Coupled Changes in the Arctic Carbon Cycle Between the Land, Marine, and Social Domains

The Arctic has experienced rapid change associated with warming since the 1970s. The rapid retreat of the terrestrial cryosphere can release a large amount greenhouse gas from the permafrost regions into the air, and the sea ice decline will affect the CO2 and CH4 balance in the ocean. Changes in the Arctic provide feedback mechanisms that can also impinge on the global ocean’s thermohaline circulation. During the past years, the overall natural processes in the Arctic have been studied although the magnitude and timing of carbon release from the cryosphere changes require further investigation. However, few studies have been conducted to link the natural and social systems in the Arctic. Scientists and policymakers must consider the coupled Arctic land, ocean, and social systems in their decisions for coping with climate change.


Arctic Cryosphere Change and Its Impacts on Carbon Cycle
The Arctic cryosphere changes and their impacts have raised many concerns (AMAP, 2017).Currently, major attentions are concentrated on the mass loss of Greenland Ice and mountain glaciers, snow cover changes, permafrost degradation, and demise of multiyear sea ice.The most recognizable consequences of Arctic warming are sea level rise, the energy budget, and carbon cycle.
The Greenland Ice Sheet and mountain glaciers mass loss is presently the main contributor to sea level rise (IPCC, 2019).For example, during 1986-2016, the area of mountain glaciers on the Kenai Peninsula in south central Alaska shrunk by about 12% (Yang et al., 2020).The mass loss of the Greenland Ice Sheet commits at least 274 ± 68 mm sea level rise within 2000-2019 (Box et al., 2022).In contrast with the rapid retreat of glaciers and Greenland Ice Sheet, snow cover metrics fluctuated considerably in the high Arctic, with overall decreasing trends in snow cover extent and snow water equivalent since the 1930s, particularly in the spring and summer (Mohammadzadeh Khani et al., 2022).The loss of snow and ice cover decreases albedo and thus increases the amount of energy absorbed over both land and ocean (IPCC, 2019).
Permafrost degradation, including active layer thickening, increasing ground temperatures, and thermokarst development, has been widely detected in the Arctic (Biskaborn et al., 2019).The greatest concern about permafrost degradation is the potential release of a large amount of carbon as greenhouse gases (CO 2 , CH 4 ), which will further amplify climate warming (Schuur et al., 2015).Soil organic carbon storage in high-latitude permafrost regions is estimated to be between 1,330 and 1,580 Gt, or more, which is much higher than the carbon pool presently held by the atmosphere (∼870 Gt, 410 ppm for CO 2 concentration; Ballantyne et al., 2012).The stability of the permafrost carbon pool is closely associated with cold temperatures.Recent evidences suggest that overall, the permafrost is losing capacity as a sink for atmospheric CO 2 and may shift to a carbon source in the foreseeable future (Schuur et al., 2015).
Since the 1980s, the extent of Arctic sea ice in September has been decreasing at a rate of 13% per decade (IPCC, 2019).During 1975-2012, the average sea ice thickness in the central Arctic Ocean has declined from 3.6 to 1.3 m (Lindsay & Schweiger, 2015).It has been projected that a nearly sea ice-free Arctic will occur in the 2030s (Docquier & Koenigk, 2021;Overland & Wang, 2013).The changes in sea ice can affect the ocean carbon cycle.The Arctic Ocean is a carbon sink, and the CO 2 uptake by the Arctic Ocean and its adjacent seas during 1997-2013 was estimated to be up to 180 Tg C year −1 (Yasunaka et al., 2016).A longer ice-free period, larger open areas, and increased primary production can enhance CO 2 uptake (Ouyang et al., 2022), while the increasing CO 2 in the surface water can slow down the carbon uptake capacity (Cai et al., 2010).This uptake of atmospheric CO 2 can lead to rapid acidification of the Arctic Ocean, particularly at higher latitudes where sea ice retreat is active (Qi et al., 2017(Qi et al., , 2022)).

Impacts of Arctic Cryosphere Change on Atmospheric Circulation and Marine Environment in a Border Area
Arctic cryosphere changes can affect atmospheric circulation far south of the Arctic.Arctic sea ice loss can lead to a weakening of the stratospheric polar vortex, which may subsequently result in extreme events in middle latitudes (Cohen et al., 2014).The disruption of the polar vortex can encourage the polar jet stream to move south, creating counterintuitive extreme cold events (Kretschmer et al., 2018).
Arctic cryosphere changes impact the marine environment far beyond the Arctic as well.During 2000-2010, the river runoff around the Arctic Ocean and net precipitation was about 4,200 and 2,200 km 3 year −1 , and these values have been projected to increase by 31% and 14% by 2100, respectively (Haine et al., 2015).The rapidly increasing freshwater input to the Arctic Ocean must eventually be exported to the North Atlantic Ocean which has the potential to weaken thermohaline circulation and impede meridional overturning (Callaghan, Johansson, Key, et al., 2011).It has been proposed that the major cooling during the Younger Dryas Period (∼12,900-11,600 years BP) was a result of the slowing or shutting down of the North Atlantic thermocline circulation, possibly due to a catastrophic release of freshwater from the North American landmass (Murton et al., 2010).
Melting of ice sheets, glaciers, and permafrost thaw can increase the riverine exports of freshwater, nutrients, and sediments to the Arctic Ocean.The discharge of Arctic rivers increased by 9%-12% during 1975-2015 (Box et al., 2019).Although the Greenland Ice Sheet is a significant source of reactive nanoparticulate iron to the Arctic Ocean (Hawkings et al., 2014), the over material fluxes of the glaciers and the Greenland Ice Sheet melting water are much lower than those of riverine exports (Hood et al., 2015).The increasing freshwater export is an important reason for the Arctic Ocean's freshening (Box et al., 2019).River discharge affects the upwelling of nutrient-rich nearshore bottom waters, further influencing the productivity of fjord and nearshore waters (Laufer-Meiser et al., 2021;Meire et al., 2017).

Complicated Processes and Effects of Arctic Cryosphere Changes
The direct cause of Arctic cryosphere change is the rising air temperature due to the anthropogenic climate warming.Obviously, the response of cryosphere components to climate warming is not linear.Currently, the most important nonlinear processes related to the Arctic cryosphere changes are thermokarst development, coastline erosion, wildfires, and subsea permafrost degradation.
Thermokarst terrains include thermokarst lakes, thaw slumps, ground subsidence, thermal erosion, and surface ablation (Jorgenson et al., 2010).Thermokarst terrain develops when certain forms of permafrost thaw, and the soil subsides under its own mass.Thermokarst development is a natural phenomenon (Bowden, 2010), while the area of thermokarst terrains in Arctic regions expanded rapidly recently (Lewkowicz & Way, 2019).Thaw slumps and permafrost collapse can expose organic-rich layers and thus accelerating microbial decomposition and lateral export of organic carbon (Abbott & Jones, 2015).The thermokarst lakes and ponds are hotpots of greenhouse gas emissions, and the expansion of thermokarst lakes and ponds will contribute to more greenhouse gas emissions in the Arctic regions (Liebner & Welte, 2020).
The erosion of Arctic coastlines has recently attracted much attention.The Arctic coastlines are particularly vulnerable to erosion because they consist of ice-rich permafrost (Grigoriev et al., 2004;Zimov et al., 2006).The Arctic coastlines exhibit the highest erosion rates in the world (Reimnitz et al., 1988).Arctic coastal erosion could significantly contribute to the Arctic carbon cycle as large quantities of organic carbon stored in permafrost are directly exported to the ocean (Fritz et al., 2017).Coastal erosion can breach thermokarst lakes, leading to the initial draining of the lakes followed by marine flooding.Additionally, coastal erosion and land loss poses a considerable threat to native, industrial, scientific, and even military communities (Ding et al., 2021).The Arctic coastal erosion is expected to increase drastically in the future due to the permafrost thaw, declining summer sea ice cover, longer and warmer thawing seasons, increasing seawater temperature, and rising sea level (Fritz et al., 2017;Gunther et al., 2015).
Wildfires can release large amounts of carbon into the atmosphere from the plants and soils, change riverine exports of carbon and nitrogen, and thus alter the Arctic carbon cycle.Wildfires in the Arctic regions have seen increases (Bruhwiler et al., 2021), and the fire season begins earlier (McCarty et al., 2020), occurring in previously fire-resistant ecosystems such as tundra bogs, fens, and marshes (McCarty et al., 2020).Wildfires also emits black carbon to the air, contributing to glacier and sea ice melting (Kim et al., 2005;Randerson et al., 2006).
The subsea permafrost is widely distributed in the continental shelves of the Arctic Ocean and surrounding areas, with an estimated extent of about 2 million km 2 (Sayedi et al., 2020).In the subsea permafrost domain, the overlying seawater temperature is only slightly below 0°C.This temperature causes rapid subsea permafrost degradation.Drilling data indicate that subsea permafrost degradation rates were much faster during 1983-2013 than that before1982 (Shakhova et al., 2017).The subsea permafrost domain is thought to contain about 560 (170-740, 90% confidence interval) Gt organic carbon and 45 (10-110) Gt carbon in methane, and the CO 2 and CH 4 release will likely increase along with subsea permafrost degradation (Sayedi et al., 2020).

Major Knowledge Gaps
Changes in patterns and extent of snow cover, permafrost, sea ice, mountain glaciers, and the Greenland Ice Sheet mass have been quantitatively assessed.The subsea permafrost degradation rates are largely unknown due to the fact that observational data are scarce.The changing rates in cryosphere components in the past decades have been summarized based on published literature (Table 1), while these values varied considerably at different time and spatial scales.For example, the Greenland Ice Sheet mass loss in 2019 was −532 ± 58 Gt (Sasgen et al., 2020).The Arctic sea ice volume, including seasonal and perennial ice, decreased by about 62% during 1982-2020 (Wang et al., 2022).In comparison with the uncertainties in the past changes, future changes in the Arctic cryosphere and their impacts are even more poorly understood.
Numerous models have been developed to project future change in the cryosphere.There is still ample room for improvement of these models, both in terms of the physical processes and parameterization schemes.Observations suggest that most current models underestimate the rates at which sea ice declines (Meier et al., 2014), and there are large uncertainties in the snow cover prediction (Mudryk et al., 2020).The processes and timescales over which permafrost within 3 m might become thawed are not well known (Biskaborn et al., 2019).Therefore, a better understanding of the physical processes involved in thawing and improved parameterization schemes are needed to enhance the performance of the cryosphere models (Hock et al., 2017).For example, more studies are required to understand the interactions among rainfall, snow, ice, and the degradation and aggradation of permafrost (Oliva & Fritz, 2018).The Arctic's large pools of stored carbon, including organic matter and methane hydrate, have been long recognized as risk factors for exacerbating climate warming.With an area accounting for 3% of the global ocean, the present-day carbon assimilation from annual air-sea flux in the Arctic Ocean is 160 ± 60 Tg C year −1 , which is about 10%-12% of global oceanic carbon uptake (MacGilchrist et al., 2014).Although an increasing partial pressure of CO 2 can slow down the carbon uptake rates (Cai et al., 2010;DeGrandpre et al., 2020), the Arctic Ocean carbon sink will likely increase because the sea ice loss increases the open water area for sea-air gas exchange (Anderson & Macdonald, 2015).However, the Arctic Ocean is a source of CH 4 , and the ebullitive and diffusive CH 4 rate of the eastern seas of the Arctic Ocean is estimated up to ∼3 Tg CH 4 year −1 (Thornton et al., 2020).CH 4 has a global warming potential 28 times greater than CO 2 over a 100 years period.Therefore, a large amount of CH 4 release from the subsea permafrost and methane hydrates will offset the CO 2 uptake, making Arctic Ocean a positive force for climate warming in the future.Presently, the amount reaching the atmosphere from these two sources appears minimal in comparison with other CH 4 sources (Ruppel & Kessler, 2017;Sayedi et al., 2020).For the terrestrial Arctic, it remains uncertain whether the permafrost regions are carbon sinks or sources in the future (Schuur et al., 2015).

New Emerging Challenges
There are several major scientific questions that remain unsolved for understanding the Arctic changes and their impacts (Bruhwiler et al., 2021).Currently, new challenges pertaining to the rapidly changing cryosphere are emerging, and the most important issues include thermokarst terrains, wildfires, coastal erosion, and subsea methane emissions.
Thermokarst terrains can greatly affect the carbon budget both in release rates and gas forms.The CO 2 and CH 4 fluxes and thaw slump development have been assessed at regional scales, while we still lack a comprehensive evaluation of greenhouse gases flux in thermokarst terrains (Turetsky et al., 2020).Relationships between thermokarst history on land and preservation of subsea permafrost after sea level rise are an example of some of the complex interactions that can affect the fate of carbon stores (Angelopoulos et al., 2021).Recently, the forma tion of exploding craters in Siberia has created many public concerns.These craters may be related to thermokarst processes, and they seem to turn into thermokarst lakes within a few years, while little is known about the greenhouse gas emissions from these craters (Chuvilin et al., 2020).
The Arctic accounts for 34% of Earth's coasts and has some of the fastest eroding coastlines (Fritz et al., 2017).Although the effects of coastal erosion on carbon cycle and infrastructure have been recognized, quantitative assessments of the effects of coastal erosion are largely lacking due to the sparse observational data.Continuous monitoring of Arctic coastal change and development of process-based models are essential for engineering construction in the coastal regions and mitigating the hazards.
Wildfires have been long recognized as important disturbances of the Arctic ecology and carbon cycle (Mack et al., 2011).The current carbon budget in the terrestrial Arctic is mainly calculated from carbon uptake and emission, with little consideration of the contribution from wildfires.Presently, detection and monitoring of below ground biomass burning, and evaluation of the resultant carbon loss remains a challenge (McCarty et al., 2020).
The subsea permafrost may play an important role in the carbon budget in the Arctic region.However, the distribution, storage of the organic matter and CH 4 , greenhouse gases release, and future trajectories largely remain unknown (Malakhova & Golubeva, 2014;Sayedi et al., 2020).Due to the potentially vast distribution of subsea permafrost and the huge stocks of organic carbon and CH 4 , future studies are required to take subsea permafrost into consideration in the Arctic carbon cycle.
We stress that the topics of thermokarst, coastal erosion, wildfire, and subsea permafrost are challenges in Arctic cryosphere change, and these processes should be included in the Earth System Models (ESMs).However, the challenges in the Arctic science are not limited to above ones, and other disturbances (e.g., biological invasion, energy extraction; Chan et al., 2019;Hovelsrud et al., 2011) and effects (e.g., pollutants release, ancient virus breakout, cultural heritage damage; Miner et al., 2021;Nicu et al., 2021) should not be neglected.

Call to Action
The alarm bell of the Arctic rapid changes was rung long ago, and yet here we are each year adding another point to the time series graph.In order to establish strategies to cope with Arctic change, international collaborations are required among science communities, policymakers, and the indigenous peoples who will be affected most by the changes.We need to plug major knowledge gaps, and we have three recommendations.
Improve models for Arctic science.There are many reasons for the growing literature but relatively stagnant science in Arctic changes (Christensen, 2014).Improving ESMs is a critical step toward understanding the Arctic changes, but many findings, such as parameterization and mechanisms, have not been incorporated into the models.Currently, coastal erosion, thermokarst development, and subsea permafrost are not incorporated into ESMs.Although wildfires are part of ESMs, our ability to model the precise details of fire regimes is limited (Hantson et al., 2016).Many efforts have been made to build models to simulate the processes and/or effects of coastal erosion (Nielsen et al., 2020), thermokarst development (Chen et al., 2021), wildfires (Descals et al., 2022), and subsea permafrost (Wilkenskjeld et al., 2022).Coupling these processes into ESMs will be helpful to obtain better predictions of Arctic changes and their impacts.
Obtain more original data between Arctic land and ocean.There are several networks pertaining to permafrost carbon and permafrost monitoring, but the original field data remain insufficient for such a large area.Some missing and discontinuous data have resulted from the failure of regular maintenance.Nowadays, many projects implemented in the Arctic regions collect data, make field observations, and develop or apply models.However, due to limited budget, some of them only obtained very short-term data, and these data are difficult to use for meta-analysis or modeling.
Currently, there are only some meteorological sites maintained by monitoring agencies in the Arctic, and many other data such as ground temperature data from boreholes and carbon fluxes data from eddy covariance towers (Figure 1) are collected by the scientific communities.Many data from these sites are not continuous due to budget limitations.The monitoring agencies should pay more attention to long-term field observations in the Arctic regions.Additionally, considering the physical diversity of the Arctic, the scientific community should encourage projects which can obtain original field data in previous data-scarce areas, and these data would be useful for meta-data analysis in the future.Moreover, potential cooperation and data-sharing among industries (e.g., oil extraction companies, shipping companies) and research centers in different Arctic countries can benefit scientific research in the Arctic changes.
Enhance international cooperation and involve Arctic residents' welfares in scientific activities.The Arctic's vast expanses of ocean and land remain a difficult region to obtain field data due to logistics and cost.There are already many international cooperative Arctic research projects, while the research topic is still not a top priority (Yamanouchi & Takata, 2020).In the Arctic regions, people are aware and conscious of the rapid climate change from landscape changes and infrastructure damages (Kaltenborn et al., 2020).Meanwhile, changes in the permafrost carbon cycle are associated with several important contaminant releases such as mercury and chemical materials from permafrost regions (Basu et al., 2022).Their biomagnification in food chains poses a significant threat to indigenous people.In addition, the release of biological pathogens such as smallpox and influenza virus can raise the prospect of disease emergence (Miner et al., 2021).To involve the risks and welfares of Arctic residents, it is necessary to foster data and sample sharing in multidisciplinary research communities.Although climate change brings some opportunities, the Arctic is unique in many aspects and has a long history of human adaptation to the cold environment (Ford et al., 2021).A comprehensive approach including the carbon cycle and these risks may provide mitigating and adaptive strategies for Arctic residents (Biresselioglu et al., 2020).Eight countries have territories within the Arctic circle and several other countries have regions with high-latitude permafrost, while the Arctic appears to be the first domino to fall, and the other dominoes point south.The responsibility must be shouldered by all the countries as stakeholders.The only way to attenuate the rapid changes in the Arctic is to decrease greenhouse gases in the air, which requires global efforts (Schuur et al., 2022).It will be years before this goal is achieved.Therefore, we should try to notify and forecast where, when, and what is likely to happen, to enable scientists and policymakers to work together to protect indigenous people and our future earth.

Conclusion
The Arctic has been experiencing rapid change during the past decades, and the changing Arctic can have important feedback to climate system.Arctic warming trends, ocean thermohaline circulation, and terrestrial and oceanic carbon cycle have been extensively studied.Results suggested that permafrost degradation will release a large amount of organic carbon, and sea ice decline may increase the CO 2 uptake from the air.However, there are still major knowledge gaps in the Arctic carbon cycle including the effects of thermokarst, coastal erosion, wildfires, and subsea permafrost.Most studies pertaining to the carbon cycle in the Arctic mainly focused on the natural processes, while the effects of Arctic changes on social domains received less attention.Although the schematic framework of the Arctic carbon cycle has been recognized, our cooperative research in Arctic science is far from sufficient for coping with climate change.Both the scientific communities and policymakers must consider the coupled Arctic land, ocean, and social systems in their decisions in the future.data in Figure 1a are from the global terrestrial network for permafrost (http://gtnpdatabase.org/boreholes), in the Arctic permafrost areas.The permafrost map in Figure 1a is from National Snow & Ice Data Center (Brown et al., 2002).The Arctic sea ice concentration (SIC) changes from 1979 to 2019 data are from the Copernicus Climate Change Service Climate Data Store (Copernicus, 2020).The annual average discharge data from 1981 to 2010 in Figure 1b are from Arctic program (Holmes et al., 2021).

Figure 1 .
Figure 1.Field observation sites in the Arctic land and Arctic sea ice concentrations from 1979 to 2019.The arrows in the right panel indicate the riverine exports from the six largest rivers emptying into the Arctic Ocean (Holmes et al., 2013).The arrow size represents the sizes of annual average discharge from 1981 to 2010.The coastal erosion, groundwater discharge, and other rivers are not shown.