Increased N2O Production from Soil Organic Matter Following a Simulated Fall-Freeze-Thaw Cycle: Effects of Fall Urea Addition, Soil Moisture, and History of Manure Applications


 Adding nitrogen substrates to soils can induce short-term changes in soil organic matter (SOM) transformations – a response termed the ‘priming effect’. However, it is unknown how priming effects on nitrous oxide (N2O) emissions can be altered following a strong freeze-thaw cycle. A mesocosm experiment evaluated two soil managements: with and without history of manure applications. These soils were subjected to three moisture regimes: Low, Medium and High. Apart from the controls, which received no N, we banded 15N-labelled urea into these soils representing a typical fall fertilization, and subsequently simulated a wide fall-freeze-thaw cycle, with temperatures from + 2, to -18, and finally + 23°C, respectively. The overall highest N2O production was observed 1 day after thawing. At that time, measurements of N2O site preference indicated that denitrification produced 83% of the N2O flux. Relative to the unamended controls (baseline), adding urea consistently triggered a 24% greater cumulative N2O production specifically originated from SOM following thawing (245 vs. 305 µg N2O-N kg− 1 soil, P = 0.022). This substantiates a positive priming of SOM that manifested shortly after the rapid, wet thawing of the soils. Soils having a manure history or higher moisture also exhibited an augmented production of N2O from SOM (Ps < 0.01). Although the overall priming of SOM was positive, two weeks after thawing, negative priming of daily N2O fluxes also occurred, but only in soils under High moisture. Besides urea additions, the propensity for primed N2O emissions from SOM after thawing was influenced by increasing moisture and earlier manure applications.


Introduction 37
Nitrous oxide (N2O) is a potent greenhouse gaswith even 300-fold higher global warming 38 potential than carbon dioxide (CO2) on mass basis (Parry et al. 2007 As climate change continues to take place, extreme fluctuations in the weather conditions can 54 occur with increased frequency. Although soil freezing and thawing are already common 55 phenomena in cold regions with relatively high latitude and altitude, the intensity and frequency 56 of freeze-thaw cycles are gradually increasing as a feedback to escalating climate change 57 8 two soil managements (i.e., CT and SW), three moisture regimes and two N additions were 138 applied as experimental treatments. The three soil moisture regimes were Low (i.e., WFPS of 139 45% over the fall, reaching 70% during freezing, and falling to 55% by the end of the thawing 140 phase), Medium (i.e., 55-80-65% WFPS) and High (65-90-75% WFPS). The N addition 141 treatments were urea (5 atom% 15 N) (Sigma-Aldrich, St. Louis, MO, US) and control (without 142 urea addition). The experimental design was a factorial with three replicates. In sum, the three 143 experimental factors were: history of field manure injection (i.e., CT and SW), three moisture 144 regimes (Low, Medium and High), and N addition (urea and unamended control). A total of 36 145 experimental pots were used for flux measurement during the experiment. 146 The moisture and N addition treatments were established on Day 0 of the fall phase which lasted 147 for 27 days. The N addition consisted of 0.29 g of powder consistency 5 atom% 15 N-urea per pot 148 placed at 5 cm depth to represent fertilizer banding. This rate was equivalent to 85 kg N ha -1 , 149 which simulates a common fall fertilization for a canola crop in the subsequent growing season. 150 After applying the N treatment, room-temperature deionized (DI) water was added to achieve fall 151 moisture levels of 45 (Low), 55 (Medium) and 65% (High) WFPS. As necessary, the WFPS was 152 maintained by weighting the pots and adding DI water every day throughout the fall phase. All 153 pots and glass flasks with DI water were kept at 2°C. Cardboard was placed 3-5 cm above the 154 top of the pots to prevent rapid evaporation while still allowing air circulation. 155

Simulated freezing phase 156
On Day 28 after the beginning of the experiment, a freezing phase was started by moving all pots 157 from a temperature of 2 to -18 °C. This freezing phase lasted for 27 days (i.e., Days 28 to 55 158 following the urea addition) which assured that the soil columns became completely frozen. 159 Additionally, to simulate multiple water inputs that accumulate over a typical winter in Central 160 Alberta (snow and ice precipitation), DI water (at 2 °C) was added during the freezing phase in 161 three successive increments. These water additions were done incrementally in amounts 162 Jersey, USA). Temperature (20 °C) and sample flow rate (1.5 standard L min -1 ) were kept 187 constant in the instrumentation. TDLWintel software provided system control as well as data 188 acquisition and recording at 1 Hz resolution (Daly and Hernandez-Ramirez 2020). 189 Aerodyne analyzer was coupled with a flow-through, recirculation, non-steady-state chamber. A 190 custom-made cylindrical polyvinyl chloride chamber system consisted of a chamber base and a 191 chamber top. The cross-sectional area of the chamber was 184 cm 2 (15.3 cm diameter). The 192 chamber base was installed 3 cm inside the soil, leaving 7 cm above the soil surface. These 193 chamber bases were installed at the center of each pot at the beginning of the fall phase. The 194 chamber top (5 cm in height) was equipped with two tubing connection ports for gas 195 11 recirculation (one for inlet and another one for outlet), a stainless capillary tubing (3/16" in inner 196 diameter, 10 cm in length) on the wall for the purpose of pressure equilibration, and rubber seals 197 fitted to the chamber top to ensure headspace closure. The total chamber headspace was 2.2 L. 198 The chamber enclosure and sample recirculation with the Aerodyne lapsed for 3 minutes. 199 During every flux measurement, air temperature and pressure were recorded by a HOBO 200 UX100-001 data logger and a Testo 511 barometer (Testo Inc., Lenzkirch, Germany), 201 respectively. 202

Measurements of CO2 fluxes 203
During the fall and freezing phases, CO2 fluxes from the same soil pots were determined by a 204 simple system, which included a Picarro G2508 cavity ring-down spectroscope (CRDS) with a 205 105 mL analytical cell at a constant 140 Torr pressure and at a temperature of 45 °C (Picarro, 206 Santa Clara, CA, USA), a low-leak diaphragm A0702 pump (Picarro, Santa Clara, CA, USA) 207 and the custom-made chamber described above. Similar to the N2O measurements with the 208 Aerodyne, a vacuum pump enabled the re-circulation of gas sample flow through the chamber 209 headspace at a rate of 240 standard mL min -1 during an enclosure time of 3 min. 210 After soil thawing (following Day 56 after the urea addition), CO2 fluxes were measured 211 with an automated chamber system, which included the CRDS described above, and an eosMX 212 multiplexer connected to 12 eosAC automated chambers (Eosense Inc., Dartmouth, NS, Canada) 213 (Roman-Perez and Hernandez-Ramirez 2021). The total headspace of the automated chamber 214 system was 2.8 L. Each flux measurement lapsed 10 min. 215

Flux calculation 216
The daily fluxes of N2O and CO2 were calculated as follows: 217 where F is the gaseous flux (µg kg -1 d -1 ), dC/dt is the slope of a simple linear regression or as the 218 first derivative of a quadratic regression at t0 (µL L -1 s -1 ), V is the headspace volume of the gas 219 chamber (L); S is the dry soil weight (kg), P is the pressure in the chamber headspace during 220 measurement (atm), R is the gas constant (atm µL K -1 µmol -1 ), T is the temperature at chamber 221 headspace during measurement (K), M is the molar mass of N within N2O (28 g mol -1 ), or C 222 within CO2 (12 g mol -1 ), and k is a conversion factor for the flux unit (from µg kg -1 s -1 to µg kg -1 223 d -1 ). 224

Calculations of N2O derived nitrification and denitrification 225
With the aim of examining the contributions of nitrification and bacterial denitrification 226 processes to the total N2O production, the N2O measurements conducted 1 day after thawing 227 were used to estimate the site preference (SP) under natural abundance. This is because the large 228 N2O production on this day improved the accuracy of isotopic ratio measurements (Waechter et 229 al. 2008). 230 Calculations for 15α R, 15β R, δ 15α N2O, δ 15β N2O, and δ 15bulk N2O were as follows: The contributions of nitrification and bacterial denitrification to N2O production were calculated 245 as follows: 246 where Fni and Fdeni are the proportional contributions of nitrification and denitrification, 247 respectively. This assumes that the SPs of the nitrification and denitrification sources are 0 and 248 33 ‰, respectively (Sutka et al. 2006). 249

Calculation of the N2O derived from SOM-N and the priming effects 250
As our study used urea labelled with 15 N, a mass balance based on isotopic composition of 251 the emitted N2O (atom%) was conducted to separate the contributions of two N pools (i.e., added 252 urea-N vs. existing SOM-N sources) to the overall N2O flux using the entire dataset over the 253 thawing phase. Atom% 15 N2O is the isotopic percentage of 15  Similar as for the SP derivation described above, the atom% 15 N2O emitted from each 255 experimental soil pot during each chamber enclosure was obtained from the Keeling plot 256 intercepts. 257 The fractions of N2O production derived from added 15 N-urea and from SOM-N were 258 calculated as follows: 259 where FN2O15N-urea and FN2OSOM are the fractions of N2O production derived from added 15 N-260 urea and from existing SOM-N, respectively; Atom% 15 N2O15N-urea and Atom% 15 N2Ocontrol are the 261 isotopic percentages of 15 N in N2O emitted from the experimental pots with and without added 262 urea, respectively; N2Ocontrol is the N2O flux from the control soils (without urea). 263 The priming effect of daily N2O fluxes was calculated as follows: 264 In the Eq.
[12], N2O priming effect >0 corresponds to a positive priming effect caused by 265 added urea, whereas <0 indicates a negative priming effect. More specifically, daily negative 266 16 priming was identified when the mean daily SOM-derived N2O flux from a urea-amended soil 267 was one standard error below the zero baseline (which was defined as the corresponding control 268 without urea addition). Results of N2O priming were expressed as magnitude and also in relative 269 basis as a percentage of the total flux for each soil pot receiving urea. With the aim of measuring 15 N isotopic composition in the soil at natural abundance (without 284 addition of labelled urea), soil samples were oven-dried at 60 °C and ball-ground to a fine 285 consistency to ensure homogeneity for isotope analysis. The soil δ 15 N was determined by using a 286 Flash 2000 Elemental Analyzer (Thermo Fisher Scientific, Delft, Netherlands) to dry combust 287 the soil sample converting all N to N2. This analyzer was interfaced online to a Finnigan Delta V 288 Plus isotopic ratio mass spectrometer (Thermo Electron, Bremen, Germany) to detect the 15 N 289 isotope composition. 290 Based on the 15 N isotopic compositions of soil N and N2O emitted from control soils, isotope 291 discrimination (ε) was calculated as follows: 292 where 15 RN2O is the isotopic ratio of N2O emitted on Day 57 (1 day after thawing) and 15 RsoilN is 293 the isotopic ratio of soil N. A positive ε implies the enrichment of 15 N during the processes of 294 transforming soil N to N2O emissions; a negative ε implies a depletion of 15 N during this 295 conversion from soil N to emitted N2O. This ε estimation was based on the premise that the 296 transformation from the SOM-N pools into the emitted N2O pool was unidirectional. 297

Statistical analyses 298
Statistical analyses were performed in R 3.1.3 (R Core Team 2014) at alpha critical value of 299 0.05. The data were transformed to meet the assumptions of normality and homoscedasticity as 300 necessary. The effects of manure history (CT vs. SW soils), N (urea vs. control) and soil water 301 content (Low, Medium vs. High) treatments on soil NH4 + , NO3 -, cumulative N2O, SOM-derived 302 N2O and cumulative CO2 were examined by three-way analysis of variance (ANOVA) for a 303 fixed-effect model with interaction analysis. We run two-way analysis ANOVA for a fixed-304 18 effect model to determine the effects of manure history and water content on primed N2O and 305 urea-derived N2O as well as the differences in the contributions of nitrification and 306 denitrification to the N2O emitted 1 day after thawing. Tukey's Honest Significant Difference 307 (HSD) test was used to compare the difference further in cases where the treatment effects 308 described above were significant. 309

N2O production derived specifically from SOM-N and priming effects 328
In parallel with the results of total N2O emissions, the main effects of urea addition, moisture 329 content, and history of manure applications showed separate, significant impact on the 330 cumulative N2O production derived from SOM during the period after thawing (Table 2 Specifically, there was a tendency for the SW soil to have a numerically higher positive priming 341 20 effect compared with the CT soil in both magnitude (Fig. 2b) and relative (+17 vs. +6% Fig. 2c) 342

basis. 343
Most of the daily N2O fluxes following soil thawing showed positive priming (Fig. 3). Soils 344 under Low moisture regime showed a peak of primed daily N2O fluxes 1 day after thawing (Fig.  345   3a). In the case of soils under Medium and High moisture regimes, the peak of positive priming 346 in daily N2O fluxes occurred 1 day later (i.e., 2 days after thawing) (Fig. 3b and Fig. 3c). Across 347 the three moistures in the SW soil, the peak of daily positive priming was greater at the two 348 higher moisture contents (29.44 µg N2O-N kg -1 d -1 at Low vs. 62.95 at Medium and 52.82 at 349 High moisture content). Overall, peak primings were greater for SW than for CT soil under both 350 Medium and High soil moisture regimes (Fig. 3). In general, following these early peaks, primed 351 N2O fluxes gradually dropped back to approximately the zero baseline. 352 Interestingly, about 2 weeks after soil thawing, negative priming of daily N2O fluxes was clearly 353 observed (i.e., SOM-derived N2O << control N2O). These episodes of evident negative primings 354 occurred only under High soil moisture for both SW and CT soils (Fig. 3c). The negative 355 priming effect began slightly earlier in the SW soil (Day 70 of the experiment) than in the CT 356 soil (Day 71) (Fig. 3c). The last day that registered negative priming effect in SW soil was Day 357 86 of the experiment; in CT soil, it was Day 85. Towards the end of the experiment, the 358 magnitude of the priming effects returned to zero or became minor. Collectively, the results 359 indicate that higher moisture generated more dynamic priming activity. 360 21 Contrary to the wide responses of cumulative and daily SOM-derived N2O fluxes to urea 361 addition, history of manure applications and soil moisture as aforementioned, the direct 362 contribution of the urea-N source to cumulative N2O fluxes (urea-derived N2O) was consistent 363 across all assessed experimental factors and treatment combinations, with no significant effects 364 caused by manure history or soil moisture (Table 2, Fig. 2a). 365

Contributions of denitrification to the peak of N2O fluxes 366
The very large N2O fluxes that occurred 1 day after soil thawing provided the opportunity to 367 measure and allocate the N2O produced from nitrification and denitrification sources in all 368 unamended control soils (i.e., under natural 15 N abundance conditions) and across the three 369 moisture contents (Fig. 4). The results for 15 N-N2O SP ranged from 1.0 ‰ in the CT soil under 370 High moisture to 5.7 ‰ in the SW soil under Medium moisture (P> 0.05) (Fig. 4c). This suggest 371 that denitrification dominated the vigorous N2O production, with 83% contribution in the case of 372 the SW soil under Medium moisture and up to nearly all the N2O produced in the case of the CT 373 soil under High moisture (97%) (Fig. 4b). When averaging across the three moistures, the 374 relative contributions of denitrification to N2O production in the CT soil were marginal-375 significantly larger than those in the SW soil (P= 0.06; Fig. 4b). 376

Isotopic depletion of 15 N-N2O relative to soil N 377
There was a consistently negative depletion of 15 N (ε) during the transformation from the soil N 378 pool to the emitted N2O pool across all soil management histories and water regimes (Table 3). 379

Inorganic soil N concentrations 380
There were no significant effects of experimental factors on the NH4 + concentrations (Table 2). 381 There was a significant interaction of manure history and water content on soil NO3 -382 concentrations (Table 2, Fig. 5). Specifically, irrespective of urea addition, soil NO3 -383 concentration was significantly lower in the treatment combination of CT soil at Low moisture 384 than most of other treatments, with the only exception of the treatment combination of CT soil at 385 Medium moisture (data not shown). The NO3concentration apparently increased with increasing 386 soil moisture content in CT soils, but this pattern was not found in SW soils (Fig. 5b). The NO3 -387 concentration was in general greater in the SW soil than the CT soil (Fig. 5b). As expected, soils 388 receiving added urea had greater increments in the NO3concentration than the soil without urea 389 (i.e., CT + urea > CT control; SW + urea > SW control, Fig. 5b). These increased nitrate 390 concentrations indicate the occurrence of nitrification in these soils. Furthermore, both NH4 + and 391 NO3concentrations increased over time from the beginning to the end of the experiment, 392 including in the control soils; therefore, this indicates that active mineralization and 393 ammonification from SOM-N also took place over the experimental period. 394

Soil CO2 fluxes 395
Within most of the fall and freezing phases, CO2 fluxes were generally low and relatively stable 396 across all treatment combinations. Over the fall phase, CO2 fluxes averaged 1.29 ± 0.13 µg CO2-397 C kg -1 d -1 (Fig. 1c). Afterwards, CO2 fluxes steadily decreased to 0.57 ± 0.14 µg CO2-C kg -1 d -1 398 on Day 6 of the freezing phase, and then became negligible (Fig. 1c). 399 23 Similar to N2O fluxes, most of the dynamics of CO2 fluxes took place shortly after soil thawing 400 (Fig. 1). Three days after thawing, the average CO2 flux across all treatments sharply peaked at 401 8.65 ± 0.29 µg CO2-C kg -1 d -1 . Thereafter, CO2 fluxes slowly decreased over time, reaching 1.94 402 ± 0.16 µg CO2-C kg -1 d -1 on the last day of the experiment (Fig. 1c). It is noted that there was a 403 strong correlation between daily CO2 and N2O fluxes following thawing (r=0.968, P<0.001; 404 Supplementary Fig. 1 and Fig. 1), with the exception of the first 3 days after thawing when the 405 N2O fluxes were decoupled and disproportionally larger than the measured CO2 fluxes. 406 Over the entire experiment and specifically in the period after thawing, the cumulative CO2 407 emissions significantly increasing with higher soil moisture in the SW soil that had not received 408 fall-urea (data not shown; Table 2). 409 This is the first time in the literature that the direction and magnitude of potential priming effects 415 on augmented N2O emissions shortly after thawing has been quantified (Fig. 2, Fig. 3 is plausible that the manured soils (SW) in our study showed a more intense response of primed 433 N2O dynamics to the fall-applied urea because the previous field manure injections in this soil 434 had increased the easily decomposable SOM. It is noted that the SW soil showed a tendency for 435 higher organic C concentrations than CT (Table 1) organic C could also reduce the N2O priming in urea-amended soils. This is explained by the 439 25 increased conversion of N2O into N2 as driven by heterotrophic utilization of organic C that 440 enhances the last step of bacterial denitrification (Daly and Hernandez-Ramirez 2020). Future 441 research focusing on these drivers of N2O priming would help to deepen our understanding of C 442 and N turnover in soils, particularly in agricultural systems that experience high, frequent 443 nutrient outputs and inputs such as croplands that receive heavy manure additions. We 444 hypothesize that in environments that are N-rich and even N-saturated, coupling availabilities of 445 C and N could reduce and even cancel the potential priming effects on N2O emissions derived 446 from SOM-N. 447

448
In addition to the effects of contrasting manure history on N2O priming following thawing, soil 449 moisture clearly affected the dynamics of primed N2O fluxes as well. Although the overall 450 priming was positive across all experimental combinations, only soils under High moisture 451 experienced negative N2O priming of daily fluxes and also longer-lasting priming effects as 452 noted above (Fig. 3). The temporal shift of daily priming effects from positive to negative and 453 eventually back to zero priming at High moisture could be explained by the mechanism of 454 preferential substrate utilization. The hypothesis of preferential substrate utilization states that 455 when given a variety of nutrient supplies, microorganisms prefer easily available and highly 456 accessible substrates over recalcitrant substrates (Cheng 1999, Cheng andKuzyakov 2005, 457 Blagodatskaya and Kuzyakov 2008). Within the context of our study, it could be postulated that 458 at the onset of thawing, soil microbes initially utilized the easily available substrates, and then 459 switched to consuming more complex substrates (e.g., wheat straw residues and roots) in  Table 1. Soil physical and chemical properties at the 0-15 cm depth increment of the soils 663 with (SW) and without (CT) history of liquid manure injections. Note that only organic 664 carbon showed a magnitude difference between these two soil managements, with SW 665 slightly higher than CT (P > 0.05). 666    (a) Cumulative N2O emissions allocated to urea and soil organic matter (SOM) sources, (b) magnitude priming and (c) relative priming caused by urea addition following soil thawing. SW and CT stand for soils with and without a history of manure additions, respectively. Low, Med, and High correspond to moisture regimes where Med stands for Medium. In Panel a, N and C acronyms correspond to the urea-N addition treatment and the zero-N addition (control) treatment, respectively. In Panel a, different letters indicate signi cant difference in total cumulative N2O (uppercase), SOM-derived N2O (lowercase) and urea-derived (italic) N2O emissions after thawing (P < 0.05). In Panels b and c, N2O primings were respectively shown as magnitudes and also in relative basis as percentages of the total uxes (shown in Panel a) of soil pots receiving urea. Error bars correspond to one standard error.

Figure 3
Primed daily N2O uxes following soil thawing. SW and CT stand for soils with and without a history of manure additions, respectively. Low, Med, and High correspond to moisture regimes where Med stands for Medium. Positive and negative primed daily N2O uxes represent positive and negative priming effects, respectively. Error bars correspond to one standard error of the mean. Figure 4 (a) Magnitude and (b) relative contributions of nitri cation and denitri cation, as well as (c) site preference for the N2O uxes emitted 1 day after thawing (Day 57 of the experiment). SW and CT stand for soils with and without a history of manure additions, respectively. Low, Med, and High correspond to moisture regimes where Med stands for Medium. In Panels a and b, numbers in the columns are respectively the ux magnitude and percentage of N2O emissions produced via denitri cation or nitri cation. Error bars correspond to standard error of the mean. Figure 5 Soil (a) ammonium and (b) nitrate concentrations at the end of the experiment for the soils with (SW) and without (CT) history of manure additions at Low, Medium (Med) and High moisture regimes. Horizontal lines (with one standard error) across moisture contents are the concentrations of (a) ammonium and (b) nitrate of the two soils at the beginning of the experiment (prior to urea addition and establishment of the three moisture regimes). Different letters indicate signi cant differences among treatment combinations (P < 0.05). Error bars correspond to one standard error of the mean. n.s. = not signi cant.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. SupplementarydatamesocosmwoCopyghr.docx