Abrupt

We'll break it to you gently: abrupt derives from abruptus, the past participle of the Latin verb abrumpere, meaning "to break off." Abrumpere combines the prefix ab- with rumpere, which means "to break" and which forms the basis for several other words in English that suggest a kind of breaking, such as interrupt, rupture, and bankrupt. Whether being used to describe a style of speaking that seems rudely short (as in "gave an abrupt answer"), something with a severe rise or drop ("abrupt temperature change"), or something that seems rash and unprecipitated ("made the abrupt decision to quit college"), abrupt, which first appeared in English in the 16th century, implies a kind of jarring unexpectedness that catches people off guard.

Abrupt

Thermokarst, the most widespread form of abrupt permafrost thaw13, occurs when soil warming melts ground ice, causing land surface collapse14,15. Water pooling in collapsed areas leads to formation of taliks (unfrozen thaw bulbs) beneath expanding lakes, accelerating permafrost thaw far faster and deeper than would be predicted from changes in air temperature alone16,17,18,19. Remote sensing and field observations reveal that localized abrupt thaw features, including thermokarst lakes, thermo-erosional gullies, thaw slumps, and peat-plateau collapse scars, are extensive across northern landscapes with ice-rich permafrost13. Despite two decades of observations showing that thermokarst lakes20,21 and other abrupt thaw features22 are hotspots of 14C-depleted permafrost-derived CH4 emissions (Fig. 1), the impact of abrupt thaw on state-of-the-art land-model predicted PCF6,7,8,9,10 remains unknown.

Here we re-analyzed model output from an earth system model, the Community Land Model version 4.5 (CLM4.5BGC)7, to identify the carbon-emissions fraction originating from gradual permafrost thaw on land, and to quantify the increase to 21st-century circumpolar permafrost-carbon emissions by including warming-induced abrupt thaw beneath thermokarst lakes. Our abrupt thaw emissions are derived from a vertically- and latitudinally-resolved box model of permafrost carbon processes, including a novel component conceptually describing circum-Arctic thermokarst-lake dynamics and related carbon release23, which we refer to here as the Abrupt Thaw (AThaw) model (Supplementary Table 1).

Whether the warming Arctic will become wetter or drier will impact future PCF strength according to abrupt thaw lake abundance43 and gradual-thaw CH4/CO2 emission ratios46. However, state-of-the-art CMIP5 models consistently predict an increase in precipitation relative to evapotranspiration in the Arctic, especially in summer51, favoring hydrological conditions for enhanced thermokarst-lake development43. Many newly formed lakes will ultimately be subject to drainage23,43,44 when they intersect topographical drainage gradients by lateral expansion52, from elevated water levels53, or when taliks penetrate permafrost, allowing the potential for internal drainage to the groundwater system54 (Supplementary Figs 4 and 5). While AThaw does not project fluxes in drained lake basins, we consider the implications of lake drainage on landscape-scale fluxes. Present day areal-based carbon fluxes in drained lake basins are one to three orders of magnitude lower than abrupt thaw lake emissions due to refreezing of taliks and colonization of drained basins by plants, whose CO2 uptake offsets emissions55,56 (Supplementary Table 5). It is conceivable that this difference could be smaller by the end of the century, particularly for RCP8.5, when temperatures are warm enough to prevent refreezing of taliks following lake drainage49. Methanotrophy57 will offset emissions of CH4 produced in drained-lake-basin taliks. However, ecosystem-scale microbial studies show a higher temperature response by methanogenesis than by methanotrophy or by CO2 fluxes attributable to respiration and photosynthesis58,59. This indicates that in a warmer world, CH4 emissions and the ratio of CH4 to CO2 emissions from individual ecosystems will increase59,60. This also implies that our estimate of AThaw contributions to late-century CPCRE is conservative, particularly for RCP8.5, and would be higher if fluxes in drained lake basins were also taken into account.

AThaw is a conceptual model which projects carbon release from abrupt thaw by accounting for the full chain of processes from formation of new thermokarst lakes under global warming and talik deepening in sub-lake sediments to eventual carbon release to the atmosphere following anaerobic microbial degradation of organic matter and CH4 oxidation. AThaw also accounts for lake drainage; although, carbon fluxes in drained lake basins are not modeled24,63,64,65. AThaw is incorporated into a multi-box permafrost-carbon release model which allocates soil organic matter into latitudinally and vertically gridded boxes of differing conditions regarding soil physics, carbon quantity and quality, and biogeochemistry23.

The formation/expansion of new thermokarst-lake areas and subsequent sub-lake talik growth is not the only mechanism of abrupt permafrost degradation making newly thawed permafrost carbon available for microbial decomposition. Further contribution comes from talik growth of present-day lakes which have not yet formed a deep talik and still store large amounts of labile carbon in sub-lake sediments. For instance, on the Alaska North Slope, Arp et al.17 demonstrated that a rapid decrease in lake-ice thickness and duration has already led to many shallow lakes transitioning from bed-ice fast lakes underlain by permafrost to floating ice lakes that have started to develop taliks. This talik formation takes place about 70 years before talik formation is projected for the adjacent terrestrial environment by top-down permafrost models in this cold continuous permafrost zone, a feedback process also not accounted for in AThaw or other models of permafrost degradation. These and other non-lake modes of abrupt thaw13, including coastal and river erosion, thermoerosional gully formation, thaw slumps, collapse of permafrost peat plateaus, and talik formation in upland terrestrial environments, which may occur sooner than predicted by large-scale models when finer resolution soil and vegetation properties are taken into account49, are not explicitly accounted for in the AThaw model description; hence, we consider our assumed FTKLmax values a conservative estimate of the abrupt thaw extent in the 21st century.

To avoid double counting CLM4.5BGC emissions from land areas that become thermokarst lakes in AThaw, and to account only for the increase in emissions and CPCRE caused by abrupt thaw, we have subtracted from the AThaw emissions and CPCRE, in our total permafrost landscape calculations, the quantity of carbon emissions and associated radiative forcing already assumed to be emitted from those land surfaces from CLM4.5BGC. We used the AThaw thermokarst-lake area fraction at each time step for each of the four soil classes (Supplementary Fig. 4) and weighted those fractions by the areal extent of the soil classes according to Hugelius et al.80 and Strauss et al.81 based on the implicit assumption of homogenous soil carbon distribution within each of the soil classes. Our calculations consider that this fraction of the land surface, subject to gradual thaw in CLM4.5BGC, undergoes abrupt thaw instead.

We calculated the radiative forcing due to atmospheric perturbations in CH4 and CO2 concentration for CLM and AThaw permafrost-soil-carbon flux trajectories following Frolking & Roulet87 (Supplementary Methods). Methodology underlying our remote-sensing based quantification of abrupt thaw in Alaska and field-based estimates of carbon emissions from abruptly-formed thermokarst areas of lakes in Alaska and Siberia are also provided in Supplementary Methods. Field and lab measurements include bubble-trap observations of ebullition fluxes, aerial and ground-based ebullition seep-mapping, and quantification of CH4 and CO2 concentrations and radiocarbon dating.

To test differences between gradual thaw and abrupt thaw net ecosystem exchange (Fig. 2b) we used the two-sided Kolmogorov-Smirnov test. We used this test also to compare radiocarbon ages of permafrost soil carbon respiration in gradual thaw versus abrupt thaw environments (Fig. 2c) Radiocarbon statistical analysis was performed on percent modern carbon data. All statistical analyses were performed in R88.

Quantifying the abruptness of shifts: (a) Time series of April temperature at one grid point in MPI-ESM-LR. The black line shows the original data, the blue line is the smoothed data (in space and time), and the red vertical line indicates the year with the largest time gradient in the smoothed data. (b) Calculation of abruptness based on two data chunks (black and red lines) around the transition shown in (a). Straight lines show the linear regression lines for both separate chunks and their extrapolation up to the transition point (dotted part of the lines). The blue arrow indicates the difference between the intercepts; black and red arrows indicate standard deviations σ of both data chunks. Abruptness is calculated by using Eq. (2) and the three quantities given in (b). 041b061a72