Search

The carbon cycle component of the newly developed Earth system model of intermediate complexity CLIMBER-X is presented. The model represents the cycling of carbon through the atmosphere, vegetation, soils, seawater and marine sediments. Exchanges of carbon with geological reservoirs occur through sediment burial, rock weathering and volcanic degassing. The state-of-the-art HAMOCC6 model is employed to simulate ocean biogeochemistry and marine sediment processes. The land model PALADYN simulates the processes related to vegetation and soil carbon dynamics, including permafrost and peatlands. The dust cycle in the model allows for an interactive determination of the input of the micro-nutrient iron into the ocean. A rock weathering scheme is implemented in the model, with the weathering rate depending on lithology, runoff and soil temperature. CLIMBER-X includes a simple representation of the methane cycle, with explicitly modelled natural emissions from land and the assumption of a constant residence time of CH4 in the atmosphere. Carbon isotopes 13C and 14C are tracked through all model compartments and provide a useful diagnostic for model–data comparison. A comprehensive evaluation of the model performance for the present day and the historical period shows that CLIMBER-X is capable of realistically reproducing the historical evolution of atmospheric CO2 and CH4 but also the spatial distribution of carbon on land and the 3D structure of biogeochemical ocean tracers. The analysis of model performance is complemented by an assessment of carbon cycle feedbacks and model sensitivities compared to state-of-the-art Coupled Model Intercomparison Project Phase 6 (CMIP6) models. Enabling an interactive carbon cycle in CLIMBER-X results in a relatively minor slow-down of model computational performance by ∼ 20 % compared to a throughput of ∼ 10 000 simulation years per day on a single node with 16 CPUs on a high-performance computer in a climate-only model set-up. CLIMBER-X is therefore well suited to investigating the feedbacks between climate and the carbon cycle on temporal scales ranging from decades to >100 000 years.

The current generation of Earth system models exhibits large inter-model differences in the simulated climate of the Arctic and subarctic zone, with differences in model structure and parametrizations being one of the main sources of uncertainty. One particularly challenging aspect in modelling is the representation of terrestrial processes in permafrost-affected regions, which are often governed by spatial heterogeneity far below the resolution of the models' land surface components. Here, we use the Max Planck Institute (MPI) Earth System Model to investigate how different plausible assumptions for the representation of permafrost hydrology modulate land–atmosphere interactions and how the resulting feedbacks affect not only the regional and global climate, but also our ability to predict whether the high latitudes will become wetter or drier in a warmer future. Focusing on two idealized setups that induce comparatively “wet” or “dry” conditions in regions that are presently affected by permafrost, we find that the parameter settings determine the direction of the 21st-century trend in the simulated soil water content and result in substantial differences in the land–atmosphere exchange of energy and moisture. The latter leads to differences in the simulated cloud cover during spring and summer and thus in the planetary energy uptake. The respective effects are so pronounced that uncertainties in the representation of the Arctic hydrological cycle can help to explain a large fraction of the inter-model spread in regional surface temperatures and precipitation. Furthermore, they affect a range of components of the Earth system as far to the south as the tropics. With both setups being similarly plausible, our findings highlight the need for more observational constraints on the permafrost hydrology to reduce the inter-model spread in Arctic climate projections.

Arctic permafrost degradation and carbon decomposition do not occur homogeneously across Arctic ecosystems due to the rich landscape diversity and the high amount of small-scale heterogeneities. Traditionally, Earth system models (ESM) are deployed to investigate future climate change in the northern permafrost areas. The typical heterogeneous landscape characteristics of the Arctic are however in scale well below the usual ESM resolutions of several hundred kilometers. To take in-depth account of small-scale heterogeneous landscapes, a higher land surface model resolution is advantageous.To investigate whether and why resolution matters in simulating the interactions of soil physics, hydrology, and vegetation in the Arctic, we develop a high-resolution version of the land surface model (LSM) JSBACH3 on the scale of 5 km for a case study in the Chersky region in eastern Siberia. We then compare the results with the output of the same model in a low ESM resolution of about 200 km. The LSM simulations are performed in standalone mode (without feedbacks to climate) using the same climate forcing for both, high- and low- resolution setups. Our analysis shows that small-scale soil characteristics are more relevant regarding resolution than vegetation properties. We found that the formulation of supercooled water processes in the soil has a major impact on the differences between low and fine resolutions, as well as soil organic matter fractions. Other soil parameters such as hydraulic conductivity, soil porosity or heat conductivity have relatively minor effects on differences between model resolutions.We show the relevance of model resolution in the simulation of Arctic land physical and biogeochemical interactions and thus argue that the development of a high-resolution pan-Arctic LSM would be a major advancement in modelling future Arctic permafrost and carbon projections.

Climate policies such as the Paris Agreement are framed in terms of global warming levels. Based on past warming and past CO2 emissions, the amount of future cumulative CO2 emissions allowed to keep global warming at or below a global warming level can be estimated. Yet, global warming scenarios in the successive Coupled Model Intercomparison Projects are framed in terms of prescribed atmospheric CO2 concentration or emissions, yielding a wide range of warming levels per CO2 pathway in response to the different transient climate responses to cumulative emissions in the coupled climate models. Based on these scenarios and the latest model projections, the IPCC Sixth Assessment Report assessed climatic impacts of different warming levels. These impacts are thus evaluated in simulations where the warming targets are passed transiently, at different points in time, and not stabilized, as opposed to how climate agreements are framed.Here, we propose a new Model Intercomparison Project AERA-MIP building on an adaptive approach - the Adaptive Emissions Reduction Approach - that successively calculates the compatible emissions to stabilize global warming at the required temperature target. Earth System Models (ESMs) are run forward in emission-driven mode, with prescribed, model-specific emissions successively calculated every five years, so that all models reach the same warming target and thereafter stabilize at this warming level. The warming uncertainty is thus side-stepped, while different emissions pathways emerge out of the variety of participating ESMs. The approach is based on the TCRE framework and successively adapting for any changes in the Earth System that might affect global mean surface temperature, including the zero emissions commitment as emissions approach zero.Simulations of the first participating modelling centers already reveal a panel of emissions pathways that successfully stabilize global warming at 1.5ºC and 2ºC. This includes the decline rate from peak emissions, the timing of having to reach net-zero emissions, and the magnitude of negative emissions needed to stabilize the climate. These different emissions pathways result in a range of atmospheric CO2 concentration evolution (350 to 450 ppm at year 2100 in the 1.5°C stabilization scenario) and distribution of anthropogenic carbon in the Earth System components. Unlike concentration-driven projections, these AERA simulations provide an uncertainty range for impacts that are directly affected by atmospheric CO2 concentration such as ocean acidification. The project also includes temporary temperature overshoot simulations using the AERA approach.

Current strategies to hold surface warming below a certain level, e.g., 1.5 or 2 °C, advocate limiting total anthropogenic carbon emissions to ∼0.9 or ∼1.25 Eg C (1018 grams carbon), respectively1,2. These allowable emission budgets are based on a near-linear relationship between cumulative emissions and warming identified in various modeling efforts3–7. The IPCC assesses this near-linear relationship with high confidence in its Summary for Policymakers8 (§D1.1 and Figure SPM.10). Here we test this proportionality in specially designed simulations with a latest-generation Earth system model that includes an interactive carbon cycle with updated terrestrial ecosystem processes, and a suite of CMIP simulations. We find that atmospheric CO2 concentrations differ by ∼100 ppmv and surface warming by ∼0.3 °C (0.46 °C over land) for the same cumulated emissions (∼1.2 Eg C, 2 °C approximate carbon budget1). CO2 concentration and warming per 1 Eg of emitted carbon (Transient Climate Response to Cumulative Carbon Emissions6,9,10; TCRE) depend not just on total emissions, but also on the duration and time sequence of emissions, which heretofore have been ignored. The time dependency clearly arises due to lagged carbon sequestration processes on land, viz., ecological succession, land-cover, and demographic changes, etc., which are still poorly represented in most ESMs. This implies a temporally evolving state of the carbon system, but one which surprisingly apportions carbon into land and ocean sinks in a manner that is independent of the emission pathway. Therefore, even though TCRE differs for different pathways with the same total emissions, it is roughly constant when related to the state of the carbon system, i.e., the amount of carbon stored in surface sinks. Efforts to better understand the state of the carbon system with observations and refined models are needed to accurately project the impact of future emissions.

Earth System Models (ESMs) are the best available tools to project the coupled dynamics of the climate and biogeochemistry under future emission scenarios. However, the future trajectories simulated by individual ESMs vary substantially with most pronounced differences in the high northern latitudes. As recently demonstrated (de Vrese et al., 2022), a significant part of this uncertainty might result from the different approaches and parametrizations of surface and soil hydrology in the permafrost regions. However, this study did not account for sub-grid lateral fluxes. To make a step forward, we further develop ICON-Land/JSBACH4, the land surface model (LSM) used within the ICON-ESM. Our recent efforts are focused on improving the simulation of Arctic hydrology by accounting for lateral water flows on small spatial scales, i.e. within the grid cells of the LSM. For this, we apply a tiling structure in which we define the spatial relation between parts of the grid cell in terms of water exchange. In this way, the model gets information about the source and sink tiles of surface runoff (based on topography) but also of lateral soil water exchange (based on proximity and soil moisture gradient).This approach results in a redistribution of surface and soil water within the grid cells with drier upland and wetter lowland regions and the correspondent changes in evapotranspiration. Comparing coupled land-atmosphere simulations with different prescribed fractions of upland and lowland areas, we see strong impacts of the tiling structure. Setups with a larger lowland-to-upland ratios lead to higher cloud cover and by up to 2K lower summer surface temperature over larger parts of the boreal regions. This result emphasizes the importance of representing the complex processes of Arctic hydrology, but also the need for detailed information about Arctic land surface properties. References: de Vrese, P., Georgievski, G., Gonzalez Rouco, J. F., Notz, D., Stacke, T., Steinert, N. J., Wilkenskjeld, S., and Brovkin, V.: Representation of soil hydrology in permafrost regions may explain large part of inter-model spread in simulated Arctic and subarctic climate, The Cryosphere Discuss., https://doi.org/10.5194/tc-2022-150, in review, 2022.

Most earth system models (ESMs) neglect climate feedbacks arising from carbon release from thawing permafrost, especially from thawing of subsea permafrost (SSPF). To assess the fate of SSPF in the next 1000 years, we implemented SSPF into JSBACH, the land component of the Max Planck Institute Earth System Model (MPI-ESM). This is the first implementation of SSPF processes in an ESM component. We investigate three extended scenarios from the 6th phase of the Coupled Model Intercomparison Project (CMIP6). In the 21st century only small differences are found among the scenarios, but in the upper-end emission scenario SSP5–8.5 (shared socio-economic pathway), especially in the 22nd century, SSPF ice melting is more than 15 times faster than in the pre-industrial period. In this scenario about 35 % of total SSPF volume and 34 % of SSPF area are lost by the year 3000 due to climatic changes. In the more moderate scenarios, the melting rate maximally exceeds that of pre-industrial times by a factor of 4, and the climate change induced SSPF loss (volume and area) by the year 3000 does not exceed 14 %. Our results suggest that the rate of melting of SSPF ice is related to the length of the local open-water season and thus that the easily observable sea ice concentration may be used as a proxy for the change in SSPF.

The fluctuations of atmospheric CO2 concentrations over the preindustrial Common Era are generally attributed to changes in land carbon storage, caused primarily by changes in surface air temperature but also by changes in land use. This dominant influence of the land carbon cycle is consistent with the negative correlation between atmospheric CO2 concentrations and δ13CO2 variations recorded in ice cores. By performing an ensemble of sensitivity experiments with the LOVECLIM model, we confirm the potentially large role that temperature changes have on the land carbon cycle. However, this process alone cannot explain the magnitude of the reconstructed atmospheric CO2 and δ13CO2 variations. In particular, even when the model is constrained to follow reconstructed temperature changes by data assimilation, and when applying relatively large values of the climate-carbon feedback parameter, it can only explain about 50% of the atmospheric CO2 decrease between the 12th and the seventeenth century. We find that land use changes are likely responsible for most of the observed long term atmospheric CO2 trend over the first millennium of the Common Era, and for up to 30% of the decrease observed after 1600 CE. In addition, in our experiments, changes in southern hemisphere westerly winds induce slightly smaller changes in atmospheric CO2 concentrations than those associated with land use change, and variations in δ13CO2 of the same order of magnitude as the observed ones. Combining the effects of changes in temperature, land use and winds over the Southern Ocean provides a reasonable agreement with reconstructions for atmospheric CO2 concentrations and δ13CO2, especially for the low CO2 values observed during the seventeenth century. This underlines the important contribution of both land and ocean carbon processes. Nevertheless, some uncertainties remain on the origin of the relatively high CO2 concentrations reconstructed during the eleventh and sixteenth centuries.

The prediction of atmospheric CO2 concentrations is limited by the high interannual variability (IAV) of terrestrial gross primary productivity (GPP). However, there are large uncertainties in the drivers of GPP IAV among Earth system models (ESMs). Here, we evaluate the impact of these uncertainties on the predictability of atmospheric CO2 in six ESMs. We use regression analysis to determine the role of environmental drivers on (i) the patterns of GPP IAV, and (ii) the predictability of GPP. There are large uncertainties in the spatial distribution of GPP IAV. Although all ESMs agree on the high IAV in the tropics, several ESMs have unique hotspots of GPP IAV. The main driver of GPP IAV is temperature in the ESMs using the Community Land Model, and soil moisture in IPSL-CM6A-LR and MPI-ESM-LR, revealing underlying differences in the source of GPP IAV among ESMs. Between 13 % and 24 % of the GPP IAV is predictable one year ahead, with four out of six ESMs between 19 % and 24 %. Up to 32 % of the GPP IAV induced by soil moisture is predictable, while only 7 % to 13 % of the GPP IAV induced by radiation. The results show that while ESMs are fairly similar in their ability to predict themselves, their predicted contribution to the atmospheric CO2 variability originates from different regions and is caused by different drivers. A higher coherence in atmospheric CO2 predictability could be achieved by reducing uncertainties of GPP sensitivity to soil moisture, and by accurate observational products for GPP IAV.

The Arctic is warming at an above-average rate, and small, shallow waterbodies such as ponds are vulnerable to this warming due to their low thermal inertia compared to larger lakes. While ponds are a relevant landscape-scale source of methane under the current climate, the response of pond methane emissions to warming is uncertain. We employ a new, process-based model for methane emissions from ponds (MeEP) to investigate the methane emission response of polygonal-tundra ponds in northeastern Siberia to warming. MeEP is the first dedicated model of pond methane emissions which differentiates between the three main pond types of the polygonal-tundra, ice-wedge, polygonal-center, and merged polygonal ponds and resolves the three main pathways of methane emissions – diffusion, ebullition, and plant-mediated transport. We perform idealized warming experiments, with increases in the mean annual temperature of 2.5, 5, and 7.5 ∘C on top of a historical simulation. The simulations reveal an approximately linear increase in emissions from ponds of 1.33 g CH4 yr−1 ∘C−1 m−2 in this temperature range. Under annual temperatures 5 ∘C above present temperatures, pond methane emissions are more than 3 times higher than now. Most of this emission increase is due to the additional substrate provided by the increased net productivity of the vascular plants. Furthermore, plant-mediated transport is the dominating pathway of methane emissions in all simulations. We conclude that vascular plants as a substrate source and efficient methane pathway should be included in future pan-Arctic assessments of pond methane emissions.

Satellite data reveal widespread changes in Earth's vegetation cover. Regions intensively attended to by humans are mostly greening due to land management. Natural vegetation, on the other hand, is exhibiting patterns of both greening and browning in all continents. Factors linked to anthropogenic carbon emissions, such as CO2 fertilization, climate change, and consequent disturbances such as fires and droughts, are hypothesized to be key drivers of changes in natural vegetation. A rigorous regional attribution at the biome level that can be scaled to a global picture of what is behind the observed changes is currently lacking. Here we analyze different datasets of decades-long satellite observations of global leaf area index (LAI, 1981–2017) as well as other proxies for vegetation changes and identify several clusters of significant long-term changes. Using process-based model simulations (Earth system and land surface models), we disentangle the effects of anthropogenic carbon emissions on LAI in a probabilistic setting applying causal counterfactual theory. The analysis prominently indicates the effects of climate change on many biomes – warming in northern ecosystems (greening) and rainfall anomalies in tropical biomes (browning). The probabilistic attribution method clearly identifies the CO2 fertilization effect as the dominant driver in only two biomes, the temperate forests and cool grasslands, challenging the view of a dominant global-scale effect. Altogether, our analysis reveals a slowing down of greening and strengthening of browning trends, particularly in the last 2 decades. Most models substantially underestimate the emerging vegetation browning, especially in the tropical rainforests. Leaf area loss in these productive ecosystems could be an early indicator of a slowdown in the terrestrial carbon sink. Models need to account for this effect to realize plausible climate projections of the 21st century.

Atmospheric methane (CH4) has changed considerably in the time between the last glacial maximum (LGM) and the preindustrial period (PI). We investigate these changes in transient experiments with an Earth System Model, focusing on the rapid changes during the deglaciation, especially pronounced in the Bølling Allerød (BA) and Younger Dryas (YD) periods. We consider all relevant natural sources and sinks of methane and examine the drivers of changes in methane emissions as well as in the atmospheric lifetime of methane. We find that the evolution of atmospheric methane is largely driven by emissions from tropical wetlands, while variations in atmospheric lifetime are not negligible but small. Our model reproduces most changes in atmospheric methane very well, with the exception of the mid-Holocene decrease in methane, though the timing of ice sheet meltwater fluxes needs to be adjusted slightly in order to exactly reproduce the variations of the BA and YD.

Terrestrial ecosystems respond to changes in environmental conditions, mainly via key climatic controls of precipitation and temperature on vegetation activities and decomposition processes (Taylor et al., 2017). Yet, the relationship between climate and the overall spatiotemporal dynamics and uncertainties of the global carbon cycle, i.e., gross primary productivity (GPP), effective ecosystem carbon turnover times (τ), and consequently the total ecosystem carbon stock (Ctotal), are unclear (Anav et al., 2013; Friedlingstein et al., 2014; Friend et al., 2014; Jones et al., 2013). Using a global observation-based synthesis, we first show that the apparent partial spatial climate sensitivities of GPP and τ are associated with relative availability of precipitation and temperature, and are therefore modulated by aridity. The apparent sensitivity of GPP to temperature increases from arid to humid climatic regions. In contrast, its sensitivity to precipitation is invariant throughout different climatic regions. Simultaneously, the τ-precipitation response is strongly non-linear resulting in ~2 times longer τ in arid regions compared to humid regions for a given temperature. Compared with these observed patterns, the offline carbon cycle simulations of seven European Earth System Models (ESMs), that participated in CMIP6, perform relatively better for climate sensitivities of GPP than those of τ. This leads to a large spread and bias in Ctotal in both warm and cold semi-arid and arid regions where only a few models capture the observed τ-precipitation relationship. The emergence of the hydrological controls, modulated by aridity, on global carbon cycle implies that the changes in precipitation may moderate the temperature-driven climate feedback of the global carbon cycle under climate change.

Climate archives reveal alternating arid and humid conditions in North Africa during the last several million years. Most likely the dry phases resembled current hyper-arid landscapes, whereas the wet phases known as African humid periods (AHPs) sustained much more surface water and greater vegetated areas that “greened” large parts of the Sahara region. Previous analyses of sediment cores from the Mediterranean Sea showed the last five AHPs differed in strength, duration and rate of change. To understand the causes of such differences we perform transient simulations of the past 190 000 years with the Earth system model of intermediate complexity CLIMBER-2. We analyse the amplitude and rate of change of the modelled AHP responses to changes in orbital parameters, greenhouse gases (GHGs) and ice sheets. In agreement with estimates from Mediterranean Sea sapropels, we find the model predicts a threshold in orbital forcing for Sahara greening and occurrence of AHPs. Maximum rates of change in simulated vegetation extent at AHP onset and termination correlate strongly with the rate of change of the orbital forcing. As suggested by available data for the Holocene AHP, the onset of modelled AHPs usually happens faster than termination. A factor separation analysis confirms the dominant role of the orbital forcing in driving the amplitude of precipitation and vegetation extent for past AHPs. Forcing due to changes in GHGs and ice sheets is only of secondary importance, with a small contribution from synergies with the orbital forcing. Via the factor separation we detect that the threshold in orbital forcing for AHP onset varies with GHG levels. To explore the implication of our finding from the palaeoclimate simulations for the AHPs that might occur in a greenhouse-gas-induced warmer climate, we extend the palaeoclimate simulations into the future. For the next 100 000 years the variations in orbital forcing will be smaller than during the last 100 millennia, and the insolation threshold for the onset of late Quaternary AHPs will not be crossed. However, with higher GHG concentrations the predicted threshold drops considerably. Thereby, the occurrence of AHPs in upcoming millennia appears to crucially depend on future concentrations of GHGs.

The present study investigates the response of the high-latitude carbon cycle to changes in atmospheric greenhouse gas (GHG) concentrations in idealized climate change scenarios. To this end we use an adapted version of JSBACH – the land surface component of the Max Planck Institute for Meteorology Earth System Model (MPI-ESM) – that accounts for the organic matter stored in the permafrost-affected soils of the high northern latitudes. The model is run under different climate scenarios that assume an increase in GHG concentrations, based on the Shared Socioeconomic Pathway 5 and the Representative Concentration Pathway 8.5, which peaks in the years 2025, 2050, 2075 or 2100, respectively. The peaks are followed by a decrease in atmospheric GHGs that returns the concentrations to the levels at the beginning of the 21st century, reversing the imposed climate change. We show that the soil CO2 emissions exhibit an almost linear dependence on the global mean surface temperatures that are simulated for the different climate scenarios. Here, each degree of warming increases the fluxes by, very roughly, 50 % of their initial value, while each degree of cooling decreases them correspondingly. However, the linear dependence does not mean that the processes governing the soil CO2 emissions are fully reversible on short timescales but rather that two strongly hysteretic factors offset each other – namely the net primary productivity and the availability of formerly frozen soil organic matter. In contrast, the soil methane emissions show a less pronounced increase with rising temperatures, and they are consistently lower after the peak in the GHG concentrations than prior to it. Here, the net fluxes could even become negative, and we find that methane emissions will play only a minor role in the northern high-latitude contribution to global warming, even when considering the high global warming potential of the gas. Finally, we find that at a global mean temperature of roughly 1.75 K (±0.5 K) above pre-industrial levels the high-latitude ecosystem turns from a CO2 sink into a source of atmospheric carbon, with the net fluxes into the atmosphere increasing substantially with rising atmospheric GHG concentrations. This is very different from scenario simulations with the standard version of the MPI-ESM, in which the region continues to take up atmospheric CO2 throughout the entire 21st century, confirming that the omission of permafrost-related processes and the organic matter stored in the frozen soils leads to a fundamental misrepresentation of the carbon dynamics in the Arctic.

Variations in the Earth's orbit are recognised as the main trigger for the hydrological changes that led to the periodic 'greening' of the Sahara region over the late Quaternary. However, the frequency and amplitude of the greening events as seen in the geological records cannot be predicted from orbital theory alone. To understand the changes in the proxy data it is also important to consider feedback mechanisms that arise from the complexity of the interactions between the vegetation, land, atmosphere and ocean components in the region. Yet discrepancies between state-of-the-art computer simulations of greening during African Humid Periods (AHPs) and proxy data still remain. We hypothesize that the effects of additional internal forcing from other climate drivers like atmospheric levels of greenhouse gases (GHGs) and extension of ice sheets may have had a greater impact than previously thought. Using two climate models of varying complexity and spatial resolution, CLIMBER-2 and MPI-ESM, we simulate several of the greening events in the Sahara within the last glacial cycle and study the effects of the orbital, GHGs and ice sheets forcings for every greening response. The results from CLIMBER-2 suggest that the critical insolation at the Tropics required for AHPs onset depends on atmospheric levels of GHGs, while the results from MPI-ESM show that the spatial pattern that develops during AHPs varies with all three forcing factors. These findings highlight the role that GHGs may play for the future of Saharan climate, when low--eccentricity orbits concur with high levels of atmospheric GHGs.

The representation of the terrestrial thermal and hydrological states in current-generation climate models is crucial to have a realistic simulation of the subsurface physical processes and land-atmosphere coupling. This is particularly important for high-latitude permafrost regions since these areas are prone to the release of substantial amounts of carbon from degrading permafrost under climate-change conditions. Many current-generation climate models still have deficiencies in the representation of terrestrial structure and physical mechanisms, such as too shallow land depth or insufficient hydro-thermodynamic coupling. We therefore introduce a deeper bottom boundary into the JSBACH land surface model. The associated changes in the simulated terrestrial thermal state influence the near-surface hydroclimate when sufficient coupling between the thermodynamic and hydrological regimes is present. Hence, we also assess the influence of introducing various physical modifications for the representation of soil hydro-thermodynamic processes in climate projections of the 21st century. The results show significant impacts on terrestrial energy uptake, as well as changes in global near-surface ground temperatures when introducing the physical modifications. The resulting simulation of high-latitude permafrost extent is subject to large variations depending on the model configuration, reflecting the uncertainty of carbon release from permafrost degradation. We further use the modified model to assess the sensitivity of simulated high-latitude climate dynamics to different hydrological configurations in the coupled MPI-ESM. The differences in soil hydrological representation in permafrost regions could explain a large part of CMIP6 inter-model spread in simulated Arctic climate, with remote effects on subarctic dynamical systems.

By performing an ensemble of sensitivity experiments with the LOVECLIM model, we confirm the earlier results showing that temperature changes had a large influence on the land carbon cycle over the preindustrial Common Era. However, this process alone cannot explain the magnitude of the reconstructed atmospheric CO2 and δ13CO2 variations. In particular, even when the model is constrained to follow reconstructed temperature changes by data assimilation, and when applying relatively large values of the climate-carbon feedback parameter, it can only explain about 50% of the atmospheric CO2 decrease between the 12th and the 17th century. We find that land use changes are likely responsible for most of the observed long term atmospheric CO2 trend over the first millennium of the Common Era, and for up to 30 % of the decrease observed after 1600 CE. In addition, in our experiments, changes in southern hemisphere westerly winds induce slightly smaller changes in atmospheric CO2 concentrations than those associated with land use change, and variations in δ13CO2 of the same order of magnitude as the observed ones. Those wind changes also have a strong impact on the difference in 14C between the northern and southern hemisphere, presenting strong similarities with observed changes. Combining the effects of changes in temperature, land use and winds over the Southern Ocean provides a reasonable agreement with reconstructions for atmospheric CO2 concentrations and δ13CO2, especially for the low CO2 values observed during the 17th century. This underlines the important contribution of both land and ocean carbon processes. Nevertheless, some uncertainties remain on the origin of the relatively high CO2 concentrations reconstructed during the 11th and 16th centuries.

The carbon budget concept (TCRE; Transient Climate Response to cumulative carbon Emissions) emerged as a major concept in climate research since the late 2000s. Due to its simplicity, it is intensively utilized in the international policy arena. It is based on the claim that one can derive the global mean temperature increase solely from the knowledge of historical cumulative emissions by observing the linear relationship between the two, regardless of the emission pathway that preceded ('pathway independence').Here, we ask for the maximally possible deviations from the TCRE ideal across emission scenario space. While there has been an extensive focus on quantifying the carbon budget using highly complex climate models, there seems to be a lesser focus on the pathway independence and possibly related deviations from the budget. Furthermore, few analytical examinations have been presented, for highly stylized settings only. This study contributes to filling that gap, utilizing the energy balance model FAIR. FAIR incorporates climate feedbacks and correctly emulates the temperature response to an emission pulse.If the carbon budget approach was perfectly valid, the temperature response to an emitted unit of carbon should be a perfect step function. The actual temperature evolution following the emission pulse is reinterpreted as a Green's function and as such, utilized to calculate the total temperature increase at any given point. The novelty in this work is that the emission pathway is not assumed, but generated by maximizing (minimizing) the temperature output.With the boundary conditions being the fixed total cumulative emissions and the maximal allowed mitigation efforts, two associated pathways are generated with the temperature increase in a given year acting as an objective value. The deviation from the budget is then extracted as a temperature difference between the upper and the lower bound of the optimization process. The results show that the absolute value of the deviation is less than the standard deviation of climate variability, confirming the fundamentals of the carbon budget approach. We also present an analytical upper bound of the deviation from path independence. The result shows that the deviation is a function of the allowed maximum emission slope.The advantage of this method is that it can utilize the impulse response properties already published for highly complex models. The current limitation of the presented approach lies in the assumption that the pulse response is assumed constant even though the climate changes. The implications of a changing pulse remain to be explored. We see our work as a twofold contribution: (i) to predict maximally possible TCRE deviations from already published impulse response experiments, and (ii), to generate analytic understanding for the driving variables.

Numerous climate archives reveal alternating arid and humid conditions in North Africa during the last several million years. Most likely the dry phases resembled current hyper-arid landscapes, whereas the wet phases known as African Humid Periods (AHPs) sustained much more surface water and greater vegetated areas that "greened" a large part of the Sahara region. Previous analyses of sediment cores from the Mediterranean Sea showed the last five AHPs differed in strength, duration and rate of change. To understand the causes of such differences we perform transient simulations of the past 190,000 years with Earth system model of intermediate complexity CLIMBER-2. We analyse amplitude and rate of change of the modelled AHPs responses to changes in orbital parameters, greenhouse gases (GHGs) and ice sheets. In agreement with estimates from Mediterranean sapropels, we find the model predicts a threshold in orbital forcing for Sahara greening and occurrence of AHPs. Maximum rates of change in simulated vegetation extent at AHP onset and termination correlate well with the rate of change of the orbital forcing. As suggested by available data for the Holocene AHP, the onset of modelled AHPs happens usually faster than termination. A factor separation analysis confirms the dominant role of the orbital forcing in driving the amplitude of precipitation and vegetation extent for past AHPs. Forcing due to changes in GHGs and ice sheets is only of secondary importance, with a small contribution from synergies with the orbital forcing. Via the factor separation we detect that the threshold in orbital forcing for AHP onset varies with GHGs levels. To explore the implication of our finding from the palaeoclimate simulations for the AHPs that might occur in a greenhouse gas-induced warmer climate, we extend the palaeoclimate simulations into the future. For the next 100,000 years the variations in orbital forcing will be smaller than during the last hundred millennia, and the insolation threshold for the onset of late Quaternary AHPs will not be crossed. However, with higher GHGs concentrations the predicted threshold drops considerably. Thereby, the occurrence of AHPs in upcoming millennia appears to crucially depend on future concentrations of GHGs.

Northern peatlands have been a persistent natural carbon sink since the Last Glacial Maximum. The continued growth and expansion of these carbon-rich ecosystems could offset a large portion of anthropogenic carbon emissions before the end of the present interglacial period. Here we used an impeded drainage model and gridded data on the depth to bedrock and the fraction of histosol-type soils to evaluate the limits to the growth of northern peatland carbon stocks. Our results show that the potential carbon stock in northern peatlands could reach a total of 875±125 Pg C before the end of the present interglacial, which could, as a result, remove 330±200 Pg C of carbon from the atmosphere. We argue that northern peatlands, together with the oceans, will potentially play an important role in reducing the atmospheric carbon dioxide concentration over the next 5000 years.

Most Earth System Models (ESMs) neglect climate feedbacks arising from carbon release from thawing permafrost, especially from thawing of sub-sea permafrost (SSPF). To assess the fate of SSPF in the next 1000 years, we implemented SSPF into JSBACH, the land component of the Max Planck Institute Earth Model (MPI-ESM). This is the first implementation of SSPF processes in an ESM-component. We investigate three extended scenarios from the 6th phase of the Coupled Model Intercomparison Project (CMIP6). In the 21st century only small differences are found among the scenarios, but in the upper-end emission scenario SSP5-8.5, especially in the 22nd century SSPF ice melting is more than 15 times faster than in the preindustrial period. In this scenario about 35 % of total SSPF volume and 34 % of SSPF area is lost by year 3000 due to climatic changes. In the more moderate scenarios, the melting maximally exceeds the preindustrial rate by a factor of 4 and the climate change induced SSPF loss (volume and area) by year 3000 does not exceed 14 %. Our results suggest that the rate of melting of SSPF ice is related to the length of the local open water season, and thus that the easily observable sea ice concentration may be used as a proxy for the change of SSPF.

Arctic coastal erosion damages infrastructure, threatens coastal communities, and releases organic carbon from permafrost. However, the magnitude, timing and sensitivity of coastal erosion increase to global warming remain unknown. Here, we project the Arctic-mean erosion rate to roughly double by 2100 and very likely exceed its historical range of variability by mid-21st century. The sensitivity of erosion to warming also doubles, reaching 0.4-0.5 m year-1 oC-1 and 2.3-2.8 TgC year-1 oC-1 by the end of the century under moderate and high-emission scenarios. Our first 21st-century pan-Arctic coastal erosion rate projections should inform policy makers on coastal conservation and socioeconomic planning. Our organic carbon flux projections also lay out the path for future work to investigate the impact of Arctic coastal erosion on the changing Arctic Ocean, on its role as a global carbon sink, and on the permafrost-carbon feedback.

The atmospheric CO2 concentration increased by about 20 ppm from 6000 BCE to the pre-industrial period (1850 CE). Several hypotheses have been proposed to explain mechanisms of this CO2 growth based on either ocean or land carbon sources. Here, we apply the Earth system model MPI-ESM-LR for two transient simulations of climate and carbon cycle dynamics during this period. In the first simulation, atmospheric CO2 is prescribed following ice-core CO2 data. In response to the growing atmospheric CO2 concentration, land carbon storage increases until 2000 BCE, stagnates afterwards, and decreases from 1 CE, while the ocean continuously takes CO2 out of the atmosphere after 4000 BCE. This leads to a missing source of 166 Pg of carbon in the ocean–land–atmosphere system by the end of the simulation. In the second experiment, we applied a CO2 nudging technique using surface alkalinity forcing to follow the reconstructed CO2 concentration while keeping the carbon cycle interactive. In that case the ocean is a source of CO2 from 6000 to 2000 BCE due to a decrease in the surface ocean alkalinity. In the prescribed CO2 simulation, surface alkalinity declines as well. However, it is not sufficient to turn the ocean into a CO2 source. The carbonate ion concentration in the deep Atlantic decreases in both the prescribed and the interactive CO2 simulations, while the magnitude of the decrease in the prescribed CO2 experiment is underestimated in comparison with available proxies. As the land serves as a carbon sink until 2000 BCE due to natural carbon cycle processes in both experiments, the missing source of carbon for land and atmosphere can only be attributed to the ocean. Within our model framework, an additional mechanism, such as surface alkalinity decrease, for example due to unaccounted for carbonate accumulation processes on shelves, is required for consistency with ice-core CO2 data. Consequently, our simulations support the hypothesis that the ocean was a source of CO2 until the late Holocene when anthropogenic CO2 sources started to affect atmospheric CO2 .

Seasonality in photosynthetic activity is a critical component of seasonal carbon, water and energy cycles in the Earth system. This characteristic is a consequence of plant's adaptive evolutionary processes to a given set of environmental conditions. Changing climate in northern lands (>30°N) alters the state of climatic constraints on plant growth, and therefore, changes in the seasonality and carbon accumulation are anticipated. However, how photosynthetic seasonality evolved to its current state, and what role climatic constraints and their variability played in this process and ultimately in carbon cycle is still poorly understood due to its complexity. Here, we take the 'laws of minimum' as a basis and introduce a new framework where the timing (Day of Year) of peak photosynthetic activity (DOYP max ) acts as a proxy for plant's adaptive state to climatic constraints on its growth. Our analyses confirm that spatial variations in DOYP max reflect spatial gradients in climatic constraints as well as seasonal maximum and total productivity. We find a widespread warming-induced advance in DOYP max (-1.66 ± 0.30 days decade-1 , P < 0.001) across northern lands, indicating a spatio-temporal dynamism of climatic constraints to plant growth. We show that the observed changes in DOYP max are associated with an increase in total gross primary productivity through enhanced carbon assimilation early in the growing season, which leads to an earlier phase shift in land-atmosphere carbon fluxes and an increase in their amplitude. Such changes are expected to continue in the future based on our analysis of Earth System Model (ESM) projections. Our study provides a simplified, yet realistic framework based on first principles for the complex mechanisms by which various climatic factors constrain plant growth in northern ecosystems.

AIM: Previous analyses of remotely sensed data detected the multimodality of the tree‐cover distribution of the boreal forest, and identified areas with potentially alternative tree‐cover states. This paper aims at investigating the causes of multimodality and multistability of the boreal forest, their influence on the asymmetric tree species distribution between Eurasia and North America, and whether multistability could be associated with recent greening trends in leaf area index (LAI) and normalized difference vegetation index (NDVI). LOCATION: Eurasian and North American boreal forests. TIME PERIOD: 2000–2010. MAJOR TAXA STUDIED: Boreal forest plant functional types. METHODS: We employ a conceptual model based on tree species competition to simulate the sensitivity of tree cover to stochastic disturbances and to changes in environmental factors. We include different plant functional types based on survival adaptations, and force the model with remotely sensed environmental data. We analyse the model as a dynamical system. We use metrics from statistics and information theory to compare the detection of alternative tree‐cover states and greening trends in LAI and NDVI. RESULTS: We find that multimodality and multistability can emerge through competition between different plant functional types. Additionally, our model is able to reproduce the asymmetry in tree species distribution between Eurasia and North America. Moreover, changes in permafrost distribution can be associated with phenomenological bifurcation points of the model. Finally, we find that the detection of multistable areas is not affected by recent vegetation trends, whereas shifts between alternative states could have affected the greening trends. MAIN CONCLUSIONS: Tree‐cover multistability in the boreal region can emerge through competition between species subject to periodic disturbances. Changes in permafrost thaw and distribution could be responsible for the asymmetry in tree species distribution between North America and Eurasia. Climate change and permafrost degradation could cause shifts in tree‐cover states and dominant species. Recent vegetation greening trends in multistable areas could have been affected by shifts between alternative states.

Dynamic vegetation models simulate global vegetation in terms of fractional coverage of a few plant functional types (PFTs). Although these models often share the same concept, they differ with respect to the number and kind of PFTs, complicating the comparability of simulated vegetation distributions. Pollen-based vegetation reconstructions are initially only available in the form of time series of individual taxa that are not distinguished in the models. Thus, to evaluate simulated vegetation distributions, the modelling results and pollen-based vegetation reconstructions have to be converted into a comparable format. The classical approach is the method of biomisation, but hitherto PFT-based biomisation methods were only available for individual models. We introduce and evaluate a simple, universally applicable technique to harmonise PFT distributions by assigning them into nine mega-biomes, using only assumptions on the minimum PFT cover fractions and few bioclimatic constraints (based on the 2 m temperature). These constraints mainly follow the limitation rules used in the classical biome models (here BIOME4). We test the method for six state-of-the-art dynamic vegetation models that are included in Earth system models based on pre-industrial, mid-Holocene and Last Glacial Maximum simulations. The method works well, independent of the spatial resolution or the complexity of the models. Large biome belts (such as tropical forest) are generally better represented than regionally confined biomes (warm–temperate forest, savanna). The comparison with biome distributions inferred via the classical biomisation approach of forcing biome models (here BIOME1) with the simulated climate states shows that the PFT-based biomisation is even able to keep up with the classical method. However, as the new method considers the PFT distributions actually calculated by the Earth system models, it allows for a direct comparison and evaluation of simulated vegetation distributions which the classical method cannot do. Thereby, the new method provides a powerful tool for the evaluation of Earth system models in general.

Satellite data show increasing leaf area of vegetation due to direct factors (human land-use management) and indirect factors (such as climate change, CO2 fertilization, nitrogen deposition and recovery from natural disturbances). Among these, climate change and CO2 fertilization effects seem to be the dominant drivers. However, recent satellite data (2000–2017) reveal a greening pattern that is strikingly prominent in China and India and overlaps with croplands world-wide. China alone accounts for 25% of the global net increase in leaf area with only 6.6% of global vegetated area. The greening in China is from forests (42%) and croplands (32%), but in India is mostly from croplands (82%) with minor contribution from forests (4.4%). China is engineering ambitious programmes to conserve and expand forests with the goal of mitigating land degradation, air pollution and climate change. Food production in China and India has increased by over 35% since 2000 mostly owing to an increase in harvested area through multiple cropping facilitated by fertilizer use and surface- and/or groundwater irrigation. Our results indicate that the direct factor is a key driver of the ‘Greening Earth’, accounting for over a third, and probably more, of the observed net increase in green leaf area. They highlight the need for a realistic representation of human land-use practices in Earth system models.

Simple and complex climate models suggest a hard snowball – a completely ice-covered planet – is one of the steady-states of Earth’s climate. However, a seemingly insurmountable challenge to the hard-snowball hypothesis lies in the difficulty in explaining how the planet could have exited the glaciated state within a realistic range of atmospheric carbon dioxide concentrations. Here, we use simulations with the Earth system model MPI-ESM to demonstrate that terminal deglaciation could have been triggered by high dust deposition fluxes. In these simulations, deglaciation is not initiated in the tropics, where a strong hydrological cycle constantly regenerates fresh snow at the surface, which limits the dust accumulation and snow aging, resulting in a high surface albedo. Instead, comparatively low precipitation rates in the mid-latitudes in combination with high maximum temperatures facilitate lower albedos and snow dynamics that – for extreme dust fluxes – trigger deglaciation even at present-day carbon dioxide levels. Snowball Earth could have thawed at atmospheric CO2-levels comparable to the present as a result of low surface albedo in mid-latitudes from a combination dust deposition and low precipitation rates, according to Earth System Model simulations

During the last million years, northern Africa has alternated between arid and humid conditions, as recorded by different kinds of climate archives, including fossil pollen, lake sediments, marine sediments and archaeological remains. Variations occur at millennial scale, with dry phases being similar to the current desert state in the region, and with wet phases, known as African Humid Periods (AHPs), characterised by a strong summer monsoon which can carry enough moisture inland to support rivers, lakes and lush vegetation further north than seen today. Recent sediment records from the Mediterranean Sea revealed that the previous five AHPs had different intensities, in relation to rainfall and vegetation extent. Motivated by these findings, our work focuses on explaining what caused such differences in intensity. To this end, we use the CLIMBER-2 climate model to study the AHP response to changes in three drivers of atmospheric dynamics: Earth's orbit variations, atmospheric concentration of CO2 and inland ice extent. Global transient simulations of the last 190,000 years are used in new factorisation analyses, which allow us to separate the individual contributions of the forcings to the AHP intensity, as well as those of their synergies. We confirm the predominant role of the orbital forcing in the strength of the last five AHPs, and our simulations agree with previous estimates of a threshold in orbital forcing above which an AHP develops. Moreover, we show that atmospheric CO2 and the extent of ice sheets can also add up to be as important as the orbital parameters. High values of CO2, past a 205 ppm threshold, and low values of ice sheets extent, below an 8 % of global land surface threshold, yield the AHPs with the most precipitation and vegetation. Additionally, our results show that AHPs differ not only in amplitude, but also in their speed of change, and we find that the non-linear vegetation response of AHPs does not correlate with a single forcing and that the vegetation growth response is faster than its subsequent decline. In regards to future change, an extension of the simulations until the next 50,000 years, shows CO2 to be the main driver of AHPs, with orbital forcing only setting the pace and their intensities being scenario-dependent.

Permafrost ponds are a steady source of methane. However, it is difficult to assess the sensitivity of pond methane emissions to ongoing warming and climate-change-induced drainage, because pond methane emissions show large temporal and spatial variability already on local scale.We study this sensitivity on the landscape level with a new process-based model for Methane Emissions from Ponds (MeEP model), which simulates the three main pathways of methane emissions (diffusion, plant-mediated transport and ebullition) as well as the temperature profile of the water column and the surrounding soils. The model was set up for the polygonal tundra in the Lena River Delta. Due to a temporal resolution of one hour, it is capable of capturing the diurnal, day-to-day and seasonal variability in methane fluxes. MeEP also considers one of the main drivers of spatial variability - ground heterogeneity. Depending on where ponds form in the polygonal tundra, they can be classified as ice-wedge, polygonal-centre or merged-polygonal ponds. In MeEP, each of these pond types is simulated separately and the representation of these ponds was informed by dedicated measurements.The model performance is validated against eddy-covariance measurements of methane fluxes and against in-situ measurements of the aqueous methane concentration, both obtained on Samoylov Island. We will present results regarding the sensitivity of modeled methane emissions from ponds to warming and drainage on the landscape scale.