed in the region (e.g. Karpoff and Roscoe, 2005; Byambaa and Todo, 2011), and heavy metals accumulate in biota and in sediments of the Selenga River delta and Lake Baikal (Boyle et al., 1998; Rudneva et al., 2005; Khazheeva et al., 2008).

In addition to water chemistry, the role of aquatic systems as a net sink or source for atmospheric CO2 is presently under debate. When precipitation or other processes transport large volumes of organic matter from land into nearby lakes and streams, the carbon of this matter effectively disappears from the carbon budget of the terrestrial ecosystem (Huotari et al., 2011). The enhanced decomposition of soil organic matter may significantly affect the transport of terrestrial carbon to rivers, estuaries, and the coastal ocean. The contribution of this process to the global and regional carbon budgets is unknown. Thus, the biological processes taking place in the terrestrial ecosystem (e.g. photosynthesis, respiration, and decomposition) and in the aquatic ecosystem are interlinked. The observed higher temperature response of aquatic ecosystems compared to terrestrial ecosystems indicates that a substantial part of the carbon respired or emitted from the aquatic system must be of terrestrial origin (Yvon-Durocher et al., 2012). Long-term measurements carried out during all seasons in the littoral zone of Lake Baikal showed that maximum CO2 sink and emission rates are observed in August and December (during the pre-ice period), respectively, and the total CO2 flux from the atmosphere into the littoral zone of Lake Baikal was estimated to be 3–5 g CO2 m-2 (Domysheva et al., 2013).

The Siberian lakes situated in tundra and forest–tundra zones are in general poorly studied. In their natural state, their productivity is low, but their ecosystems are highly sensitive to external influences. Profuse blooming of cyanobacteria is usually associated with urban and industrial effluents and nutrient run-off. An assessment is needed of the impact of climate change in the northern Eurasian region on eutrophication, accompanied by blooms of cyanobacteria. Besides, the northern Eurasian region is characterized by thaw lakes, which comprise 90 % of the lakes in the Russian permafrost zone (Romanovsky et al., 2002). These lakes, which are formed in melting permafrost, have long been known to emit CH4. The latest observations of the lakes in the permafrost zone of northern Siberia indicate that they are releasing much more CH4 into the atmosphere than previously thought. Rather than being emitted in a constant flow, 95 % of CH4 comes from random bubbling in disperse locations. In coming decades, this could become a more significant factor in global climate change (Walter et al., 2006).

One direct consequence of climate change is the explosive reproduction of toxic cyanobacteria (Nodularia, Microcystis, Anabaena, Aphanizomenon, Planktothrix) and diatoms (Pseudo-nitzschia) (Moore et al., 2008; Paerl and Huisman, 2009). These blooms occur in ponds, lakes, reservoirs, and bays of the sea. Cyanobacteria and diatoms excrete especially dangerous carcinogens and neurotoxins into the water. The toxicity of some cyanotoxins exceeds the toxicity of currently banned warfare agents. Antidotes to these toxins do not exist at the moment.

Water conservation has received an increasing attention in China, and multiple new projects have been initiated recently. Especially the construction of water transfer, reservoir, and irrigation schemes have received much attention, because the central and western regions of China are suffering from water shortages. These projects are expected to improve water usage and security, especially for agricultural activities, and to provide sufficient water resources for local societies. In China, the river systems are dominated by rivers flowing from the Tibetan plateau to the Pacific Ocean. The Yangtze is the longest river in China, and flows from the Tibetan plateau to Shanghai. The Yellow river is the second longest in China, and it is characterized by seasonal flooding, which causes great economic and societal losses. The Amur River forms the northern border with Russia. The Haihe River flows through Beijing to Tianjin, and is under heavy stress from the highly populated and industrialized capital metropolitan region. Only one river from China flows to the Arctic Ocean: the Ertix River, which flows to the north through Kazakhstan, across Siberian Russia, finally joining the Ob River, which flows to the Arctic Ocean.

The fundamental large-scale task is to estimate how human actions such as land use changes, energy production, the use of natural resources, changes in energy efficiency, and the use of renewable energy sources will influence the environments and societies of the northern Eurasian region. For example, the industrial development of Siberia should be considered as one of most important drivers of future land use and land cover changes in Russia. Siberia is a treasure chest of natural resources of Russia, containing 85 % of its prospected gas reserves, 75 % of its coal reserves, and 65 % of its oil reserves. Siberia has more than 75 % of Russia's lignite, 95 % of its lead, approximately 90 % of its molybdenum, platinum, and platinoids, 80 % of its diamonds, 75 % of its gold, and 70 % of its nickel and copper (Korytnyi, 2009).

During the 20th century, a considerable transformation of landscapes in the tundra and taiga zones in northern Eurasia has occurred as a result of various industrial, socio-economic and demographic processes, leading to the industrial development of previously untouched territories (Bergen et al., 2013). This has led to a decrease in the rural population and, mostly after the 1990s, to decrease in agricultural activities. There has also been a significant reduction in agricultural land use, and its partial replacement by zonal forest ecosystems (Lyuri et al., 2010). According to recent estimates, the total area of abandoned agricultural land in Russia in the 1990s to 2010s is at about 57 million ha, of which 18 million ha have been restored by forests and 6 million ha of this are located in Asian Russia (Schepaschenko et al., 2015). As a result, these areas have become active accumulators of atmospheric CO2 (Kalinina et al., 2009). These new forests (substituting resources) could form the basis for sustainable development in these regions, in case relevant management programs for the forests re-established on abandoned lands are going to be implemented.

The dynamics of land cover, particularly forests, have been documented since 1961 when the results of the first complete inventory of Russian forests were published. According to official statistics, the area of forests in Asian Russia increased by around 80 million ha during 1961–2009, mostly before the middle of the 1990s. This large increase is explained by improved quality of forest inventories in remote territories, natural reforestation, mostly during the Soviet era as a result of forest fire suppression, and encroaching forest vegetation in previously non-forested land. Based on official statistics, the area of cultivated agricultural land in the region decreased by around 10 million ha between 1990 and 2009. After the year 2000, the forested area in Siberia decreased, mostly due to fire and the impacts of industrial transformations in high latitudes (Shvidenko and Schepaschenko, 2014). A critical decrease in the forest area has also been observed in the most populated areas with intensive forest harvesting particularly in the southern part of Siberia and the Far East. For example, in the Krasnoyarsk Krai, the total area of forests decreased by 5 %, while that of mature coniferous forests decreased by 25 %. Overall, the typical processes in these regions are a dramatic decline in the quality of forests, unsustainable use of forest resources, and insufficient governance and forest management in the region, including frequent occurrence of illegal logging, natural, and human-induced disturbances (Shvidenko et al., 2013a).

Future land use and land cover changes will crucially depend on how successfully the strategy of sustainable development of northern territories is developed and implemented. An effective system for the adaptation of boreal forests to global change needs to be developed and implemented in the region. An “ecologization” of the current practices of industrial development of previously untouched territories would allow a substantial decrease in the physical destruction of landscapes, and halt the decline of surrounding ecosystems due to air pollution and water and soil contamination (Kotilainen et al., 2008).

The expected changes in the climate and environment will have multiple and complicated impacts on ecosystems, with consequent land cover changes. The alteration of fire regimes and the thawing of permafrost will intensify the process of “green desertification” in large areas. Climate warming will have multiple effects on soil–vegetation–snow interactions. For example, in a warmer climate, mosses and other vegetation grow faster, providing a better thermal insulation of the permafrost in summer, and better feeding conditions for reindeer. However, snow can also more easily accumulate on thicker vegetation, thus protecting the deeper soil from cooling during the winter (Tishkov, 2012).

Both north and east Russia possess abundant mineral resources (Korytnyi, 2009). The resource orientation of northern and eastern Russia's economy, which has not changed for centuries, increased in the post-Soviet period, and has been influenced primarily by the product market. It is also expected that the natural resource development sector will continue to dominate the economy in the majority of these territories for the next decades.

A crucial factor in greenhouse gas emission dynamics is the fuel balance. In Russia, features of the fuel balance have led to an increased pollution. On average, specific emissions in the northern and eastern cities of Russia, where coal accounts for most of the power generation, are 3 times higher than in cities where power is generated mainly from gas or fuel oil (Bondur, 2011a). The geographical location, undeveloped infrastructure, harsh climate, and coal burning are the main reasons for increased levels of anthropogenic pollution in these areas (Bondur and Vorobev, 2015; Bondur, 2014). In small towns, low-capacity boiler rooms are the main source of emissions. Usually, the lack of financial resources leads to the use of low-quality coal and obsolete boilers. In the steppe zone of Asian Russia, Mongolia, Kazakhstan, and Buryatia, the main source of emissions is the burning of harvest residues.

The dynamics of GHG emissions in Russia are largely determined by the economic conditions of production. The economic crisis in 1990–1998 slowed down environmental degradation to some extent: emissions generally decreased by 40 %. However, the underlying environmental problems not only remained unresolved, but significantly deepened, and turned into systemic problems. The most polluting industries were more resistant to the decline in production. Technological degradation took place, cleaning systems were eliminated, and production shifted to part-time, leading to inefficient capacity utilization. Significant amounts of pollution continued to be emitted from the domestic sector. Emissions decreased in most regions of the country, and in 83 % of the cities, but much more slowly than production. As a result, the specific emissions (per product cost at comparable prices) had grown by the end of the 1990s in all categories of cities, except cities with more than 1 million inhabitants (Bityukova et al., 2010). All this can cause negative impacts on ecosystems. For example, there are about 2 million ha of technogenic deserts around Norilsk. Norilsk is probably the biggest smelter in the world, and produces more than 2 million t of pollutants per year (Groisman et al., 2013).

The frequency and intensity of weather extremes have increased substantially during the last decades in Europe, Russia, and China. Further acceleration is expected in the future (IPCC, 2013). The evolving impacts, risks, and costs of weather extremes on population, environment, transport, and industry have so far not been properly assessed in the northern latitudes of Eurasia. New knowledge is needed for improving the forecasting of extreme weather events, for understanding the effect of wildfires on radiative forcing and atmospheric composition in the region, for estimating the impacts of weather extremes on major biogeochemical cycles, and for understanding the effects of disturbances in forests on the emissions of BVOC and VON (volatile organic nitrogen) (Bondur, 2011b, 2015; Bondur and Ginsburg, 2016). How do changes in the physical, chemical and biological state of the different ecosystems and the inland, water, and coastal areas affect the economies and societies in the region, and vice versa?

The number of large hydrometeorological events in Russia that cause substantial economic and social losses has increased by more a factor of 2 from 2001 to 2013 (State Report, 2011). The main hazards are related to atmospheric processes on various temporal and spatial scales, including strong winds, floods and landslides caused by heavy precipitation, and fires caused by drought and extreme temperatures. High temperatures and long droughts can substantially decrease the productivity and cause high dieback in dark coniferous forests. Hurricanes occur fairly often in the forest zone. For example, a hurricane destroyed about 78 000 ha of forest in the Irkutsk region in July 2004 (Vaschuk and Shvidenko, 2006). However, there are no reliable statistics on many types of natural hazards.

In order to build scenarios of the future frequency and properties of weather-related hazards, one should first analyse the atmospheric mechanisms behind the circulation structures responsible for these hazards: the cyclones related to strong winds and heavy precipitation and the anticyclones related to drought and fires episodes. Studying the cyclone/anticyclone tracks, frequency and intensity can provide a statistical basis for understanding the geographical distribution and properties of the major atmospheric hazards and extremes (e.g. Shmakin and Popova, 2006). For future climate projections, atmospheric hazards and extremes should be interpreted from the viewpoint of cyclone/anticyclone statistics, and possible changes in the cyclone/anticyclone geography and frequency should be analysed.

Fires are the most important natural disturbances in the boreal forests. Fires strongly determine the structure, composition, and functioning of the forest. Each year, about 0.5–1.5 % of the boreal forest burns. Since boreal forests cover 15 % of the Earth's land surface, this is a significant area (Kasischke, 2000; Conard et al., 2002; Bondur, 2011b, 2015). Climate change already substantially impacts fire regimes in northern Eurasia. More frequent and severe catastrophic (mega-) fires have become a typical feature of the fire regimes. Such fires envelope areas of up to a hundred thousand hectares within large geographical regions, lead to the degradation of forest ecosystems, decrease the biodiversity, may spread to usually unburned wetlands, cause large economic losses, deteriorate life conditions and health of local populations, and lead to “green desertification”, which is an irreversible transformation of the forest cover for long periods (Shvidenko and Schepaschenko, 2013; Bondur, 2011b, 2015). Megafires also lead to specific weather conditions over the affected areas that are comparable in size to large-scale pressure systems (∼ 30 million ha and more). The annually burned area in the Russian territory was estimated to be 8.2±0.8×106 ha during 1998–2010, and about two-thirds of this area consisted of boreal forests. For this period, the fire carbon balance (total amount of carbon in the burnt fuel) was estimated to be 121 ± 28 Tg C yr-1 (Shvidenko et al., 2011). Current model projections suggest that the number of fires will double by the end of this century. The extent of catastrophic fires escaping from the control and fire intensity are projected to increase. Due to increased severity of fire and deeper soil, carbon emissions from fires are predicted to increase by a factor of 2 to 4 (Gromtsev, 2002; Malevsky-Malevich et al., 2008; Flanningan et al., 2009; Shvidenko et al., 2011). During and after fires, significant changes take place in the forest ecosystems, including the soil. These changes include (i) a significant amount of biomass is combusted, and large amounts of carbon and nitrogen are released to the atmosphere in the form of carbon dioxide and other gases or particles (Harden et al., 2000; Andreae and Merlet, 2001; Kaiser et al., 2012; Konovalov et al., 2014; L. Kulmala et al., 2014); (ii) fire alters the microbial community structure in the soil as well as the structure of the vegetation (Dooley and Treseder, 2012; Sun et al., 2015); (iii) fires determine the structure of the vegetation, succession dynamics and the fragmentation of forest cover, tree species composition, and the productivity of boreal forests (Gewehr et al., 2014), and (iv) fire is one of the crucial drivers controlling the dynamics of the carbon stock of boreal forests (Jonsson and Wardle, 2010; Köster et al., 2014).

Disturbances resulting from fire, pest outbreaks, and diseases also have substantial effects on the emissions of BVOCs and volatile organic nitrogen compounds (Isidorov, 2001), and consequently on atmospheric aerosol formation. The acceleration of fire regimes will also affect the amount of black carbon in the atmosphere, and thus has an effect on the albedo of the cryosphere.

The degradation of permafrost will cause serious damage both to infrastructure and to ecosystems and water systems in the northern Eurasian region. This includes, for example, damage to pipelines and buildings, deformation of roads and railroads in Russia, Mongolia, and China, variations in the ion distribution in soil water in young and ancient landslides, cryogenic landslides, spatial and temporal changes of grass and willow vegetation, saline water accumulation in local depressions of the permafrost table, and formation of highly saline lenses of groundwater called “salt traps”.

Due to the large extent of permafrost-covered areas in northern Eurasia (for ecosystem effects, see Sect. 2.1.1 and 2.1.2), there are numerous infrastructural issues related to possible changes in the thickness and temperature of the frozen part of the subsurface, and thus in the mechanical soil properties. Climate change-induced changes in the cryosphere are probably among the most dramatic issues affecting the infrastructure in northern Eurasia, as this infrastructure is literally standing on permafrost. Moreover, an interesting coupling may be related to the decreasing ice-cover of the Arctic Ocean, which results in increased humidity and precipitation on the continent, and thus a further thickening and longer duration of the annual snow cover. Snow is a good thermal insulator, and influences the average ground surface temperature, thus playing a potentially important role in speeding up the thawing of permafrost.

The increased risk of damage to local infrastructure, such as buildings and roads, can cause significant social problems, and exerts pressure on the local economies. Thawing permafrost is structurally weak, and places a variety of infrastructure at risk. For example, the failure of buildings, roads, pipelines, or railways can have dramatic environmental consequences, as seen in the 1994 breakdown of the pipeline to the Vozei oilfield in northern Russia, which resulted in a spill of 160 000 t of oil – the world's largest terrestrial oil spill (United Nations Environment Program, 2013). Maintenance and repair costs related to permafrost thaw and degradation of infrastructure in northern Eurasia have recently increased, and will most probably increase further in the future. This is an especially prominent problem in discontinuous permafrost regions, where even small changes in the permafrost temperature can cause significant damage to infrastructure. Most settlements in permafrost zones are located on the coast, where strong erosion places structures and roads at risk. After damage to the infrastructure, local residents and indigenous communities are often forced to relocate. This can cause changes in, or even disappearances of, local societies, cultures, and traditions (United Nations Environment Program, 2013).

In northern Eurasia, from the eastern part of the Barents Sea to the Bering Sea, the permafrost is located directly on the seacoast. In many of these coastal permafrost areas, sea level rise and continuing permafrost degradation leads to significant coastal erosion, and to the possibility of a collapse of coastal constructions, lighthouses, ports, houses, etc. In this region, the sea level rise is coupled to the permafrost degradation in a complex way, and should be focused on in future studies.

Understanding and measuring artificial radionuclides in marine ecosystems is needed for improving emergency preparedness capabilities, and for developing risk assessments of potential nuclear accidents. The awareness of the general public and associated stakeholders across the region should also be raised concerning the challenges and risks associated with nuclear technologies, environmental radioactivity, and emergency preparedness. The current state of radioactive contamination in terrestrial and marine ecosystems in the European Arctic region will be studied by examining environmental samples collected from Finnish Lapland, Finnmark, and Troms in Norway, the Kola Peninsula, and the Barents Sea. The results will provide updated information on the present levels, occurrence and fate of radioactive substances in the Arctic environments and food chains. The results will also allow us to estimate where the radioactive substances originate from, and what risks they may pose in case of accidents.

Annual expeditions for sample collection are needed for the development of models to predict the distribution of radionuclides in the northern marine environment, and for the assessment of the current state of radioactive contamination in marine ecosystems in the European Arctic region. In view of recent developments and increased interests in the European Arctic region for oil and gas extraction, special attention needs to be given to the analysis of norms (naturally occurring radioactive materials) in order to understand current levels. The future focus should be put on atmospheric modelling, and on the assessment of radionuclide distributions in the case of accidents leading to the release of radioactive substances to the environment in the European Arctic region. This includes the assessment of nuclear accident scenarios for dispersion modelling.

Climate and weather strongly affect the living conditions, mostly in the Eastern part of the northern Eurasian societies, influencing people's health, incidence of diseases and adaptive capacity. The vulnerability of societies, including their adaptive capacity, varies greatly depending on both their physical environment, and on their demographic structure and economic activities. There is a need to analyses the scientific background and robustness of the adaptation and mitigation strategies (AMS) of the region's societies, and their resilience capacity, with special emphasis on the forest sector and agriculture. The future research needs are in understanding what ways populated areas are vulnerable to climate change; how their vulnerability can be reduced and their adaptive capacities improved; what responses should be identified to mitigate and adapt to climate changes.

Health issues are also important in multi-disciplinary studies of northern Eurasia, as the living conditions of both humans and livestock are changing dramatically. SLCF, such as black carbon, ozone, and aerosol particles, are important players in both air quality and Arctic climate change and their impacts are not yet quantified. Black carbon has a special role when designing future emission control strategies, since it is the only major aerosol component whose reduction is likely to be beneficial to both climate and human health. These changes can be expressed through complex parameters combining the direct effects of, e.g., temperature and wind speed, with indirect effects of several climatic and non-climatic factors such as the atmospheric pressure variability, or the frequency of unfavourable weather events, e.g. heat waves or strong winds. During the last decades, living conditions in northern Eurasia have generally improved, but with a significant regional and seasonal variation (Zolotokrylin et al., 2012).

Both northern and eastern Eurasia have small and diminishing populations, mainly due to the migration outflow started in the 1990s due to severe and unfavourable living conditions combined with changing state policies with respect to the development of the northern territories. This reversed the previous long-standing pattern of migration inflow. The combination of outflow and natural population decrease (with some regional exceptions in several ethnic republics and autonomous regions (okrugs) with oil and gas industry) led to a steady population decline in most regions in northern and eastern Russia from 1990s. In the post-soviet period, the population of eastern Russia decreased by 2.7 million, while the population of Russia's Arctic zone decreased by nearly by one-third (500 000 people), in contrast to the majority of the world's Arctic territories (Glezer, 2007a, b). The population change in north-eastern Russia was particularly remarkable: the Chukotka Autonomous Okrug lost 68 % of its population, the Magadan Oblast lost 59 %, and the Kamchatka Krai lost 33 %.

Geographical and ethnic factors influence the demography and settlement pattern in the region. Geographical factors include environmental conditions and the mixture of urban and rural territories. Areas with a large proportion of indigenous people employed in traditional nature management were exposed to relatively small post-soviet transformations in the 1990s and 2000s. In contrast, the largest transformations occurred in areas with a larger proportion of Russian people and developed mining industries. The differences in the transformations between settlements with predominantly indigenous and predominantly Russian populations are evident. For example, in the Chukotka Autonomous Okrug, the former remained mostly intact, with only small decreases in population, while the latter disappeared entirely or were significantly depopulated (Litvinenko, 2012, 2013).

When assessing the impacts of climate change and other environmental changes on human societies, it should be taken into account that the urban environments in northern Eurasian cities and towns situated in the less favoured regions are currently incapable of mitigating unfavourable impacts. The impact of climate parameters, such as temperature (including seasonal, weekly, and daily cycles, and extreme values), strong winds, snowfall, snowstorms, and precipitation should be investigated. Both the frequency and the duration of weather events should be considered. These climate parameters influence human health, incidence of diseases, adaptation potential, and economic development in general. Furthermore, it is important to explore the interactions between the environmental change and post-soviet transformations of natural resource utilization in northern Eurasia in order to assess the complexity of their socio-ecological consequences at regional and local levels (Litvinenko, 2012; Tynkkynen, 2010). The population dynamics of the northern Russian regions in 1990–2012, and the linkage between intra-regional differences in population dynamics, spatial transformations of natural resources utilization, and ethnic composition of the populations should be clarified. It would be desirable to develop an “early warning system” for the timely mitigation of the negative socio-ecological effects of both environmental changes, and changes in the availability of natural resources as well as accident like leakages in gas and oil pipelines. Such systems would be useful for federal, regional, and local authorities as well as for local communities.

It should also be taken into account that the majority of the world's ethnic groups are small and engaged in culturally specialized methods of subsistence, so any change in their immediate environment may lead to their traditional way of life becoming unsustainable. These changes may be due to rising sea levels, warming seawater, melting ice cover, thawing permafrost, flooding rivers, changing rain patterns, or moving vegetational zones. These are direct effects of climate change and environmental deterioration on ethnodiversity. However, even more threatening are the indirect effects. The immediate environment of small ethnic groups is often vulnerable to the adverse impact of majority populations representing governments and nations. The effects of climate change may lead to a rapid and massive transfer of majority populations to areas previously inhabited by small ethnic groups.