This is the accepted version of the article: Marañón Jiménez, Sara; Peñuelas, Josep; Richter, Andreas; [et al.]. «Coupled carbon and nitrogen losses in response to seven years of chronic warming in subarctic soils». Soil biology and biochemistry, Vol. 134 (July 2019), p. 152-161. DOI 10.1016/j.soilbio.2019.03.028 This version is avaible at https://ddd.uab.cat/record/204026 under the terms of the license *Manuscript with continous line numbering Click here to view linked References 1 Coupled carbon and nitrogen losses in response to seven years of chronic warming 2 in subarctic soils 3 4 Running title: Coupled losses of C and N from subarctic soils 5 Marañón-Jiménez S.1,2,3, Peñuelas J.1,2,4, Richter A.5, Sigurdsson B. D.6, Fuchslueger, 6 L.3, Leblans N.I.W.3, Janssens I. A.3 7 1 8 CREAF, Cerdanyola del Vallès, 08193 Barcelona, Spain 9 2 10 CSIC, Global Ecology Unit CREAF-CSIC-UAB, Bellaterra, 08193 Barcelona, Spain 11 3 12 Centre of Excellence PLECO (Plant and Vegetation Ecology), Department of Biology, 13 University of Antwerpen, Campus Drie Eiken, Universiteitsplein 1, C. 203, BE- 2610, 14 Wilrijk, Belgium. 15 4 16 Institut d’Estudis Catalans, 08001 Barcelona, Spain 17 5 18 Department of Microbiology and Ecosystem Science, University of Vienna, 19 Althanstraße 14, 1090 Wien, Austria 20 6 21 Agricultural University of Iceland, Hvanneyri 311, Borgarnes, Iceland 22 23 24 ‡ 25 Address correspondence to S. Marañón-Jiménez, email: s.maranon@creaf.uab.es 26 27 Research Article 2 28 Abstract 29 Increasing temperatures may alter the stoichiometric demands of soil microbes and 30 impair their capacity to stabilize carbon (C) and retain nitrogen (N), with critical 31 consequences for the soil C and N storage at high latitude soils. Geothermally active 32 areas in Iceland provided wide, continuous and stable gradients of soil temperatures to 33 test this hypothesis. In order to characterize the stoichiometric demands of microbes 34 from these subarctic soils, we incubated soils from ambient temperatures after the 35 factorial addition of C, N and P substrates separately and in combination. In a second 36 experiment, soils that had been exposed to different in situ warming intensities (+0, 37 +0.5, +1.8, +3.4, +8.7, +15.9 ºC above ambient) for seven years were incubated after 38 the combined addition of C, N and P to evaluate the capacity of soil microbes to store 39 and immobilize C and N at the different warming scenarios. The seven years of chronic 40 soil warming triggered large and proportional soil C and N losses (4.1 ± 0.5 % °C-1 of 41 the stocks in unwarmed soils) from the upper 10 cm of soil, with a predominant 42 depletion of the physically accessible organic substrates that were weakly sorbed in soil 43 minerals up to 8.7 ºC warming. Soil microbes met the increasing respiratory demands 44 under conditions of low C accessibility at the expenses of a reduction of the standing 45 biomass in warmer soils. This together with the strict microbial C:N stoichiometric 46 demands also constrained their capacity of N retention, and increased the vulnerability 47 of soil to N losses. Our findings suggest a strong control of microbial physiology and 48 C:N stoichiometric needs on the retention of soil N and on the resilience of soil C stocks 49 from high-latitudes to warming, particularly during periods of vegetation dormancy and 50 low C inputs. 51 52 3 53 Keywords: Substrate induced respiration, microbial biomass, temperature increase, 54 nitrogen immobilization, microbial carbon and nutrients limitation, nitrogen loss 4 55 1. Introduction 56 Global warming is expected to accelerate the decomposition of soil organic matter 57 (SOM) more than its production, causing large releases of CO2 to the atmosphere and 58 positive feedbacks to the climatic system (Davidson and Janssens et al. 2006, Jenkinson 59 et al. 1991). Soils at northern latitudes store more than half of the surface-soil carbon 60 (C) (Tarnocai et al. 2009). As their SOM decomposition has been strongly limited by 61 low temperatures and they are warming more rapidly, they are particularly vulnerable to 62 temperature driven C losses (Smith et al. 2015, Crowther et al. 2016). As such, warming 63 of northern soils may potentially increase global concentrations of atmospheric CO2 64 (McGuire et al. 2009). Model predictions for future CO2 emissions and climate change 65 projections by the Intergovernmental Panel on Climate Change (IPCC) remain, 66 nonetheless, largely uncertain (Friedlingstein et al. 2006, Todd-Brown et al. 2013), 67 partly due to the lack of accurate representation of vegetation and soil microbial 68 feedbacks (Bardgett et al. 2013, Friedlingstein et al. 2006) and interactions between C 69 and nutrient cycles (Bardford et al. 2016, Friedlingstein et al. 2006). 70 71 The coupling between C and nitrogen (N) biogeochemical cycles is especially tight in 72 northern ecosystems. Low temperatures constrain the depolymerization and 73 mineralization rates of soil organic N and the release of N-monomers and mineral N, 74 thus limiting plant productivity (Hobbie et al. 2002, Schimel and Bennett 2004, Todd- 75 Brown et al. 2013). Rising temperatures are expected to accelerate the rates of microbial 76 N transformations and alleviate the plant N limitations in these ecosystems, thus 77 increasing plant productivity and C inputs to the soil (Dormann and Woodin 2002, 78 Natali et al. 2012, Wu et al. 2011). Increases in vegetation productivity at warmer 79 temperatures can even offset the soil C losses associated with the accelerated SOM 80 mineralization rates from soil microbes (Melillo et al. 2002, Sistla et al. 2013, IPCC 81 2013). The vulnerability of soil C stocks to warming will therefore depend on the 82 capacity of soils to retain nutrients and ultimately on the ability of plants to profit from 83 the enhanced nutrient availability. 84 85 Soil microbial biomass plays a fundamental role in the stabilization of soil C (Liang et 86 al. 2017, Miltner et al. 2012) and as a short- and long-term N reservoir in soils at high 87 latitudes (Bardgett et al. 2003, Zogg et al. 2000). A large fraction of the N pool in these 88 cold ecosystems is contained in microbial biomass (Jonasson et al. 1996, Xu et al. 5 89 2013). This large N storage potential and the low N mineralization rates imply that 90 microbes successfully compete with plants for the limiting N pools during the growing 91 season (Dunn et al. 2006, Skouw Haugwitz et al. 2011), but also that microbial turnover 92 and N release may represent a major pathway for plant N uptake during periods of 93 declining microbial populations (Bardgett et al. 2003). Microbial N retention becomes 94 even more crucial in ecosystems with a period of vegetation dormancy or senescence, 95 such as at high latitudes, when the short photoperiod and low temperatures prevent 96 vegetation productivity and N uptake (Bardgett et al. 2005). Microbial immobilization 97 then becomes a crucial mechanism to minimize potential N losses from the system 98 during relatively long winter periods (Groffman et al. 2011, Jonasson et al.1996, Kaiser 99 et al. 2011). Warming can, however, desynchronize the intimate seasonal coupling 100 between microbial N immobilization and vegetation uptake in these ecosystems 101 (Bardgett et al. 2005, Jaeger et al. 1999, Lipson et al. 1999), leading to potential soil N 102 and C losses. 103 104 The physiological response of soil microbes to warmer temperatures may elicit shifts in 105 their resource demands, and cause disequilibria on plant-microbial interactions. 106 Although vegetation growth is generally N limited at high latitude ecosystems, C has 107 been found to limit soil microbial growth and biomass even at these high latitudes (Wild 108 et al. 2015). Warmer temperatures may cause persistent increases in microbial 109 respiratory demands and the depletion of the most physically accessible organic 110 substrates in soil (Marañón-Jiménez et al. 2018), thus compromising the C available to 111 maintain constant levels of standing biomass. According to the ecological 112 stoichiometric theory, soil microbes regulate their elemental composition by retaining 113 elements in which they are limited and releasing those in excess (Sterner and Elser 114 2002). This implies a predominance of microbial N mineralization to N immobilization 115 in strongly C-limited microbes. Warming-induced increases in N mineralization during 116 periods of inactive plant N uptake and accessible C inputs may consequently lead to 117 potential losses of soil N by dissimilatory pathways, either by nitrate leaching or 118 gaseous N fluxes (Turner and Henry 2010). Temperature-driven N losses may account 119 for the smaller increase in plant productivity compared to net N mineralization and soil 120 respiration rates frequently observed in experimental warming experiments (Bai et al. 121 2013, Lu et al. 2013, Rustad et al. 2001), causing divergences between observed and 122 predicted soil C losses for high latitudes (Todd-Brown et al. 2013, McGuire et al. 2018). 6 123 The potential changes in the capacity of subarctic soils to retain N have not been 124 explored mechanistically yet, even though this information is fundamental to constrain 125 the climate change projections of productivity and soil organic C (SOC) of northern 126 ecosystems. 127 128 Geothermally active areas in Iceland provide stable, continuous and wide gradients of 129 soil temperature (Sigurdsson et al. 2016) that encompass the full range of warming 130 scenarios projected by the IPCC for the northern region (IPCC, 2013). This allow 131 testing for non-linear responses to soil warming and the inference of realistic 132 predictions of soil biogeochemical processes. Previous studies at the same experimental 133 plots from these soil temperature gradients found a linear reduction of 1.28 ± 0.16 ton 134 SOC ha-1 per ºC degree of warming from the upper 10 cm of soil (Leblans et al. 2016). 135 Warming increased C losses by accelerating the mass-specific C mineralization rates of 136 soil microorganisms (Marañón-Jiménez et al. 2018, Walker et al. 2018). Surprisingly, 137 enhanced N mineralization in these N-limited soils did not lead to higher vegetation 138 productivity according to the predictions of most ecosystem models (Todd-Brown et al. 139 2013). On the contrary, aboveground and belowground plant biomass did not change. 140 Vegetation apparently did not benefit from the N released at higher temperatures, 141 probably due to ecosystem N losses. Despite the large and rapid loss of soil C, soil C:N 142 stoichiometry indeed remained unaltered (Leblans et al. 2016), implying a proportional 143 loss of N. 144 In order to assess the mechanisms underlying this coupled soil C and N loss, we 145 incubated soils that had been exposed for seven years to a range of warming intensities 146 in the field due to geothermal activity (0 - 15.9 ºC above ambient, hereafter “in situ 147 temperatures”). In a first set of soil incubations, the factorial addition of C, N and P 148 substrates separately and in combination to soils from ambient temperatures allowed us 149 to characterize the stoichiometric demands of the microbes from these subarctic soils 150 (hereafter “experiment of stoichiometric demands characterization”). In a second set of 151 soil incubations, the combined addition of C, N and P to the warmed soils along the 152 geothermal gradient allowed us to evaluate the capacity of soil microbes to store and 153 immobilize C and N as affected by different warming scenarios, both at ambient 154 nutrient conditions and when C, N and P are plentiful (hereafter “experiment of 155 warming impacts on soil C and N retention”). Regarding the microbial stoichiometric 7 156 demands from these subarctic soils, we hypothesized that soil microbes have strong C 157 limitation due to the short growing period for vegetation (low C inputs) and the high 158 clay content of these soils (high physical protection). We also hypothesized that this C 159 limitation and a restricted C:N stoichiometric plasticity of soil microbes limit the 160 immobilization of mineralized N. Regarding the warming impacts on soil C and N 161 retention, the total losses of C from these (Leblans et al. 2016, Poeplau et al. 2016) and 162 many other soils (Hicks Pries et al. 2017, Crowther et al. 2016, Melillo et al. 2017) 163 exposed to warmer temperatures, and the increasing mass-specific respiration rates of 164 soil microbes (Marañón-Jiménez et al. 2018), led us to hypothesize a depletion of the 165 most physically accessible substrates in soil. We also hypothesized that these C scarcity 166 conditions in warmer soils impair the C retention by microbial biomass and the 167 immobilization of the mineralized N that is released from SOM at warmer temperatures. 168 These two complementary experiments will therefore contribute to elucidate the causes 169 of the divergences on the soil C losses between field warming experiments and model 170 predictions at high latitude ecosystems. 171 172 2. Methods 173 2.1. Study site 174 Soils were collected at the ForHot research site in the Hengil geothermal area, 40 km 175 east of Reykjavik, Iceland (64°00′01″N, 21°11′09″W; 83-168 m a.s.l.), which has been 176 described in detail by Sigurdsson et al. (2016). Mean annual air temperature, annual 177 precipitation and wind speed were 5.2 °C, 1460 mm and 6.6 m s-1, respectively 178 (Synoptic Station, 9 km south of Hveragerdi, Icelandic Meteorological Office, 2016). 179 The mean temperatures of the warmest and coldest months, July and December, were 180 12.2 and -0.1 °C, respectively. The growing season normally starts in late May and ends 181 in late August. Snow cover is not permanent during winter due to the mild oceanic 182 climate, but the soil typically freezes for at least two months during mid-winter. The 183 main vegetation type is unmanaged grassland, dominated by Agrostis capillaris, 184 Ranunculus acris and Equisetum pratense, all perennial species with short-lived 185 aboveground parts that regrow each year from underground stems or rhizomes. Sites 186 had been grazed by sheep for centuries (low-intensity grazing), but this practice was 187 ceased in the 1970´s (Sigurdsson et al. 2016). 188 8 189 The soil in the area has been subjected to warming since May 2008 due to geothermal 190 activity, when an earthquake shifted geothermal systems to previously unwarmed soils. 191 Hot groundwater warmed the underlying bedrock and surfaced along faults in the soil 192 crust. Soil temperatures were highest near these faults and declined perpendicular to 193 them. No signs of soil contamination by geothermal byproducts, such as exchangeable 194 sulfur, were found (Sigurdsson et al. 2016). The soils are Andosols with a silty-loamy 195 texture. 196 197 2.2. Experimental design and soil sampling 198 Five replicate transects were established in 2012, each covering six levels of in situ soil 199 warming: 0, 0.5, 1.8, 3.4, 8.7 and 15.9 °C above ambient (mean annual temperatures in 200 the upper 10 cm of soil). A 0.5 × 0.5 m plot was established for each warming level for 201 soil sampling (n = 6 in situ temperatures × 5 replicate transects = 30 plots). Soil 202 temperature was monitored hourly at 10 cm soil depth using TidbiT v2 HOBO Data 203 Loggers (Onset Computer Corporation, Bourne, USA). Despite the seasonal and daily 204 oscillations of soil temperatures, the temperature increases above ambient were rather 205 constant along the year and vertically down to ca. 20-25 cm depth (Sigurdsson et al. 206 2016). The mean annual soil temperatures and main soil parameters are indicated in 207 Table 1. Plant community composition showed no changes in dominant plant species up 208 to +8.7 °C warming (Gudmundsdóttir et al. 2014, Michielsen 2014). At the most 209 extreme warming level (15.9 ºC above ambient) the vegetation community shifted 210 towards a higher dominance of non-vascular plants (mosses) (Leblans, personal 211 communication). 212 213 After seven years of soil warming (August 2015), samples from the upper 10 cm of 214 mineral soil were collected from all plots. The mean soil temperature in unwarmed plots 215 two weeks prior to sampling was 11.9±0.3 °C. Soils from each warming level were 216 sieved to 2 mm, mixed and homogenized to constitute a composite sample. The samples 217 were then stored at 5 °C, which is approximately the mean annual temperature of the 218 ambient unwarmed soil, until the analyses and incubations. 219 220 2.3. Initial soil parameters 221 Three subsamples of 15, 7.5 and 7 g of fresh soil were extracted with 2 M KCl, 0.5 M 222 NaHCO3 and 0.5 M K2SO4, respectively, within 24 h of sampling. Ammonium (NH4+) 9 223 and nitrate (NO3-) were determined from the KCl extracts (Bremner and Keeney 1965). 224 Half of the NaHCO3 and K2SO4 extract volume was digested at 400 ºC with H2SO4 with 225 selenium as a catalyst. Total phosphorus (P) and total extractable N (TNK2SO4) were 226 determined from the digested NaHCO3 and K2SO4 extracts, respectively. Available 227 inorganic P (Pinorg) was determined from the undigested NaHCO3 extracts (Olsen et al. 228 1954) and dissolved organic C (DOCK2SO4) and NH4+ from the undigested K2SO4 229 extracts. Organic P (Porg) and dissolved organic N (DONK2SO4) were determined as the 230 difference between digested and undigested NaHCO3 and K2SO4 extracts, respectively 231 (Jones and Willett 2006). Two other pools of soluble organic C were quantified using 232 extractants of different ionic strengths. For this, two subsamples of 10 g of fresh soil 233 were extracted with deionized water (DOCwater), which is a common measure of readily- 234 soluble C, and a weak phosphate buffer at 10 mM (0.33 mM KH2PO4 and 6.67 mM 235 Na2HPO4) adjusted to pH 7.0 (DOCbuffer), which extracts both the readily-soluble C and 236 weakly adsorbed C in clay minerals (Nelson et al. 1994, Kaiser and Zech 1999). The 237 lower ionic strength and pH of the buffer solution compared to the 0.5 M K2SO4 238 solution reduces the flocculation of organic colloids and the re-adsorption of the 239 solubilized C onto the diffuse double layer surrounding clay particles (Haney et al. 240 2001). The relative accessibility of extractable soil C pools (DOCK2SO4, DOCwater, 241 DOCbuffer) was calculated as the ratio of DOC to SOC pools. 242 243 Another set of subsamples of the same mass of fresh soil were also extracted as 244 described above for determining microbial biomass C and total microbial N and P by 245 fumigation-extraction (Jenkinson and Powlson 1976). Microbial biomass C (Cmicro), 246 total microbial N (Nmicro) and total microbial P (Pmicro) were determined as the 247 difference in DOCK2SO4, TNK2SO4 and total P between the fumigated and unfumigated 248 subsamples, respectively. All analyses were performed by colorimetric detection with a 249 San++ Continuous Flow Analyzer (Skalar Analytical B.V., Breda, The Netherlands). 250 NO3- was determined after reduction to NO2- and formation of the diazo complex at 540 251 nm wavelength (EN-ISO 13395). NH4+ was determined after reaction with salicylate, a 252 catalyst and active chlorite solution to form a green colored complex at 660 nm 253 wavelength (ISO 11732). TNK2SO4 and NH4+ in digested and undigested K2SO4 extracts 254 respectively, were determined colorimetrically at 660 nm wavelength. DOCK2SO4 was 255 determined after reaction with phenolphthalein at 550 nm wavelength (ISO 5667-3). 256 Pinorg was determined colorimetrically as phospho-molybdic complex at 880 nm 10 257 wavelength in both digested and undigested extracts (ISO 15681-2). Total soil organic 258 C and total soil N (SOC and TN, respectively) were determined from dry soils by dry 259 combustion at 850 °C with a Thermo Flash 2000 NC Analyser 260 (Thermo Fisher Scientific, Delft, The Netherlands). Inorganic C is not detectible in 261 these volcanic soils (Arnalds 2015), so total C can be considered as organic C. Soil pH 262 was determined by stirring and settling in deionized water in a ratio 1:5 (Pansu and 263 Gautheyrou 2006). 264 265 We calculated the stoichiometric C:N imbalance between soil organic pools and 266 microbial biomass following Mooshammer et al. 2014a, as the ratio of C:N in the SOM 267 pools (SOC:TN and DOCK2SO4:TNK2SO4) over microbial biomass C:N (Cmicro:Nmicro). 268 The C:N imbalance is then a measure of the divergence between the C:N stoichiometry 269 of soil microbes and soil organic substrates, where C:N imbalance < 1 thus reflects a 270 lack of C in SOM pools for soil microbes. 271 272 2.4. Substrate addition and soil incubation 273 Subsamples of 40 g (dry equivalent) of fresh soil from the unwarmed ambient plots 274 (hereafter “incubation flasks”) were distributed into flasks within 72 h after sampling. In 275 order to determine the stoichiometric demands of soil microorganisms and their 276 capacity of C storage and N immobilization (experiment of stoichiometric demands 277 characterization), a 1-ml of deionized water solution with a source of C, N, P or their 278 combinations (hereafter “addition”) was added to each flask. We hypothesized that 279 losses of soil N were associated with a restricted capacity of microbial N 280 immobilization, so we tested the effect of two levels of N addition instead of the CP 281 combination. C was added as glucose (1.73 mg of glucose g-1 dry soil, that is, 0.69 mg 282 C g-1), N was added as NH4NO3 (0.1 mg of NH4NO3 g-1, that is, 34 µg N g-1 for the N 283 addition level and 0.05 mg of NH4NO3 g-1, 17 µg N g-1 for the “half-N” addition level) 284 and P was added as KH2PO4 (0.101 mg KH2PO4 g-1, 23 µg P g-1). The amount of 285 substrates added accounted for ca. 1 % of the initial soil C content and 0.7 and 0.35 % 286 of the initial soil N content for N and “half-N”, respectively (Table 1). Phosphorous 287 retention is generally >90 % for Icelandic Andosols (Arnalds et al. 1995), so that the P 288 added was ca. ten times the initial available inorganic P soil content to ensure that 289 enough P was accessible to soil microbes. These amounts of substrates were chosen to 290 ensure the alleviation of potential C and nutrient limitations of soil microbes while 11 291 avoiding potential changes in soil pH. The corresponding combination of the above C, 292 N and P concentrations were used for the CN, NP and CNP addition levels, equivalent 293 to a weight ratio of 20:1:0.67 for the CNP addition level. A set of incubation flasks was 294 also incubated after the addition of 1 ml of deionized water without substrate (hereafter 295 “water-only”). 296 297 The response of microbial biomass to soil warming and the capacity of the warmed soils 298 to retain N in presence of available nutrients (experiment of warming impacts on soil C 299 and N retention) was determined by incubating the samples from each in situ warming 300 level with “water-only” and with added C, N and P in combination (CNP) as a single 301 addition level, using the same soil mass and substrate concentrations as above (see 302 Marañón-Jiménez et al. 2018 for further details). Soil moisture was adjusted to 60 % 303 water-holding capacity in all incubation flasks, and the soil was mixed to ensure an even 304 distribution of the solution. 305 306 The soils were then incubated at the mean annual soil temperature in the field (5 °C) 307 and allowed to equilibrate for 12 h. This time lapse was determined in a preliminary 308 assay using the same soils based on the time needed to obtain acceptable coefficients of 309 variability (<20 %) of microbial respiration. Microbial respiration (i.e. substrate induced 310 respiration) was then measured in all samples using an infrared gas analyzer (EGM- 311 4/SRC-1, PP-Systems, Hitchin, UK) coupled to a custom-made chamber with a fan and 312 vent. Incubation flasks were partially closed during the incubation to prevent drying but 313 allow the gas exchange. The flasks were ventilated with a fan for ca. 2 minutes prior 314 each respiration measurement to release the accumulated CO2 in soil pores and in the air 315 layer closed to the soil surface. Flasks were immersed in a water bath at a constant 316 temperature of 5 ºC to maintain the targeted temperature during the respiration 317 measurements. Temperature was continuously monitored during the measurements and 318 incubation using TidbiT v2 HOBO Data Loggers (Onset Computer Corporation, 319 Bourne, USA). Gravimetric soil moisture stayed constant at 60 % water-holding 320 capacity throughout the experiment. 321 322 The incubation temperature of the soil samples was then increased progressively to 30 323 °C over 6 days (4.6 ºC per day) in an incubator with adjustable temperature, allowing us 324 to discard any potential limitation of low incubation temperatures on the microbial 12 325 substrate uptake and growth (Nedwell 1999). Cmicro, Nmicro and the remaining DOCK2SO4, 326 NH4+ and DONK2SO4 in the soil were determined for all incubated samples as described 327 above six days after the C and nutrient additions to allow soil microbes to take up the 328 substrates. We were only interested in relative differences among treatments, so the 329 concentrations in the microbial fraction presented here were not corrected for extraction 330 efficiency. All fractions are presented relative to soil dry mass. 331 332 2.5. Data analyses 333 The effect of in situ soil warming on initial soil and microbial C and nutrient contents 334 and ratios prior to the incubations was tested using one-way ANOVAs, and differences 335 among warming levels were further tested by post hoc tests with Tukey correction for 336 multiple testing. The effects of C, N and P substrate additions on microbial respiration, 337 Cmicro, Nmicro, microbial C:N ratios; the remaining DOCK2SO4, NH4+ and DONK2SO4 and 338 the DOCK2SO4:TNK2SO4 ratio in unwarmed soils (experiment of stoichiometric demands 339 characterization) after the incubation were tested using one-way ANOVAs, and 340 differences among addition levels were further tested by post hoc tests with Tukey 341 correction for multiple testing. The differences from soils without any addition were 342 also tested usin post hoc unnett s tests, using the “water-only” unamended soils as 343 control. The effect of soil warming, substrate addition (C, N and P combined) and their 344 interaction on microbial respiration, Cmicro, Nmicro, the microbial C:N ratio (experiment 345 of warming impacts on soil C and N retention) were tested using two-ways ANOVAs, 346 with “addition” and “in situ soil warmin ” as fixed factors. Differences among in situ 347 warming levels were further tested by post hoc tests with Tukey correction for multiple 348 testing. The effect of substrate addition on the above variables was also tested for each 349 warming level separately by one-way ANOVAs. Data were transformed when required 350 to improve normality and homoscedasticity (Quinn and Keough, 2009). Stoichiometric 351 ratios were calculated on a mass basis. Statistical analyses and model construction were 352 performed using JMP 13.0 (SAS Institute). All results are presented as means ± 353 standard errors. 354 355 3. Results 356 3.1. Response of microbial biomass C and respiration of ambient soils to the addition of 357 C, N and P 13 358 Microbial biomass C (non-corrected for extraction efficiency) constituted only 0.63 % 359 of the SOC in this subarctic soil but contained four times more C than the DOCK2SO4 360 pool (Table 1). Microbial respiration increased ca. 12 h after the C addition (P<0.001), 361 but N addition and P addition did not cause any significant changes in the rate of 362 microbial respiration (Fig. 1a), either alone or in combination with C. 363 364 Microbial biomass C responded to the additions very similarly to microbial respiration. 365 It increased 29-47 % approximately six days after the addition of a labile C substrate 366 (glucose) (Fig. 1b, P<0.001), while it even decreased in response to the N and P 367 additions alone. Microbial biomass C, however, increased after the combined addition 368 of N and P either alone or in combination with C. 369 370 3.2. Response of microbial N of ambient soils to the addition of C, N and P 371 The microbial N pool represented ca. three times the total extractable N in the soil 372 (Table 1). Most of this extractable soil N (79 %) was in an organic form, while NH4+ 373 and NO3- represented only 17 % and 3 % of this pool, respectively (Table 1). Total 374 microbial N only increased significantly in response to the combined addition of C and 375 N (Fig. 2a, P=0.02), although values also increased, but not significantly, in all the rest 376 of the addition levels. Consequently, the C addition also caused a depletion of the NH4+ 377 in soil (Fig. 2b, P<0.001). Circa 82 and 72 % of the NH4+ initially available was 378 depleted from the soil when C and N were added in combination in the CN and CNP 379 addition levels, respectively (Fig. 2b), while a large proportion (86, 81 and 111 % for 380 “half-N”, N and NP, respectively) still remained in the soil otherwise. In contrast, soil 381 DONK2SO4 decreased in response to N-only addition (Fig. 2c, P=0.007). 382 383 3.3. Response of microbial C:N ratios of ambient soils to the addition of C, N and P 384 The C:N ratios of K2SO4-extractable soil organic substrates decreased to lower values 385 than in microbial biomass after six days of incubation (C:N imbalance <1, Fig. 3). 386 Microbial C:N ratios increased significantly in response to the CNP addition and 387 decreased after the addition of N and P only (P<0.001). 388 389 3.4. Response of easily accessible soil C pools and C:N ratios to warming 390 Seven years of continuous warming provoked a substantial depletion of the pools of 391 DOC extracted with K2SO4 and with phosphate buffer (DOCK2SO4 and DOCbuffer, 14 392 respectively, Fig. 4a), while the most readily-available DOC pool (DOCwater) did not 393 show a consistent decreasing pattern with soil temperatures in situ. Moreover, the 394 relative accessibility of the DOCbuffer pool, calculated as the ratio of DOCbuffer to SOC 395 pools, decreased with the intensity of soil warming up to 8.7 ºC above ambient 396 (P<0.001, Fig. 4b), while the relative accessibility of the DOCK2SO4 pool was not 397 substantially affected below this soil warming intensity. Nonetheless, the non- 398 extractable C pools (SOC) were depleted in a higher proportion at the highest warming 399 level (15.9 ºC above ambient, Table 1), contributing to increase the relative accessibility 400 of both the DOCK2SO4 and DOCbuffer pools. The relative accessibility of the DOCwater 401 pool remained however unaffected by in situ soil warming. 402 403 Soil warming also decreased the pools of soil DOCK2SO4 and TNK2SO4 proportionally, 404 without any significant shifts in DOCK2SO4:TNK2SO4 ratios along the in situ temperature 405 gradient (Fig. 4c). Even though the C:N ratios of soil organic matter (SOC:TN) were 406 2.3 times higher than the C:N ratios of microbial biomass, the imbalance from the C:N 407 of the extractable fraction of organic substrates (DOCK2SO4:TNK2SO4) was initially close 408 to one (Fig. 4c), since the C:N ratios of the extractable organic pools were much lower 409 than the ratios of the total organic matter pools. Warming did not cause shifts in the 410 stoichiometric imbalance between the extractable organic substrates and microbial 411 biomass, given the coupled and proportional losses of C and N from both biomass and 412 soil (Fig. 4c). 413 414 3.5. Response of soil microbes to warming and to the addition of C, N and P 415 Despite the depletion of the easily accessible soil C pools, microbial respiration only 416 decreased slightly with in situ warming (P=0.04, Fig. 5a), and this decrease was only 417 si nificant at unamended samples (“water-only”, P=0.03). In situ soil warming however 418 decreased substantially both microbial biomass C and N (P<0.001 for both variables), 419 with the largest changes between 1.8 and 3.4 °C above ambient (Fig. 5b, c). Microbial 420 C:N ratios thus did not change significantly with in situ soil warming, although variance 421 increased at the warmest soils (Fig. 5d, P=0.13). 422 423 The addition of a substrate containing a labile source of C, N and P (CNP) increased 424 microbial respiration in a similar magnitude across all in situ warming levels (P<0.001 15 425 for “addition” effect, P=0.87 for “addition” and “in situ soil warmin ” interactions, Fig. 426 5a). In contrast, the substrate added increased microbial biomass C only in soils from 427 moderate warming levels <3.4 °C (P<0.001, Fig. 5b), but it did not increase at higher 428 warming levels (P<0.001 for “addition” and “in situ soil warmin ” interactions), even 429 though the amount of remaining DOC was still higher than in unamended soils 430 (P<0.01). Microbial N showed very similar response (P<0.001 for “addition” effects, 431 Fi . 5c), but the interaction between “addition” and “in situ soil warmin ” was not 432 significant in this case (P=0.18). Microbial C:N ratios, however, did not change 433 substantially in response to the added CNP substrate (P=0.10, Fig. 5d), although they 434 tended to increase in response to the addition at in situ warming levels ≤3.4 °C (P=0.05 435 for “addition” and “in situ soil warmin ” interactions), indicating a proportionally 436 higher retention of C than N. 437 438 4. Discussion 439 Nitrogen was lost in the same proportion as C in these subarctic soils (Table 1, Fig. 4c), 440 so that the C:N ratios did not change substantially along the in situ soil temperature 441 gradient. This is in contrast to the increase in the availability of soil mineral N and 442 vegetation productivity generally observed in field warming experiments (Dieleman et 443 al. 2012, Dormann and Woodin 2002, Wu et al. 2011). The proportional loss of both 444 elements points to the tight C:N stoichiometric coupling as a mechanism. Soil C losses 445 in response to warmer temperatures have frequently been observed, but experimental 446 results do not always match model predictions for high-latitude ecosystems (Todd- 447 Brown et al. 2013, McGuire et al. 2018). Overlooking the relevance of the C and N 448 stoichiometric needs of soil microbes for soils to retain these elements can be a potential 449 cause of these divergences. Soil warming provoked the depletion of a large fraction of 450 the easily accessible C pools in these soils (Fig. 4), where microbial C limitation was 451 already strong (Fig. 1), leading to substantial reductions in microbial biomass and in the 452 capacity of N retention of soil microbes. The strict C and N stoichiometric needs of soil 453 microbes may have determined the coupled losses of C and N from warmed soils, 454 accounting for the constant soil C:N ratios. 455 456 4.1. C, N and P limitation of microbes in high-latitude soils 457 Nutrient immobilization by soil microbes can strongly control biogeochemical cycling 458 in ecosystems where temperatures limit the release of nutrients from SOM (Skouw 16 459 Haugwitz et al. 2011). In these subarctic soils, most of the soil N was in organic form 460 and the microbial N pool represented ca. three times the total extractable N, pointing to 461 the high sensitivity of N biogeochemical fluxes and soil N storage capacity to changes 462 in microbial biomass N. The soils in our incubations have been exposed in situ to 463 constant temperature increases relative to ambient temperatures (Sigurdsson et al. 464 2016), so an increase in mineralization rates and N release to the soil are expected 465 throughout the year. Litter decomposition and mass-specific mineralization rates of the 466 microbes from the same study site were accordingly higher in warmer soils (Leblans et 467 al. 2016, Marañón-Jiménez et al. 2018). The short photoperiod and low temperatures, 468 however, limited vegetation productivity and nutrient uptake during winter dormancy 469 (Leblans et al. 2017). The role of soil microbes in nutrient immobilization for 470 preventing nutrient leaching is therefore crucial during this period, and particularly 471 during winter thaws (Yano et al. 2015). 472 473 Soil microorganisms in these subarctic soils were strongly C limited even at ambient 474 temperatures, indicated by a large and equivalent increase in respiration and biomass in 475 response to C addition (Fig. 1). By contrast, microbial respiration was not altered by the 476 N or P additions, and microbial biomass even decreased after the addition of these 477 nutrients alone (Fig 1b). Besides the low vegetation inputs during prolonged winter 478 periods, the strong C limitation can be also partly associated with the low accessibility 479 of most organic substrates, which are sorbed by soil minerals of high specific surface 480 area in these volcanic-ash soils. The large differences between SOC and DOC pools 481 points to a high proportion of non-extractable C strongly occluded (Poeplau et al. 2016). 482 More than ten times organic C was extracted by phosphate buffer than by water in the 483 ambient soils, which also indicates a high proportion of soil C weakly adsorbed to 484 colloidal surfaces (Hayes, 1985). The high adsorption capacity of the fine-textured soils 485 may promote a long-lasting microbial C limitation that, most likely, aggravate in winter, 486 when plant C inputs decrease. 487 488 The relationship between the C:N stoichiometry of soil microorganisms and SOM 489 substrates governs the predominant biogeochemical pathways by which microbes meet 490 their stoichiometric needs using available resources (Mooshammer et al. 2014b). 491 Accordingly, soil microorganisms retain limited elements and release those in excess 17 492 (Sterner and Elser 2002). The microbial C:N ratios in the soils at ambient temperatures 493 (C:N=5.41±0.15, Fig. 4c) were slightly lower than those reported for grassland soils 494 (C:N=6.6) and global averages (C:N=7.6) (Xu et al. 2013). The SOC:TN ratios of SOM 495 (C:N=11.97±0.07) were also lower than for grasslands (C:N=13.3) and globally 496 (C:N=16.4), and the ratios were even lower in the pool of extractable SOM 497 (C:N=6.02±0.72 , Fig. 4c). The relatively low microbial C:N ratios in these subarctic 498 soils and a C:N imbalance in relation to the extractable organic pools close to one (Fig. 499 4c) indicate that N immobilization was not required in large amounts to meet their 500 stoichiometric needs. On the contrary, a net mineralization occurred during the soil 501 incubation in non-amended soils (Fig. 2b), while the immobilization of mineral N was 502 conditioned by the supply of an accessible C pool and the production of new microbial 503 biomass (Figs. 1b and 2). 504 505 Carbon limitation and the strict C:N stoichiometric needs of soil microbes 506 (Zeichmeister-Boltenstern et al. 2015) actually constrained microbial N immobilization. 507 Only the C addition provoked a significant increase in microbial N (Fig. 2a), and N 508 immobilization was highest when C and N were added in combination, although the 509 addition of inorganic N alone also stimulated microbial N immobilization slightly. A 510 86, 81 and 111 % of the total NH4+ initially available still remained in the soil six days 511 after addition for the “half-N”, N and NP addition levels, respectively (Fig. 2b), while 512 only 18 to 28 % remained when C was also added for the CN and CNP additions. The 513 decrease of microbial biomass (Fig. 1b) and the predominant use of DON as C source 514 when only N was added (Fig. 2c) are further evidences of C limitation for microbial 515 growth and N immobilization (Farrell et al. 2014). Similar C constraints of microbial N 516 demands have been observed in Siberia (Wild et al. 2015), reminding the need to frame 517 the concept of C or nutrient limitation to specific ecosystem components or processes 518 rather than generalizing to entire ecosystems. Sub-surface soils (>5 cm depth) also 519 showed no capacity for net retention of increased N inputs after 20 years of fertilization 520 experiment in Alaska, leading to a net C loss (Mack et al. 2004). Soils with relatively 521 low C:N ratios may also present a secondary microbial P limitation. The addition of P in 522 these soils may fuel the synthesis of P-rich mRNA for protein transcription (Elser et al. 523 1996), enhancing immobilization of soil DON for protein synthesis up to certain level, 524 where the N immobilization is again saturated and limited by C availability (Hessen et 525 al. 2007). This limitation was evidenced by the decrease in microbial biomass in 18 526 response to the P addition (Fig. 1b). In contrast, the simultaneous supply of N and P 527 needed for protein synthesis may have promoted the allocation of soil organic substrates 528 for microbial growth, resulting in increases in microbial biomass (Fig. 1b). Soils with 529 low C:N ratios where the N storage function of soil microbes is not supported by a 530 continuous supply of easily accessible C will be therefore vulnerable to N losses. 531 532 4.2. Response of microbial cycling to soil warming 533 Seven years of continuous soil warming led to a substantial loss of total soil C and N 534 from the upper 10 cm (Table 1), but not all pools of SOC were depleted equally. Both 535 DOCK2SO4 and DOCbuffer pools decreased significantly with in situ soil warming, while 536 the DOCwater pool did not show a consistent decreasing pattern (Fig. 4a). In relative 537 terms, soil warming provoked a predominant depletion of the DOCbuffer pool in relation 538 to the total SOC up to +8.7 °C warming (Fig. 4b), indicating a proportional decrease of 539 the soil organic C adsorbed within the soil minerals. Water-extractable C is known as 540 the most readily-available C pool for soil microbes, but it has also shown a lower 541 biodegradability compared to the buffer-extractable C pool when both pools are fully 542 accessible to soil microbes (Nelson et al. 1994, Wagai and Sollings 2002). Soil 543 microbes may have resorted on the weakly-adsorbed C fraction, the largest DOC pool in 544 these soils, as a predominant C source as the water-extractable C pool was depleted at 545 increasing soil temperatures. Increasing rates of depolymerization and solubilization 546 from the weakly-adsorbed SOM fraction may have also contributed to increase the 547 water-extractable C inputs, compensating the microbial consumption of this pool. 548 Nonetheless, the non-extractable C pools (SOC) also experienced a predominant 549 depletion at the most extreme warming level (15.9 ºC above ambient), probably causing 550 a decrease in the surface of organic colloidal surfaces, which contributed to increase the 551 relative accessibility of both the DOCK2SO4 and DOCbuffer pools. Therefore, soil microbes 552 may have satisfied their increasing energy demands at warmer temperatures by a 553 proportionally higher solubilization of the C adsorbed in soil mineral surfaces. 554 555 Microbes increased their respiratory demands per unit of biomass in warmer soils 556 (Marañón-Jiménez et al. 2018, Walker et al. 2018), probably as a consequence of 557 increasing energy costs for metabolic maintenance and for the solubilization of 558 adsorbed organic substrates. Soil warming did, however, not cause substantial shifts in 19 559 the C:N imbalance between SOM and microbial biomass (Fig. 4c) and the response of 560 respiration to the substrate (C, N and P) addition was also equivalent across warming 561 levels (Fig. 5a). Rather than increasing their C demands at the ecosystem level, 562 microbes maintained accelerated rates of C consumption under conditions of low C 563 accessibility by a reduction of the standing biomass (Walker et al. 2018, Fig. 5b), which 564 provoked a coupled and equivalent loss of microbial N (Fig. 5c, d). These results again 565 highlight the strict C:N stoichiometric needs of soil microbes and the tight coupling 566 between N immobilization and biomass production. Warming can therefore lead to 567 proportional soil C and N losses when increased N mineralization rates are not 568 compensated by rapid plant N uptake and plant-derived C inputs to the soil. 569 570 5. Conclusions 571 Seven years of chronic exposure to warmer temperatures led to large and proportional 572 losses of C and N from these high-latitude soils. These findings point to the strict C:N 573 stoichiometric needs of soil microbes and the tight coupling between microbial N 574 immobilization and biomass production as a key mechanism. The continuous exposure 575 of soil microbes to higher temperatures for seven years increased their respiratory 576 demands and provoked the depletion of a large fraction of the easily accessible C pools 577 of these subarctic soils, where microbial C limitation was already strong. Soil warming 578 constrained, as a result, the C retention in microbial biomass and the immobilization of 579 mineralized N. A release of mineral N that is not rapidly compensated by plant N 580 uptake is vulnerable to be lost through leaching in case of nitrification and gaseous 581 fluxes in case of denitrification. The loss of N storage capacity of microbial biomass 582 likely provoked a shift from a close to a leakier N cycle with a detrimental effect on soil 583 N availability and C storage capacity. This mechanism may be key in soils where the 584 low C availability can compromise the maintenance of microbial biomass under a 585 warmer climate, particularly during periods of limited plant C inputs and N uptake. Our 586 results also highlight the need to change the frequent misconception of the ubiquitous N 587 limitation in high latitude ecosystems by a better framed concept of limitation for each 588 specific process or ecosystem component. Accordingly, our findings suggest a strong 589 control of microbial physiology and C:N stoichiometric needs on the retention of soil N 590 and ultimately on the resilience of high-latitude soil C stocks to warming. Overlooking 591 this may be the cause of the large divergences between the predicted response of soil C 592 stocks from models and observations at high latitudes. 20 593 594 6. Acknowledgements 595 This research was supported by the European Union’s Seventh Framework Pro ramme 596 the Ministry of Economy, Innovation, Science and Employment of the Junta de 597 Andalucía (postdoctoral fellowship of the Andalucía Talent Hub Program, Marie 598 Skłodowska-Curie actions, COFUND – Grant agreement No. 291780, to SMJ), the 599 European Union’s Horizon 2020 Research and Innovation Programme (Marie 600 Skłodowska-Curie grant agreement No. 676108 to SMJ), the European Research 601 Council Synergy grant 610028 (IMBALANCE-P), the Research Council of the 602 University of Antwerp (FORHOT TOP-BOF project; Methusalem grant) and the 603 Institut d’Estudis Catalans (IEC-FORHOT project). This work contributes to FSC-Sink, 604 CAR-ES and the ClimMani COST Action (ES1308). The Agricultural University of 605 Iceland and Mogilsá – the Icelandic Forest Research provided logistical support. We 606 thank Matthias Meys, Sara Diels, Johan De Gruyter, Giovanni Dalmasso, Fabiana 607 Quirós and Nadine Calluy for their invaluable help in the laboratory and thank Jennifer 608 Soong, Sara Vicca and James Weedon for their constructive suggestions. We also thank 609 Anne Cools and Tom Van Der Spiet for their assistance with the lab chemical analyses. 610 611 References 612 Arnalds, O., Hallmark, C.T., Wilding, L.P., 1995. Andisols from four different regions 613 of Iceland. Soil Science Society of America Journal 58, 161–169. 614 Arnalds, O., 2015. The soils of Iceland, 1st edn. Springer, Dordrecht. 615 Bai, E., Li, S., Xu, W., Li, W., Dai, W., Jiang, P., 2013. A meta‐analysis of 616 experimental warming effects on terrestrial nitrogen pools and dynamics. New 617 Phytologist 199, 441–451. 618 Bardgett, R. D., Streeter, T. C., Bol, R., 2003. Soil microbes compete effectively with 619 plants for organic‐nitrogen inputs to temperate grasslands. Ecology 84, 1277- 620 1287. 621 Bardgett, R., Bowman, W., Kaufmann, R., Schmidt, S., 2005. A temporal approach to 622 linking aboveground and belowground ecology. Trends in Ecology & Evolution 623 20, 634–641. 624 Bardgett, R., Manning, P., Morriën, E., Vries, F., 2013. Hierarchical responses of plant– 625 soil interactions to climate change: consequences for the global carbon cycle. 626 Journal of Ecology 101, 334–343 21 627 Bradford, M., Wieder, W., Bonan, G., Fierer, N., Raymond, P., Crowther, T., 2016. 628 Managing uncertainty in soil carbon feedbacks to climate change. Nature 629 Climate Change 6, 751–758. 630 Bremner, J.M., Keeney, D.R., 1965. Steam distillation methods for determination of 631 ammonium nitrate and nitrite. Anal Chim Acta 32(5):485. 632 Crowther, T.W., Todd-Brown, K.E.O., Rowe, C.W., Wieder, W.R., Carey, J.C., 633 Machmuller, M.B., Snoek, B.L., Fang, S., Zhou, G., Allison, S.D., Blair, J.M., 634 Bridgham, S.D., Burton, A.J., Carrillo, Y., Reich, P.B., Clark, J.S., Classen, 635 A.T., Dijkstra, F.A., Elberling, B., Emmett, B.A., Estiarte, M., Frey, S.D., Guo, 636 J., Harte, J., Jiang, L., Johnson, B.R., Kröel-Dulay, G., Larsen, K.S., Laudon, H., 637 Lavallee, J.M., Luo, Y., Lupascu, M., Ma, L.N., Marhan, S., Michelsen, A., 638 Mohan, J., Niu, S., Pendall, E., Peñuelas, J., Pfeifer-Meister, L., Poll, C., 639 Reinsch, S., Reynolds, L.L., Schmidt, I.K., Sistla, S., Sokol, N.W., Templer, 640 P.H., Treseder, K.K., Welker, J.M., Bradford, M.A., 2016. Quantifying global 641 soil carbon losses in response to warming. Nature 540, 104-108. 642 Davidson, E.A., Janssens, I.A., 2006. Temperature sensitivity of soil carbon 643 decomposition and feedbacks to climate change. Nature 440, 165-173. 644 Dormann, C.F., Woodin, S.J., 2002. Climate change in the Arctic: using plant 645 functional types in a meta‐analysis of field experiments. Functional Ecology 16, 646 4-17. 647 Dunn, R.M., Mikola, J., Bol, R., Bardgett, R.D., 2006. Influence of microbial activity 648 on plant–microbial competition for organic and inorganic nitrogen. Plant and 649 Soil 289, 321-334. 650 Farrell, M., Prendergast-Miller, M., Jones, D.L., Hill, P.W., Condron L.M., 2014. Soil 651 microbial organic nitrogen uptake is regulated by carbon availability. Soil 652 Biology Biochemistry 77, 261–267. 653 Friedlingstein, P., Cox, P., Betts, R., Bopp, L., von Bloh, W., Brovkin, V., Cadule, P., 654 Doney, S., Eby, M., Fung, I., Bala, G., John, J., Jones, C., Joos, F., Kato, T., 655 Kawamiya, M., Knorr, W., Lindsay, K., Matthews, H.D., Raddatz, T., Rayner, 656 P., Reick, C., Roeckner, E., Schnitzler, G.K., Schnur, R., Strassmann, K., 657 Weaver, A.J., Yoshikawa, C., Zeng, N., 2006. Climate–Carbon Cycle Feedback 658 Analysis: Results from the C4MIP Model Intercomparison. Journal of Climate 659 19, 3337-3353. 22 660 Groffman, P., Hardy, J., Fashu-Kanu, S., Driscoll, C., Cleavitt, N., Fahey, T., Fisk, M., 661 2011. Snow depth, soil freezing and nitrogen cycling in a northern hardwood 662 forest landscape. Biogeochemistry 102, 223–238. 663 Gudmundsdóttir, E., Óskarsson, Ú., Elmarsdóttir, Á. 2014. Effects of soil warming on 664 forest and grassland vegetation at Ölufsi. Rit Mógilsár, 31, 73–80. 665 Haney, R.L., Franzluebbers, A.J., Hons, F.M., Hossner, L.R., Zuberer, D.A., 2001. 666 Molar concentration of K2SO4 and soil pH affect estimation of extractable C 667 with chloroform fumigation–extraction. Soil Biology and Biochemistry 33, 668 1501-1507. 669 Haugwitz, M., Michelsen, A., Schmidt, I., 2011. Long-term microbial control of 670 nutrient availability and plant biomass in a subarctic-alpine heath after addition 671 of carbon, fertilizer and fungicide. Soil Biology and Biochemistry 43, 179–187. 672 Hayes, M.H.B., 1985. Extraction of humic substances from soil. In: Aiken, G. R., 673 McKnight, D. M.,Wershaw, R. L., MacCarthy, P. (Eds), Humic Substances. 674 Soil, Sediment and Water; Geochemistry, Isolation and Characterisation. Wiley, 675 New York, pp. 329-362. 676 Hessen, D.O., Jensen, T.C., Kyle, M., Elser, J.J., 2007. RNA responses to N- and P- 677 limitation; reciprocal regulation of stoichiometry and growth rate in Brachionus. 678 Functional Ecology 21, 956-962. 679 Hicks Pries, C.E., Castanha, C., Porras, R.C., Torn, M.S., 2017. The whole-soil carbon 680 flux in response to warming. Science 355, 1420-1423. 681 Hobbie, S.E., Nadelhoffer, K.J., Högberg, P., 2002. A synthesis: The role of nutrients as 682 constraints on carbon balances in boreal and arctic regions. Plant and Soil 242, 683 163-170. 684 IPCC, 2013. Climate Change 2013: The Physical Science Basis. Contribution of 685 Working Group I to the Fifth Assessment Report of the Intergovernmental Panel 686 on Climate Change. In: Stocker, T.F., Qin D., Plattner G.-K., Tignor M., Allen 687 S.K., Boschung J., Nauels A., Xia Y., Bex V., Midgley P.M. (Eds.) Cambridge 688 University Press, Cambridge, United Kingdom and New York. 689 Jaeger, C., Monson, R., Fisk, M., Schmidt, S., 1999. Seasonal partitioning of nitrogen 690 by plants and soil microorganisms in an alpine ecosystem. Ecology 80, 1883– 691 1891. 23 692 Jenkinson, D.S., Adams, D.E., Wild, A., 1991. Model estimates of CO2 emissions from 693 soil in response to global warming. Nature, 351, 304-306, 694 doi:10.1038/351304a0. 695 Jenkinson, D.S., Powlson, D.S., 1976. Effects of biocidal treatments on metabolism in 696 soil. 5. Method for measuring soil biomass. Soil Biology Biochemistry 8:209– 697 213. 698 Jonasson, S., Michelsen, A., Schmidt, I.K., Nielsen, E.V., Callaghan, T.V., 1996. 699 Microbial biomass C, N and P in two arctic soils and responses to addition of 700 NPK fertilizer and sugar: implications for plant nutrient uptake. Oecologia 106, 701 507-515. 702 Jones, D.L., Willett, V.B., 2006. Experimental evaluation of methods to quantify 703 dissolved organic nitrogen (DON) and dissolved organic carbon (DOC) in soil. 704 Soil Biology Biochemistry 38:991–999. 705 Kaiser, C., Fuchslueger, L., Koranda, M., Gorfer, M., Stange, C., Kitzler, B., Rasche, 706 F., Strauss, J., Sessitsch, A., Zechmeister-Boltenstern, S., Richter, A., 2011. 707 Plants control the seasonal dynamics of microbial N cycling in a beech forest 708 soil by belowground C allocation. Ecology 92, 1036–1051. 709 Kaiser, K., Zech, W., 1999. Release of Natural Organic Matter Sorbed to Oxides and a 710 Subsoil. Soil Sci. Soc. Am. J., 63, 1157-1166. 711 Leblans, N.I.W., 2016. Natural gradients in temperature and nitrogen: Iceland 712 represents a unique environment to clarify long-term global change effects on 713 carbon dynamics. Doctoral thesis. Faculty of Agricultural and Environmental 714 Sciences, Iceland. 715 Leblans, N.I.W., Sigurdsson, B.D., Vicca, S., Fu, Y., Penuelas, J., Janssens, I.A., 2017. 716 Phenological responses of Icelandic subarctic grasslands to short-term and long- 717 term natural soil warming. Global Change Biology 23, 4932-4945. 718 Liang, C., Schimel, J.P., Jastrow, J.D., 2017. The importance of anabolism in microbial 719 control over soil carbon storage. Nature Microbiology 2, 17105. 720 Lipson, D., Schmidt, S., Monson, R., 1999. Links between microbial population 721 dynamics and nitrogen availability in an alpine ecosystem. Ecology 80, 1623– 722 1631. 723 Mack, M. C., Schuur, E. A. G., Bret-Harte, M. S., Shaver, G. R., Chapin F. S. I., 724 2004. Ecosystem carbon storage in arctic tundra reduced by long-term nutrient 725 fertilization. Nature 431, 440–443. 24 726 Marañón-Jiménez, S., Soong, J.L., Leblans, N.I.W., Sigurdsson, B.D., Peñuelas, J., 727 Richter, A., Asensio, D., Fransen, E., Janssens, I.A., 2018. Geothermally 728 warmed soils reveal persistent increases in the respiratory costs of soil microbes 729 contributing to substantial C losses. Biogeochemistry 138 (3), 245-260. 730 McGuire, A., Anderson, L., Christensen, T., Dallimore, S., Guo, L., Hayes, D., 731 Heimann, M., Lorenson, T., Macdonald, R., Roulet, N., 2009. Sensitivity of the 732 carbon cycle in the Arctic to climate change. Ecological Monographs 79, 523– 733 555. 734 McGuire, A.D., Lawrence, D.M., Koven, C., Clein, J.S., Burke, E., Chen, G., Jafarov, 735 E., MacDougall, A.H., Marchenko, S., Nicolsky, D., Peng, S., Rinke, A., Ciais, 736 P., Gouttevin, I., Hayes, D.J., Ji, D., Krinner, G., Moore, J.C., Romanovsky, V., 737 Schädel, C., Schaefer, K., Schuur, E.A.G., Zhuang, Q., 2018. Dependence of the 738 evolution of carbon dynamics in the northern permafrost region on the trajectory 739 of climate change. Proceedings of the National Academy of Sciences. 740 Melillo, J.M., Frey, S.D., DeAngelis, K.M., Werner, W.J., Bernard, M.J., Bowles, F.P., 741 Pold, G., Knorr, M.A., Grandy, A.S., 2017. Long-term pattern and magnitude of 742 soil carbon feedback to the climate system in a warming world. Science 358, 743 101-105. 744 Melillo, J.M., Steudler, P.A., Aber, J.D., Newkirk, K., Lux, H., Bowles, F.P., Catricala, 745 C., Magill, A., Ahrens, T., Morrisseau, S., 2002. Soil warming and carbon-cycle 746 feedbacks to the climate system. Science, 298, 2173-2176. 747 Michielsen, L. 2014. Plant communities and global change: Adaptation by changes in 748 present species composition or adaptation in plant traits. A case study in Iceland. 749 Unpublished Master University of Antwerp, Antwerp. 750 Miltner, A., Bombach, P., Schmidt-Brücken, B., Kästner, M., 2012. SOM genesis: 751 microbial biomass as a significant source. Biogeochemistry 111, 41-55. 752 Mooshammer M, Wanek W, Hämmerle I, Fuchslueger L, Hofhansl F, Knoltsch A, 753 Schnecker J, Takriti M, Watzka M, Wild B, Keiblinger K, Zechmeister- 754 Boltenstern S, Richter A. 2014a. Adjustment of microbial nitrogen use 755 efficiency to carbon: nitrogen imbalances regulates soil nitrogen cycling. Nat 756 Commun 5:3694. 757 Mooshammer, M., Wanek, W., Zechmeister-Boltenstern, S., Richter, A., 2014b. 758 Stoichiometric imbalances between terrestrial decomposer communities and 25 759 their resources: Mechanisms and implications of microbial adaptations to their 760 resources. Frontiers in Microbiology 5, 1-10. 761 Natali, S., Schuur, E., Rubin, R., 2012. Increased plant productivity in Alaskan tundra 762 as a result of experimental warming of soil and permafrost. Journal of Ecology 763 100, 488–498. 764 Nedwell, D., 1999. Effect of low temperature on microbial growth: lowered affinity for 765 substrates limits growth at low temperature. FEMS Microbiology Ecology 30: 766 101-111. 767 Nelson, P.N., Dictor, M.C., Soulas, G., 1994. Availability of organic carbon in soluble 768 and particle-size fractions from a soil profile. Soil Biology and Biochemistry 26, 769 1549-1555. 770 Olsen, S.R., Cole C.V., Watanabe, W.S., Dean, L.A., 1954. Estimation of available 771 phosphorus in soils by extraction with sodium bicarbonate. United States 772 Department of Agriculture Circular no. 939 773 Pansu, M., Gautheyrou, J., 2006. Handbook of soil analysis. Mineralogical, organic and 774 inorganic methods. Springer, Montpellier 775 Poeplau, C., Kätterer, T., Leblans, N.I.W., Sigurdsson, B.D., 2016. Sensitivity of soil 776 carbon fractions and their specific stabilisation mechanisms to extreme soil 777 warming in a subarctic grassland. Global Change Biology 23, 1316-1327. 778 Quinn, G.P., Keough, M.J., 2009. Experimental design and data analysis for biologists. 779 Cambridge University Press, Cambridge. 780 Rustad, L., Campbell, J., Marion, G., Norby, R., Mitchell, M., Hartley, A., Cornelissen, 781 J., Gurevitch, J., Available, N., 2001. A meta-analysis of the response of soil 782 respiration, net nitrogen mineralization, and aboveground plant growth to 783 experimental ecosystem warming. Oecologia 126, 543-562. 784 Sigurdsson, B.D., Leblans, N.I.W., Dauwe, S., Gudmundsdottir, E., Gundersen, P., 785 Gunnarsdottir, G.E., Holmstrup, M., Ilieva-Makulec, K., Katterer, T., 786 Marteinsdottir, B.-S., Maljanen, M., Oddsdottir, E.S., Ostonen, I., Penuelas, J., 787 Poeplau, C., Richter, A., Sigurdsson, P., Van Bodegom, P., Wallander, H., 788 Weedon, J., Janssens, I., 2016. Geothermal ecosystems as natural climate change 789 experiments: The ForHot research site in Iceland as a case study. Icelandic 790 Agricultural Sciences, 53-71. 791 Schimel, J. P., Bennett, J., 2004. Nitrogen mineralization: challenges of a changing 792 paradigm. Ecology, 85(3), 591–602. 26 793 Sistla, S.A., Moore, J.C., Simpson, R.T., Gough, L., Shaver, G.R., Schimel, J.P., 2013. 794 Long-term warming restructures Arctic tundra without changing net soil C 795 storage. Nature, 497, 615-618. 796 Smith, S., Edmonds, J., Hartin, C., Mundra, A., Calvin, K., 2015. Near-term 797 acceleration in the rate of temperature change. Nature Climate Change 5, 333- 798 336. 799 Sterner, R.W., Elser, J.J., 2002. Ecological Stoichiometry: Biology of Elements from 800 Molecules to the Biosphere. Princeton University Press, Princeton, NJ, USA. 801 Tarnocai, C., Canadell, J.G., Schuur, E.A.G., Kuhry, P., Mazhitova, G., Zimov, S., 802 2009. Soil organic carbon pools in the northern circumpolar permafrost region. 803 Global Biogeochemical Cycles 23, GB2023. 804 Todd-Brown, K., Randerson, J., Post, W., Hoffman, F., Tarnocai, C., Schuur, E., 805 Allison, S., 2013. Causes of variation in soil carbon simulations from CMIP5 806 Earth system models and comparison with observations. Biogeosciences 10, 807 1717–1736. 808 Turner, M.M., Henry, H.A.L., 2010. Net nitrogen mineralization and leaching in 809 response to warming and nitrogen deposition in a temperate old field: the 810 importance of winter temperature. Oecologia 162, 227-236. 811 Wagai, R., Sollins, P., 2002. Biodegradation and regeneration of water-soluble carbon 812 in a forest soil: leaching column study. Biology and Fertility of Soils 35, 18-26. 813 Walker, T.W.N., Kaiser, C., Strasser, F., Herbold, C.W., Leblans, N.I.W., Woebken, D., 814 Janssens, I.A., Sigurdsson, B.D., Richter, A., 2018. Microbial temperature 815 sensitivity and biomass change explain soil carbon loss with warming. Nature 816 Climate Change 8, 885-889. 817 Wild, B., Schnecker, J., Knoltsch, A., Takriti, M., Mooshammer, M., Gentsch, N., 818 Mikutta, R., Alves, R., Gittel, A., Lashchinskiy, N., Richter, A., 2015. Microbial 819 nitrogen dynamics in organic and mineral soil horizons along a latitudinal 820 transect in western Siberia. Global Biogeochemical Cycles 29, 567–582. 821 Wu, Z., Dijkstra, P., Koch, G., Peñuelas, J., Hungate, B., 2011. Responses of terrestrial 822 ecosystems to temperature and precipitation change: a meta‐analysis of 823 experimental manipulation. Global Change Biology 17, 927–942. 824 Xu, X., Thornton P.E., Post W.M., 2013. A global analysis of soil microbial biomass 825 carbon, nitrogen and phosphorus in terrestrial ecosystems. Global Ecology and 826 Biogeography 22, 737-749. 27 827 Yano, Y., Brookshire, E., Holsinger, J., Weaver, T., 2015. Long-term snowpack 828 manipulation promotes large loss of bioavailable nitrogen and phosphorus in a 829 subalpine grassland. Biogeochemistry 124, 319–333. 830 Zogg, G., Zak, D., Pregitzer, K., Burton, A., 2000. Microbial immobilization and the 831 retention of anthropogenic nitrate in a northern hardwood forest. Ecology 81, 832 1858–1866. 833 28 834 Figure captions: 835 Figure 1: A) Microbial respiration and B) microbial biomass C in unwarmed soils in 836 response to the C, N and P additions. Microbial respiration was measured 12 h after the 837 additions at the mean annual soil temperature (5 °C). Microbial biomass was measured 838 six days after the substrate additions. Different letters indicate significant differences by 839 Tukey´s post hoc tests at α=0.05. 840 841 Figure 2: A) Total microbial N, B) remaining NH4+ and C) remaining dissolved 842 organic N in unwarmed soils six days after the C, N and P additions. Triangles indicate 843 the initial NH4+ concentration in soil prior to the soil incubation. Different letters 844 indicate si nificant differences by Tukey´s post hoc tests at α=0.05. 845 846 Figure 3: C:N ratios in A) soil microbes and B) K2SO4-extractable organic pools from 847 unwarmed soils six days after to the C, N and P additions. Different letters indicate 848 si nificant differences by Tukey´s post hoc tests at α=0.05. 849 850 Figure 4: A) Dissolved organic C pools, B) their relative accessibility and C) C:N ratios 851 of K2SO4-extractable organic pools, microbial biomass and the C:N imbalance between 852 these at the different intensities of soil warming. Data correspond to the initial values in 853 soils before the incubation or substrates addition. The relative accessibility of 854 extractable soil C pools was calculated as their ratio to the total organic C pool. The 855 C:N imbalance was calculated as the ratio of C:N of soil organic pools over microbial 856 C:N. Different letters indicate significant differences by Tukey´s post hoc tests at 857 α=0.05. 858 859 Figure 5: A) Microbial respiration, B) microbial biomass C, C) total microbial N and D) 860 microbial C:N ratios in response to the C, N and P addition at the different intensities of 861 soil warming. Microbial respiration was measured 12 h after the additions at the mean 862 annual soil temperature (5 °C). Microbial biomass C and N were measured six days 863 after the additions. Different letters indicate significant differences among the soil 864 warming intensities according to two-way ANOVAs and Tukey´s post hoc tests. * and 865 ** indicate significant differences between substrate addition levels within each soil 866 warming intensity according to one-way ANOVAs: *0.01