Datasets:

---

LLMDH

/

OpenScience

like 0

Follow

LLM - Digital Humanities 15

ERROR: type should be string, got "https://doi.org/10.5194/bg-2021-338\nPreprint. Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. Soil carbon loss in warmed subarctic grasslands is rapid and\nrestricted to topsoil Niel Verbrigghe1, Niki I. W. Leblans1,2, Bjarni D. Sigurdsson3, Sara Vicca1, Chao Fang1,4,5,\nLucia Fuchslueger1,6, Jennifer L. Soong1,7, James T. Weedon8, Christopher Poeplau9,\nCristina Ariza-Carricondo1, Michael Bahn10, Bertrand Guenet11, Per Gundersen12, Gunnhildur E. Gunnarsdóttir13, Thomas Kätterer14, Zhanfeng Liu15, Marja Maljanen16, Sara Marañón-Jiménez17,18,\nKathiravan Meeran10, Edda S. Oddsdóttir19, Ivika Ostonen20, Josep Peñuelas17,18, Andreas Richter6,21,\nJordi Sardans17,18, Páll Sigurðsson3, Margaret S. Torn22, Peter M. Van Bodegom23, Erik Verbruggen1,\nTom W. N. Walker24, Håkan Wallander25, and Ivan A. Janssens1 Niel Verbrigghe1, Niki I. W. Leblans1,2, Bjarni D. Sigurdsson3, Sara Vicca1, Chao Fang1,4,5, 1Research Group Plants and Ecosystems, University of Antwerp, Antwerp, Belgium. 2 2Climate Impacts Research Centre, Umeå University, Umeå, Sweden. 3Agricultural University of Iceland, Hvanneyri, Borgarnes, Iceland. 4Institute of Ecology, School of Applied Meteorology, Nanjing University of Information Science and Technology, Nanjing,\nChina. 4Institute of Ecology, School of Applied Meteorology, Nanjing University of Information Science and Technology, Nanjing,\nChina. 5State Key Laboratory of Grassland Agro-ecosystems, Institute of Arid Agroecology, School of Life Sciences, Lanzhou 5State Key Laboratory of Grassland Agro-ecosystems, Institute of Arid Agroecology, School of Life Sciences, Lanzhou\nUniversity, Lanzhou, China 5State Key Laboratory of Grassland Agro-ecosystems, Institute of Arid Agroecology, School of Life Sciences, Lanzhou\nUniversity, Lanzhou, China. y\n6Centre for Microbiology and Environmental Systems Science, University of Vienna, Vienna, Austria 6Centre for Microbiology and Environmental Systems Science, University of Vienna, Vienna, Au 7Soil and Crop Sciences Department, Colorado State University, Fort Collins, Colorado, USA. p\np\ny\n8Systems Ecology, Department of Ecological Science, Vrije Universiteit Amsterdam, Amsterdam, The Netherlands. 9 Department of Ecological Science, Vrije Universiteit Amsterdam, Amsterdam, The Netherlands. 8Systems Ecology, Department of Ecological Science, Vrije Universiteit Amsterdam, Amsterdam 9Thünen Institute of Climate-Smart Agriculture, Braunschweig, Germany 10Department of Ecology, University of Innsbruck, Innsbruck, Austria. 11Laboratoire de Géologie, École normale supérieure/CNRS, PSL Research University, Paris, France. Laboratoire de Géologie, École normale supérieure/CNRS, PSL Research University, Paris, France 12Department of Geosciences and Natural Resource Management, University of Copenhagen, Frederiksberg C, Denmark. 13Soil Conservation Service of Iceland, Gunnarsholt, Hella, Iceland. 14 Department of Geosciences and Natural Resource Management, University of Copenhagen, Frederi Department of Geosciences and Natural Resource Management, University of Copenhagen, Frederiksberg C, Denmark. 13Soil Conservation Service of Iceland, Gunnarsholt, Hella, Iceland. p\ng\ny\np\ng\ng\n13Soil Conservation Service of Iceland, Gunnarsholt, Hella, Iceland. 13Soil Conservation Service of Iceland, Gunnarsholt, Hella, Iceland. Department of Ecology, Swedish University of Agricultural Sciences, Uppsala, Sweden. Correspondence: Niel Verbrigghe (Niel.Verbrigghe@UAntwerpen.be) Soil carbon loss in warmed subarctic grasslands is rapid and\nrestricted to topsoil This rapid equilibration of SOC\nobserved in Andosol suggests a critical role for ecosystem adaptations to warming and could imply short-lived soil carbon-\nclimate feedbacks. Our data further revealed that the soil C loss occurred in all aggregate size fractions, and that SOC losses\nonly occurred in topsoil (0-10 cm). SOC stocks in subsoil (10-30 cm), where plant roots were absent, remained unaltered, even after >50 years of warming. The observed depth-dependent warming responses indicate that explicit vertical resolution is a\n10\nprerequisite for global models to accurately project future SOC stocks for this soil type and should be investigated for soils\nwith other mineralogies. after >50 years of warming. The observed depth-dependent warming responses indicate that explicit vertical resolution is a\n10\nprerequisite for global models to accurately project future SOC stocks for this soil type and should be investigated for soils\nwith other mineralogies. Soil carbon loss in warmed subarctic grasslands is rapid and\nrestricted to topsoil 14Department of Ecology, Swedish University of Agricultural Sciences, Uppsala, Sweden. 15Key Laboratory of Vegetation Restoration and Management of Degraded Ecosystems & CAS Engineering Laboratory for\nVegetation Ecosystem Restoration on Islands and Coastal Zones, South China Botanical Garden, Chinese Academy of\nSciences, Guangzhou, China. partment of Environmental and Biological Sciences, University of Eastern Finland, Kuopio, Finland 17CREAF, Cerdanyola del Vallès, Barcelona, Catalonia, Spain. 18CSIC, Global Ecology Unit CREAF-CSIC-UAB, Bellaterra, Barcelona, Spain. 19Icelandic Forest Research, Mógilsá, Reykjavík, Iceland. 20Institute of Ecology and Earth Sciences, University of Tartu, Tartu, Estonia 21International Institute for Applied Systems Analysis (IIASA), Laxenburg, Austria. 22Climate and Ecosystem Sciences Division, Berkeley Lab, Berkeley, CA, USA. 23Environmental Biology Department, Institute of Environmental Sciences, CML, Leiden University, Leiden, The\nNetherlands. 23Environmental Biology Department, Institute of Environmental Sciences, CML, Leiden University, Leiden, The\nNetherlands. 24Department of Environmental Systems Science, ETH Zürich, Zürich, Switzerland. 25MEMEG, Department of Biology, Lund University, Lund, Sweden. Correspondence: Niel Verbrigghe (Niel.Verbrigghe@UAntwerpen.be) Correspondence: Niel Verbrigghe (Niel.Verbrigghe@UAntwerpen.be) 1 1 https://doi.org/10.5194/bg-2021-338\nPreprint. Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. Abstract. Global warming may lead to carbon transfers from soils to the atmosphere, yet this positive feedback to the cli-\nmate system remains highly uncertain, especially in subsoils (Ilyina and Friedlingstein, 2016; Shi et al., 2018). Using natural\ngeothermal soil warming gradients of up to +6.4 ◦C in subarctic grasslands (Sigurdsson et al., 2016), we show that soil organic\ncarbon (SOC) stocks decline strongly and linearly with warming (−2.8 ton ha−1 ◦C−1). Comparison of SOC stock changes following medium-term (5 and 10 years) and long-term (>50 years) warming revealed that all SOC loss occurred within the\n5\nfirst five years of warming, after which continued warming no longer reduced SOC stocks. This rapid equilibration of SOC\nobserved in Andosol suggests a critical role for ecosystem adaptations to warming and could imply short-lived soil carbon-\nclimate feedbacks. Our data further revealed that the soil C loss occurred in all aggregate size fractions, and that SOC losses\nonly occurred in topsoil (0-10 cm). SOC stocks in subsoil (10-30 cm), where plant roots were absent, remained unaltered, even following medium-term (5 and 10 years) and long-term (>50 years) warming revealed that all SOC loss occurred within the\n5\nfirst five years of warming, after which continued warming no longer reduced SOC stocks. 1\nIntroduction Soils store more carbon (C) than the atmosphere and vegetation biomass combined (Batjes, 2 Soils store more carbon (C) than the atmosphere and vegetation biomass combined (Batjes, 2016; Scharlemann et al., 2014). Global warming has been hypothesised to lead to increased soil CO2 emissions that may lead to large reductions in soil organic\n15\ncarbon (SOC) stocks, constituting a positive feedback to the climate system (Davidson and Janssens, 2006; Jenkinson et al.,\n1991). The strength and even sign of this carbon cycle-climate feedback are, however, highly uncertain (Crowther et al., 2016;\nTodd-Brown et al., 2018; van Gestel et al., 2018). Accordingly, the World Climate Research Programme has acknowledged it\nas one of the \"Grand Challenges\" of climate research (Ilyina and Friedlingstein, 2016). 15 In situ soil warming studies provide ideal tools to study the response of soil SOC stocks to warming (Batjes, 2016), yet\n20\nchallenges remain great. First, ecosystem responses to warming may take decades to stabilise (Walker et al., 2020; Melillo\net al., 2017), implying that extrapolations of responses from-, or model parametrisation based on short-term experiments\nmay lead to erroneous estimation of the future evolution of SOC stocks. Second, SOC stock changes are rarely studied in\nsubsoils. The high cost and labour requirements of SOC research, combined with the fact that most biological activity and In situ soil warming studies provide ideal tools to study the response of soil SOC stocks to warming (Batjes, 2016), yet\n20\nchallenges remain great. First, ecosystem responses to warming may take decades to stabilise (Walker et al., 2020; Melillo\net al., 2017), implying that extrapolations of responses from-, or model parametrisation based on short-term experiments\nmay lead to erroneous estimation of the future evolution of SOC stocks. Second, SOC stock changes are rarely studied in\nsubsoils. The high cost and labour requirements of SOC research, combined with the fact that most biological activity and In situ soil warming studies provide ideal tools to study the response of soil SOC stocks to warming (Batjes, 2016), yet\n20\nchallenges remain great. First, ecosystem responses to warming may take decades to stabilise (Walker et al., 2020; Melillo\net al., 2017), implying that extrapolations of responses from-, or model parametrisation based on short-term experiments\nmay lead to erroneous estimation of the future evolution of SOC stocks. Second, SOC stock changes are rarely studied in\nsubsoils. 1\nIntroduction The high cost and labour requirements of SOC research, combined with the fact that most biological activity and SOC mineralisation occur in topsoils, explains why soil biology and ecology, including SOC cycling, are rarely studied below\n25\na depth of 20-30 cm (Yost and Hartemink, 2020). This is a major issue, because more SOC is stored below this threshold\nthan above (Shi et al., 2020) and therefore the carbon cycle-climate feedback does not stop at 20-30 cm depth. Unfortunately,\nthe very few soil warming experiments that also warmed subsoils and quantified SOC stock changes yielded very different\nwarming responses, ranging from declining to increasing subsoil SOC stocks (Soong et al., 2020a; Hanson et al., 2020). SOC mineralisation occur in topsoils, explains why soil biology and ecology, including SOC cycling, are rarely studied below\n25\na depth of 20-30 cm (Yost and Hartemink, 2020). This is a major issue, because more SOC is stored below this threshold\nthan above (Shi et al., 2020) and therefore the carbon cycle-climate feedback does not stop at 20-30 cm depth. Unfortunately,\nthe very few soil warming experiments that also warmed subsoils and quantified SOC stock changes yielded very different\nwarming responses, ranging from declining to increasing subsoil SOC stocks (Soong et al., 2020a; Hanson et al., 2020). SOC mineralisation occur in topsoils, explains why soil biology and ecology, including SOC cycling, are rarely studied below\n25\na depth of 20-30 cm (Yost and Hartemink, 2020). This is a major issue, because more SOC is stored below this threshold\nthan above (Shi et al., 2020) and therefore the carbon cycle-climate feedback does not stop at 20-30 cm depth. Unfortunately,\nthe very few soil warming experiments that also warmed subsoils and quantified SOC stock changes yielded very different\nwarming responses, ranging from declining to increasing subsoil SOC stocks (Soong et al., 2020a; Hanson et al., 2020). To address both these challenges, we determined SOC stock changes along natural geothermal gradients at the ForHot\n30\nresearch site in Iceland (Sigurdsson et al., 2016) encompassing the full warming range projected for Northern regions (up\nto +6.4 ◦C), throughout the topsoil (0-10 cm) and the subsoil (10-30 cm). 1\nIntroduction We further hypothesised similar subsoil and topsoil SOC loss, given that subsoils were exposed to the same\nwarming intensity and duration as topsoils. soils would still be losing SOC over time, while the long-term warmed soils would have reached a new equilibrium at lower\n45\nSOC content. We further hypothesised similar subsoil and topsoil SOC loss, given that subsoils were exposed to the same\nwarming intensity and duration as topsoils. 1\nIntroduction 35\nNext to measuring SOC stocks, we gathered data about soil aggregates, carbon inputs to the soil by plants and arbuscular\nmycorrhizal fungi and carbon flux from topsoil to subsoil. This allowed us to elaborate about the possible mechanisms behind\nSOC stock changes along the warming gradient. The recently warmed grassland we investigated at the ForHot site has been warmed since 2008 when a major earthquake The recently warmed grassland we investigated at the ForHot site has been warmed since 2 The recently warmed grassland we investigated at the ForHot site has been warmed since 2008, when a major earthquake\nshifted geothermal systems to previously unwarmed soils, causing increased temperature in the soil above by radiative heating\n40\n(Halldórsson and Sigbjörnsson, 2009; O’Gorman et al., 2014). In contrast, the long-term warmed grassland had been warmed\nfor at least 45 years at the time of the earthquake in 2008 (Sigurdsson et al., 2016). The soil type on both study sites is Andosol,\nand they are covered by the same grassland type (Sigurdsson et al., 2016). W h\nth\ni\nd th t b\nf th\nl\nti\nf\nt\nt d SOC\nl t\nt\nt\nh\ndi\nt\nd shifted geothermal systems to previously unwarmed soils, causing increased temperature in the soil above by radiative heating\n40\n(Halldórsson and Sigbjörnsson, 2009; O’Gorman et al., 2014). In contrast, the long-term warmed grassland had been warmed\nfor at least 45 years at the time of the earthquake in 2008 (Sigurdsson et al., 2016). The soil type on both study sites is Andosol,\nand they are covered by the same grassland type (Sigurdsson et al., 2016). We hypothesised that because of the slow reaction of protected SOC pools to temperature change, medium-term warmed We hypothesised that because of the slow reaction of protected SOC pools to temperature change, medium-term warmed\nsoils would still be losing SOC over time, while the long-term warmed soils would have reached a new equilibrium at lower\n45\nSOC content. We further hypothesised similar subsoil and topsoil SOC loss, given that subsoils were exposed to the same\nwarming intensity and duration as topsoils. soils would still be losing SOC over time, while the long-term warmed soils would have reached a new equilibrium at lower\n45\nSOC content. 1\nIntroduction We compared topsoil and subsoil SOC dynamics\nalong replicate warming gradients exposed to medium-term (5 and 10 years) and long-term (>50 years, but possibly centuries) To address both these challenges, we determined SOC stock changes along natural geothermal gradients at the ForHot\n30\nresearch site in Iceland (Sigurdsson et al., 2016) encompassing the full warming range projected for Northern regions (up\nto +6.4 ◦C), throughout the topsoil (0-10 cm) and the subsoil (10-30 cm). We compared topsoil and subsoil SOC dynamics\nalong replicate warming gradients exposed to medium-term (5 and 10 years) and long-term (>50 years, but possibly centuries) To address both these challenges, we determined SOC stock changes along natural geothermal gradients at the ForHot\n30\nresearch site in Iceland (Sigurdsson et al., 2016) encompassing the full warming range projected for Northern regions (up\nto +6.4 ◦C), throughout the topsoil (0-10 cm) and the subsoil (10-30 cm). We compared topsoil and subsoil SOC dynamics\nalong replicate warming gradients exposed to medium-term (5 and 10 years) and long-term (>50 years, but possibly centuries) 2 https://doi.org/10.5194/bg-2021-338\nPreprint. Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. warming by sampling permanent study plots twice in a six-year period (2013 and 2018). This enabled us to characterise the warming by sampling permanent study plots twice in a six-year period (2013 and 2018). This enabled us to characterise the\nmagnitude, shape and temporal dynamics of the temperature response of SOC stocks in these northern, non-permafrost, soils. 35\nNext to measuring SOC stocks, we gathered data about soil aggregates, carbon inputs to the soil by plants and arbuscular\nmycorrhizal fungi and carbon flux from topsoil to subsoil. This allowed us to elaborate about the possible mechanisms behind warming by sampling permanent study plots twice in a six-year period (2013 and 2018). This enabled us to characterise the\nmagnitude, shape and temporal dynamics of the temperature response of SOC stocks in these northern, non-permafrost, soils. 35\nNext to measuring SOC stocks, we gathered data about soil aggregates, carbon inputs to the soil by plants and arbuscular\nmycorrhizal fungi and carbon flux from topsoil to subsoil. This allowed us to elaborate about the possible mechanisms behind\nSOC stock changes along the warming gradient. magnitude, shape and temporal dynamics of the temperature response of SOC stocks in these northern, non-permafrost, soils. 2\nLarge, linear and fast topsoil SOC loss Topsoil (0-10 cm; comprising the A horizon and rooting zone) SOC stocks linearly declined by 2 Topsoil (0-10 cm; comprising the A horizon and rooting zone) SOC stocks linearly declined by 2.8 ± 0.5 (± SE) ton SOC ha−1 ◦C−1\nsoil warming (or 8.8 ± 2.1 ◦C−1; both P < 0.001) for mass-corrected SOC stocks (fig. 1a, 2). These topsoil temperature re-\n50\nsponses did not differ between the medium-term and the long-term warmed grassland, i.e., the warming:grassland interaction\nterm was not significant (P = 0.47). This clearly suggests that warming induced SOC losses only during the initial five years of\nexposure and that SOC stocks did not change thereafter. However, several sources of variation such as sampling errors and the large heterogeneity inherent to soils, induced quite soil warming (or 8.8 ± 2.1 ◦C−1; both P < 0.001) for mass-corrected SOC stocks (fig. 1a, 2). These topsoil temperature re-\n50\nsponses did not differ between the medium-term and the long-term warmed grassland, i.e., the warming:grassland interaction\nterm was not significant (P = 0.47). This clearly suggests that warming induced SOC losses only during the initial five years of\nexposure and that SOC stocks did not change thereafter. However, several sources of variation such as sampling errors and the large heterogeneity inherent to soils, induced quite broad uncertainty intervals that reduced the potential to detect statistically significant changes in SOC stocks or in their tempera-\n55\nture response. Hence, to demonstrate that the warming-induced SOC stock loss had indeed stabilised within five years of warm-\ning and did not further loose SOC, we calculated what SOC stock decline could have remained undetected given the variability\nin our samples using a one-sided 95 % confidence interval on the soil warming regression coefficient. This shows that average\nadditional SOC stock losses smaller than 0.88 ton C ha−1 ◦C−1 would not have been detected at p<0.05, implying that in the five year period following the initial warming response (i.e., 2013-2018), annual declines of up to 0.18 ton C ha−1 ◦C−1 year−1 would\n60\nhave remained undetected. In the subsequent time span of >50 years, only changes smaller than 0.018 ton C ha−1 ◦C−1 year−1\nwould have remained undetected, i.e., a rate 30-fold less than the SOC loss observed in the initial 5 years of soil warming. 2\nLarge, linear and fast topsoil SOC loss Also when not corrected for warming-induced density changes, SOC stocks and soil C concentrations declined with warming\n(fig. B1; fig. B2) and did not further decrease after five years of soil warming. In contrast to our hypothesis, our data thus\nrevealed that in topsoil, a stepwise increase in temperature caused a fast SOC stock loss that stabilised within five years of\n65 year period following the initial warming response (i.e., 2013-2018), annual declines of up to 0.18 ton C ha−1 ◦C−1 year−1 would\n60\nhave remained undetected. In the subsequent time span of >50 years, only changes smaller than 0.018 ton C ha−1 ◦C−1 year−1\nwould have remained undetected, i.e., a rate 30-fold less than the SOC loss observed in the initial 5 years of soil warming. Also when not corrected for warming induced density changes SOC stocks and soil C concentrations declined with warming 3 3 https://doi.org/10.5194/bg-2021-338\nPreprint. Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. warming, despite the sustained higher temperatures. Even grasslands that had been warmed at least 55 years exhibited no\nlarger SOC loss than that observed after 5 years of soil warming. warming, despite the sustained higher temperatures. Even grasslands that had been warmed at least 55 years exhibited no\nlarger SOC loss than that observed after 5 years of soil warming. y = −2.8x + 32.1\nR² = 0.29; P < 0.001\na)\nP = 0.63\nb)\nTopsoil\nSubsoil\n0\n1\n2\n3\n4\n5\n10\n20\n30\n40\n50\n10\n20\n30\n40\n50\nSoil warming (°C)\nSOC stock (ton C ha−1)\nYr−grassland\n2013 − LTW\n2013 − MTW\n2018 − LTW\n2018 − MTW\nFigure 1. Soil organic carbon (SOC) stocks (ton C ha−1) along soil warming gradients, a) in the topsoil (0-10 cm) and b) in the subsoil\n(10-30 cm) after a soil mass correction. The regression (solid line; dashed lines represent the 95 % confidence interval; regression details\nprovided in the inset) for medium-term warmed (MTW) and long-term warmed (LTW) grassland (in topsoil) for both 2013 and 2018 were\ncombined, since no soil temperature x warming-duration interaction effect, nor a main effect for warming-duration or sampling year was\nfound. The soil mass correction is visualised in fig. B8. The uncorrected SOC stocks yield qualitatively similar conclusions (fig. B1), as did\nthe C percentage in top- and subsoil (fig. B2). 2\nLarge, linear and fast topsoil SOC loss https://doi.org/10.5194/bg-2021-338\nPreprint. Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. https://doi.org/10.5194/bg-2021-338\nPreprint. Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. 0\n5\n10\n>50\n>55\n0−1.4°C\n1.4−2.7°C\n2.7−4.1°C\n4.1−5.5°C\nYears of warming\nSoil warming category\n15\n20\n25\n30\nSOC stocks\n(kg ha−1)\n0−1.4°C\n1.4−2.7°C\n2.7−4.1°C\n4.1−5.5°C\n10\n20\n30\n40\n50\n0\n5\n10\n>50\n>55\nYears of warming\nSOC stock (ton ha−1)\nFigure 2. Warming effect on topsoil (0-10 cm) SOC stocks, as observed during repeated sampling campaigns. Stocks after 5 and 10 years\nof warming are sampled in the medium-term warmed grasslands, stocks after >50 and >55 years in the long-term warmed grasslands. The\ndata at the start of warming is interpolated from the ambient plots in grasslands combined. Soils are divided in four warming categories for\nrepresentation. The colours on the heatmap and the smoother lines are based on a linear regression equation per sampling event. 0\n5\n10\n>50\n>55\n0−1.4°C\n1.4−2.7°C\n2.7−4.1°C\n4.1−5.5°C\nYears of warming\nSoil warming category\n15\n20\n25\n30\nSOC stocks\n(kg ha−1) Years of warming 0−1.4°C\n1.4−2.7°C\n2.7−4.1°C\n4.1−5.5°C\n10\n20\n30\n40\n50\n0\n5\n10\n>50\n>55\nYears of warming\nSOC stock (ton ha−1) Years of warming Figure 2. Warming effect on topsoil (0-10 cm) SOC stocks, as observed during repeated sampling campaigns. Stocks after 5 and 10 years\nof warming are sampled in the medium-term warmed grasslands, stocks after >50 and >55 years in the long-term warmed grasslands. The\ndata at the start of warming is interpolated from the ambient plots in grasslands combined. Soils are divided in four warming categories for\nrepresentation. The colours on the heatmap and the smoother lines are based on a linear regression equation per sampling event. Figure 2. Warming effect on topsoil (0-10 cm) SOC stocks, as observed during repeated sampling campaigns. Stocks after 5 and 10 years\nof warming are sampled in the medium-term warmed grasslands, stocks after >50 and >55 years in the long-term warmed grasslands. The\ndata at the start of warming is interpolated from the ambient plots in grasslands combined. Soils are divided in four warming categories for\nrepresentation. The colours on the heatmap and the smoother lines are based on a linear regression equation per sampling event. 2\nLarge, linear and fast topsoil SOC loss (n = 78 & 40 for topsoil and subsoil respectively) y = −2.8x + 32.1\nR² = 0.29; P < 0.001\na)\nP = 0.63\nb)\nTopsoil\nSubsoil\n0\n1\n2\n3\n4\n5\n10\n20\n30\n40\n50\n10\n20\n30\n40\n50\nSoil warming (°C)\nSOC stock (ton C ha−1)\nYr−grassland\n2013 − LTW\n2013 − MTW\n2018 − LTW\n2018 − MTW Figure 1. Soil organic carbon (SOC) stocks (ton C ha−1) along soil warming gradients, a) in the topsoil (0-10 cm) and b) in the subsoil\n(10-30 cm) after a soil mass correction. The regression (solid line; dashed lines represent the 95 % confidence interval; regression details\nprovided in the inset) for medium-term warmed (MTW) and long-term warmed (LTW) grassland (in topsoil) for both 2013 and 2018 were\ncombined, since no soil temperature x warming-duration interaction effect, nor a main effect for warming-duration or sampling year was\nfound. The soil mass correction is visualised in fig. B8. The uncorrected SOC stocks yield qualitatively similar conclusions (fig. B1), as did\nthe C percentage in top- and subsoil (fig. B2). (n = 78 & 40 for topsoil and subsoil respectively) Figure 1. Soil organic carbon (SOC) stocks (ton C ha−1) along soil warming gradients, a) in the topsoil (0-10 cm) and b) in the subsoil\n(10-30 cm) after a soil mass correction. The regression (solid line; dashed lines represent the 95 % confidence interval; regression details\nprovided in the inset) for medium-term warmed (MTW) and long-term warmed (LTW) grassland (in topsoil) for both 2013 and 2018 were\ncombined, since no soil temperature x warming-duration interaction effect, nor a main effect for warming-duration or sampling year was\nfound. The soil mass correction is visualised in fig. B8. The uncorrected SOC stocks yield qualitatively similar conclusions (fig. B1), as did\nthe C percentage in top- and subsoil (fig. B2). (n = 78 & 40 for topsoil and subsoil respectively) 4 4 0\n5\n10\n>50\n>55\n0−1.4°C\n1.4−2.7°C\n2.7−4.1°C\n4.1−5.5°C\nYears of warming\nSoil warming category\n15\n20\n25\n30\nSOC stocks\n(kg ha−1)\n0−1.4°C\n1.4−2.7°C\n2.7−4.1°C\n4.1−5.5°C\n10\n20\n30\n40\n50\n0\n5\n10\n>50\n>55\nYears of warming\nSOC stock (ton ha−1)\nFigure 2. Warming effect on topsoil (0-10 cm) SOC stocks, as observed during repeated sampling campaigns. Stocks after 5 and 10 years\nhttps://doi.org/10.5194/bg-2021-338\nPreprint. Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. 2\nLarge, linear and fast topsoil SOC loss of warming are sampled in the medium-term warmed grasslands, stocks after >50 and >55 years in the lon\ndata at the start of warming is interpolated from the ambient plots in grasslands combined. Soils are divided\nrepresentation. The colours on the heatmap and the smoother lines are based on a linear regression equation To gain insight in the warming-induced soil physical changes and their effects on SOC stocks, soils were fractionated\ninto different size classes that were analysed separately (see Methods). Aggregate fractionation showed that with increasing o ga\ns g t\nt e wa\ng\nduced so\np ys ca c a ges a d t e\ne ects o\nSOC stoc s, so s we e\nact o ated\ninto different size classes that were analysed separately (see Methods). Aggregate fractionation showed that with increasing\nwarming intensity, the mass of >2 mm fraction declined significantly, in favour of the >250 µm and >63 µm fractions. No\n70\nsignificant change was detected in the mass of the smallest (<63 µm) fraction (fig. 3). Opposed to this contrasting response of\nrelative mass, all soil fractions exhibited similar soil C % declines with soil warming (fig. 3). The relative mass increase of the\nsmaller fractions was compensated for by the soil C % decline, resulting in a stable amount of C in the >250 µm and >63 µm\nfractions. As a result, all of the warming-induced SOC stock decline we observed in the bulk soil, was attributable to C losses\nin the >2 mm fraction. 75 70 5 5 https://doi.org/10.5194/bg-2021-338\nPreprint. Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. MTW\nLTW\nRel. mass (%)\nSoil C (%)\nSoil C (g C 100 g−1 dw soil)\n0\n5\n10\n15\n20 0\n5\n10\n15\n20\n0\n20\n40\n60\n2\n4\n6\n0\n1\n2\n3\n4\nSoil warming (°C)\nFraction\n>2mm\n>250µm\n>63µm\n<63µm\nigure 3. Relative mass, soil C % and absolute soil C amount of soil aggregate fractions originating from topsoil in the medium-term warmed\nMTW) and long-term warmed (LTW) grassland. Darker lines indicate smaller fractions. Fractions significantly affected by soil warming\nre represented with solid lines, non-significant relations upon soil warming are shown with dashed lines. (n = 17 & 16 for each fraction in\nMTW and LTW grassland respectively) MTW\nLTW\nRel. 3\nStable subsoil SOC stocks Higher subsoil than topsoil SOC losses were reported in two forests (Lin et al., 2018;\nSoong et al., 2020a), while unresponsive (this study) subsoil SOC stocks or even increases in subsoil SOC stocks were observed in grasslands (Jia et al., 2019). Further research is needed to unravel the drivers of these contrasting subsoil SOC responses to\n105\nwarming among experiments, which may be related to differences in soil properties, aggregate dynamics or rooting depths. in grasslands (Jia et al., 2019). Further research is needed to unravel the drivers of these contrasting subsoil SOC responses to\n105\nwarming among experiments, which may be related to differences in soil properties, aggregate dynamics or rooting depths. 2\nLarge, linear and fast topsoil SOC loss Aggregate fractionation suggests C % all size\n85\nfractions were impacted similarly by warming (fig. 3). This likely indicates both particulate organic matter, often occluded in\nlarge-size aggregates, and mineral-associated organic matter, present in all aggregate sizes, decreased with warming. affected by more intense (>9 ◦C) long-term warming (Radujkovi´c et al., 2018). Aggregate fractionation suggests C % all size\n85\nfractions were impacted similarly by warming (fig. 3). This likely indicates both particulate organic matter, often occluded in\nlarge-size aggregates, and mineral-associated organic matter, present in all aggregate sizes, decreased with warming. affected by more intense (>9 ◦C) long-term warming (Radujkovi´c et al., 2018). Aggregate fractionation suggests C % all size\n85\nfractions were impacted similarly by warming (fig. 3). This likely indicates both particulate organic matter, often occluded in\nlarge-size aggregates, and mineral-associated organic matter, present in all aggregate sizes, decreased with warming. 2\nLarge, linear and fast topsoil SOC loss mass (%)\nSoil C (%)\nSoil C (g C 100 g−1 dw soil)\n0\n5\n10\n15\n20 0\n5\n10\n15\n20\n0\n20\n40\n60\n2\n4\n6\n0\n1\n2\n3\n4\nSoil warming (°C)\nFraction\n>2mm\n>250µm\n>63µm\n<63µm Soil warming (°C) Figure 3. Relative mass, soil C % and absolute soil C amount of soil aggregate fractions originating from topsoil in the medium-term warmed\n(MTW) and long-term warmed (LTW) grassland. Darker lines indicate smaller fractions. Fractions significantly affected by soil warming\nare represented with solid lines, non-significant relations upon soil warming are shown with dashed lines. (n = 17 & 16 for each fraction in\nMTW and LTW grassland respectively) Figure 3. Relative mass, soil C % and absolute soil C amount of soil aggregate fractions originating from topsoil in the medium-term warmed\n(MTW) and long-term warmed (LTW) grassland. Darker lines indicate smaller fractions. Fractions significantly affected by soil warming\nare represented with solid lines, non-significant relations upon soil warming are shown with dashed lines. (n = 17 & 16 for each fraction in\nMTW and LTW grassland respectively) We suggest that the rapid topsoil SOC loss observed under warming, as well as its attenuation in the medium-term, emerged\nfrom the interplay between soil microbial biomass and activity. Warming at the same study site accelerated microbial growth\nand respiration (Marañón-Jiménez et al., 2018; Walker et al., 2020), which, in the absence of increased plant inputs to soil (fig. B3), caused the initial SOC loss observed here (fig. B1a). In turn, the warming-induced SOC loss caused a decline in micro- 6 6 https://doi.org/10.5194/bg-2021-338\nPreprint. Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. bial biomass, creating a negative feedback on microbial activity that we presume prevented further SOC loss (Walker et al.,\n80\n2018, 2020). Alternatively, ephemeral SOC loss under warming may have resulted from physiological adaptations (Allison\net al., 2010; Bradford et al., 2019) or compositional shifts (Melillo et al., 2017) in the microbial community, but we found\nno evidence this occurred here. Previous research in these grasslands showed that soil microbial carbon use efficiency (CUE)\nremained constant under short- and long-term warming (Walker et al., 2020), and microbial community composition was only g\ng (\n)\ny\np\ny\naffected by more intense (>9 ◦C) long-term warming (Radujkovi´c et al., 2018). 3\nStable subsoil SOC stocks We hypothesised that the similar warming intensity across the soil profile (fig. B4) would elicit similar declines in subsoil\nSOC stocks than those in topsoil. In contrast, SOC remained constant in the subsoil, even under 50 years of soil warming\n90\nin the long-term warmed grasslands (P = 0.63; fig. 1b, 2). This lack of SOC loss from the subsoil may be explained by\nthree, potentially co-occurring mechanisms. First, limited fresh C inputs from litter and root exudates below the rooting zone\n(<10 cm deep; fig. B3) could be a critical factor (Tian et al., 2016) explaining the lack of a positive warming effect on subsoil\ndecomposition, as a plausible positive priming effect often elicited by fresh C inputs would have been restricted to the topsoil. Second, a large fraction of SOC in the topsoil is particulate organic matter protected in aggregates, whereas most subsoil SOC\n95\nis associated with minerals (Fontaine et al., 2007; Rumpel and Kögel-Knabner, 2011). As such, by accelerating mineralisation\nof plant litter that is deposited only in topsoil (Walker et al., 2018), and thereby reducing aggregate stability and breaking\nup macroaggregates (fig. 3) (Poeplau et al., 2020), warming may have had much greater effect on SOC losses in topsoil than\nin subsoil, where mineral protection dominates. Third, although unlikely given the absence of increased dissolved organic C (DOC) with warming in subsoil (fig. B6), it can also not be excluded that SOC stocks in subsoils only appear stable, because\n100\nincreased losses are compensated for by increased inputs from above (Osher et al., 2003). Only very few studies have assessed subsoil SOC stocks responses to deep soil warming, and observed responses differ\nstrongly in magnitude and even direction. Higher subsoil than topsoil SOC losses were reported in two forests (Lin et al., 2018;\nSoong et al., 2020a), while unresponsive (this study) subsoil SOC stocks or even increases in subsoil SOC stocks were observed (DOC) with warming in subsoil (fig. B6), it can also not be excluded that SOC stocks in subsoils only appear stable, because\n100\nincreased losses are compensated for by increased inputs from above (Osher et al., 2003). Only very few studies have assessed subsoil SOC stocks responses to deep soil warming, and observed responses differ\nstrongly in magnitude and even direction. 4\nImplications for carbon-climate feedbacks Earth System Model (ESM) inter-comparison studies (Eyring et al., 2016) have revealed large variability in both contemporary\nglobal SOC stock estimates and future SOC stock projections, underlining the need for empirical observations to better con-\nstrain the response of SOC to temperature change (Nishina et al., 2014). Long-term warming experiments like this study are\n110 110 7 7 https://doi.org/10.5194/bg-2021-338\nPreprint. Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. thus needed to reduce the uncertainty on model projections (Abramoff et al., 2019). Although geothermally active areas offer\nlong-lasting, continuous and large soil temperature gradients and overcome the technical challenges and high costs associated\nwith warming manipulation experiments (Sigurdsson et al., 2016; O’Gorman et al., 2014), their use as a proxy for climate\nchange has some drawbacks of its own, such as limited aboveground warming and a stepwise increase in soil temperature at thus needed to reduce the uncertainty on model projections (Abramoff et al., 2019). Although geothermally active areas offer\nlong-lasting, continuous and large soil temperature gradients and overcome the technical challenges and high costs associated\nwith warming manipulation experiments (Sigurdsson et al., 2016; O’Gorman et al., 2014), their use as a proxy for climate\nchange has some drawbacks of its own, such as limited aboveground warming and a stepwise increase in soil temperature at\nthe initiation of the geothermal gradient (De Boeck et al., 2015). Also the Andosol, covering only ± 0.8 % of the earth’s surface\n115\n(Baillie, 2001), makes that one should be cautious extrapolating the results to the entire sub-arctic region. Nonetheless, this\nsite offers a unique opportunity to study the direct versus long-term response of SOC stocks to temperature change and the\nresults from this study and other deep soil warming experiments clearly indicate that introducing vertically resolved plant- and\nmicrobial dynamics in ESMs is a necessity for more accurate projections of the carbon-climate feedback. the initiation of the geothermal gradient (De Boeck et al., 2015). Also the Andosol, covering only ± 0.8 % of the earth’s surface\n115\n(Baillie, 2001), makes that one should be cautious extrapolating the results to the entire sub-arctic region. 4\nImplications for carbon-climate feedbacks Nonetheless, this\nsite offers a unique opportunity to study the direct versus long-term response of SOC stocks to temperature change and the\nresults from this study and other deep soil warming experiments clearly indicate that introducing vertically resolved plant- and\nmicrobial dynamics in ESMs is a necessity for more accurate projections of the carbon-climate feedback. the initiation of the geothermal gradient (De Boeck et al., 2015). Also the Andosol, covering only ± 0.8 % of the earth’s surface\n115\n(Baillie, 2001), makes that one should be cautious extrapolating the results to the entire sub-arctic region. Nonetheless, this\nsite offers a unique opportunity to study the direct versus long-term response of SOC stocks to temperature change and the\nresults from this study and other deep soil warming experiments clearly indicate that introducing vertically resolved plant- and\nmicrobial dynamics in ESMs is a necessity for more accurate projections of the carbon-climate feedback. In conclusion, warming caused a large but rapidly equilibrating SOC loss in the topsoil that increased linearly with warming\n120\nintensity, while no SOC loss was observed in the subsoil in our subarctic grasslands exposed to decades of soil warming. Future work should focus on understanding whether these observed temporal dynamics are consistent throughout the northern\nnon-permafrost region. Improved understanding of the variation in subsoil SOC responses to warming is also critical for\nconstraining Earth System Models and obtaining reliable climate projections. In conclusion, warming caused a large but rapidly equilibrating SOC loss in the topsoil that increased linearly with warming\n120\nintensity, while no SOC loss was observed in the subsoil in our subarctic grasslands exposed to decades of soil warming. Future work should focus on understanding whether these observed temporal dynamics are consistent throughout the northern\nnon-permafrost region. Improved understanding of the variation in subsoil SOC responses to warming is also critical for\nconstraining Earth System Models and obtaining reliable climate projections. 8 8 https://doi.org/10.5194/bg-2021-338\nPreprint. Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. Appendix A: Material and methods\n125\nThis study was conducted at the ForHot research site, located in the Hengill geothermal area, 40 km east of Reykjavík, Iceland\n(64°00’01\" N, 21°11’09\" W; 100 – 225 m a.s.l. (Sigurdsson et al., 2016). Appendix A: Material and methods\n125 First in these soils, the A horizon that is enriched with SOC is maximum 10 cm deep (Arnalds,\n2015). Second, 95.7 ± 0.4 (SE) % of the fine root biomass sampled in the top 30 cm layer, was found in upper 10 cm (fig. B5). Third bulk density in subsoil is significantly higher than in topsoil (P < 0 001) (fig B7)\n135 Third, bulk density in subsoil is significantly higher than in topsoil (P < 0.001) (fig. B7). 135\nThe site comprises two areas that have been subjected to geothermal soil warming for different periods of time (Sigurdsson\net al., 2016). One area (hereafter “medium-term warmed grassland”) has been warmed since May 2008, when a large earthquake\nshifted geothermal systems to previously unwarmed soils. The second area (2.5 km North-east from the first area; hereafter\n“long-term warmed grassland”) was already mentioned to be warmed in the early 18th century (Magnússon and Vídalín, 1708) y\ng\ny\ng\np\ng\nThe site comprises two areas that have been subjected to geothermal soil warming for different periods of time (Sigurdsson\net al., 2016). One area (hereafter “medium-term warmed grassland”) has been warmed since May 2008, when a large earthquake\nshifted geothermal systems to previously unwarmed soils. The second area (2.5 km North-east from the first area; hereafter\n“long-term warmed grassland”) was already mentioned to be warmed in the early 18th century (Magnússon and Vídalín, 1708) and has thus likely been warmed for centuries. For sure, the warming was registered in a census during the 1960s, and no change\n140\nin the location of the hotspots has been recorded during the past 50 years (Kristján Sæmundsson, personal communication). The soil warming increment at both sites is relatively constant throughout the year and extreme deviations are rare (Sigurdsson\net al., 2016). Soil warming is caused by horizontal heat conduction through the soil, causing fairly homogeneous warming with\ndepth and inducing a fairly natural temperature depth profile (fig. B4). This homogeneous soil warming is in line with CMIP5 predictions of rapid transfer of the temperature signal from air to shallow and deeper soils (Soong et al., 2020b). The geothermal\n145\nwater is confined within the bedrock and no signs of soil contamination by geothermal byproducts have been found (Shi et al.,\n2020). Appendix A: Material and methods\n125 This study was conducted at the ForHot research site, located in the Hengill geothermal area, 40 km east of Reykjavík, Iceland\n(64°00’01\" N, 21°11’09\" W; 100 – 225 m a.s.l. (Sigurdsson et al., 2016). The mean annual temperature between 2006 and 2016\nwas 5.2 ± 0.1 (± SE) °C, and mean annual daily minimum and maximum temperatures were 2.2 ± 0.2 (SE) and 8.6 ± 0.2 (±\nSE) °C. The mean annual precipitation during the same period was 1413 ± 57 (± SE) mm (Icelandic Meteorological Office; Eyrarbakki weather station which closed in 2017). The main vegetation type is unmanaged grassland, dominated by Agrostris\n130\ncapillaris, Ranunculus acris and Equisetum pratense and the underlying soil is classified as Brown Andosol (Arnalds, 2015). In this study, we define the 0-10 cm layer as topsoil, and the 10-30 cm layers as subsoil. This subsoil layer strongly differed\nfrom the topsoil in many ways. First in these soils, the A horizon that is enriched with SOC is maximum 10 cm deep (Arnalds,\n2015). Second, 95.7 ± 0.4 (SE) % of the fine root biomass sampled in the top 30 cm layer, was found in upper 10 cm (fig. B5). Eyrarbakki weather station which closed in 2017). The main vegetation type is unmanaged grassland, dominated by Agrostris\n130\ncapillaris, Ranunculus acris and Equisetum pratense and the underlying soil is classified as Brown Andosol (Arnalds, 2015). In this study, we define the 0-10 cm layer as topsoil, and the 10-30 cm layers as subsoil. This subsoil layer strongly differed\nfrom the topsoil in many ways. First in these soils, the A horizon that is enriched with SOC is maximum 10 cm deep (Arnalds,\n2015). Second, 95.7 ± 0.4 (SE) % of the fine root biomass sampled in the top 30 cm layer, was found in upper 10 cm (fig. B5). Third, bulk density in subsoil is significantly higher than in topsoil (P < 0.001) (fig. B7). 135 Eyrarbakki weather station which closed in 2017). The main vegetation type is unmanaged grassland, dominated by Agrostris\n130\ncapillaris, Ranunculus acris and Equisetum pratense and the underlying soil is classified as Brown Andosol (Arnalds, 2015). In this study, we define the 0-10 cm layer as topsoil, and the 10-30 cm layers as subsoil. This subsoil layer strongly differed\nfrom the topsoil in many ways. Appendix A: Material and methods\n125 Soil pH (mean: 5.5 ± 0.1 (SE)) and soil moisture did not show major changes along the soil warming gradients, with\nsoil moisture very rarely approaching the permanent wilting point and no relation between soil temperature and the frequency\nof drought events (Leblans et al., 2017). Further, the plant species composition was very similar between the medium-term and predictions of rapid transfer of the temperature signal from air to shallow and deeper soils (Soong et al., 2020b). The geothermal\n145\nwater is confined within the bedrock and no signs of soil contamination by geothermal byproducts have been found (Shi et al.,\n2020). Soil pH (mean: 5.5 ± 0.1 (SE)) and soil moisture did not show major changes along the soil warming gradients, with\nsoil moisture very rarely approaching the permanent wilting point and no relation between soil temperature and the frequency\nof drought events (Leblans et al., 2017). Further, the plant species composition was very similar between the medium-term and the long term warmed grassland and no drastic changes in dominant plant species occurred up to +6.4 ◦C warming (Leblans\n150\net al., 2017) (which is the RCP8.5 projected annual warming level for high northern latitudes for the year 2100) (IPCC, 2013). More detailed information on the site characteristics can be found in Sigurdsson et al. (2016). We established five replicate transects in each area (the medium-term and the long-term warmed grassland) in 2012, around\ntwo and four geothermal heat sources respectively. In the medium-term warmed grassland, all transects were located on south- the long term warmed grassland and no drastic changes in dominant plant species occurred up to +6.4 ◦C warming (Leblans\n150\net al., 2017) (which is the RCP8.5 projected annual warming level for high northern latitudes for the year 2100) (IPCC, 2013). More detailed information on the site characteristics can be found in Sigurdsson et al. (2016). We established five replicate transects in each area (the medium-term and the long-term warmed grassland) in 2012, around\ntwo and four geothermal heat sources respectively. In the medium-term warmed grassland, all transects were located on south- west facing slopes, three with the geothermal heat source at the bottom of the slope, and two with the geothermal heat source\n155\nat the top, to eliminate effects of topography and downward transport of groundwater and nutrients from introducing a bias\nin the SOC stocks. 4\nImplications for carbon-climate feedbacks The mean annual temperature between 2006 and 2016\nwas 5.2 ± 0.1 (± SE) °C, and mean annual daily minimum and maximum temperatures were 2.2 ± 0.2 (SE) and 8.6 ± 0.2 (±\nSE) °C. The mean annual precipitation during the same period was 1413 ± 57 (± SE) mm (Icelandic Meteorological Office; Appendix A: Material and methods\n125\nThis study was conducted at the ForHot research site, located in the Hengill geothermal area, 40 km east of Reykjavík, Iceland\n(64°00’01\" N, 21°11’09\" W; 100 – 225 m a.s.l. (Sigurdsson et al., 2016). The mean annual temperature between 2006 and 2016\nwas 5.2 ± 0.1 (± SE) °C, and mean annual daily minimum and maximum temperatures were 2.2 ± 0.2 (SE) and 8.6 ± 0.2 (±\nSE) °C. The mean annual precipitation during the same period was 1413 ± 57 (± SE) mm (Icelandic Meteorological Office; Appendix A: Material and methods\n125 In the long-term warmed grassland, all transects were located on level ground. Within the long-term and\nmedium-term warmed grassland, all measurement plots had similar microtopography, soil depth and grazing history. west facing slopes, three with the geothermal heat source at the bottom of the slope, and two with the geothermal heat source\n155\nat the top, to eliminate effects of topography and downward transport of groundwater and nutrients from introducing a bias\nin the SOC stocks. In the long-term warmed grassland, all transects were located on level ground. Within the long-term and\nmedium-term warmed grassland, all measurement plots had similar microtopography, soil depth and grazing history. 9 https://doi.org/10.5194/bg-2021-338\nPreprint. Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. Each transect consists of six 2 x 2 m permanent measurement plots distributed along the soil temperature gradient, including\nunwarmed soil (MAT: 5.7 ± 0.1 ◦C), yielding 60 plots in total. Each 2 x 2 m permanent measurement plot was accompanied by\n160\ntwo adjacent 0.5 x 0.5 m subplots for destructive measurements. Plot-specific soil warming was recorded hourly at 10 cm soil\ndepth using HOBO TidbiT v2 Water Temperature Data Loggers (Onset Computer Corporation, USA). Because the permanent\nplots occurred at different warming intensities in the different transects, we adopted a regression approach (see statistics below). More detailed information on the experimental design is provided in Sigurdsson et al. (2016). In July 2013 and 2018, two 0-10 cm soil cores (corer ø= 5.12 cm) were taken within each subplot. In the medium-term\n165\nwarmed grassland, soils were too shallow to sample deeper, but additional 10-30 cm cores were taken in the long-term warmed\ngrassland. Cores were analysed for: (1) soil C concentrations; (2) pH (topsoil only); (3) soil bulk density (BD); and (4) SOC\nstocks. From the first core we obtained fine roots (<2 mm) and soil particles (>2 mm) (necessary to calculate BD) by washing the In July 2013 and 2018, two 0-10 cm soil cores (corer ø= 5.12 cm) were taken within each subplot. In the medium-term\n165\nwarmed grassland, soils were too shallow to sample deeper, but additional 10-30 cm cores were taken in the long-term warmed\ngrassland. Cores were analysed for: (1) soil C concentrations; (2) pH (topsoil only); (3) soil bulk density (BD); and (4) SOC\nstocks. Appendix A: Material and methods\n125 Further installation details of the lysimeters are described in Edlinger (2016), as well as the sampling procedure. The C-input data for arbuscular mycorrhizae originates from Zhang et al. (2020), where the sampling\n180\nprocedure is described. Aboveground biomass was sampled by placing a 20x40 cm frame on the plot, after which all vegetation\nwas clipped. The samples were taken to the lab and sorted by hand in a grass and moss fraction. Both fractions were dried for\n48 h at 70 ◦C, weighed and milled. The samples were analysed for C concentration (%) by dry combustion (Macro Elemental\nAnalyser, model vario MAX CN, Hanau, Germany). Aggregate fractionation was done in 2018 only. Per plot, a 0-10 cm soil the sampling procedure. The C-input data for arbuscular mycorrhizae originates from Zhang et al. (2020), where the sampling\n180\nprocedure is described. Aboveground biomass was sampled by placing a 20x40 cm frame on the plot, after which all vegetation\nwas clipped. The samples were taken to the lab and sorted by hand in a grass and moss fraction. Both fractions were dried for\n48 h at 70 ◦C, weighed and milled. The samples were analysed for C concentration (%) by dry combustion (Macro Elemental\nAnalyser, model vario MAX CN, Hanau, Germany). Aggregate fractionation was done in 2018 only. Per plot, a 0-10 cm soil the sampling procedure. The C-input data for arbuscular mycorrhizae originates from Zhang et al. (2020), where the sampling\n180\nprocedure is described. Aboveground biomass was sampled by placing a 20x40 cm frame on the plot, after which all vegetation\nwas clipped. The samples were taken to the lab and sorted by hand in a grass and moss fraction. Both fractions were dried for\n48 h at 70 ◦C, weighed and milled. The samples were analysed for C concentration (%) by dry combustion (Macro Elemental\nAnalyser, model vario MAX CN, Hanau, Germany). Aggregate fractionation was done in 2018 only. Per plot, a 0-10 cm soil core was taken (corer ø= 5.12 cm) and dried at room temperature for a some weeks. Stones were removed and aggregates\n185\nlarger than 8 mm were broken up by dry-sieving on a 8 mm soil sieve. The dry-sieved soil was then slaked for 5 min with DI\nwater, after which it was wet-sieved on a 2 mm, 250 µm and 63 µm, to separate into four size fractions. Appendix A: Material and methods\n125 From the first core we obtained fine roots (<2 mm) and soil particles (>2 mm) (necessary to calculate BD) by washing the\ncores over two sieves with mesh sizes 2 mm and 0.5 mm. Roots and >2 mm particles were dried and weighed to gain fine root\n170\nbiomass (g m−2) and the volume of >2 mm particles (g cm−3) was measured by the water displacement method. The second\nsoil core was first dried and weighed (as for aboveground vegetation), and soil was then sieved to obtain soil particles <2 mm\nand split into three aliquots. One aliquot of 2 g was milled (Retsch MM301 Mixer Mill, Haan, Germany) and analysed for\nC concentration (%) by dry combustion (Macro Elemental Analyser, model vario MAX CN, Hanau, Germany). Finally, BD (g cm−3) and SOC stocks (ton ha−1) were calculated according to the approach described in Bárcena et al. (2014). 175\nTo measure dissolved organic matter (DOC), teflon suction cup lysimeters (Prenart Super Quartz, Prenart Equipment Aps,\nFrederiksberg, Denmark) were placed at about 30-40 cm depth in the medium-term and long-term warmed grassland, in Octo-\nber 2014. Samples were taken during summer 2015, 2016 and 2017. The DOC was analysed with a combined Total Organic\nCarbon (Shimadzu, Kyoto, Japan). Further installation details of the lysimeters are described in Edlinger (2016), as well as (g cm−3) and SOC stocks (ton ha−1) were calculated according to the approach described in Bárcena et al. (2014). 175\nTo measure dissolved organic matter (DOC), teflon suction cup lysimeters (Prenart Super Quartz, Prenart Equipment Aps,\nFrederiksberg, Denmark) were placed at about 30-40 cm depth in the medium-term and long-term warmed grassland, in Octo-\nber 2014. Samples were taken during summer 2015, 2016 and 2017. The DOC was analysed with a combined Total Organic\nCarbon (Shimadzu, Kyoto, Japan). Further installation details of the lysimeters are described in Edlinger (2016), as well as (g cm−3) and SOC stocks (ton ha−1) were calculated according to the approach described in Bárcena et al. (2014). 175\nTo measure dissolved organic matter (DOC), teflon suction cup lysimeters (Prenart Super Quartz, Prenart Equipment Aps,\nFrederiksberg, Denmark) were placed at about 30-40 cm depth in the medium-term and long-term warmed grassland, in Octo-\nber 2014. Samples were taken during summer 2015, 2016 and 2017. The DOC was analysed with a combined Total Organic\nCarbon (Shimadzu, Kyoto, Japan). Appendix A: Material and methods\n125 Each fraction was dried\nat 70 ◦C for 72 h, after which all fractions were ground with a ball mill to homogenise and analysed for C concentration (%)\nby dry combustion (Macro Elemental Analyser, model vario MAX CN, Hanau, Germany). Relative mass of the fractions was calculated by dividing the fraction mass by the sum of all fraction masses of initial sample. The absolute soil C-amount of each\n190\nfraction was calculated by multiplying the fraction soil mass per 100 g of dry soil with the soil C % (fig. 3). A soil mass correction of the SOC stocks as described in Ellert and Bettany (1995) was necessary to compare stock changes\nacross the soil warming gradient as soil compaction increased with warming in the upper soil layers (increasing BD; fig. B7), 10 https://doi.org/10.5194/bg-2021-338\nPreprint. Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. implying that soil depths in unwarmed soil corresponded to shallower soil depths at warmer soils. The calculation method is\nexplained in detail in fig. B8. 5 implying that soil depths in unwarmed soil corresponded to shallower soil depths at warmer soils. The calculation method is\nexplained in detail in fig. B8. 195\nThe soil warming dependence of bulk soil SOC stocks (corrected and uncorrected for soil compaction), BD, DOC and soil C\n% were tested with a linear mixed effects model (Pinheiro et al., 2021). Soil warming and warming-duration (medium-term vs. long-term) were included as main effects, while sampling year was used as random effect to account for sampling differences\nand interannual variabilities between the two sampling campaigns. In all cases, criteria for normality and homoscedasticity\nwere met. For all tests, the dataset was reduced to cover only the warming levels captured by the projections for high northern\n200\nlatitudes for the year 2100 (0 – 6.4 ◦C warming) (IPCC, 2013). All tests were performed using R software (R Development\nCore Team, 2011). implying that soil depths in unwarmed soil corresponded to shallower soil depths at warmer soils. The calculation method is\nexplained in detail in fig. B8. 195\nThe soil warming dependence of bulk soil SOC stocks (corrected and uncorrected for soil compaction), BD, DOC and soil C\n% were tested with a linear mixed effects model (Pinheiro et al., 2021). Soil warming and warming-duration (medium-term vs. Appendix A: Material and methods\n125 (n = 78 & 40 for topsoil and subsoil respectively) y = −1.7x + 32.8\nR² = 0.11; P= 0.0037\nP= 0.47\nTopsoil\nSubsoil\n0\n1\n2\n3\n4\n5\n20\n30\n40\n50\n20\n30\n40\n50\nSoil warming (°C)\nSOC stock (ton ha−1)\nYr−grassland\n2013 − LTW\n2013 − MTW\n2018 − LTW\n2018 − MTW P= 0.47\nTopsoil\nSubsoil\n0\n1\n2\n3\n4\n5\n20\n30\n40\n20\n30\n40\n50\nSoil warming (°C)\nSOC stock (ton ha−1)\nYr−grassland\n2013 − LTW\n2013 − MTW\n2018 − LTW\n2018 − MTW P= 0.47\nSubsoil\n0\n1\n2\n3\n4\n5\n20\n30\n40\n50\nSoil warming (°C)\nSOC stock\n2013 − MT\n2018 − LTW\n2018 − MT Figure B1. Reduction of soil organic carbon (SOC) stock with soil warming, a) in the topsoil (0-10 cm); b) in the subsoil (10-30 cm). All soil\nsamples were taken in July 2013 or July 2018. The regression for medium-term warmed (MTW) and long-term warmed (LTW) grassland\n(in topsoil) for both 2013 and 2018 were combined, since no soil temperature x warming-duration interaction effect, nor a main effect for\nwarming-duration was found. Soil warming is expressed relative to ambient soil temperature (both at 10 cm depth). In topsoil, a linear relation\nwas observed, while no significant effect was present in subsoil. The 95 % confidence bounds are shown around the topsoil regression slope. (n = 78 & 40 for topsoil and subsoil respectively) Figure B1. Reduction of soil organic carbon (SOC) stock with soil warming, a) in the topsoil (0-10 cm); b) in the subsoil (10-30 cm). All soil\nsamples were taken in July 2013 or July 2018. The regression for medium-term warmed (MTW) and long-term warmed (LTW) grassland\n(in topsoil) for both 2013 and 2018 were combined, since no soil temperature x warming-duration interaction effect, nor a main effect for\nwarming-duration was found. Soil warming is expressed relative to ambient soil temperature (both at 10 cm depth). In topsoil, a linear relation\nwas observed, while no significant effect was present in subsoil. The 95 % confidence bounds are shown around the topsoil regression slope. (n = 78 & 40 for topsoil and subsoil respectively) Figure B1. Reduction of soil organic carbon (SOC) stock with soil warming, a) in the topsoil (0-10 cm); b) in the subsoil (10-30 cm). Appendix A: Material and methods\n125 long-term) were included as main effects, while sampling year was used as random effect to account for sampling differences\nand interannual variabilities between the two sampling campaigns. In all cases, criteria for normality and homoscedasticity\nwere met. For all tests, the dataset was reduced to cover only the warming levels captured by the projections for high northern\n200\nlatitudes for the year 2100 (0 – 6.4 ◦C warming) (IPCC, 2013). All tests were performed using R software (R Development\nCore Team, 2011). were met. For all tests, the dataset was reduced to cover only the warming levels captured by the projections for high northern\n200\nlatitudes for the year 2100 (0 – 6.4 ◦C warming) (IPCC, 2013). All tests were performed using R software (R Development\nCore Team, 2011). 11 11 Appendix B: Supplementary Appendix B: Supplementary y = −1.7x + 32.8\nR² = 0.11; P= 0.0037\nP= 0.47\nTopsoil\nSubsoil\n0\n1\n2\n3\n4\n5\n20\n30\n40\n50\n20\n30\n40\n50\nSoil warming (°C)\nSOC stock (ton ha−1)\nYr−grassland\n2013 − LTW\n2013 − MTW\n2018 − LTW\n2018 − MTW\nFigure B1. Reduction of soil organic carbon (SOC) stock with soil warming, a) in the topsoil (0-10 cm); b) in the subsoil (10-30 cm). All soil\nsamples were taken in July 2013 or July 2018. The regression for medium-term warmed (MTW) and long-term warmed (LTW) grassland\n(in topsoil) for both 2013 and 2018 were combined, since no soil temperature x warming-duration interaction effect, nor a main effect for\nwarming-duration was found. Soil warming is expressed relative to ambient soil temperature (both at 10 cm depth). In topsoil, a linear relation\nwas observed, while no significant effect was present in subsoil. The 95 % confidence bounds are shown around the topsoil regression slope. Appendix A: Material and methods\n125 (topsoil n = 42 & 40 for MTW and LTW grassland\nrespectively; subsoil n = 40 for LTW grassland) 13 P = 0.2\nP = 0.7\nP = 0.09\nP = 0.6\nP = 0.4\nP = 0.9\nP = 0.9\nP = 0.2\nMTW\nLTW\nAGB\nFine roots\nAMF\nTotal\n1\n2\n3\n4\n5\n0\n1\n2\n3\n4\n5\n0\n4\n8\n12\n4\n8\n12\n16\n0.5\n1.0\n1.5\n2.0\n10\n15\n20\n25\nSoil warming (°C)\nC−input (ton ha−1) P = 0.2\nP = 0.7\nP = 0.09\nP = 0.6\nP = 0.4\nP = 0.9\nP = 0.9\nP = 0.2\nMTW\nLTW\nAGB\nFine roots\nAMF\nTotal\n1\n2\n3\n4\n5\n0\n1\n2\n3\n4\n5\n0\n4\n8\n12\n4\n8\n12\n16\n0.5\n1.0\n1.5\n2.0\n10\n15\n20\n25\nSoil warming (°C)\nC−input (ton ha−1)\nxies for annual soil C-inputs from aboveground biomass (AGB), fine root biomass and arbuscular myco\narmed (MTW) and the long-term warmed (LTW) grasslands. Vascular plant aboveground biomass wa\ninputs; vascular plant fine root biomass for belowground C-inputs and C sequestered by AMF (Cnew;\nnputs by arbuscular mycorrhizae. For the observed soil warming range (0-6.4 ◦C warming), no change\nwere obtained by a linear regression analysis. (n = 20 & 17 for MTW and LTW inputs respectively) P = 0.7\nP = 0.09\nP = 0.6\nP = 0.9\nP = 0.9\nP = 0.2\nGB\nFine roots\nAMF\nTotal\n1\n2\n3\n4\n5\n0\n1\n2\n3\n4\n5\n0\n4\n4\n8\n12\n16\n0.5\n1.0\n1.5\n2.0\n10\n15\n20\n25\nSoil warming (°C)\nC−input (ton ha−1)\nFigure B3. Proxies for annual soil C-inputs from aboveground biomass (AGB), fine root biomass and arbuscular mycorrhizae (AMF) in the\nmedium-term warmed (MTW) and the long-term warmed (LTW) grasslands. Vascular plant aboveground biomass was used as a proxy for\naboveground C-inputs; vascular plant fine root biomass for belowground C-inputs and C sequestered by AMF (Cnew; data from Zhang et al. (2020)) for C-inputs by arbuscular mycorrhizae. For the observed soil warming range (0-6.4 ◦C warming), no change in C-inputs could be\nfound. P-values were obtained by a linear regression analysis. (n = 20 & 17 for MTW and LTW inputs respectively) Soil warming (°C) Figure B3. Appendix A: Material and methods\n125 All soil\nsamples were taken in July 2013 or July 2018. The regression for medium-term warmed (MTW) and long-term warmed (LTW) grassland\n(in topsoil) for both 2013 and 2018 were combined, since no soil temperature x warming-duration interaction effect, nor a main effect for\nwarming-duration was found. Soil warming is expressed relative to ambient soil temperature (both at 10 cm depth). In topsoil, a linear relation\nwas observed, while no significant effect was present in subsoil. The 95 % confidence bounds are shown around the topsoil regression slope. (n = 78 & 40 for topsoil and subsoil respectively) 12 https://doi.org/10.5194/bg-2021-338\nPreprint. Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. y = −0.42x + 5.74\nR² = 0.16; P= 0.0081\ny = −0.49x + 6.72\nR² = 0.6; P= 0.00022\nP= 0.16\nMTW\nLTW\nTopsoil\nSubsoil\n0\n2\n4\n0\n2\n4\n0\n3\n6\n9\n0\n3\n6\n9\nSoil warming (°C)\nSoil C (%)\nYear\n2013\n2018 y = −0.42x + 5.74\nR² = 0.16; P= 0.0081\ny = −0.49x + 6.72\nR² = 0.6; P= 0.00022\nP= 0.16\nMTW\nLTW\nTopsoil\nSubsoil\n0\n2\n4\n0\n2\n4\n0\n3\n6\n9\n0\n3\n6\n9\nSoil warming (°C)\nSoil C (%)\nYear\n2013\n2018\nFigure B2. Percentage of carbon in topsoil and subsoil, for the long-term warmed (LTW) and medium-term warmed (MTW) grassland. The 95 % confidence bounds are shown around the topsoil linear regression slopes. (topsoil n = 42 & 40 for MTW and LTW grassland\nrespectively; subsoil n = 40 for LTW grassland) LTW Figure B2. Percentage of carbon in topsoil and subsoil, for the long-term warmed (LTW) and medium-term warmed (MTW) grassland. The 95 % confidence bounds are shown around the topsoil linear regression slopes. (topsoil n = 42 & 40 for MTW and LTW grassland\nrespectively; subsoil n = 40 for LTW grassland) Figure B2. Percentage of carbon in topsoil and subsoil, for the long-term warmed (LTW) and medium-term warmed (MTW) grassland. The 95 % confidence bounds are shown around the topsoil linear regression slopes. Appendix A: Material and methods\n125 Proxies for annual soil C-inputs from aboveground biomass (AGB), fine root biomass and arbuscular mycorrhizae (AMF) in the\nmedium-term warmed (MTW) and the long-term warmed (LTW) grasslands. Vascular plant aboveground biomass was used as a proxy for\naboveground C-inputs; vascular plant fine root biomass for belowground C-inputs and C sequestered by AMF (Cnew; data from Zhang et al. (2020)) for C-inputs by arbuscular mycorrhizae. For the observed soil warming range (0-6.4 ◦C warming), no change in C-inputs could be\nfound. P-values were obtained by a linear regression analysis. (n = 20 & 17 for MTW and LTW inputs respectively) 14 −5\n−10\n−20\n−30\n−2\n−1\n0\n1\n2\nTemperature difference from 10 cm (°C)\nSoil depth (cm) −5\n−10\n−20\n−30\n−2\n−1\n0\n1\n2\nSoil depth (cm) Temperature difference from 10 cm (°C) Figure B4. Soil warming along the vertical soil profile, based on the reference temperature measured at 10 cm which is used throughout the\npaper. The median temperature is about 1 ◦C higher at 5 cm depth, while being slightly lower at 20 cm and 30 cm depth. Due to the shallow\nsoil in the medium-term warmed grassland, the warming profile is given for the long-term warmed grassland only. 15 topsoil\nsubsoil\nP = 0.8\nP = 0.6\n0\n10\n20\n30\n40\n0\n1\n2\n3\n4\n5\nSoil warming (°C)\nFine root density (mg roots cm−3)\nSoil layer\nTopsoil\nSubsoil\nFigure B5. Fine root density in long-term warmed (LTW) grassland in top-and subsoil sampled in July 2018. For the observed soil warming\nrange (0-6.4 ◦C warming), no change in C-inputs could be found. P-values were obtained by a linear regression analysis. In the top 30 cm\nsoil layer, 95.7 ± 0.4 (SE) % of the fine root biomass was found in upper 10 cm, which is why we define the 0-10 cm layer as topsoil, and\nthe 10-30 cm layer as subsoil. (n = 17 & 15 for topsoil and subsoil respectively)\nPreprint. Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. topsoil\nsubsoil\nP = 0.8\nP = 0.6\n0\n10\n20\n30\n40\n0\n1\n2\n3\n4\n5\nSoil\narming (°C)\nFine root density (mg roots cm−3)\nSoil layer\nTopsoil\nSubsoil Figure B5. Fine root density in long-term warmed (LTW) grassland in top-and subsoil sampled in July 2018. Appendix A: Material and methods\n125 For the observed soil warming\nrange (0-6.4 ◦C warming), no change in C-inputs could be found. P-values were obtained by a linear regression analysis. In the top 30 cm\nsoil layer, 95.7 ± 0.4 (SE) % of the fine root biomass was found in upper 10 cm, which is why we define the 0-10 cm layer as topsoil, and\nthe 10-30 cm layer as subsoil. (n = 17 & 15 for topsoil and subsoil respectively) Figure B5. Fine root density in long-term warmed (LTW) grassland in top-and subsoil sampled in July 2018. For the observed soil warming\nrange (0-6.4 ◦C warming), no change in C-inputs could be found. P-values were obtained by a linear regression analysis. In the top 30 cm\nsoil layer, 95.7 ± 0.4 (SE) % of the fine root biomass was found in upper 10 cm, which is why we define the 0-10 cm layer as topsoil, and\nthe 10-30 cm layer as subsoil. (n = 17 & 15 for topsoil and subsoil respectively) Figure B5. Fine root density in long-term warmed (LTW) grassland in top-and subsoil sampled in July 2018. For the observed soil warming\nrange (0-6.4 ◦C warming), no change in C-inputs could be found. P-values were obtained by a linear regression analysis. In the top 30 cm\nsoil layer, 95.7 ± 0.4 (SE) % of the fine root biomass was found in upper 10 cm, which is why we define the 0-10 cm layer as topsoil, and\nthe 10-30 cm layer as subsoil. (n = 17 & 15 for topsoil and subsoil respectively) 16 P = 0.87\nP = 0.8\nMWT\nLTW\n0\n1\n2\n3\n4\n5\n1\n2\n0.0\n0.5\n1.0\n1.5\nSoil warming °C\nDOC (ppm)\nYear\n2015\n2016\n2017\nFigure B6. Dissolved organic carbon (DOC) in medium-term warmed (MTW) and long-term warmed (LTW) grassland during summer 2015,\n2016 and 2017. No DOC change was observed with soil warming in the SWT (P = 0.95), nor in the LTW (P = 0.8), meaning that carbon-inputs\ninto deeper soil layers remained constant over the whole soil temperature gradient. Statistical analysis was done with a linear mixed effects\nmodel with soil warming as explanatory variable and sampling year as a random factor. Criteria for normality and homoscedasticity were\nmet. Appendix A: Material and methods\n125 Dissolved organic carbon was sampled at a approximate depth of 30 cm with Prenart Super Quartz (Prenart, Fredriksberg, Denmark)\nsoil water samplers, installed around 1 October 2014. The samples was always inserted from ‘downslope,’ where the soil was deep enough. Analyses of the samples was done at the IGN Biochemistry Lab, University of Copenhagen. (n = 25 & 29 for MTW and LTW grassland\nrespectively)\nhttps://doi.org/10.5194/bg-2021-338\nPreprint. Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. P = 0.87\nP = 0.8\nMWT\nLTW\n0\n1\n2\n3\n4\n5\n1\n2\n0.0\n0.5\n1.0\n1.5\nSoil warming °C\nDOC (ppm)\nYear\n2015\n2016\n2017\nFigure B6. Dissolved organic carbon (DOC) in medium-term warmed (MTW) and long-term warmed (LTW) grassland during summer 2015,\nhttps://doi.org/10.5194/bg-2021-338\nPreprint. Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. P = 0.87\nP = 0.8\nMWT\nLTW\n0\n1\n2\n3\n4\n5\n1\n2\n0.0\n0.5\n1.0\n1.5\nSoil warming °C\nDOC (ppm)\nYear\n2015\n2016\n2017\n⃝\n( ) Figure B6. Dissolved organic carbon (DOC) in medium-term warmed (MTW) and long-term warmed (LTW) grassland during summer 2015,\n2016 and 2017. No DOC change was observed with soil warming in the SWT (P = 0.95), nor in the LTW (P = 0.8), meaning that carbon-inputs\ninto deeper soil layers remained constant over the whole soil temperature gradient. Statistical analysis was done with a linear mixed effects\nmodel with soil warming as explanatory variable and sampling year as a random factor. Criteria for normality and homoscedasticity were\nmet. Dissolved organic carbon was sampled at a approximate depth of 30 cm with Prenart Super Quartz (Prenart, Fredriksberg, Denmark)\nsoil water samplers, installed around 1 October 2014. The samples was always inserted from ‘downslope,’ where the soil was deep enough. Analyses of the samples was done at the IGN Biochemistry Lab, University of Copenhagen. (n = 25 & 29 for MTW and LTW grassland\nrespectively) 17 https://doi.org/10.5194/bg-2021-338\nPreprint. Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. Appendix A: Material and methods\n125 y = 0.027x + 0.6\nR² = 0.12; P = 1.5e−03\na)\ny = 0.031x + 0.85\nR² = 0.1; P = 0.0096\nb)\nTopsoil\nSubsoil\n0\n1\n2\n3\n4\n5\n0\n1\n2\n3\n4\n5\n0.4\n0.6\n0.8\n1.0\n1.2\n1.4\nSoil warming (°C)\nBulk density (g cm−3)\nGrassland; Year\nLTW; 2013\nLTW; 2018\nMTW; 2013\nMTW; 2018 Soil warming (°C) Figure B7. Changes in bulk density with soil warming in the long-term warmed (LTW; circles) and the medium-term warmed (MTW;\ntriangles) grassland. Soil warming is expressed relative to ambient soil temperature (at 10 cm depth). Bulk density is separated for the topsoil\n(0-10 cm) and the subsoil (10-30 cm). The regression for medium-term warmed (MTW) and long-term warmed (LTW) grassland (in topsoil)\nfor both 2013 and 2018 were combined, since no soil temperature x warming-duration interaction effect, nor a main effect for warming-\nduration or sampling year was found. The 95 % confidence bounds are shown around the regression slopes. (n = 78 & 40 for topsoil and\nsubsoil respectively) 18 https://doi.org/10.5194/bg-2021-338\nPreprint. Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. Figure B8. Soil mass correction. Due to significant soil compaction (increasing bulk density (BD)) with increasing soil temperature in\nthe upper soil layers (fig. B7), a certain soil depth in unwarmed soil corresponds to ever shallower soil depths at warmer soil temperatures. Therefore, the SOC stocks were corrected for soil compaction, i.e., the corrected SOC stocks were calculated on the same mass of soil. Figure B8. Soil mass correction. Due to significant soil compaction (increasing bulk density (BD)) with increasing soil temperature in\nthe upper soil layers (fig. B7), a certain soil depth in unwarmed soil corresponds to ever shallower soil depths at warmer soil temperatures. Therefore, the SOC stocks were corrected for soil compaction, i.e., the corrected SOC stocks were calculated on the same mass of soil. For topsoil, we calculated a corrected thickness of a warmed soil layer, corresponding to the same core mass as the the For topsoil, we calculated a corrected thickness of a warmed soil layer, correspond For topsoil, we calculated a corrected thickness of a warmed soil layer, corresponding to the same core mass as the the\nunwarmed soils. Using the ratio of corrected and uncorrected layer thickness, we calculated a corrected SOC stock for the\n205\nwarmed topsoil. Appendix A: Material and methods\n125 For warmed subsoils, we calculated the corrected thickness in the same way as for topsoil, but subtracted a\nsurplus thickness of the above topsoil. The corrected subsoil SOC stock was then calculated as the sum of the surplus topsoil\nSOC stock and the SOC stock in the corrected subsoil layer. The detailed calculation method is shown below. unwarmed soils. Using the ratio of corrected and uncorrected layer thickness, we calculated a corrected SOC stock for the\n205\nwarmed topsoil. For warmed subsoils, we calculated the corrected thickness in the same way as for topsoil, but subtracted a\nsurplus thickness of the above topsoil. The corrected subsoil SOC stock was then calculated as the sum of the surplus topsoil\nSOC stock and the SOC stock in the corrected subsoil layer. The detailed calculation method is shown below. Corrections 0-10cm soil layer First, the corrected soil layer thickness of the 0-10 cm layer (C.Th0−10) was calculated for the\n10 C.Th0−10 = U.Th0−10 × U.BD0−10\nC.BD0−10\n(B1) C.Th0−10 = U.Th0−10 × U.BD0−10\nC.BD0−10\n(B1) C.Th0−10 = U.Th0−10 × U.BD0−10\nC.BD0−10 (B1) Where U.Th0−10 is the uncorrected soil layer thickness of the 0-10 cm layer (10 cm), U.BD0−10 is the uncorrected BD\nof the 0-10 cm layer (which corresponds to the BD at ambient soil temperature) and C.BD0−10 is the measured BD for the\n0-10 cm layer. Then, the corrected SOC stocks of the 0-10 depth layer (C.SOC0−10) were calculated: C.SOC0−10 = U.SOC0−10 × C.Th0−10\nU.Th0−10\n215 (B2) 19 https://doi.org/10.5194/bg-2021-338\nPreprint. Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. Where U.SOC0−10 is the uncorrected SOC stock in the 0-10 cm depth layer, and C.Th0−10/U.Th0−10 corresponds to the\nproportional thickness of the corrected layer compared to the uncorrected layer. Author contributions. BDS, NIWL, SV, JLS, LF, JTW, HW, PMVB, NV and IAJ designed the study. PG, HW, CF, NIWL, BDS, IAJ, SV,\nKVDV, EV, ZFL, SM-J, NV and MM provided the data. All authors contributed substantially to the analysis and the writing of the manuscript. Corrections 10-30 cm soil layer (only applicable to the long-term warmed grassland) First, the thickness of the surplus soil layer from the 0-10 cm layer (S.Th0−10) was calculated: First, the thickness of the surplus soil layer from the 0-10 cm layer (S.Th0−10) was calculated: S.Th0−10 = (U.Th0−10 −C.Th0−10) × U.BD0−10\nC.BD0−10\n220 (B3) S.Th0−10 = (U.Th0−10 −C.Th0−10) × U.BD0−10\nC.BD0−10\n(B3)\n220\nThe second term is a correction factor for the soil compaction of the surplus soil layer. Then, the corrected thickness of the\n10-30 cm soil layer, not yet taking the surplus soil sampled from the 0-10 cm layer (S.Th0−10) into account, (c.Th10−30) was\ncalculated: C.BD0−10 The second term is a correction factor for the soil compaction of the surplus soil layer. Then, the corrected thickness of the\n10-30 cm soil layer, not yet taking the surplus soil sampled from the 0-10 cm layer (S.Th0−10) into account, (c.Th10−30) was\ncalculated: c.Th10−30 = U.Th10−30 × U.BD10−30\nC.BD10−30\n(B4) c.Th10−30 = U.Th10−30 × U.BD10−30\nC.BD10−30 (B4) Where U.Th10−30 is the uncorrected soil layer thickness of the 10-30 cm layer (20 cm), U.BD10−30 is the uncorrected BD\n225\nof the 10-30 cm layer (which corresponds to the BD at ambient soil temperature) and C.BD10−30 is the measured BD for the\n10-30 cm layer. Where U.Th10−30 is the uncorrected soil layer thickness of the 10-30 cm layer (20 cm), U.BD10−30 is the uncorrected BD\n225\nof the 10-30 cm layer (which corresponds to the BD at ambient soil temperature) and C.BD10−30 is the measured BD for the\n10-30 cm layer. Where U.Th10−30 is the uncorrected soil layer thickness of the 10-30 cm layer (20 cm), U.BD10−30 is the uncorrected BD\n225\nof the 10-30 cm layer (which corresponds to the BD at ambient soil temperature) and C.BD10−30 is the measured BD for the\n10-30 cm layer. Subsequently, we took into account the thickness of the surplus soil sampled from the 0-10 cm layer to calculate the final\ncorrected soil thickness of the 10-30 cm soil layer (C.Th10−30). Hence, C.Th10−30 is the part of the 10-30 cm layer that Subsequently, we took into account the thickness of the surplus soil sampled from the 0-10 cm layer to calculate the final\ncorrected soil thickness of the 10-30 cm soil layer (C.Th10−30). Corrections 10-30 cm soil layer (only applicable to the long-term warmed grassland) Hence, C.Th10−30 is the part of the 10-30 cm layer that Subsequently, we took into account the thickness of the surplus soil sampled from the 0-10 cm layer to calculate the final\ncorrected soil thickness of the 10-30 cm soil layer (C.Th10−30). Hence, C.Th10−30 is the part of the 10-30 cm layer that\nremains after (i) correcting for soil compaction and (ii) subtracting the thickness of the surplus soil sampled at the 0-10 cm\n230\nlayer: corrected soil thickness of the 10 30 cm soil layer (C.Th10−30). Hence, C.Th10−30 is the part of the 10 30 cm layer that\nremains after (i) correcting for soil compaction and (ii) subtracting the thickness of the surplus soil sampled at the 0-10 cm\n230\nlayer: remains after (i) correcting for soil compaction and (ii) subtracting the thickness of the surplus soil sampled at the 0-10 cm\n230\nlayer: C.Th10−30 = c.Th10−30 −S.Th0−10 Th10−30 = c.Th10−30 −S.Th0−10 h10−30 −S.Th0−10 (B5) h0−10\n(B5) ected SOC stock for the 10-30 cm layer (C.SOC10−30) was calculated: Subsequently, the corrected SOC stock for the 10-30 cm layer (C.SOC10−30) was calculated: OC stock for the 10-30 cm layer (C.SOC10−30) was calculated: Subsequently, the corrected SOC stock for the 10-30 cm layer (C.SOC10−30) was calculated: Subsequently, the corrected SOC stock for the 10-30 cm layer (C.SOC10−30) was calculated: Subsequently, the corrected SOC stock for the 10-30 cm layer (C.SOC10−30) was calculated: C.SOC10−30 = (U.SOC0−10 −C.SOC0−10) + U.SOC10−30 × C.Th10−30\nU.Th10−30\n(B6) C.SOC10−30 = (U.SOC0−10 −C.SOC0−10) + U.SOC10−30 × C.Th10−30\nU.Th10−30 (B6) Where U.SOC10−30 is the uncorrected SOC stock in the 10-30 cm depth layer, and C.Th10−30/U.Th10−30 corresponds to\n235\nthe proportional thickness of the corrected layer compared to the uncorrected layer. Where U.SOC10−30 is the uncorrected SOC stock in the 10-30 cm depth layer, and C.Th10−30/U.Th10−30 corresponds to\n235\nthe proportional thickness of the corrected layer compared to the uncorrected layer. Where U.SOC10−30 is the uncorrected SOC stock in the 10-30 cm depth layer, and C.Th10−30/U.Th10−30 corresponds to\n35\nthe proportional thickness of the corrected layer compared to the uncorrected layer. Data availability. All data used in this manuscript has been made available on Zenodo. DOI: 10.5281/zenodo.4745479 Data availability. All data used in this manuscript has been made available on Zenodo. DOI: 10.5281/zenodo.4745479 Data availability. All data used in this manuscript has been made available on Zenodo. DOI: 10.5281/zenodo.4745479 Author contributions. Corrections 10-30 cm soil layer (only applicable to the long-term warmed grassland) BDS, NIWL, SV, JLS, LF, JTW, HW, PMVB, NV and IAJ designed the study. PG, HW, CF, NIWL, BDS, IAJ, SV,\nKVDV, EV, ZFL, SM-J, NV and MM provided the data. All authors contributed substantially to the analysis and the writing of the manuscript. Author contributions. BDS, NIWL, SV, JLS, LF, JTW, HW, PMVB, NV and IAJ designed the study. PG, HW, CF, NIWL, BDS, IAJ, SV,\nKVDV, EV, ZFL, SM-J, NV and MM provided the data. All authors contributed substantially to the analysis and the writing of the manuscript. 20 https://doi.org/10.5194/bg-2021-338\nPreprint. Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. Competing interests. The authors declare that they have no conflict of interest. 240 Competing interests. The authors declare that they have no conflict of interest. 240 Competing interests. The authors declare that they have no conflict of interest. 240 Acknowledgements. This research was supported by a joint Fonds Wetenschappelijk Onderzoek Flanders (FWO) and Fonds zur Förderung\nder wissenschaftlichen Forschung (FWF) grant with nos. FWO-G0F2217N & FWF-I-3237, awarded to I.A.J and M.B., the European Re-\nsearch Council Synergy grant 610028 (IMBALANCE-P), and the Research Council of the University of Antwerp (FORHOT TOP-BOF\nproject). This work contributes to the FSC-Sink, CAR-ES and the ClimMani COST Action (ES1308). Reykir - the Icelandic state gardening\nschool, Keldnaholt – the Agricultural University of Iceland and Mogilsá – the Icelandic Forest Research, provided logistical support for\n245\nthe present study. Further, we thank the Lorentz Center in Leiden. We thank Iris Janssens, Jochen Janssens, Inge Van De Putte, Maxime\nSepelie, Jana Vynckier, Alexander Meire, Lieven Michielsen, Freja Dreesen, Sebastien Leys, Elín Gudmundsdóttir, Annemie Vinck, Paul\nLeblans, Kwinten Leblans, Sigvatur, Már Gudmundsson, Elías Óskarsson and Simon Arnar Pálsson for their help in the field. We thank Brita\nBerglund, Baldur Vigfusson, Nadine Calluy, Tom Van Der Spiet, Anne Cools, Marijke Van den Bruel, Els Oosterbos, Saad El-Rawi, Miguel\nPortillo Estrada and Wannes Kiebooms for their assistance with the lab analyses. 250 school, Keldnaholt – the Agricultural University of Iceland and Mogilsá – the Icelandic Forest Research, provided logistical support for\n245\nthe present study. Further, we thank the Lorentz Center in Leiden. Corrections 10-30 cm soil layer (only applicable to the long-term warmed grassland) We thank Iris Janssens, Jochen Janssens, Inge Van De Putte, Maxime\nSepelie, Jana Vynckier, Alexander Meire, Lieven Michielsen, Freja Dreesen, Sebastien Leys, Elín Gudmundsdóttir, Annemie Vinck, Paul\nLeblans, Kwinten Leblans, Sigvatur, Már Gudmundsson, Elías Óskarsson and Simon Arnar Pálsson for their help in the field. We thank Brita\nBerglund, Baldur Vigfusson, Nadine Calluy, Tom Van Der Spiet, Anne Cools, Marijke Van den Bruel, Els Oosterbos, Saad El-Rawi, Miguel\nPortillo Estrada and Wannes Kiebooms for their assistance with the lab analyses. 250 21 21 https://doi.org/10.5194/bg-2021-338\nPreprint. Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. https://doi.org/10.5194/bg-2021-338\nPreprint. Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. References Davidson, E. A. and Janssens, I. A.: Temperature Sensitivity of Soil Carbon Decomposition and Feedbacks to Climate Change, Nature, 440,\n165–173, https://doi.org/10.1038/nature04514, 2006. Davidson, E. A. and Janssens, I. A.: Temperature Sensitivity of Soil Carbon Decomposition and Feedbacks to Climate Change, Nature, 440,\n165–173, https://doi.org/10.1038/nature04514, 2006. De Boeck, H. J., Vicca, S., Roy, J., Nijs, I., Milcu, A., Kreyling, J., Jentsch, A., Chabbi, A., Campioli, M., Callaghan,\nT., Beierkuhnlein, C., and Beier, C.: Global Change Experiments: Challenges and Opportunities, BioScience, 65, 922–931,\n275\nhttps://doi.org/10.1093/biosci/biv099, 2015. ,\n,\n,\n,\ny,\n,\nj ,\n,\n,\n,\ny\ng,\n,\n,\n,\n,\n,\np\n,\n,\ng\n,\nT., Beierkuhnlein, C., and Beier, C.: Global Change Experiments: Challenges and Opportunities, BioScience, 65, 922–931,\n275\nhttps://doi.org/10.1093/biosci/biv099, 2015. Edlinger, A.: Changes of belowground processes of the carbon and nitrogen cycle in response to soil warming. Will global warming increase\nthe turnover of soil C and N above plant demands and lead to increased leaching losses Master’s thesis M Sc thesis Copenhagen T., Beierkuhnlein, C., and Beier, C.: Global Change Experiments: Challenges and Opportunities, BioScience, 65, 922–931,\n275\nhttps://doi.org/10.1093/biosci/biv099, 2015. Edlinger, A.: Changes of belowground processes of the carbon and nitrogen cycle in response to soil warming. Will global warming increase\nthe turnover of soil C and N above plant demands and lead to increased leaching losses, Master’s thesis, M. Sc. thesis. Copenhagen\nUniversity, Denmark & University of Natural Resources and Applied Sciences (BOKU), http://forhot.is/wp-content/uploads/2018/01/ https://doi.org/10.1093/biosci/biv099, 2015. Edlinger, A.: Changes of belowground processes of the carbon and nitrogen cycle in response to soil warming. Will global warming increase\nthe turnover of soil C and N above plant demands and lead to increased leaching losses, Master’s thesis, M. Sc. thesis. Copenhagen\nUniversity, Denmark & University of Natural Resources and Applied Sciences (BOKU), http://forhot.is/wp-content/uploads/2018/01/ Edlinger, A.: Changes of belowground processes of the carbon and nitrogen cycle in response to soil warming. Will global warming increase\nthe turnover of soil C and N above plant demands and lead to increased leaching losses, Master’s thesis, M. Sc. thesis. Copenhagen\nUniversity, Denmark & University of Natural Resources and Applied Sciences (BOKU), http://forhot.is/wp-content/uploads/2018/01/\nEdlinger-Anna-Theresa.-2016.-Changes-of-belowground-processes-of-the-carbon-and-nitrogen.xps, 2016. 280 Edlinger-Anna-Theresa.-2016.-Changes-of-belowground-processes-of-the-carbon-and-nitrogen.xps, 2016. 280\nEllert, B. H. and Bettany, J. R.: Calculation of Organic Matter and Nutrients Stored in Soils under Contrasting Management Regimes,\nCanadian Journal of Soil Science, 75, 529–538, 1995. Eyring, V., Bony, S., Meehl, G. A., Senior, C. References Abramoff, R. Z., Torn, M. S., Georgiou, K., Tang, J., and Riley, W. J.: Soil Organic Matter Temperature Sensitivity Cannot Be Directly\nInferred from Spatial Gradients, Global Biogeochemical Cycles, 33, 761–776, 2019. Abramoff, R. Z., Torn, M. S., Georgiou, K., Tang, J., and Riley, W. J.: Soil Organic Matter Temperature Sensitivity Cannot Be Directly\nInferred from Spatial Gradients, Global Biogeochemical Cycles, 33, 761–776, 2019. Allison, S. D., Wallenstein, M. D., and Bradford, M. A.: Soil-Carbon Response to Warming Dependent on Microbial Physiology, Nature\nGeoscience, 3, 336–340, https://doi.org/10.1038/ngeo846, 2010. 255\nArnalds, O.: The Soils of Iceland, World Soils Book Series, Springer Netherlands, Dordrecht, https://doi.org/10.1007/978-94-017-9621-7,\n2015. Baillie, I.: Soil survey staff 1999, soil taxonomy: a basic system of soil classification for making and interpreting soil surveys, agricultural\nhandbook 436, Natural Resources Conservation Service, USDA, Washington DC, USA, pp. 869, 2001. Allison, S. D., Wallenstein, M. D., and Bradford, M. A.: Soil-Carbon Response to Warming Dependent on Microbial Physiology, Nature\nGeoscience, 3, 336–340, https://doi.org/10.1038/ngeo846, 2010. 255\nArnalds, O.: The Soils of Iceland, World Soils Book Series, Springer Netherlands, Dordrecht, https://doi.org/10.1007/978-94-017-9621-7,\n2015 Allison, S. D., Wallenstein, M. D., and Bradford, M. A.: Soil-Carbon Response to Warming Dependent on Microbial Physiology, Nature\nGeoscience, 3, 336–340, https://doi.org/10.1038/ngeo846, 2010. 5 Arnalds, O.: The Soils of Iceland, World Soils Book Series, Springer Netherlands, Dordrecht, https://doi.org/10.1007/978-94-017-9621-7,\n2015. Baillie, I.: Soil survey staff 1999, soil taxonomy: a basic system of soil classification for making and interpreting soil surveys, agricultural\nhandbook 436, Natural Resources Conservation Service, USDA, Washington DC, USA, pp. 869, 2001. Arnalds, O.: The Soils of Iceland, World Soils Book Series, Springer Netherlands, Dordrecht, https://doi.org/10.1007/978-94-017-9621-7,\n2015. Baillie, I.: Soil survey staff 1999, soil taxonomy: a basic system of soil classification for making and interpreting soil surveys, agricultural\nhandbook 436, Natural Resources Conservation Service, USDA, Washington DC, USA, pp. 869, 2001. Bárcena, T. G., Gundersen, P., and Vesterdal, L.: Afforestation Effects on SOC in Former Cropland: Oak and Spruce Chronosequences\n260\nResampled after 13 Years, Global Change Biology, 20, 2938–2952, 2014. Bárcena, T. G., Gundersen, P., and Vesterdal, L.: Afforestation Effects on SOC in Former Cropland: Oak and Spruce Chronosequences\n260\nResampled after 13 Years, Global Change Biology, 20, 2938–2952, 2014. Batjes, N.: Harmonized Soil Property Values for Broad-Scale Modelling (WISE30sec) with Estimates of Global Soil Carbon Stocks, Geo-\nderma, 269, 61–68, https://doi.org/10.1016/j.geoderma.2016.01.034, 2016. References Batjes, N.: Harmonized Soil Property Values for Broad-Scale Modelling (WISE30sec) with Estimates of Global Soil Carbon Stocks, Geo-\nderma, 269, 61–68, https://doi.org/10.1016/j.geoderma.2016.01.034, 2016. Bradford, M. A., McCulley, R. L., Crowther, T. W., Oldfield, E. E., Wood, S. A., and Fierer, N.: Cross-Bi Bradford, M. A., McCulley, R. L., Crowther, T. W., Oldfield, E. E., Wood, S. A., and Fierer, N.: Cross-Biome Patterns in Soil Microbial\nRespiration Predictable from Evolutionary Theory on Thermal Adaptation, Nature ecology & evolution, 3, 223, 2019. 265 Respiration Predictable from Evolutionary Theory on Thermal Adaptation, Nature ecology & evolution, 3, 223, 2019. 265\nCrowther, T. W., Todd-Brown, K. E. O., Rowe, C. W., Wieder, W. R., Carey, J. C., Machmuller, M. B., Snoek, B. L., Fang, S., Zhou, G.,\nAllison, S. D., Blair, J. M., Bridgham, S. D., Burton, A. J., Carrillo, Y., Reich, P. B., Clark, J. S., Classen, A. T., Dijkstra, F. A., Elberling, B.,\nEmmett, B. A., Estiarte, M., Frey, S. D., Guo, J., Harte, J., Jiang, L., Johnson, B. R., Kröel-Dulay, G., Larsen, K. S., Laudon, H., Lavallee,\nJ. M., Luo, Y., Lupascu, M., Ma, L. N., Marhan, S., Michelsen, A., Mohan, J., Niu, S., Pendall, E., Peñuelas, J., Pfeifer-Meister, L., Poll, Crowther, T. W., Todd-Brown, K. E. O., Rowe, C. W., Wieder, W. R., Carey, J. C., Machmuller, M. B., Snoek, B. L., Fang, S., Zhou, G., Emmett, B. A., Estiarte, M., Frey, S. D., Guo, J., Harte, J., Jiang, L., Johnson, B. R., Kröel-Dulay, G., Larsen, K. S., Laudon, H., Lavallee, M., Frey, S. D., Guo, J., Harte, J., Jiang, L., Johnson, B. R., Kröel-Dulay, G., Larsen, K. S., Laudon, H., Lavalle M., Ma, L. N., Marhan, S., Michelsen, A., Mohan, J., Niu, S., Pendall, E., Peñuelas, J., Pfeifer-Meister, L., Po J. M., Luo, Y., Lupascu, M., Ma, L. N., Marhan, S., Michelsen, A., Mohan, J., Niu, S., Pendall, E., Peñuelas, J., Pfeifer-Meister, L., Poll, C., Reinsch, S., Reynolds, L. L., Schmidt, I. K., Sistla, S., Sokol, N. W., Templer, P. H., Treseder, K. K., Welker, J. M., and Bradford, M. A.:\n270\nQuantifying Global Soil Carbon Losses in Response to Warming, Nature, 540, 104–108, https://doi.org/10.1038/nature20150, 2016. C., Reinsch, S., Reynolds, L. L., Schmidt, I. K., Sistla, S., Sokol, N. W., Templer, P. H., Treseder, K. K., Welker, J. M., and Bradford, M. A.:\n270\nQuantifying Global Soil Carbon Losses in Response to Warming, Nature, 540, 104–108, https://doi.org/10.1038/nature20150, 2016. Fontaine, S., Barot, S., Barré, P., Bdioui, N., Mary, B., and Rumpel, C.: Stability of organic carbon in deep soil layers controlled by fresh\ncarbon supply, Nature, 450, 277–280, 2007. References Contribution of Working Group I to the Fifth Assessment Report of the Intergov-\nernmental Panel on Climate Change, Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 2013. 295 IPCC: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergov\nernmental Panel on Climate Change, Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 2013. 295 ernmental Panel on Climate Change, Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 2013. 295\nJenkinson D S Adams D E and Wild A : Model Estimates of CO 2 Emissions from Soil in Response to Global Warming Nature 351 ernmental Panel on Climate Change, Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 2013. 295\nJenkinson, D. S., Adams, D. E., and Wild, A.: Model Estimates of CO 2 Emissions from Soil in Response to Global Warming, Nature, 351,\n304–306, 1991. ernmental Panel on Climate Change, Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 2013. 295\nJenkinson, D. S., Adams, D. E., and Wild, A.: Model Estimates of CO 2 Emissions from Soil in Response to Global Warming, Nature, 351,\n304–306, 1991. Jia, J., Cao, Z., Liu, C., Zhang, Z., Lin, L., Wang, Y., Haghipour, N., Wacker, L., Bao, H., Dittmar, T., Simpson, M. J., Yang, H., Crowther,\nT. W., Eglinton, T. I., He, J.-S., and Feng, X.: Climate Warming Alters Subsoil but Not Topsoil Carbon Dynamics in Alpine Grassland, ernmental Panel on Climate Change, Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 2013. 295\nJenkinson, D. S., Adams, D. E., and Wild, A.: Model Estimates of CO 2 Emissions from Soil in Response to Global Warming, Nature, 351,\n304–306, 1991. Jia J Cao Z Liu C Zhang Z Lin L Wang Y Haghipour N Wacker L Bao H Dittmar T Simpson M J Yang H Crowther Jenkinson, D. S., Adams, D. E., and Wild, A.: Model Estimates of CO 2 Emissions from Soil in Response to Global Warming, Nature, 351,\n304–306, 1991. Jia, J., Cao, Z., Liu, C., Zhang, Z., Lin, L., Wang, Y., Haghipour, N., Wacker, L., Bao, H., Dittmar, T., Simpson, M. J., Yang, H., Crowther,\nT. W., Eglinton, T. I., He, J.-S., and Feng, X.: Climate Warming Alters Subsoil but Not Topsoil Carbon Dynamics in Alpine Grassland,\nGlobal Change Biology, 25, 4383–4393, https://doi.org/10.1111/gcb.14823, 2019. References A., Stevens, B., Stouffer, R. J., and Taylor, K. E.: Overview of the Coupled Model In-\ntercomparison Project Phase 6 (CMIP6) Experimental Design and Organization, Geoscientific Model Development, 9, 1937–1958, Ellert, B. H. and Bettany, J. R.: Calculation of Organic Matter and Nutrients Stored in Soils under Contrasting Management Regimes,\nCanadian Journal of Soil Science, 75, 529–538, 1995. Canadian Journal of Soil Science, 75, 529–538, 1995. Eyring, V., Bony, S., Meehl, G. A., Senior, C. A., Stevens, B., Stouffer, R. J., and Taylor, K. E.: Overview of the Coupled Model In-\ntercomparison Project Phase 6 (CMIP6) Experimental Design and Organization, Geoscientific Model Development, 9, 1937–1958,\nh\n//d i\n/10 5194/\nd 9 193\n2016 2016\n285 Eyring, V., Bony, S., Meehl, G. A., Senior, C. A., Stevens, B., Stouffer, R. J., and Taylor, K. E.: Overview of the Coupled Model In-\ntercomparison Project Phase 6 (CMIP6) Experimental Design and Organization, Geoscientific Model Development, 9, 1937–1958,\nhttps://doi.org/10.5194/gmd-9-1937-2016, 2016. 285 https://doi.org/10.5194/gmd-9-1937-2016, 2016. 285\nFontaine, S., Barot, S., Barré, P., Bdioui, N., Mary, B., and Rumpel, C.: Stability of organic carbon in deep soil layers controlled by fresh\ncarbon supply, Nature, 450, 277–280, 2007. 22 https://doi.org/10.5194/bg-2021-338\nPreprint. Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. https://doi.org/10.5194/bg-2021-338\nPreprint. Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. Halldórsson, B. and Sigbjörnsson, R.: The Mw6.3 Ölfus Earthquake at 15:45 UTC on 29 May 2008 in South Iceland: ICEARRAY Strong-\nMotion Recordings, Soil Dynamics and Earthquake Engineering, 29, 1073–1083, https://doi.org/10.1016/j.soildyn.2008.12.006, 2009. Hanson, P. J., Griffiths, N. A., Iversen, C. M., Norby, R. J., Sebestyen, S. D., Phillips, J. R., Chanton, J. P., Kolka, R. K., Malhotra, A.,\n290\nOleheiser, K. C., et al.: Rapid net carbon loss from a whole-ecosystem warmed Peatland, AGU Advances, 1, e2020AV000 163, 2020. Ilyina, T. and Friedlingstein, P.: WCRP Grand Challenge-Carbon Feedbacks in the Climate System, Tech. rep., Tech. rep., WCRP,\nhttps://www. wcrp-climate. org/JSC37/Documents . .., 2016. IPCC Cli\nCh\n2013 Th Ph\ni\nl S i\nB\ni\nC\nib i\nf W ki\nG\nI\nh Fif h A\nR\nf h I IPCC: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Asse IPCC: Climate Change 2013: The Physical Science Basis. References A.: Geothermally Warmed Soils Reveal Persistent Increases in the Respiratory Costs of Soil Microbes Contributing to Substantial C Marañón-Jiménez, S., Soong, J. L., Leblans, N. I. W., Sigurdsson, B. D., Peñuelas, J., Richter, A., Asensio, D., Fransen, E., and Janssens,\nI. A.: Geothermally Warmed Soils Reveal Persistent Increases in the Respiratory Costs of Soil Microbes Contributing to Substantial C\nLosses, Biogeochemistry, 138, 245–260, https://doi.org/10.1007/s10533-018-0443-0, 2018. 310 Losses, Biogeochemistry, 138, 245–260, https://doi.org/10.1007/s10533-018-0443-0, 2018. 310\nMelillo, J. M., Frey, S. D., DeAngelis, K. M., Werner, W. J., Bernard, M. J., Bowles, F. P., Pold, G., Knorr, M. A., and Grandy, A. S.:\nLong-Term Pattern and Magnitude of Soil Carbon Feedback to the Climate System in a Warming World, Science, 358, 101–105,\nhttps://doi.org/10.1126/science.aan2874, 2017. Nishina, K., Ito, A., Beerling, D. J., Cadule, P., Ciais, P., Clark, D. B., Falloon, P., Friend, A. D., Kahana, R., Kato, E., Keribin, R., Lucht, Losses, Biogeochemistry, 138, 245 260, https://doi.org/10.1007/s10533 018 0443 0, 2018. 310\nMelillo, J. M., Frey, S. D., DeAngelis, K. M., Werner, W. J., Bernard, M. J., Bowles, F. P., Pold, G., Knorr, M. A., and Grandy, A. S.:\nLong-Term Pattern and Magnitude of Soil Carbon Feedback to the Climate System in a Warming World, Science, 358, 101–105,\nhttps://doi.org/10.1126/science.aan2874, 2017. Nishina, K., Ito, A., Beerling, D. J., Cadule, P., Ciais, P., Clark, D. B., Falloon, P., Friend, A. D., Kahana, R W., Lomas, M., Rademacher, T. T., Pavlick, R., Schaphoff, S., Vuichard, N., Warszawaski, L., and Yokohata, T.: Quantifying Uncer-\n315\ntainties in Soil Carbon Responses to Changes in Global Mean Temperature and Precipitation, Earth System Dynamics, 5, 197–209,\nhttps://doi.org/Nishina, K.; Ito, A.; Beerling, D.J.; Cadule, P.; Ciais, P.; Clark, D.B.; Falloon, P.; Friend, A.D.; Kahana, R.; Kato, E.;\nKeribin, R.; Lucht, W.; Lomas, M.; Rademacher, T.T.; Pavlick, R.; Schaphoff, S.; Vuichard, N.; Warszawaski, L.; Yokohata, T.. 2014\nQuantifying uncertainties in soil carbon responses to changes in global mean temperature and precipitation. Earth System Dynamics, 5 (1). 197-209. https://doi.org/10.5194/esd-5-197-2014 <https://doi.org/10.5194/esd-5-197-2014>, 2014. 320\nO’Gorman, E. J., Benstead, J. P., Cross, W. F., Friberg, N., Hood, J. M., Johnson, P. W., Sigurdsson, B. D., and Woodward, G.: Climate\nChange and Geothermal Ecosystems: Natural Laboratories, Sentinel Systems, and Future Refugia, Global Change Biology, 20, 3291–\n3299, https://doi.org/10.1111/gcb.12602, 2014. Osher, L. J., Matson, P. A., and Amundson, R.: Effect of Land Use Change on Soil Carbon in Hawaii, Biogeochemistry, 65, 20, 2003. 320 (1). 197 209. References 300 Jia, J., Cao, Z., Liu, C., Zhang, Z., Lin, L., Wang, Y., Haghipour, N., Wacker, L., Bao, H., Dittmar, T., Simpson, M. J., Yang, H., Crowther,\nT. W., Eglinton, T. I., He, J.-S., and Feng, X.: Climate Warming Alters Subsoil but Not Topsoil Carbon Dynamics in Alpine Grassland, Global Change Biology, 25, 4383–4393, https://doi.org/10.1111/gcb.14823, 2019. 300\nLeblans, N. I. W., Sigurdsson, B. D., Vicca, S., Fu, Y., Penuelas, J., and Janssens, I. A.: Phenological Responses of Icelandic Subarctic Grass-\nlands to Short-Term and Long-Term Natural Soil Warming, Global Change Biology, 23, 4932–4945, https://doi.org/10.1111/gcb.13749,\n2017. Lin, W., Li, Y., Yang, Z., Giardina, C. P., Xie, J., Chen, S., Lin, C., Kuzyakov, Y., and Yang, Y.: Warming Exerts Greater Leblans, N. I. W., Sigurdsson, B. D., Vicca, S., Fu, Y., Penuelas, J., and Janssens, I. A.: Phenological Responses of Icelandic Subarctic Grass-\nlands to Short-Term and Long-Term Natural Soil Warming, Global Change Biology, 23, 4932–4945, https://doi.org/10.1111/gcb.13749,\n2017. 2017. Lin, W., Li, Y., Yang, Z., Giardina, C. P., Xie, J., Chen, S., Lin, C., Kuzyakov, Y., and Yang, Y.: Warming Exerts Greater\nImpacts on Subsoil than Topsoil CO2 Efflux in a Subtropical Forest, Agricultural and Forest Meteorology, 263, 137–146,\n305 Lin, W., Li, Y., Yang, Z., Giardina, C. P., Xie, J., Chen, S., Lin, C., Kuzyakov, Y., and Yang, Y.: Warming Exerts Greater\nImpacts on Subsoil than Topsoil CO2 Efflux in a Subtropical Forest, Agricultural and Forest Meteorology, 263, 137–146,\n305\nhttps://doi.org/10.1016/j.agrformet.2018.08.014, 2018. g\ny\ng\ng\nImpacts on Subsoil than Topsoil CO2 Efflux in a Subtropical Forest, Agricultural and Forest Meteorology, 263, 137–146,\n305\nhttps://doi.org/10.1016/j.agrformet.2018.08.014, 2018. Á Impacts on Subsoil than Topsoil CO2 Efflux in a Subtropical Forest, Agricultural and Forest Meteorology, 263, 137–146,\n305\nhttps://doi.org/10.1016/j.agrformet.2018.08.014, 2018. Magnússon, Á. and Vídalín, P.: Jarðabók. Annað bindi [The farm register of Iceland], 1708. Marañón-Jiménez, S., Soong, J. L., Leblans, N. I. W., Sigurdsson, B. D., Peñuelas, J., Richter, A., Asensio, D., Fransen, E., and Janssens,\nI. A.: Geothermally Warmed Soils Reveal Persistent Increases in the Respiratory Costs of Soil Microbes Contributing to Substantial C Magnússon, Á. and Vídalín, P.: Jarðabók. Annað bindi [The farm register of Iceland], 1708. Magnússon, Á. and Vídalín, P.: Jarðabók. Annað bindi [The farm register of Iceland], 1708. Marañón-Jiménez, S., Soong, J. L., Leblans, N. I. W., Sigurdsson, B. D., Peñuelas, J., Richter, A., Asensio, D., Fransen, E., and Janssens,\nI. Osher, L. J., Matson, P. A., and Amundson, R.: Effect of Land Use Change on Soil Carbon in Hawaii, Biogeochemistry, 65, 20, 2003. References E., Holmstrup, M., Ilieva-Makulec, K.,\nK\nT M\ni\ndó i\nB\nl\nG\nh\nl E\nN\nl Cli\nCh\nE\ni\nTh F\nH\nR\nh Si\ni Soong, J. L., Castanha, C., Hicks Pries, C. E., Ofiti, Porras, R. C., and Riley: Five Years of Whole-Soil Warming Led to Loss of Subsoil\n345\nCarbon Stocks and Increased CO2 Efflux, Science Advances, 0, 2020a. Soong, J. L., Phillips, C. L., Ledna, C., Koven, C. D., and Torn, M. S.: CMIP5 models predict rapid and deep soil warming over the 21st Soong, J. L., Castanha, C., Hicks Pries, C. E., Ofiti, Porras, R. C., and Riley: Five Years of Whole-Soil Warming Led to Loss of Subsoil\n345\nCarbon Stocks and Increased CO2 Efflux, Science Advances, 0, 2020a. Soong, J. L., Phillips, C. L., Ledna, C., Koven, C. D., and Torn, M. S.: CMIP5 models predict rapid and deep soil warming over the 21st\ncentury, Journal of Geophysical Research: Biogeosciences, 125, e2019JG005 266, 2020b. Tian, Q., Yang, X., Wang, X., Liao, C., Li, Q., Wang, M., Wu, Y., and Liu, F.: Microbial Community Mediated Response of Or- Soong, J. L., Castanha, C., Hicks Pries, C. E., Ofiti, Porras, R. C., and Riley: Five Years of Whole-Soil Warming Led to Loss of Subsoil\n345\nCarbon Stocks and Increased CO2 Efflux, Science Advances, 0, 2020a. Soong, J. L., Phillips, C. L., Ledna, C., Koven, C. D., and Torn, M. S.: CMIP5 models predict rapid and deep soil warming over the 21st\ncentury, Journal of Geophysical Research: Biogeosciences, 125, e2019JG005 266, 2020b. Tian, Q., Yang, X., Wang, X., Liao, C., Li, Q., Wang, M., Wu, Y., and Liu, F.: Microbial Community Mediated Response of Or- Q., Yang, X., Wang, X., Liao, C., Li, Q., Wang, M., Wu, Y., and Liu, F.: Microbial Community Med Tian, Q., Yang, X., Wang, X., Liao, C., Li, Q., Wang, M., Wu, Y., and Liu, F.: Microbial Community Mediated Response of Or\nganic Carbon Mineralization to Labile Carbon and Nitrogen Addition in Topsoil and Subsoil, Biogeochemistry, 128, 125–139,\n350\nhttps://doi.org/10.1007/s10533-016-0198-4, 2016. ganic Carbon Mineralization to Labile Carbon and Nitrogen Addition in Topsoil and Subsoil, Biogeochemistry, 128, 125–139,\n350\nhttps://doi.org/10.1007/s10533-016-0198-4, 2016. Todd-Brown, K. E., Zheng, B., and Crowther, T. W.: Field-Warmed Soil Carbon Changes Imply High 21st-Century Modeling Uncertainty,\nBi\ni\n15 3659 3671 2018 https://doi.org/10.1007/s10533-016-0198-4, 2016. Todd-Brown, K. E., Zheng, B., and Crowther, T. References https://doi.org/10.5194/esd 5 197 2014 <https://doi.org/10.5194/esd 5 197 2014>, 2014. 320\nO’Gorman, E. J., Benstead, J. P., Cross, W. F., Friberg, N., Hood, J. M., Johnson, P. W., Sigurdsson, B. D., and Woodward, G.: Climate\nChange and Geothermal Ecosystems: Natural Laboratories, Sentinel Systems, and Future Refugia, Global Change Biology, 20, 3291–\n3299, https://doi.org/10.1111/gcb.12602, 2014. ,\np\ng\ng\n,\nOsher, L. J., Matson, P. A., and Amundson, R.: Effect of Land Use Change on Soil Carbon in Hawaii, Biogeochemistry, 65, 20, 2003. 23 https://doi.org/10.5194/bg-2021-338\nPreprint. Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. https://doi.org/10.5194/bg-2021-338\nPreprint. Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D., and R Core Team: nlme: Linear and Nonlinear Mixed Effects Models, https://CRAN.R-project. 325\norg/package=nlme, r package version 3.1-152, 2021. Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D., and R Core Team: nlme: Linear and Nonlinear Mixed Effects Models, https://CRAN.R-project. 325\norg/package=nlme, r package version 3.1-152, 2021. Poeplau, C., Sigurðsson, P., and Sigurdsson, B. D.: Depletion of Soil Carbon and Aggregation after Strong Warming of a Subarctic Andosol\nunder Forest and Grassland Cover, SOIL, 6, 115–129, https://doi.org/10.5194/soil-6-115-2020, 2020. R Development Core Team, R.: R: A Language and Environment for Statistical Computing, R foundation for statistical computing Vienna,\nAustria, 2011. 330\nRadujkovi´c, D., Verbruggen, E., Sigurdsson, B. D., Leblans, N. I., Janssens, I. A., Vicca, S., and Weedon, J. T.: Prolonged Exposure Does\nNot Increase Soil Microbial Community Compositional Response to Warming along Geothermal Gradients, FEMS microbiology ecology,\n94, fix174, 2018. Rumpel, C. and Kögel-Knabner, I.: Deep soil organic matter—a key but poorly understood component of terrestrial C cycle, Plant and soil,\n338 143 158 2011\n335 Poeplau, C., Sigurðsson, P., and Sigurdsson, B. D.: Depletion of Soil Carbon and Aggregation after Strong Warming of a Subarctic Andosol\nunder Forest and Grassland Cover, SOIL, 6, 115–129, https://doi.org/10.5194/soil-6-115-2020, 2020. under Forest and Grassland Cover, SOIL, 6, 115–129, https://doi.org/10.5194/soil-6-115-2020, 2020. R Development Core Team, R.: R: A Language and Environment for Statistical Computing, R foundation for statistical computing Vienna,\nAustria, 2011. 330 R Development Core Team, R.: R: A Language and Environment for Statistical Computing, R foundation for statistical computing Vienna,\nAustria, 2011. 330 Austria, 2011. 330\nRadujkovi´c, D., Verbruggen, E., Sigurdsson, B. D., Leblans, N. I., Janssens, I. A., Vicca, S., and Weedon, J. References T.: Prolonged Exposure Does\nNot Increase Soil Microbial Community Compositional Response to Warming along Geothermal Gradients, FEMS microbiology ecology,\n94, fix174, 2018. Rumpel, C. and Kögel-Knabner, I.: Deep soil organic matter—a key but poorly understood component of terrestrial C cycle, Plant and soil, Radujkovi´c, D., Verbruggen, E., Sigurdsson, B. D., Leblans, N. I., Janssens, I. A., Vicca, S., and Weedon, J. T.: Prolonged Exposure Does\nNot Increase Soil Microbial Community Compositional Response to Warming along Geothermal Gradients, FEMS microbiology ecology,\n94, fix174, 2018. Rumpel, C. and Kögel-Knabner, I.: Deep soil organic matter—a key but poorly understood component of terrestrial C cycle, Plant and soil,\n338, 143–158, 2011. 335 338, 143–158, 2011. 335\nScharlemann, J. P., Tanner, E. V., Hiederer, R., and Kapos, V.: Global Soil Carbon: Understanding and Managing the Largest Terrestrial\nCarbon Pool, Carbon Management, 5, 81–91, https://doi.org/10.4155/cmt.13.77, 2014. Shi, Z., Crowell, S., Luo, Y., and Moore, B.: Model Structures Amplify Uncertainty in Predicted Soil Carbon Responses to Climate Change,\nNature communications, 9, 1–11, 2018. ,\n,\n335\nScharlemann, J. P., Tanner, E. V., Hiederer, R., and Kapos, V.: Global Soil Carbon: Understanding and Managing the Largest Terrestrial\nCarbon Pool, Carbon Management, 5, 81–91, https://doi.org/10.4155/cmt.13.77, 2014. Shi, Z., Crowell, S., Luo, Y., and Moore, B.: Model Structures Amplify Uncertainty in Predicted Soil Carbon Responses to Climate Change,\nNature communications, 9, 1–11, 2018. Shi, Z., Allison, S. D., He, Y., Levine, P. A., Hoyt, A. M., Beem-Miller, J., Zhu, Q., Wieder, W. R., Trumbore, S., and Randerson, J. T.: The\n340\nAge Distribution of Global Soil Carbon Inferred from Radiocarbon Measurements, Nature Geoscience, 4, 1–5, 2020. Sigurdsson, B. D., Leblans, N., Dauwe, S., Guðmundsdóttir, E., Gundersen, P., G Gunnarsdóttir, G. E., Holmstrup, M., Ilieva-Makulec, K.,\nKätterer, T., Marteinsdóttir, B., et al.: Geothermal Ecosystems as Natural Climate Change Experiments: The ForHot Research Site in\nIceland as a Case Study, Icelandic Agricultural Sciences, 29, 53–71, https://doi.org/10.16886/IAS.2016.05, 2016. Shi, Z., Allison, S. D., He, Y., Levine, P. A., Hoyt, A. M., Beem-Miller, J., Zhu, Q., Wieder, W. R., Trumbore, S., and Randerson, J. T.: The\n340\nAge Distribution of Global Soil Carbon Inferred from Radiocarbon Measurements, Nature Geoscience, 4, 1–5, 2020. on of Global Soil Carbon Inferred from Radiocarbon Measurements, Nature Geoscience, 4, 1–5, 2020. Sigurdsson, B. D., Leblans, N., Dauwe, S., Guðmundsdóttir, E., Gundersen, P., G Gunnarsdóttir, G. References W.: Field-Warmed Soil Carbon Changes Imply High 21st-Century Modeling Uncertainty,\nBiogeosciences, 15, 3659–3671, 2018. Todd-Brown, K. E., Zheng, B., and Crowther, T. W.: Field-Warmed Soil Carbon Changes Imply High 21st-Century Modeling Uncertainty,\nBiogeosciences, 15, 3659–3671, 2018. van Gestel, N., Shi, Z., van Groenigen, K. J., Osenberg, C. W., Andresen, L. C., Dukes, J. S., Hovenden, M. J., Luo, Y., Michelsen,\nd ll\ni h\nh\nd\ndi i\nil\nb\ni h\ni\n55\n5 van Gestel, N., Shi, Z., van Groenigen, K. J., Osenberg, C. W., Andresen, L. C., Dukes, J. S., Hovende van Gestel, N., Shi, Z., van Groenigen, K. J., Osenberg, C. W., Andresen, L. C., Dukes, J. S., Hovenden, M. J., Luo, Y., Michelsen,\nA., Pendall, E., Reich, P. B., Schuur, E. A. G., and Hungate, B. A.: Predicting Soil Carbon Loss with Warming, Nature, 554, E4–E5,\n355\nhttps://doi.org/10.1038/nature25745, 2018. ,\n,\n,\n,\ng\n,\n,\ng,\n,\n,\n,\n,\n,\n,\n,\n,\n,\n,\nA., Pendall, E., Reich, P. B., Schuur, E. A. G., and Hungate, B. A.: Predicting Soil Carbon Loss with Warming, Nature, 554, E4–E5,\n355\nhttps://doi.org/10.1038/nature25745, 2018. Walker T W N Kaiser C Strasser F Herbold C W Leblans N I W Woebken D Janssens I A Sigurdsson B D and Richter A : A., Pendall, E., Reich, P. B., Schuur, E. A. G., and Hungate, B. A.: Predicting Soil Carbon Loss with Warming, Nature, 554, E4–E5,\n355\nhttps://doi.org/10.1038/nature25745, 2018. Walker, T. W. N., Kaiser, C., Strasser, F., Herbold, C. W., Leblans, N. I. W., Woebken, D., Janssens, I. A., Sigurdsson, B. D., and Richter, A.:\nMicrobial Temperature Sensitivity and Biomass Change Explain Soil Carbon Loss with Warming, Nature Climate Change, 8, 885–889,\nhttps://doi.org/10.1038/s41558-018-0259-x, 2018. Walker, T. W. N., Kaiser, C., Strasser, F., Herbold, C. W., Leblans, N. I. W., Woebken, D., Janssens, I. A., Sigurdsson, B. D., and Richter, A.:\nMicrobial Temperature Sensitivity and Biomass Change Explain Soil Carbon Loss with Warming, Nature Climate Change, 8, 885–889,\nhttps://doi.org/10.1038/s41558-018-0259-x, 2018. Walker, T. W. N., Janssens, I. A., Weedon, J. T., Sigurdsson, B. D., Richter, A., Peñuelas, J., Leblans, N. I. W., Bahn, M., Bartrons, M.,\n360\nDe Jonge, C., Fuchslueger, L., Gargallo-Garriga, A., Gunnarsdóttir, G. E., Marañón-Jiménez, S., Oddsdóttir, E. S., Ostonen, I., Poeplau,\nC., Prommer, J., Radujkovi´c, D., Sardans, J., Sigurðsson, P., Soong, J. L., Vicca, S., Wallander, H., Ilieva-Makulec, K., and Verbruggen, 24 https://doi.org/10.5194/bg-2021-338\nPreprint. Yost, J. L. and Hartemink, A. E.: How deep is the soil studied–an analysis of four soil science journals, Plant and soil, 452, 5–18, 2020.\n365\nZhang, J., Ekblad, A., Sigurdsson, B. D., and Wallander, H.: The Influence of Soil Warming on Organic Carbon Se-\nquestration of Arbuscular Mycorrhizal Fungi in a Sub-Arctic Grassland, Soil Biology and Biochemistry, 147, 107 826,\nhttps://doi.org/10.1016/j.soilbio.2020.107826, 2020. E.: A Systemic Overreaction to Years versus Decades of Warming in a Subarctic Grassland Ecosystem, Nature Ecology & Evolution, 4,\n101–108, https://doi.org/10.1038/s41559-019-1055-3, 2020.\nYost, J. L. and Hartemink, A. E.: How deep is the soil studied–an analysis of four soil science journals, Plant and soil, 452, 5–18, 2020.\n365\nZhang, J., Ekblad, A., Sigurdsson, B. D., and Wallander, H.: The Influence of Soil Warming on Organic Carbon Se-\nquestration of Arbuscular Mycorrhizal Fungi in a Sub-Arctic Grassland, Soil Biology and Biochemistry, 147, 107 826,\nhttps://doi.org/10.1016/j.soilbio.2020.107826, 2020.\nDOI = 10.5281/zenodo.4745479 E.: A Systemic Overreaction to Years versus Decades of Warming in a Subarctic Grassland Ecosystem, Nature Ecology & Evolution, 4,\n101–108, https://doi.org/10.1038/s41559-019-1055-3, 2020. References Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. https://doi.org/10.5194/bg-2021-338\nPreprint. Discussion started: 5 January 2022\nc⃝Author(s) 2022. CC BY 4.0 License. E.: A Systemic Overreaction to Years versus Decades of Warming in a Subarctic Grassland Ecosystem, Nature Ecology & Evolution, 4,\n101–108, https://doi.org/10.1038/s41559-019-1055-3, 2020. DOI = 10.5281/zenodo.4745479 25"