Proxies and feedbacks were the primary themes of workshop sessions held on the second day. Workshop participants discussed individual components of the Earth system and how they interact. They also considered the types of feedbacks that are operating in different components and the proxies that researchers can use to reconstruct those changes. Specifically, workshop participants discussed ice-sheet and high-latitude proxies and feedbacks, including sea level; ocean proxies and feedbacks, including ocean circulation and changes to biochemistry and productivity; and terrestrial proxies and feedbacks, including fire and vegetation dynamics and the terrestrial carbon cycle.
The first session considered the interaction between polar ice sheets, sea level, and the solid earth, and how they influence the response of Earth system components to climate forcings. For example, feedbacks resulting in deformational, gravitational, and rotational perturbations to sea level driven by ice-sheet growth and decay—collectively known as glacial isostatic adjustment (GIA) (Figure 8)—add uncertainty to reconstructions of the magnitude and rates of past sea-level changes and, therefore, to understanding of the response of both sea level and polar ice sheets to a warm climate.
Interdisciplinary studies of geological records of past climate changes, such as evaluations of paleo-sea-level indicators preserved in coastal sediments, have advanced understanding of sea-level variability throughout the glacial-interglacial cycles of the last several hundred thousand years, as well as affected understanding of ice and solid-earth processes resulting in spatially heterogeneous patterns of sea-level change (Dutton et al., 2015). Layers of ice evaluated in cores of the Antarctic and Greenland ice sheets, for example, provide valuable information about climate conditions at the time of formation, including precipitation, temperature, greenhouse gas concentrations, volcanic activity, and solar activity. Mountain glacier records can also reveal information about regional patterns of climate change, Earth system dynamics, and seasonal climate variability when combined with other climate proxies.
While polar ice sheets are expected to become the dominant contributor to global mean sea-level rise in the 21st century, projections of the magnitude and rate of sea-level rise include assumptions about the link between sea level and global temperatures in the past and an incomplete understanding of the complex processes involved in rapid ice-sheet retreat (Rignot et al., 2014). From ice cores and ice-proximal sediment records to paleo sea-level indicators preserved in coastal deposits around the globe, multiple proxies are considered to evaluate past sea-level and ice-sheet behavior in response to a warm climate; however, direct evidence of past ice-sheet thickness and extent is limited. This kind of paleoclimate information informs geophysical models used to constrain modern and future rates of global sea-level change by correcting for the influence of GIA on local measurements.
Sea Level, Ice Sheets, and GIA
Jacky Austermann, Columbia University, explained that the focus of the paleo sea-level community has been understanding sea level during the Holocene and last interglacial (Ferranti et al., 2006; Hibbert et al., 2016; Khan et al., 2019; Kopp et al., 2009; Pedoja et al., 2014), resulting in a first-order understanding of sea-level changes over those time periods. However, details that are important for linking paleoclimate to projections for the future remain unknown: Exactly how high was sea level during the last interglacial? How low was sea level during the LGM? In order to get to the precision and level of detail in paleo reconstructions needed to inform future climate projections, revisiting and remapping sites with better chronological technologies to get higher quality data could be useful. In the breakout session on sea level, ice sheets, and GIA, moderated by Paul Bierman, University of Vermont, and Andrea Hawkes, University of North Carolina Wilmington, participants also discussed the need to do sea-level benchmarking in projection models to better articulate sea-level changes to policy and decision makers (Figure 9).
Feedbacks between ice sheets and the solid earth are an ongoing theme that matter for understanding both the past and the future. Austermann suggested another important avenue of research is translating a local sea-level indicator to the global mean sea level, in which deformation of the solid earth needs to be accounted for. With respect to GIA, Austermann argued the biggest gap is understanding three-dimensional variability of viscosity and transient and nonlinear rheologies, and she called for more benchmarking and intercomparisons within the GIA modeling community. Participants in the breakout session discussion also suggested the need to integrate more sea-level observations and glacial chronologies into GIA models, and how model intercomparisons between different GIA and ice-sheet models may by useful. Austermann said that she thinks the key to understanding the last interglacial and other older time periods will be understanding longer-term deformation as driven by mantle convection, tectonics, and sediment loading (Figure 10).
Models of long-term deformation could be improved and local observations to estimate long-term uplift or subsidence can be used to calibrate models.
Jason Briner, State University of New York at Buffalo, highlighted opportunities to understand regional manifestations of global climate change with relevance for predicting the future. Improving the global distribution of highly detailed mountain glacier records, combined with terrestrial proxy climate data, can inform understanding of the last glacial to present Holocene transition. Mountain glacier records can reveal spatial patterns of climate change, climate system dynamics, and some seasonal information when paired with independent climate proxy reconstructions from nearby regions.
A major unknown in ice sheet-science is the size of polar ice sheets during interglacial periods, because there is limited information preserved with which to reconstruct the size of ice sheets when they were smaller than they are today. Because of the uncertainties in sea-level reconstructions and ice-sheet models, there are not good constraints on the overall sensitivity of ice sheets to present and future warming climate states, hindering the ability to test theories about feedbacks between ice sheets and the climate system. More direct observations of ice-sheet history, particularly during interglacial periods, could help answer these questions, including rock samples from beneath today’s ice sheets, drills that can access the ice-sheet bed, and offshore records adjacent to Greenland and Antarctica. Participants in the breakout session echoed the need for new, fast coring systems and offshore marine records, paired with sediment provenance data to help determine paleo ice-sheet extent and ice-sheet source and flow dynamics. Participants also discussed how more near and intermediate field records, sourced from regions proximal to paleo ice sheets, would be useful to do sea-level reconstructions and inform ice-sheet extent, as well as to calibrate GIA models to constrain both paleo and modern sea-level rise and ice-sheet mass balance.
Erich Osterberg, Dartmouth College, spoke about how Arctic ice cores can be used to develop higher-resolution understanding of the response of ecosystems and the cryosphere by focusing on past warm periods, as well as times of rapid climate change over human-relevant timescales (years–decades). Past warm periods of particular relevance include the last (Eemian) interglacial and the Holocene thermal maximum, the latter having perhaps been overlooked by the ice-core community. He noted that tools are now available to analyze and date pre-Holocene ice at the resolutions necessary to reveal processes including triggers, feedbacks, and instabilities of past abrupt
change to inform future climate change projections. In the Arctic Ocean and Arctic borderlands breakout session moderated by Beth Caissie, U.S. Geological Survey (USGS), and Joseph Ortiz, Kent State University, participants also discussed the need to focus on pre-LGM warm periods and, specifically, to explore how the Arctic and sub-Arctic outside of the North Atlantic, and their teleconnections with low latitudes, change during periods of rapid change and previous warm periods. Amelia Shevenell, University of South Florida, added that, for the Antarctic Ice Sheet, ice-sheet evolution at catchment scale during warm climate intervals remains a knowledge gap, but records of these warm periods are more likely to be preserved on the continental shelf.
Tom Cronin, USGS, discussed how sediment coring in the Arctic Ocean can provide high-resolution records to better understand chronostratigraphy in the Cenozoic on orbital timescales. Cronin discussed Cenozoic climate needs in the context of the 2004 Arctic Coring Expedition (ACEX), which has provided valuable information about Cenozoic climate (e.g., via oxygen isotope, micropaleontological, and paleo-CO2 proxies). The 2004 ACEX record includes a gap in stratigraphy from 18–44 million years ago that leaves open questions about a key interval of Cenozoic glacial, sea-ice, and ice-sheet evolution. Future International Ocean Discovery Program (IODP) cruises, such as the planned IODP 377 expedition in 2022, have the potential to collect additional records that help fill this gap and contain necessary geochemical and microfossil proxy data needed for a more complete Cenozoic climate record in the Arctic.
More generally, participants identified similar data needs in both the Arctic and Antarctic. Osterberg noted the need to develop seasonally specific proxies to resolve summertime temperatures and to understand seasonal biases of proxies at particular locations. A network of ice cores with regional climate signals to understand regional sensitivities and records that can then be used in data assimilation could also be useful. Shevenell reiterated the importance of thinking about the regional signal of ice sheets in the Antarctic and the need to develop and calibrate regional proxies. Osterberg called for faster and cheaper logistics for drilling in the Arctic, and Shevenell called for improved drilling capabilities in icy continental-shelf waters and drill-site coverage in the Antarctic.
Shevenell explained the largest unknown is the Antarctic’s contribution to sea-level rise, and particularly, one of the biggest gaps in Antarctic ice-sheet history is the evolution and behavior of the East Antarctic Ice Sheet, which may be more sensitive to future changes in climate than in the past. In order to understand these questions, Shevenell emphasized the use of ice-proximal marine sediment records, such as sediments cored from the continental shelf near the marine-terminating ice-sheet margin, that can reveal shelf development and be integrated into models to understand how Antarctica’s ice sheets may respond in a warming world. Shevenell also reiterated the need to understand past ice extent, both the maximum and minimum ice extent, to understand retreat rates and the processes that drive retreat through time, using continental-shelf records. In the Antarctica and Southern Ocean processes breakout session, moderated by Julia Wellner, University of Houston, and Liz Sikes, Rutgers University, participants also discussed the need to understand past glacial ice extent from the terrestrial perspective, and sea-ice extent from the marine perspective. For both terrestrial and marine records, more temporal and spatial data may be essential, beyond the well-studied regions of the Atlantic and West Antarctic Ice Sheet and multiple catchment and basin studies. Central Arctic paleoclimate records may also be critical to evaluating the response of arctic amplification to climate forcings over different timescales. Cronin noted that, on orbital timescales, records of glacial-interglacial variability in marine sediment cores over the last ~600,000–800,000 years include important evidence that past interglacial periods experienced summer sea-ice free conditions, but records are complicated by chronostratigraphy due to the influence of sea-ice and ice shelves, and attention to sea-ice and ice-shelf dynamics on orbital timescales may be useful.
Osterberg shared his experience working with local- and state-level policymakers in the Northeast and found it helpful to focus on extreme events that have the biggest impact on society. Additionally, focus on human-relevant timescales and regional changes specific to where stakeholders live may be
critical, which also motivates the development of higher-resolution records to examine the rate of sea-level change in more detail or the response of Greenland to small changes in temperature over decades to centuries.
Sediments preserved in ocean basins have accumulated over time and have recorded detailed information about the role of the ocean in the Earth system, the response of the ocean to forcings, and the complexity of internal system feedbacks of ice-ocean-carbon cycle interactions. Proxies that can be
derived from the marine sediment archive include biotic proxies (evaluating changes in composition of plant and animal groups in response to a changing climate) and geological or geochemical proxies (measurements of mass movements of material through the climate system, as discrete particles or in a dissolved chemical form). These proxies can shed light on past changes in ocean chemistry, circulation, temperature, exchange of mass with land-based ice sheets, and other parameters dating back millions of years. For example, a wealth of information is contained in the chemistry of calcite shells of planktic organisms and benthic foraminifera (Figure 15) that accumulate in layers of sediment on the seafloor. The ratios of oxygen isotopes present in these shells vary in response to changes in global ice volume and ocean temperature, and isotopes of carbon indicate the transfer of organic material among reservoirs on land, in the ocean, and in the atmosphere. Similar to pollen found in terrestrial sediments, these shell-forming phytoplankton and zooplankton are abundant and widely distributed, and their modern climate preferences are often well known; therefore, the assemblages of plankton populations identified in marine sediment records can also be used to reconstruct past climates in archives millions of years old.
Similar to tree rings, annual growth bands of coral growing in shallow coastal environments contain a myriad of chemical information, such as ratios of isotopes of oxygen indicating changes in temperature and precipitation over thousands of years. The age, elevation, biotic composition, and preservation of fossil coral reefs found in coastal limestone sediments that were once growing close to the sea surface is one of several proxies of sea-level position during past climates hundreds to millions of years ago. Another ocean proxy, which was discussed in this session, is biomineralization—the development of mineral structures by living organisms. While biologically controlled, the formation of hard mineral structures by marine organisms such as coral, diatoms, foraminifera, coralline algae, and mussels, can also be influenced by environmental conditions, such as temperature, salinity, and ocean chemistry (ion composition). Biominerals can serve as a proxy for climate change and seawater composition if the relative effects of biological and environmental processes can be separated. This session explored community-driven science themes that could emerge in the future from ocean proxies, evaluations of feedbacks, and related disciplines.
Andreas Schmittner, Oregon State University, identified overarching knowledge gaps that should be addressed with respect to ocean feedbacks. There does not yet exist a mechanistic model of Earth’s climate system that can accurately simulate glacial-interglacial cycles, from variability in millennial to orbital scale to high-frequency timescales, and the mechanisms that caused CO2 changes. Schmittner noted that paleo data assimilation could be used to
improve quantitative understanding of Atlantic Meridional Overturning Circulation (AMOC)5 variability, specifically the strength of the overturning that is unconstrained with existing data. The internal system feedbacks of ice-ocean-carbon cycle interactions, which are crucial, are not currently captured by models. Schmittner suggested that low-resolution Earth system models could be used to fully couple components of the system to interactive ice sheets and oxygen isotopes in those ice sheets.
Kira Lawrence, Lafayette College, emphasized the need to focus on science questions that will directly inform policy and decision making. In particular, Lawrence suggested focusing on rates of change and tipping points, with attention to warm periods, especially for the cryosphere and biosphere, to inform understanding of habitability of Earth environments, the persistence of food supplies, and potential for mass migration. People and policy makers care about the terrestrial realm, she noted, such as changes in rainfall and drought, fire regimes, and storminess in the places they live in. Schmittner pointed out that the same models are being used for paleoclimate and future projections (Figure 9), and the importance of work not only on human relevant timescales, but also the consequences of anthropogenic effects on long timescales.
Participants in the breakout group on ocean circulation and deepwater formation, moderated by Jerry McManus, Lamont-Doherty Earth Observatory, and Andrew Thompson, California Institute of Technology, suggested that the community should be thinking about global overturning circulation, with a key research question: How does the global ocean respond to perturbations in climate? Specifically, what are the physical mechanisms, over what timescales are changes in circulation communicated between different basins, and what are their signatures in the paleo record? Answering these questions would likely require observations from the Pacific, Southern Ocean, and Arctic, expanding beyond an Atlantic focus, as well as modeling tools, such as data assimilation and long model integration times.
It is understood that greenhouse gases are important in the climate system, and there is robust evidence of close coupling between CO2 and warm
5 A large system of ocean currents that carry warm water from the tropics northwards into the North Atlantic.
periods in Earth’s climate (Shakun et al., 2012), explained Mo Walczak, Oregon State University. The ocean is the largest active reservoir of CO2 on Earth, holding about 50 times as much CO2 as the atmosphere, suggesting that the ocean has a sizable role to play in the release of CO2 during warm periods (Bauska et al., 2015). It is not well understood where in the ocean CO2 exchange is occurring between the ocean and atmosphere, including whether North Atlantic intermediate or bottom water formation was stronger or weaker during past abrupt warm periods, and the role of deep overturning circulation around the Southern Ocean and in the Pacific where carbon is stored (Du et al., 2020). While evidence points to a large pool of CO2 in the deep ocean as an important and dynamic player during periods of warming, there is a lack of understanding about the global connections that drive the weakening of CO2 storage.
Walczak pointed to evidence that challenges the AMOC-centric view of the previous decade, in which destabilization of Pacific ice sheets and changes in Pacific Ocean circulation reliably precede similar changes in the North Atlantic. With global oceans being most strongly connected in the south, efforts also could be directed towards understanding the role of Antarctic climate in modulating global climate change, considering that surface temperatures in the Southern Ocean may be a major factor in variability of heat content in the whole ocean, though it is still not understood how that transfer of heat happens (Baggenstos et al., 2019; Clark et al., 2018; Shackleton et al., 2020). The breakout discussion participants agreed that understanding the timescales and magnitude of heat and carbon storage in the deep ocean will likely be highly relevant to the nexus of science and decision making.
Following the discussion on warm periods, the breakout participants shared that key intervals in the past could be used to identify the state dependence or stability of deep circulation, which could also include targeted intervals of time where the forcing is relatively well known. Walczak pitched the importance of marine sediment records containing proxy records that come with complex transport and depositional histories, requiring careful matching between research questions and cores. Complex cores can answer
many paleo environmental questions, but require knowledge of the depositional sequence using geophysical tools like computed tomography (CT) scanning, Walczak said.
Ocean Productivity and Biochemistry
The ocean productivity and biochemistry breakout group moderated by Christopher Hayes, University of Southern Mississippi, and Summer Praetorius, USGS, discussed overarching, outstanding questions of ocean productivity: How is primary productivity linked to export productivity, and what are the drivers of abrupt changes in ocean productivity, including short-term feedbacks and timescales for recovery? Because ecological changes in the ocean will have impacts on the climate system, participants discussed investigating whether abrupt changes in marine ecosystems are integral to abrupt climate changes in the past.
Biominerals form the basis for some of the most foundational records in paleoceanography from ice ages to past sea-surface temperatures, and this foundational chemical information is used to test hypotheses about the role of the ocean in past climates, explained Alex Gagnon, University of Washington. However, proxies in biominerals are complex to interpret because they can be sensitive to multiple environmental factors, and their biological bias can complicate the interpretation. This complexity offers an opportunity to separate out environmental factors; in addition, within biominerals, there are useful relationships between geochemical signatures that are correlated, reproducible, and systematic. For many proxies, their primary compositional signature is driven by physiology and the environmental factors modulating that physiology, so there may be opportunities to integrate process studies of how marine organisms build their skeletons, how proxies work, and geochemical models to translate mechanistic biomineralization understanding into understanding of proxies at scales of interest.
Gagnon noted that tools from material science and physics allow for linking different scales from the mechanics of how biominerals are incorporated into the skeletons of marine organisms to the interpretation of proxy records (Figure 15). The breakout discussion raised the possibility of using other emerging tools, such as genomics, that can reveal large-scale changes in biodiversity and trophic shifts that accompany past changes in climate. Participants also repeated the need to ground truth proxies on a fine scale, which can be done by merging modern studies of transects with paleo coring efforts.
During this session, workshop participants focused on terrestrial proxies and feedbacks related to fire dynamics, vegetation dynamics, and the global carbon cycle. Examples of proxy records that can be used to understand these systems include tree rings, charcoal, peat sediments, speleothems, and biomarkers.
Terrestrial proxies can be particularly relevant to policy making, planning, and public communication because they provide information about the places where people live. Although no time in the past is a complete analog for current or future climate, Jack Williams, University of Wisconsin–Madison, challenged the scientific community to develop and better study a series of model systems, each with elements relevant to policy makers, climate adaptation planners, and for public communication. He suggested that some example model systems that the community can deliver include processes governing the climate system during past warm states; mass extinctions and drivers of major biodiversity losses in the past; triggers of abrupt climate events and their biological consequences; carbon-cycle perturbations; and, droughts, fires, and extreme events. Williams specifically called for a renewed emphasis on climate impacts and adaptation, and suggested several science questions in this vein: How do tipping points cascade among Earth system components and what are the key feedbacks? How does past climate variability shape past and present biodiversity and how sensitive are species and ecosystems to climate change? Are ecological responses to abrupt climate change fast, slow, or abrupt, and what are the governing mechanisms? How does climate change
affect fire regimes and trigger conversions to new ecological states? Methodologically, Williams highlighted as particularly promising the combination of large community data systems that allow joint study of past climates and ecosystem changes at global scales (Box 8), new statistical techniques for integrating observations with simulations from Earth system models, and the advancement of ancient DNA extracted from sediments for new insights into past biodiversity dynamics.
Fire and Vegetation
Fire is an important source of gases and aerosols in the climate system, explained Joe McConnell, Desert Research Institute, which contributes to atmospheric chemistry, including oxidation capacity and atmospheric lifetime. Aerosols, including black carbon, the primary light-absorbing component, contribute directly and indirectly to atmospheric radiative forcing, surface albedo reduction, and ocean fertilization. Understanding past fire emissions and concentrations of atmospheric aerosols is critical for making future climate predictions. Paleo-fire and paleo-vegetation proxy records—including fire scars in trees; charcoal, biomarkers, and pollen in lake sediments, peat, and ocean sediments; and gases and aerosol tracers in glacier ice—are used to quantify past fire emissions and their drivers.
Given the importance of emissions, participants in the fire and vegetation breakout discussion, moderated by Cathy Whitlock, Montana State University, and Mary Edwards, University of Southampton, described how linking emissions information with the paleo fire record may be useful—for example, assessing emissions by comparing marine and terrestrial records to understand the scale of fire feedbacks into the atmosphere. With respect to paleo proxies, McConnell said one outstanding question is how reliable, stable, and reproducible these archives are (e.g., black carbon in lake sediments) over millennial timescales. Another open question about drivers of fire emissions is what was the role of humans—using fire as a land-management tool in preindustrial societies and using fire suppression in recent times—and are the relationships derived from the past stable and applicable for future climate? The breakout group participants also discussed interest in better separating out the human impacts on both fire regimes and land-use changes as compared to natural climate variability.
Grant Harley, University of Idaho, explained that tree-ring records are limited to where trees form consistent annual growth rings, leaving spatial gaps in Africa and especially the tropics, which could help to fill in the paleo temperature record. Participants in the breakout session added that for tree-ring studies—as well as for charcoal and other biomarkers which can reveal
connections between vegetation and climate change—there are critical spatial gaps particularly in Africa, as well as in northern Asia, Russia, Siberian forests, and the American tropics. While it can be more challenging to extract a temperature signal from tree rings in the tropics, Harley explained that it requires being savvy about site selection, finding trees growing in habitats where temperature rather than moisture is the limiting factor, and utilizing new tools and metrics, including quantitative wood anatomy and blue light intensity. Temporally, data during past warm periods and times of abrupt change, times of critical transitions in climate and vegetation states, for example 12 million years ago during the grassland transition, and in recent centuries where fire management has altered ecosystems, may be essential. Harley explained that increasing the network density of tree-ring records and maximizing tree-ring chronology lengths across sites can both reveal subregional information about hydroclimate and temperature, as well as provide a fuller picture of the range of variability of past changes. Longer records could provide information on past and current spatiotemporal characteristics of climate variability, the range and amplitude of preinstrumental conditions, the timing and direction of rate changes, and influence of ocean and atmospheric forcing mechanisms on long-term variability, all of which are important considerations for policy and decision makers.
Carbon Cycle and Feedbacks
Near-surface permafrost soils (remains of plants and animals that died tens of thousands of years ago) contain about 1,500 Pg of carbon—twice as much as there is in the atmosphere, and the equivalent of about 150 years’ worth of present-day fossil fuel emissions (Figure 16). An unknown amount of additional frozen carbon, which could be at least an order of magnitude larger, is stored in deeper sedimentary deposits, explained Katey Walter Anthony, University of Alaska Fairbanks. When permafrost thaws, soil microbes decompose the organic matter and generate greenhouse gases such as methane and CO2. These gases contribute to warming, which causes more thaw and more carbon release, known as the permafrost carbon feedback (shown in Figure 16). This process is thought to have been a normal part of glacial-interglacial cycles for the past 1.2 million years in which, during glacial periods, permafrost soils accumulate carbon in unglaciated pockets of the landscape, and when the climate warms during deglaciations and interglacial periods, some of the old carbon thaws and is in part returned to the atmosphere as greenhouse gases.
A policy-relevant question is whether a runaway permafrost carbon feedback is a threat in a warmer world. Models predict that warming in the next 80 years could release up to 170 Pg of carbon, representing approximately 10 percent of projected fossil fuel emissions (Walter Anthony et al., 2016). However, it was noted that there is evidence from paleo records that previous permafrost thaw may not have led to spikes in atmospheric greenhouse gas levels, with some likely compensation from simultaneous uptake of atmospheric carbon by new permafrost soils and peatlands in areas that previously supported glaciers and ice sheets. The rate of carbon release is a major difference between the past and future, but there is limited knowledge about the response of permafrost to past warming, making it difficult to evaluate the permafrost carbon feedback in the future. Walter Anthony highlighted knowledge gaps in the permafrost carbon feedback from the last deglaciation (21,000 years ago) to present, including: the emission rate and climate impact, permafrost thaw regime, form of carbon emissions (CO2 vs. methane), how much thawed sediment carbon was laterally transported to other environments, constraints on the timing and rates of carbon uptake by Holocene soils, and the carbon stocks and fates of subsea permafrost on continental shelves. Looking back beyond the LGM, periods of abrupt permafrost thaw over decades to centuries could be useful to consider and use as analogs for the future.
Andy Ridgwell, University of California, Riverside, described short-term feedbacks in the terrestrial carbon cycle (Figure 17), which underscore the importance of vegetation through albedo changes, evapotranspiration and water storage, and the potential for exposed soils to cause dust entrainment and transport. Currently, there are challenges reconstructing vegetation distributions in the past and maintaining modeling systems that are flexible enough to simulate vegetation changes through the Cenozoic. Walter Anthony added that having good understanding of ecosystem productivity during past climate regimes would help to better incorporate permafrost into models. Relatedly, increased understanding of lakes and wetlands during past glacial periods would require better knowledge of the topography and wetland vegetation to understand the conversion of carbon to methane in the past. Participants in the terrestrial and ocean carbon cycle breakout session, moderated by Matthew Winnick, University of Massachusetts Amherst, also discussed the need for better understanding land-atmosphere feedbacks, specifically the coupling of CO2 and water, in order to constrain tropical carbon cycle dynamics, which includes attention to reconstructing past climates in the tropics, and leveraging remote sensing of modern carbon pools.
In terms of long-term feedbacks, silicate weathering is thought to be the long-term control on atmospheric CO2 on timescales of hundreds of thousands of years, but there may also be important processes in the ocean including the alteration of the seafloor. Participants in the breakout discussion described how seafloor carbon cycle processes are underexplored. There also could be opportunities to examine carbon cycle dynamics during transitions from hot to cold states, as well as to focus on transition zones between the terrestrial and oceanic environment. Workshop participants also discussed interest in better characterizing the role of carbon cycle processes traditionally thought of as long-term or geologic on shorter timescales and under evolving climate states, particularly processes thought of as being imbalanced, and how they may have driven climate changes in the past. In terms of relevance for the future, Ridgwell explained that the year 2100 has no relevance for silicate weathering, but for timescales on the order of thousands to 10,000 years, the carbon cycle response to cumulative CO2 emissions will only decay slowly, still leading to prolonged warming that can, over millennia, melt a lot of ice mass (Winkelmann et al., 2015).
One challenge in model development is the tension between the need for high-resolution modeling for a dynamical atmosphere and the evolution of soil and regolith growth and erosion and the very long integration times needed to
simulate weathering processes. Ridgwell called for creative modeling approaches, for example, combining multiple offline or asynchronous models or creating efficient parameterizations to encompass all of the processes that are used to model weathering feedbacks. Workshop participants discussed adding increasing layers of complexity to achieve true Earth system models versus the role of simple models. They suggested several different strategies, including integrating fluxes from a complex model into simpler models, building parameterizations from snapshots of fully interactive model runs, and using coarse model resolution for computationally expensive components like the atmosphere.
This page intentionally left blank.