skip to main content

Title: Detecting the permafrost carbon feedback: talik formation and increased cold-season respiration as precursors to sink-to-source transitions

Abstract. Thaw and release of permafrost carbon (C) due to climate change is likely tooffset increased vegetation C uptake in northern high-latitude (NHL)terrestrial ecosystems. Models project that this permafrost C feedback mayact as a slow leak, in which case detection and attribution of the feedbackmay be difficult. The formation of talik, a subsurface layer of perenniallythawed soil, can accelerate permafrost degradation and soil respiration,ultimately shifting the C balance of permafrost-affected ecosystems fromlong-term C sinks to long-term C sources. It is imperative to understand andcharacterize mechanistic links between talik, permafrost thaw, andrespiration of deep soil C to detect and quantify the permafrost C feedback.Here, we use the Community Land Model (CLM) version 4.5, a permafrost andbiogeochemistry model, in comparison to long-term deep borehole data alongNorth American and Siberian transects, to investigate thaw-driven C sourcesin NHL (>55N) from 2000 to 2300. Widespread talik at depth isprojected across most of the NHL permafrost region(14million km2) by 2300, 6.2million km2 of which isprojected to become a long-term C source, emitting 10Pg C by 2100,50Pg C by 2200, and 120Pg C by 2300, with few signs ofslowing. Roughly half of the projected C source region is in predominantlywarm sub-Arctic permafrost following talik onset. This region emits only20Pg C by 2300, but more » the CLM4.5 estimate may be biased low by notaccounting for deep C in yedoma. Accelerated decomposition of deep soilC following talik onset shifts the ecosystem C balance away from surfacedominant processes (photosynthesis and litter respiration), butsink-to-source transition dates are delayed by 20–200 years by highecosystem productivity, such that talik peaks early (2050s, although boreholedata suggest sooner) and C source transition peaks late(2150–2200). The remaining C source region in cold northern Arcticpermafrost, which shifts to a net source early (late 21st century), emits5 times more C (95Pg C) by 2300, and prior to talik formation dueto the high decomposition rates of shallow, young C in organic-rich soilscoupled with low productivity. Our results provide important clues signalingimminent talik onset and C source transition, including (1) late cold-season(January–February) soil warming at depth (2m),(2) increasing cold-season emissions (November–April), and (3) enhancedrespiration of deep, old C in warm permafrost and young, shallow C in organic-rich cold permafrost soils. Our results suggest a mosaic of processes thatgovern carbon source-to-sink transitions at high latitudes and emphasize theurgency of monitoring soil thermal profiles, organic C age and content, cold-season CO2 emissions, andatmospheric 14CO2 as key indicatorsof the permafrost C feedback.

« less
Authors:
; ; ; ;
Award ID(s):
1636476
Publication Date:
NSF-PAR ID:
10085002
Journal Name:
The Cryosphere
Volume:
12
Issue:
1
Page Range or eLocation-ID:
123 to 144
ISSN:
1994-0424
Sponsoring Org:
National Science Foundation
More Like this
  1. Abstract. Ecohydrological models are powerful tools to quantify the effects that independent fluxes may have on catchment storage dynamics. Here, we adapted the tracer-aided ecohydrological model, EcH2O-iso, for cold regions with the explicit conceptualization of dynamic soil freeze–thaw processes. We tested the model at the data-rich Krycklan site in northern Sweden with multi-criterion calibration using discharge, stream isotopes and soil moisture in three nested catchments. We utilized the model's incorporation of ecohydrological partitioning to evaluate the effect of soil frost on evaporation and transpiration water ages, and thereby the age of source waters. The simulation of stream discharge, isotopes, and soil moisture variability captured the seasonal dynamics at all three stream sites and both soil sites, with notable reductions in discharge and soil moisture during the winter months due to the development of the frost front. Stream isotope simulations reproduced the response to the isotopically depleted pulse of spring snowmelt. The soil frost dynamics adequately captured the spatial differences in the freezing front throughout the winter period, despite no direct calibration of soil frost to measured soil temperature. The simulated soil frost indicated a maximum freeze depth of 0.25 m below forest vegetation. Water ages of evaporation and transpiration reflect themore »influence of snowmelt inputs, with a high proclivity of old water (pre-winter storage) at the beginning of the growing season and a mix of snowmelt and precipitation (young water) toward the end of the summer. Soil frost had an early season influence of the transpiration water ages, with water pre-dating the snowpack mainly sustaining vegetation at the start of the growing season. Given the long-term expected change in the energy balance of northern climates, the approach presented provides a framework for quantifying the interactions of ecohydrological fluxes and waters stored in the soil and understanding how these may be impacted in future.

    « less
  2. Abstract. Airborne and ground-based measurements of aerosol concentrations, chemicalcomposition, and gas-phase precursors were obtained in three valleys innorthern Utah (USA). The measurements were part of the Utah Winter FineParticulate Study (UWFPS) that took place in January–February 2017. Totalaerosol mass concentrations of PM1 were measured from a Twin Otteraircraft, with an aerosol mass spectrometer (AMS). PM1 concentrationsranged from less than 2µgm−3 during clean periods to over100µgm−3 during the most polluted episodes, consistent withPM2.5 total mass concentrations measured concurrently at groundsites. Across the entire region, increases in total aerosol mass above∼2µgm−3 were associated with increases in theammonium nitrate mass fraction, clearly indicating that the highest aerosolmass loadings in the region were predominantly attributable to an increase inammonium nitrate. The chemical composition was regionally homogenous fortotal aerosol mass concentrations above 17.5µgm−3, with 74±5% (average±standard deviation) ammonium nitrate, 18±3%organic material, 6±3% ammonium sulfate, and 2±2%ammonium chloride. Vertical profiles of aerosol mass and volume in the regionshowed variable concentrations with height in the polluted boundary layer.Higher average mass concentrations were observed within themore »first few hundredmeters above ground level in all three valleys during pollution episodes. Gas-phase measurements of nitric acid (HNO3) and ammonia (NH3) duringthe pollution episodes revealed that in the Cache and Utah valleys, partitioningof inorganic semi-volatiles to the aerosol phase was usually limited by theamount of gas-phase nitric acid, with NH3 being in excess. The inorganicspecies were compared with the ISORROPIA thermodynamic model. Total inorganicaerosol mass concentrations were calculated for various decreases in totalnitrate and total ammonium. For pollution episodes, our simulations of a50% decrease in total nitrate lead to a 46±3% decrease in totalPM1 mass. A simulated 50% decrease in total ammonium leads to a36±17%µgm−3 decrease in total PM1 mass, over the entirearea of the study. Despite some differences among locations, ourresults showed a higher sensitivity to decreasing nitric acid concentrationsand the importance of ammonia at the lowest total nitrate conditions. In theSalt Lake Valley, both HNO3 and NH3 concentrations controlledaerosol formation.

    « less
  3. Abstract. A large array of proxy recordssuggests that the “4.2ka event” marks an approximately300-year long period (∼3.9 to 4.2ka) ofmajor climate change across the globe. However, the climatic manifestation ofthis event, including its onset, duration, and termination, remains lessclear in the Indian summer monsoon (ISM) domain. Here, we present new oxygenisotope (δ18O) data from a pair of speleothems (ML.1 and ML.2)from Mawmluh Cave, Meghalaya, India, that provide a high-resolution record ofISM variability during a period (∼3.78 and 4.44ka) that fullyencompasses the 4.2ka event. The sub-annually to annually resolved ML.1δ18O record is constrained by 18 230Th dates with anaverage dating error of ±13 years (2σ) and a resolution of ∼40 years, which allows us to characterize the ISM variability withunprecedented detail. The inferred pattern of ISM variability during theperiod contemporaneous with the 4.2ka event shares broad similarities andkey differences with the previous reconstructions of ISM from the MawmluhCave and other proxy records from the region. Our data suggest that the ISMintensity, in the context of the length of our record, abruptly decreased at∼4.0ka (±13more »xlink:href="cp-14-1869-2018-ie00001.svg" width="32pt" height="10pt" src="cp-14-1869-2018-ie00001.png"/> years), marking the onset of a multi-centennialperiod of relatively reduced ISM, which was punctuated by at least twomulti-decadal droughts between ∼3.9 and 4.0ka. The latter stands outin contrast with some previous proxy reconstructions of the ISM, in which the4.2ka event has been depicted as a singular multi-centennial drought.

    « less
  4. Abstract. Urbanization and deforestation have important impacts on atmosphericparticulate matter (PM) over Amazonia. This study presents observations andanalysis of PM1 concentration, composition, and opticalproperties in central Amazonia during the dry season, focusing on theanthropogenic impacts. The primary study site was located 70 km downwind ofManaus, a city of over 2 million people in Brazil, as part of theGoAmazon2014/5 experiment. A high-resolution time-of-flight aerosol massspectrometer (AMS) provided data on PM1 composition, and aethalometermeasurements were used to derive the absorption coefficient babs,BrC ofbrown carbon (BrC) at 370 nm. Non-refractory PM1 mass concentrationsaveraged 12.2 µg m−3 at the primary study site, dominated byorganics (83 %), followed by sulfate (11 %). A decrease inbabs,BrC was observed as the mass concentration of nitrogen-containingorganic compounds decreased and the organic PM1 O:C ratio increased,suggesting atmospheric bleaching of the BrC components. The organic PM1was separated into six different classes by positive-matrix factorization(PMF), and the mass absorption efficiency Eabs associated with eachfactor was estimated through multivariate linear regression ofbabs,BrC on the factor loadings. The largest Eabs values wereassociated with urban (2.04±0.14 m2 g−1) and biomass-burning(0.82±0.04 to 1.50±0.07 m2 g−1) sources. Together, these sources contributed at least 80 % ofbabs,BrCmore »while accounting for 30 % to 40 % of the organic PM1 massconcentration. In addition, a comparison of organic PM1 compositionbetween wet and dry seasons revealed that only part of the 9-foldincrease in mass concentration between the seasons can be attributed tobiomass burning. Biomass-burning factor loadings increased by 30-fold,elevating its relative contribution to organic PM1 from about 10 % inthe wet season to 30 % in the dry season. However, most of the PM1mass (>60 %) in both seasons was accounted for by biogenicsecondary organic sources, which in turn showed an 8-fold seasonalincrease in factor loadings. A combination of decreased wet deposition andincreased emissions and oxidant concentrations, as well as a positivefeedback on larger mass concentrations are thought to play a role in theobserved increases. Furthermore, fuzzy c-means clustering identified threeclusters, namely “baseline”, “event”, and “urban” to representdifferent pollution influences during the dry season. The baseline cluster,representing the dry season background, was associated with a mean massconcentration of 9±3 µg m−3. This concentration increasedon average by 3 µg m−3 for both the urban and the event clusters.The event cluster, representing an increased influence of biomass burningand long-range transport of African volcanic emissions, was characterized byremarkably high sulfate concentrations. The urban cluster, representing theinfluence of Manaus emissions on top of the baseline, was characterized byan organic PM1 composition that differed from the other two clusters.The differences discussed suggest a shift in oxidation pathways as well asan accelerated oxidation cycle due to urban emissions, in agreement withfindings for the wet season.

    « less
  5. Abstract

    Thermokarst lakes accelerate deep permafrost thaw and the mobilization of previously frozen soil organic carbon. This leads to microbial decomposition and large releases of carbon dioxide (CO2) and methane (CH4) that enhance climate warming. However, the time scale of permafrost-carbon emissions following thaw is not well known but is important for understanding how abrupt permafrost thaw impacts climate feedback. We combined field measurements and radiocarbon dating of CH4ebullition with (a) an assessment of lake area changes delineated from high-resolution (1–2.5 m) optical imagery and (b) geophysical measurements of thaw bulbs (taliks) to determine the spatiotemporal dynamics of hotspot-seep CH4ebullition in interior Alaska thermokarst lakes. Hotspot seeps are characterized as point-sources of high ebullition that release14C-depleted CH4from deep (up to tens of meters) within lake thaw bulbs year-round. Thermokarst lakes, initiated by a variety of factors, doubled in number and increased 37.5% in area from 1949 to 2009 as climate warmed. Approximately 80% of contemporary CH4hotspot seeps were associated with this recent thermokarst activity, occurring where 60 years of abrupt thaw took place as a result of new and expanded lake areas. Hotspot occurrence diminished with distance from thermokarst lake margins. We attribute older14C ages of CH4released from hotspot seepsmore »in older, expanding thermokarst lakes (14CCH420 079 ± 1227 years BP, mean ± standard error (s.e.m.) years) to deeper taliks (thaw bulbs) compared to younger14CCH4in new lakes (14CCH48526 ± 741 years BP) with shallower taliks. We find that smaller, non-hotspot ebullition seeps have younger14C ages (expanding lakes 7473 ± 1762 years; new lakes 4742 ± 803 years) and that their emissions span a larger historic range. These observations provide a first-order constraint on the magnitude and decadal-scale duration of CH4-hotspot seep emissions following formation of thermokarst lakes as climate warms.

    « less