Microbially-mediated P fluxes in calcareous soils as a function of water-extractable phosphate

1 Title: Microbially-mediated P fluxes in calcareous soils as a function of water-extractable 2 phosphate 3 4 Authors: 5 Schneider Kimberley D1,2,3, Voroney R Paul2, Lynch Derek H4, Oberson Astrid1, Frossard 6 Emmanuel1,Bünemann Else K1,5 7 1 8 Eschikon 33, CH-8315 Lindau, Switzerland 9 2 School of Environmental Sciences, University of Guelph, Guelph, ON, Canada, N1G 2W1 10 3 (Present Address) Agriculture and Agri-Food Canada, Science and Technology Branch, 174 11 Stone Road West, Guelph, ON, Canada N1G 4S9 12 4 13 Canada, B2N 5E3 14 5 Institute of Agricultural Sciences, Swiss Federal Institute of Technology Zurich (ETH), Department of Plant, Food, and Environmental Sciences, Dalhousie University, Truro, NS, Research Institute of Organic Agriculture (FiBL), Ackerstrasse, 5070 Frick, Switzerland 15 16 Corresponding author: 17 Schneider, KD 18 Agriculture and Agri-Food Canada, Science and Technology Branch, 174 Stone Road West, 19 Guelph, ON, Canada N1G 4S9 20 Email: kschne01@uoguelph.ca 21 Phone: +1 226 217 8024 22 Fax: +1 2262178187 23 1 © 2016. This manuscript version is made available under the Elsevier user license http://www.elsevier.com/open-access/userlicense/1.0/ 33 24 Keywords:organic phosphorus mineralization, isotopic dilution method, 25 phosphorus, perennial forage, microbial P turnover 26 Abstract 27 Soil phosphorus (P) tests are designed to indicate plant-available inorganic orthophosphate (Pi), 28 but fail to account for Pi that may become available through organic phosphorus (Po) 29 mineralization. This P source may be especially important in soils with low concentrations of 30 solution and labile Pi. We assessed gross Pomineralization and immobilizationusing labeling with 31 33 32 from forage fields of dairy farmsin Ontario, Canada. Rapid microbial 33P uptake during 33 incubation was found for the soils with the lowest available Pias indicated by both Olsen soil test 34 P and water-extractable Pi. The tracer incorporation into microbial P after 8 days ranged from 7 35 to 44% of applied 33P and was negatively related to water-extractable Pifollowing a power-type 36 relationship. As concentrations of microbial P were similar in all soils, this suggests faster 37 turnover of P in the microbial biomass at water-extractable Pi below 0.1 mg P kg-1 soil. Daily 38 gross Po mineralization rates ranged from 0.2 to 2.8 mg P kg-1 soil d-1 and contributed 7 to 56% 39 of the isotopically-exchangeable P in 8 days. Based on these findings, microbial processes have 40 the potential to make a significant contribution to forage P nutrition. 41 1.0 Introduction 42 A thorough understanding of processes that govern phosphorus (P) availability in soils is 43 important as too little available orthophosphate (Pi) can limit crop yields, while excess Pi can 44 result in off-site environmental degradation (Condron, 2004). P availabilityhas generally been 45 considered to be controlledby physical and chemical processes including sorption-desorption and 46 precipitation-dissolution; however, biologically driven processes including organic P (Po) 47 mineralizationand microbial Pi immobilization can also have a significant effect onPi availability 48 (Frossard et al., 2000; Bünemann, 2015). The ability of the microbial biomass to contribute to P, available P in four calcareous Alfisolswithvarying concentrations of Olsen soil test Pthat were collected 2 49 plant available P depends on the quantity of the microbial P pool and its turnover time (time for a 50 nutrient pool to completely renew itself) (Oberson and Joner, 2005). 51 Determination of the contribution of Po mineralization to plant-available P would be useful, 52 especially in systems with low concentrations of available Pi.For example, organic agricultural 53 production systems often have low indices of plant-available P (Entz et al., 2001; Gosling and 54 Shepherd, 2005; Oberson and Frossard, 2005; Knight et al., 2010). At the same time, the ability 55 of commonly used soil tests to accurately predict plant-available P has been questioned because 56 they do not account for biological contributions to plant-available Pi(Condron, 2004; Fortune et 57 al., 2005; Roberts et al., 2008; Nash et al., 2014). It can be assumed thatrelative to physico- 58 chemical processes, Po mineralization plays a greater role in contributing to plant-available P in 59 systems that are low in solution Pi. 60 Organic P mineralization is difficult to measure due to the high reactivity of dissolved phosphate 61 anions, which are often rapidly sorbed to soil surfaces following their release from organic 62 matter (Frossard et al., 1996; López-Hernández et al., 1998). In the past, the measurement of 63 total soil Po over time was used to infer whether net mineralization occurred (Condron et al., 64 2005), but extended time periods are needed for differences to become detectable. This method 65 alsodoes not allow for the detection of the simultaneous processes that are occurring (i.e. Po may 66 be mineralized and subsequently sorbedto soil constituents and/or immobilized again by soil 67 microorganisms). Isotopic dilution techniques using either 32P or 33P have been developed to 68 measure gross Pomineralization under steady state conditions (López-Hernández et al., 1998, 69 Oehl et al., 2001a). This is done by pairing a short-term (typically 80-100 min) Isotopic 70 Exchange Kinetics (IEK) experiment with a longer term (several days or more) incubation 71 experiment.The IEK experiment uses carrier-free radio-labeled P (e.g. 33Pi) and is based on the 3 72 assumption that unlabeled soil31Piand labeled33Pi have the same fate in the system (Frossard et 73 al., 2011). From the IEK experiment,a baseline is established which estimates the amount of 74 isotopically exchangeable P (termed the E-value) due to physico-chemical processes. Using 75 equations derived by Fardeau et al. (1985), E-values are extrapolated to calculate how much Pi 76 would become available over a longer period of time. In alonger term incubation experiment, 77 however, some of the Pi entering the soil solution resultsfrombiological mineralizationprocesses 78 which release unlabeled 31Pi and thus lower the specific activity (SA), i.e. the ratio of 33Pi to 31Pi 79 of the soil solution, providing a higher E-value than the one extrapolated from IEK data. Gross 80 Po mineralization is thus derived from the difference between measured and extrapolated E- 81 values. 82 In addition to gross Po mineralization, 33P uptake by the microbial biomass can be measured 83 using isotopic techniques(Oehl et al., 2001b). Assuming that soil solution Piis the sole source of 84 microbial P uptake, microbial Pi immobilization is calculated from the SA of the microbial 85 biomass, the SA of the soil solution and the amount of microbial P (Bünemann et al., 2007). Net 86 Pomineralization can then be determined by subtracting microbial Pi immobilization from gross 87 mineralization. Oehl et al. (2001b) found that P cycled faster through the microbial biomass in 88 soils under biodynamic and organic managementwhen compared to soil under 89 conventionalmanagement. Achat et al. (2009a) and Bünemann et al. (2012) found rapid 90 microbial P uptake under P-limited conditions, suggesting that microbial P uptake and 91 subsequent turnover may occur more quickly in soils low in available Pi. 92 Bünemann et al. (2012) calculated gross Po mineralization rates in grassland soils from a long 93 term fertilization experiment in Switzerland and found that it was mainly governed by available 94 Pivia a stimulation of microbial Pi immobilization at low Pi availability. This finding is probably 4 95 related to the fact that external Pi concentrations affect the expression of P uptake and transport 96 genes in soil microorganisms (Saleh-Lakha et al., 2005). In a recent review,Bünemann (2015) 97 outlined the need for Po mineralization studies using high P-sorbing calcareous soils with low 98 soil solution Pi concentrations. Here, we present microbially-mediated P transformations in 99 calcareous Alfisols obtained from forage fields in Ontario, Canada,as assessed using isotopic 100 techniquesacross a gradient of available Pi. Our main objective was to explore the effect of 101 available Pi,as indicated by both Olsen soil test phosphorus (STP) (Olsen et al., 1954) and water- 102 extractable Pi(Pw) concentrations on Po mineralization and microbial P uptake. 103 2.0 Materials and Methods 104 2.1 Soil sampling 105 Soils from four different forage fieldsfrom southwestern Ontario dairy farms were used in this 106 experiment. The soil sampleshad been collected as part of a larger study (Schneider, 2014), 107 which related Olsen STP and other soil properties to total forage yield and P uptake on paired 108 organic and conventional dairy farm soils.The soils were all Grey-Brown Luvisols according to 109 the Canadian System of Soil Classification (Hoffman and Richards, 1952; 1954; AAFC, 1998) 110 and Alfisols according to the American soil taxonomy (Soil Survey Staff, 2014), and Luvisols 111 according to the Food and Agriculture Organization World Reference Base (FAO, 2015). 112 In this study, soils from two organic (soils 1 and 2) and two conventional (soils 3 and 4) dairy 113 farms were selected in order to provide a gradient of Pi availability. The soils (0-15 cm) were 114 sampled from second or third year forage fields dominantly comprised of 115 Medicago sativaandPhleum pratense. The forage fields ranged in size from 4-20 ha and were 116 part of a regular crop rotation that also included annual crops. Additional information on farm 117 management practices for each of the fields can be found in Supplementary material A. The 5 118 Olsen STPconcentrationsof soils used in this study ranged from 2.2 to 13.7 mg P kg-1, total 119 organic carbon ranged from 21 to 27 g kg-1 and pH ranged between 7.2 and 7.5 (Table 1). After 120 initial collection and sieving (2 mm), soilswere frozen (-20°C) until needed for the incubation 121 study. From an initial 12 soil samples collected separately per field (3 from each of 4 sub-plots), 122 4 composite samples were made representing each field at the sub-plot level. 123 124 2.2 Experimental Design 125 The experiment consisted of two parts: a short-term (100 min) isotopic exchange kinetics (IEK) 126 experiment and a longer 7-d incubation experiment. The IEK experiment allowed the 127 establishment of isotopic dilution over time as a result of physico-chemical processes. The longer 128 incubation experiment with 33P-labeled soil provided the measurement of isotopic dilution due to 129 both physico-chemical and biological processesat two sampling times. 130 Before both experiments, the frozen soils were thawed at 4°C and then pre-incubated in the dark 131 at 22°C for 28 d with gravimetric water content maintained at a level suitable for microbial 132 activity (0.20-0.24 g H2O g-1 soil) (equivalent to ~65% field capacity, as estimated using the Soil 133 Water Characteristics Hydraulic Properties Calculator version 6.02.74 [Saxton and Willey, 134 2006]). During this period, respiration was measured twice to confirm that the microbial activity 135 was constantbythe onset of the incubation experiment. This was necessary because basal gross 136 organic P mineralization can only be determined under steady-state conditions (Oehl et al., 137 2001). Due to the time needed for the IEK experiments and the number of samples to process 138 (n = 4 composite samples per field x 4 fields = 16) the incubation experiment was conducted in 139 two sets. Each set contained half (eight) of the soils with two samples from each field. 140 2.3 Isotopic exchange kinetics (IEK) experiment 6 141 The IEK experimentswere conductedin triplicate usingunlabeled soil and were conducted in 142 batches on day 5 and day 6 of the incubation experiment for each of the two sets of soils. The 143 details of IEK methodology have been described elsewhere(Fardeau et al., 1985). Briefly, a 144 soil:solution ratio of 1:10 was created by weighing moist soil equivalentto 10 g dry soil into 250 145 mL centrifuge bottles (Nalgene) and adding nanopure water to bring the total volume (including 146 soil water) to 99 ml. The soil solutions were brought to a steady state by shaking end-over-end 147 for 16 h, after which the solutions were placed on magnetic stir plates and mixed for 10 min 148 before the addition of 1mL carrier-free (no 31P) 33P solution (Hartmann Analytics GmbH, 149 Braunschweig, Germany) to start the experiment (t = 0). The targetamount of initially added 150 radioactivity (R) was 167kBq (actual range was 164 -178 kBq). While maintaining constant 151 stirring, ~ 5 mL of each solution wassampled at 1, 4, 10, 40, 60 and 80 minutes after the addition 152 of the tracer and immediately syringe-filtered through a 0.2 µm cellulose acetate membrane. The 153 concentration of Pi in the filtered solution (Cp) was determined colorimetricallyat 610 nm using 154 the malachite green method and a 1 cm cell path(Ohno and Zibilske, 1991). The radioactivity (r) 155 in the filtrate was measured by liquid scintillation counting (2500TR, Packard Instrument 156 Company Inc., Meriden, CT). 157 From the IEK experiment, the following parameters were determined: 1) Pw concentration in the 158 filtered soil extract (mg P kgsoil-1), where Pw = 10 (L kg-1) * Cp (mg L-1); 2) coefficients m and 159 n, which are determined by nonlinear regression and account for the immediate and slow 160 physico-chemical reactions, respectively, that describe the rate of decrease of radioactivity in soil 161 solution with time (Fardeau et al., 1985; Achat et al., 2009b) , and 3) the amount of isotopically 162 exchangeable P during the first minute of the batch experiment(E1min) in mg P kg soil-1, 163 determined as Pw / m. 7 164 Contrary to the common assumption that P dynamics during the short term batch experiment are 165 controlled by physico-chemical processes only, Bünemann et al. (2012) found extremely rapid P 166 uptake by the microbial biomass ina soil low in Pw, necessitating the useof HgCl2, a known 167 microbial inhibitor, during the IEK experiment. Therefore, using select samples of each of the 168 four soils, a pre-test with HgCl2 additions was made to determine whether potential rapid 169 microbial P uptake had any significant effect on IEK parameters over the 80 min of the test (as 170 described by Bünemann et al. 2012). There were no significant differences in water-extractable 171 31 172 deemedunnecessary for these soils. 173 The temporal development of radioactive P remaining in a soil-solution system at steady state is 174 described by Fardeau et al. (1985) as: 175 r(t) / R = m* (t + (m)1/n)-n + rinf/R 176 where r(t) is the radioactivity remaining in soil solution of the initial amount added (R) after the 177 time (t) of isotopic exchange (kBq) and rinf/R is the maximum possible dilution of the isotope in 178 solution at t = infinite which is approximated by the ratio of Pw to total Pi (in our study 179 calculated as total soil P minus total NaOH-EDTA extractable Po). The parameters m and n are 180 soil specific constants obtained from nonlinear regression as described above. The amount of 181 isotopically exchanged P at a given time t, i.e. the E-value (E(t)), is then calculated from r(t)/R as 182 the inverse of the specific activity in solution: 183 E(t) (mg P kg-1) = Pw / (r(t)/R) 184 under the assumption that the specific activity of isotopically exchanged P in solution is identical 185 to the specific activity of isotopically exchanged P in the whole system. P or 33Pdue to HgCl2 addition. Therefore, the use of a microbial inhibitor during the IEK was (1) (2) 186 8 187 2.4 Incubation experiment 188 In parallel with the short-term IEK experiment, a 7-d incubation experiment was conducted. A 189 seven day incubation was selected as recommended by Oehl et al. (2001b), to limit effects of re- 190 mineralizationof 33P-labeled microbial P. The soils were extracted at two sampling times and 191 analysed for Pi in solution and in the microbial biomass.The 33Pin both pools was also 192 determined to allow the calculation of gross P fluxes as well as to gain an indication of the rateof 193 microbial P uptake and release. All analytical measurements were conducted in duplicate. 194 To label each sample, the equivalent of 190 g oven-dry soil was weighed into a bowl, five mL of 195 a33P (as phosphoric acid) labeling solution was added, and the soil was mixed manually (with a 196 spoon) for 5 min. Labeling resulted in a concentration of 2.49 (set 1) and 2.72 (set 2) kBq g-1 soil 197 and increased the soil moisture content slightly to values ranging from 23 to 28% (~75% field 198 capacity).Two portions equivalent to 10 g dry soil each were removed for the concurrent 199 measurement of soil respiration in order to assess if respiration remained steady over the 7-d 200 incubation. The remainder was weighed into plastic containers and incubated at 22°C.The 201 moisture content was kept constant throughout the 7-d incubation. 202 The labeled soil was sampled (weighed and extracted) on day 0 (~7 h after labeling) and on day 203 7. Samples were analysed for31P and 33P inPwand in microbial P. All radioactivity measurements 204 were correctedfor decay to the time of soil labeling. The extraction period of16 h was added to 205 the sampling time point, since isotopic exchange continues during extraction (Bünemann et al., 206 2007). Therefore, all results from the incubation experiment are shown for 1.2 (day 1) and 8.1 d 207 (day 8) after labeling. 208 In a pre-test, different methods were compared for measuring soil microbial P. The method of 209 Morel et al. (1996) using a pulse of chloroform fumigation followed by water extraction was 9 210 compared with the method of Kuono et al. (1995) using liquid fumigation with hexanol and 211 simultaneous extraction with anion exchange membranes, and with a modified version using 212 liquid hexanol fumigation with water extraction. All methods yielded comparable estimates of 213 fumigant labile Piand the trends among samples were consistent(data not shown). 214 In this study, we chose to use the water extraction withhexanol fumigation because it allowed us 215 to use the Pwconcentrations determined for the non-fumigated samples (the control samples) as 216 the Pw concentrations required to calculate E-values. For this method, analytical duplicates, each 217 containing moist soil equivalent to 2 g dry soil and 20 mL of nanopure water were placed in 50 218 mL centrifuge tubes. Fumigated subsamples received an additional 1 mL of hexanol. Two non- 219 fumigated, spiked samples were included: the first received 25 mg P kg-1 (31P spike) and the 220 second 25 mg P kg-1plus ~ 9.4 MBq33P kg-1 soil (33P spike),in order to calculate the recovery for 221 both 31P and 33P. Results of pre-tests showed the recovery to be linear, thus one P spike addition 222 (for each 31P and 33P) was determined to be adequate. Samples were shaken horizontally for 16 h, 223 after which they were filtered through 0.2 µm cellulose acetate filters and solution Piwas 224 measured as described earlier (Section 2.3). Hexanol-labile P (Phex), a direct measure of 225 microbial biomass P was calculated as: 226 31 227 where31Pfum and 31Pnonfumare the amounts of 31P extracted from fumigated and non-fumigated 228 subsamples, respectively, and 31Prec is the fraction of the 31P spikethat is recovered in the soil 229 solution after shaking. No conversion factor was applied to account for incomplete extraction of 230 soil microbial P, as Kp factors have been shown to be soil specific and can vary widely (Oberson 231 and Joner, 2005; Yevdokimov et al., 2016) and have not been determined for these soils and this 232 method. Phex (mg P kg-1) = (31Pfum-31Pnonfum)/31Prec (3) 10 233 The radioactivity was measured in the filtrate of the non-fumigated and fumigated samples as 234 describedearlier. The release of 31P from microbial cells as a consequence of fumigation affects 235 the recovery of the label in the soil solution due to isotopic exchange reactions (McLaughlin et 236 al., 1988). In order to determine the amount of 33P in the microbial biomass, it is necessary to 237 correct for the release of 31P from the microbial P pool (which will release 33P from the soil solid 238 phase). This is simulated by the addition of a 31P spike which has been shown to increase the 239 recovery of the label (Oehl et al., 2001a). An equation was established for each soil sample using 240 the non-fumigated subsamples (no P addition) and the subsamples receiving a 25 mg P kg-1(31P 241 spike): 242 33 243 where33Pnonfum+ P represents the 33P (in kBq g-1) extracted from non-fumigated subsamples 244 receiving the 31P spike and where a and b are the slope and intercept, respectively. The amount 245 of Phex (in mg P kg-1) was then inserted as 31P spike into equation (4) and the resulting value of 246 extracted 33P was used to calculate the incorporation of 33P in Phex (33Phex) as: 247 33 248 where33Pfum is the 33P extracted from hexanol-fumigated subsamples, and 33Prec the recovery of 249 the 33P spike described above. 250 For the incubation experiment, E-values were calculated as the inverse of the specific activities 251 measured in the water extract.To calculate cumulative gross Po mineralization, E-values were 252 extrapolated from the short-term batch experiment using equations (1) and (2) for each sampling 253 time in the incubation experiment and subtracted from the respective measured E-values. As 254 these soils are calcareous, it is possible that some of this difference could be a result of microbial 255 solubilization of recalcitrant calcium phosphate minerals (Arcand and Schneider, 2005). While it Pnonfum+ P (kBq g-1) = a (31P spike (mg P kg-1)) + b Phex (kBq g-1) = (33Pfum–33Pnonfum+P) / 33Prec (4) (5) 11 256 cannot be ruled out completely, we are assuming this contribution to be low to negligible, and 257 that the difference between measured and extrapolated E-values is solely a result of gross Po 258 mineralization. 259 The quantity of P in a soil pool that has been derived from a labeled P source can be calculated 260 from the specific activities of the pool and the source, and the amount of P in the pool (Fardeau 261 et al., 1996). Assuming that Pw is the sole source of microbial P uptake, microbial Pi 262 immobilization at each sampling point was calculated according to Bünemann et al. (2007) as: 263 Pi immobilization (mg P kg-1) = (SAPhex/SAPw) * (Phex) 264 whereSAPhex(kBq ug-1 P) is the specific activity of hexanol-labile (microbial) P, SAPw(kBq ug-1 265 P) is the specific activity in Pw, and Phex is the amount of hexanol-labile (microbial) P (in mg P 266 kg-1). Net Po mineralization was then derived as gross Po mineralization minus Pi immobilization. 267 During the incubation experiment respiration was determined by trapping CO2 released from 10 268 g of soil in 0.2 M NaOH (Alef, 1995) and precipitating it as SrCO3, followed by back titration of 269 the unreacted NaOHwith 0.2 M HCl, using phenolphthalein as an indicator solution. In all cases, 270 the CO2 was collected over a period of 7 days. 271 2.5 Statistics 272 Statistical analyses were performed using the statistical software program SAS ver. 9.2 for 273 Windows (SAS Institute, Cary, NC, USA). In this study, available Pi in the soils was indicated 274 by both Olsen STP and Pw, which were positively correlated for these soils (r = 0.96,p 275 < 0.0001).Regression statistical methods were used to evaluate relationships between P flux 276 parameters andPw concentrations. Relationships were also analyzed for Olsen STP, but the data 277 are not shown since results were similar to those obtained with Pw. After visual inspection of 278 scatterplots, the SAS procedures Proc Glm or Proc Nonlinwere used for linear and nonlinear 12 (6) 279 regression, respectively (Bowley, 2008). Using the SAS Proc Univariate procedure (Fernandez 280 1992; Bowley2008), homogeneity of error (residuals) was tested for each model. The residuals 281 were also tested for normality as determined by the Shapiro–Wilk statistic. In the case of non- 282 linear regression analysis, pseudo R2 values were calculated according to the method of Efron 283 (1978). Analysis of variance (ANOVA) was used to determine significant differences among the 284 data using the Proc GlmSAS procedure (Bowley, 2008). A multiple mean comparison test was 285 performed using Tukey’s HSD (honest significance difference) test. A type I error rate of p 286 =0.05 was used for all statistical evaluations. 287 3.0 Results 288 3.1 Isotopic exchange kinetics batch experiment 289 Water-extractable Pi (Pw) concentrations ranged from a mean of 0.03 (soil 1) to 0.24 (soil 4) mg 290 P kg-1, and E1min ranged from 0.40 (soil 1) to 1.62 (soil 4) mg P kg-1 (Table 2). The Pwand 291 E1minconcentrations were significantly correlated with previously measured Olsen STP 292 concentrations (r = 0.96 and r = 0.93, respectively). The parameter mranged from 0.09 (soil 1) to 293 0.14 (soil 4), whilecalculated n values ranged from 0.32 (soil 4) to 0.38 (soil 1)(Table 2). The full 294 list of IEK parameters from individual soil samples can be foundin Supplementary Material B. 295 3.2 Incubation experiment 296 Soil respiration during the last week of the pre-incubation was similar to that during the week of 297 the incubation experiment (data not shown), indicating that the respiration rate had stabilized 298 during pre-incubation and was not increased again by soil mixing during labeling. Mean CO2 299 release rates for the four soils were not significantly different from each other, with mean 300 respiration rates during the incubation experiment ranging from 13.1 to 17.1 mg C kg-1 soil d- 301 1 (Table 2).The Pwand Phexconcentrations (Supplementary Material C)showed minimal variation 13 302 from the first sampling point (1 day after soil labeling) to the end of the incubation experiment (8 303 days after soil labeling), again supporting that the system was at a steady-state and in the absence 304 of flush effects. 305 3.2.1 Amount of 33P in different P pools 306 The proportion of the added33P that was detected in soil solution one day after soil labeling 307 ranged from 0.4 to 1.7%(Figure 1). After eight days, this proportion had decreased slightly to 0.2 308 to 1.3%. At both sampling times, the recovery of33P in Pwwas positively related to 309 Pwconcentrations (p< 0.001 for both day 1 and day 8). In the Phex pool, recovery of the initially 310 added radioisotope one day after soil labeling ranged from 1.4 to 32.4 % and increased to a range 311 of 6.9 to 43.7 % after eight days. Soil 1, with the lowest Pw concentrations, had the highest 312 recovery of33P in the Phex pool (Figure 1). One day after soil labeling, this was the only soil with 313 notably high (>20 %) amounts of the label present in the Phex pool. Eight days after soil labeling, 314 the negative relationship between the proportion of 33P in the Phex pool and the Pw concentrations 315 was significant (p <0.0001 [based on the log-log relationship where 33P in Phex = -3.45(Pw) + 316 1.59, R2= 0.82]),and 11 of 16 samples had>20% of the label in the Phexpool (Figure 1). 317 The SA of the Phex pool was notably higher in three samples from soil 1 that had very low 318 solution Pi concentrations (Figure 2, Table 3). Aftereight days, the SA of the Phex pool had 319 increased across the soils (Table 3) and the power-type relationship between the SA of the Phex 320 pool and Pwwas significant with a rapid increase below Pw concentrations of ~ 0.1 mg kg-1 321 (Figure 2). Finally, the ratio of the SAPhex to the SAPwat day 8 was significantly greater for soil 322 1 than for the other three soils (Table 3). 323 3.2.2 P fluxes 14 324 For all 16 samples, measured E-values were greater than extrapolated E-values atboth 1 and 8 d 325 after labeling(Supplementary Material D; Table 4). Across all soil samples, gross Po 326 mineralization accounted for 4.2 to 60.9 % and 7.0 to 56.2 % of the total isotopically 327 exchangeable P (Emeas) at 1 and 8 d after soil labeling, respectively. Daily gross Po mineralization 328 and Pi immobilization rates ranged from 0.2 to 2.8 and 0.5 to 1.6 mg P kg-1 d-1 across all soil 329 samples, while daily net Po mineralization rates were calculated to range from -0.8 to 2.4 mg P 330 kg-1 d-1 (Supplemental Material, Table D1). 331 There were no significant differences found among soils for cumulative or daily gross Po 332 mineralization, Pi immobilization or net Po mineralization (Table 4). However, when all 16 333 samples were included, regression analysis revealed a significant positive linear relationship 334 between gross Po mineralization and Pwconcentrations for both day 1 (p= 0.006), and day 8 (p= 335 0.0024) (Figure 3). Gross Pi immobilization was not significantly related toPwconcentrations for 336 either sampling day (p= 0.779 for day 1, and 0.057 for day 8) (Figure 3), while there was a 337 significant positive linear relationship between net Po mineralization and Pw concentrations for 338 both day 1 and day 8 (p =0.0003and 0.0009, respectively) (Figure 3). Because the daily fluxes 339 were calculated from the cumulative rates after 8 days, the relationships with Pwconcentrations 340 were similar to those at the 8 day sampling time and are therefore not shown. 341 342 4.0 Discussion 343 4.1 Isotopic Exchange Kinetics Parameters 344 Overall, the soils in this study were found to have relatively low Pwand E1min 345 concentrations.Morel et al. (1992) observed that wheat yield was responding to P fertilizer in 346 soils presenting E1min values less than 5 mg P kg-1 soil. This observation was confirmed for other 15 347 crops by Galletet al. (2003). The E1minconcentrationsin this study ranged from 0.3 to 2.2 mg P kg- 348 1 349 E1minfor a field which was not P-fertilized and 1.6 mg P kg-1 was the E1min derived from a P- 350 fertilized soil (Bünemann et al., 2012); however, the E1min concentrations found here are low 351 compared to other studies of Po mineralization using field crop soils, even in the absence of 352 mineral P fertilization (e.g. Oehl et al., 2001b).Most of the soils in this study are categorized as 353 low (<10 mg P kg-1) in Olsen STP for the growing of perennial forages (OMAFRA, 2009); 354 however, the fields they were collected from were providing satisfactory yields to the farmers 355 that were above conventional county averages (Schneider, 2014). 356 The relationships between IEK parameters and soil Pwconcentrations were in agreement with the 357 literature. For example, the finding of a positive correlationof m (r1min/R)with Pw is consistent 358 with the results of other authors (e.g. Achat et al., 2009b). Accordingly, m was lowest in soil 1 359 and greatest in soil 4, while n was greatest in soil 4 and lowest in soil 1, which indicates the 360 greatest fast and slow sorption in soil 1 and the least in soil 4 (Sen Tran et al., 1998). soil and are similar to the range from an 18-year grassland study where 0.8 mg P kg-1 was the 361 362 4.2 Incubation experiment 363 4.2.1 33Precovery in the hexanol-labile (microbial) P pool 364 The percentage of the added 33P remaining in soil solution was very low (range was 0.4 to 1.7%) 365 for all soils after just 1 day of soil labeling, decreasing further by 8 days. These results support 366 that the soils in this study probably have a strong P-sorption capacity, especially those with low 367 solution Pi concentrations (Figure 1). 368 Soil 1 with the lowest Pw concentrations had the highest recovery of33P in the Phex pool (Figure 369 1), with notably high (>20%) amounts of the label present in the Phexpool already after one day. 16 370 After eight days, the negative relationship between 33P recovery in the Phex pool and 371 Pwconcentrations was significant and 11 of 16 subsamples had>20% of the label in the Phexpool. 372 The recovery of 33P in the Phex pool is in the upper end of recovery percentages reported by 373 similar studies, where recovery of 33P in the Phex or microbial P pool in soils labeled with carrier- 374 free 33Pihave ranged from ~2 to 46% at comparable sampling times (7-8 d)(Achat et al., 2009a; 375 Bünemann et al., 2004, 2012; Oehl et al., 2001b, 2004; Oberson et al., 2001). This potentially 376 indicates a rapid and extensive uptake of solution Pi by the microbial biomass in these soils 377 under steady state conditions. 378 The proportion of label recovered in the Phex pool may depend on the size of the microbial P 379 pool. Some studies have found greater recoveries of 33P in treatments that had the highest 380 microbial biomass (e.g.Bünemann et al., 2004), but others have not (Bünemann et al., 2012).In 381 this study, greater P uptake was not significantly related to the size of the microbial P pool(p = 382 0.10 for linear regression) (Figure E1, in Supplementary Material E). In fact, some of the soils 383 with the lowest microbial biomass had the greatest recoveries of 33P in Phex, indicating an 384 increased rate of P uptake by the soil microbial biomass. 385 The soils in the present study were collected from alfalfa-grass forage fields and some of the 386 legume-derived organic matter may have helped to support rapid microbial P uptake. For 387 example, Oberson et al. (2001) found more rapid uptake of added 33P by the Pchl pool in soil from 388 a grass-legume pasture, when compared to soil from native savannah and continuous rice 389 systems. Oberson et al. (2011) stated that even in low available P soils, microbes are limited by 390 C and N rather than P and that legume residues, by supplying N-rich organic matter, may help 391 increase microbial P cycling. While this may help to explain some of the overall high levels of 17 392 microbial P uptake, it does not explain differences in 33P uptake by Phex pools across the soils, 393 which appear to be inversely dependent on Pw concentrations. 394 The relationship between the amount of 33P found in Phex and Pwconcentration follows a power 395 type relationship and it appears that rather than a gradual development, there is a threshold 396 Pwconcentration (somewhere < 0.1 mg P kg-1) where a dramatic increase in the amount of label 397 found in Phex occurs (Figure 1). These results indicate that when soil solution Pi concentrations 398 are very low, a rapid uptake of P into the microbial biomass occurs. Bünemann et al. (2012) also 399 reported rapid P uptake by the microbial biomass for a P-limited soil (fertilized only with N and 400 K). In their study, a microbial inhibitor was used during the 100 min IEK batch experiment to 401 ensure that the Pi removal effects were solely from physico-chemical processes. In this study, the 402 necessity of using a microbial P inhibitor was investigated during an IEK pre-test on 403 compositesoil samples and was determined to be unnecessary. However, the rapid tracer uptake 404 observed by three soil samples (1a, 1b and 1c) in the incubation experiment after 1 day of 405 labelingsuggests that the use of a microbial inhibitor would have been desirable for these 406 individual soil samples. Without the use of a microbial inhibitor, rapid 33P uptake by the 407 microbial P pool during the IEK experiment likely would have resulted in an overestimation of 408 the extrapolated E-values, which would in turn result in an underestimation of gross Po 409 mineralization for these three samples. The possible effects of this on our results is discussed in 410 the P fluxes section, however, it is evident that further research on the necessity and the practical 411 use of microbial inhibitors (many of which are highly toxic) is needed. 412 The observed dramatic increase in microbial P uptake could be explained by a biological 413 response of soil microbes to low Pw concentrations. It has been established that soil 414 microorganisms have developed systems to assist them in P uptake, including the induction of 18 415 phosphate-specific transport systems, production of phosphatase enzymes, and promotion of the 416 dissolution of insoluble P minerals (Saleh-Lakha et al., 2005). In this study, it is plausible 417 thathigh-affinity phosphate transport systems were activated by low available Pi conditions, 418 accelerating the uptake of Pi from solution and hence the uptake of the 33P label. It would be 419 informative for future studies to combine studies measuring gross Po mineralization with 420 molecular biology approaches to quantify gene expression and proteins related to microbial P 421 uptake. 422 As discussed earlier, the relationship observed between 33P incorporation into Phex and Pw 423 concentrations(Figure 1) supportsthat rapid P uptake is promoted under low available Pi 424 conditions. If the microbial P (Phex) pool remains constant (as was the case in this experiment), 425 then re-mineralization of Phex should be equal to the immobilizationof Pi by the microbial 426 biomass (Achat et al., 2009a).Thus, our results indicate that faster microbial P turnover occurs 427 under conditions of low soil solution Piconcentrations. This is further supported by SAPhexbeing 428 negatively and linearly related to Pw concentrations (R2=0.42, p = 0.0068; Figure 2) at the 8 day 429 sampling time. Further,Oehl et al. (2001a) discusses howthe convergence of the SAPmic 430 (represented here by SAPhex) and SAPwcorresponds with the turnover time of microbial P. Here 431 we found that the ratio of the SAPhex: SAPw for soil 1 was 0.77 after 8 days of incubation which 432 was significantly greater than for the other three soils (Table 3). 433 However, it is interesting that despite the indication of an increased rate of microbial P turnover 434 in soil 1 with the lowest Pw concentrations, we do not see differences in respiration rates among 435 the soils. Thus, it is possible that high-affinity phosphate transport systems were upregulated 436 (Saleh-Lakha et al., 2005), but did not require extra carbon costs. Further, P mineralization may 437 not be as closely linked to C mineralization as N mineralization as a result of P being 19 438 mineralized by both biological and biochemical (enzyme driven) mineralization (McGill and 439 Cole, 1981; Oehl et al., 2001a). In support of this concept, Fraser et al. (2015) have recently 440 shown that alkaline phosphatase activity and phoD gene abundance arenegatively related to soil 441 available inorganic P concentrations. 442 The significance of a rapid uptake of P by the microbial biomass under low available 443 Piconcentrations is not fully clear. Oehl et al. (2001b) suggest that a rapid P uptake could be 444 anearly sign of P deficiency for soil microbes. Faster microbial P turnover may also help to keep 445 P in a more available form for plants, especially in high P-fixing soils (Oberson et al., 2011). 446 Further research is needed to explore the significance of microbial P turnover rates in relation to 447 plant-available Pi. 448 449 4.3 P fluxes 450 Gross Po mineralization rates were found to range from 0.2 to 2.8 mg P kg-1 d-1, which is within 451 the range of rates (0.1 to 12.6 mg P kg-1d-1) determined in other studies using similar isotopic 452 dilution techniques (Bünemann, 2015). Pi immobilization rates and net Po mineralization rates 453 were alsoclose to the range of rates previously reported (Bünemann, 2015). Six out of the sixteen 454 soil sampleswere calculated to haveslightly negative net Po mineralization rates, indicating net 455 immobilization (Supplemental Material, Table D1). We acknowledge that calculatedPi 456 immobilization and net Pomineralization rates may be less accurate than gross Po mineralization 457 ratesdue to the fact that we did not apply a Kp factor to convertPhex into microbial biomass P 458 (Brookes et al., 1982). This could have resulted in the underestimation of microbial Pi 459 immobilization and relatedly an overestimation of net Po mineralization, making immobilization 460 even more important (Bünemann et al., 2012). 20 461 Gross Po mineralization was found to account for an average of 39.9% and 32.4% of the total 462 isotopically exchangeable P (Emeas), which was on average 19.2 and 30.2 mg P kg-1 soil after 1 463 and 8 days of soil labeling, respectively.Bünemann (2015) compared and summarized the 464 absolute and relative contributions of physico-chemical and biological processes to total 465 isotopically exchangeable P determined after 5-10 days of incubation from 6 different isotopic 466 dilution studies. The proportions of isotopically exchangeable Pi derived from gross Po 467 mineralization in our study, are at the upper end of values reported for arable land but at the low 468 end of those reported for permanent grassland sites (Bünemann et al., 2015). In the sole 469 permanent grassland study, after 8 days of incubation, Bünemann et al. (2012) found gross Po 470 mineralization to account for 68% (65 mg kg-1) and 51% (25 mg kg-1) in non-P-fertilized and P- 471 fertilized soils, respectively. Based on the high contributions of Po mineralization to available Pi, 472 the authors emphasized the importance of microbial processes in P dynamics in grassland soils. 473 The soils in the present study were collected from second and third year perennial forage fields, 474 and the significant biological contributions to total isotopically exchangeable P indicate that soil 475 microbial P cycling is important in such forage systems. Our results also indicate that biological 476 contributionsaccounted for a greater proportion of total isotopically exchangeable P when Pw 477 concentrations were the lowest (soil 1). Soil 1 averaged an Emeas of 23.7 mg kg-1 after eight days 478 withgross Po mineralization accounting for 10.2 mg kg-1 (43%), which was a greater proportion 479 than in the three other soils (range was 34-35%). More studies are needed to better delineate the 480 relative contributions of gross Po mineralization under different cropping systems and different 481 soil test P concentrations. 482 483 4.3.1 Relationships of P fluxes with available Pi 21 484 Overall, regression analysis revealed that gross Po mineralization increased with Pw 485 concentrations (Figure 3).This was unexpected, since the literature suggests that gross Po 486 mineralization may be inversely related to Pw concentrations because enzyme activity and/or 487 microbial P cycling may increase as available Pi decreases (Quiquampoix and Mousain, 2005; 488 Oehl et al., 2001b).Reasons for the increasing trend of gross Po mineralization with Pw 489 concentrations are not fully clear; and we did not find a positive relationship between Pw and 490 Phex or Pw and Porg, which might support greater Po mineralization at higher Pw 491 concentrations. Although regression analysis found a positive relationship, as seen in Table 4, 492 there were no significant differences in gross Po mineralization across the four soils after 8 days 493 of incubation.A larger sample size comprising a greater range of Pw concentrations might help to 494 further understand the observed relationship. As discussed above (section 4.2.1), it is possible 495 that gross Po mineralization was underestimated in three of the soils with the lowest available Pi, 496 which would weaken the positive relationship between gross Po mineralization and Pw. Further, 497 it appears that there could be a change-point at a Pw concentration ~0.10 mg kg-1 where below 498 this concentration, gross Po mineralization could be inversely related to Pw concentrations 499 (Figure 3). However, more data points would be required to substantiate this trend. 500 Although not significant, a weak negative relationship between Pi immobilization and Pwwas 501 observed. This is in agreement with the rapid uptake of the label at low Pw concentrations 502 (Figure 1). Bünemann et al. (2012) found gross P fluxes in grassland soils to be dominated by 503 microbial Pi immobilization, especially in the absence of P fertilization. 504 One day after soil labeling, net Po mineralization was shown to increase with Pw concentrations. 505 This was expected as gross Po mineralization increased and Pi immobilization decreased slightly 506 with increasing Pw concentrations. If gross Po mineralization was underestimated for samples in 22 507 the low end of the range of available Pi concentrations, thennet Po mineralization (gross 508 Pomineralization– Piimmobilization) would also be underestimated, potentially removing the 509 positive relationship between net Po mineralization and Pwconcentrations. 510 4.5The use of the isotopic dilution method in calcareous soils with low water-extractable Pi 511 concentrations 512 In a recent review, Bünemann et al. (2015) discuss how very low Pi availability can impede the 513 determination of IEK parameters needed to establish the baseline for isotopic dilution due to 514 physico-chemical processes over time, and comment that more isotopic dilution studies using 515 high-P sorbing acid or calcareous soilsare needed. LowPwconcentrations run the risk of 516 introducing more error as a small change in concentrations can lead to greater variation of 517 extrapolated E-values (Bünemann et al., 2007). In this study, some soil Pw concentrations were 518 very low, with about half of the samples being below the detection limit of 0.01 mg P L-1 or 0.1 519 mg P kg -1 for the malachite green method as determined recently (Randriamanantsoa et al., 520 2013).However, because many measurements were done per sample and at different points of the 521 IEK and incubation experiment (for example see supplemental material Tables B1 and C1), we 522 are convinced that our concentrations were consistently and reliably measured even at these low 523 concentrations. Future isotopic dilution studies with low soil Pw concentrations could consider 524 using a hexanol concentration method for determining phosphate in solution which has been 525 recently described by Randriamanantsoa et al. (2013) and applied in an isotopic dilution study 526 (Randriamanantsoa et al., 2015). 527 4.6 Conclusions 528 In this study, for the first time soil microbial P fluxes were explored across a gradient of water- 529 extractable Pi concentrations in calcareous soils. The rapid incorporation of 33P into the microbial 23 530 P pool (Phex) indicates that important P uptake occurs under steady-state conditions and that this 531 uptakeis accelerated under low available Pi conditions. A very rapid uptake of the radioisotope, 532 without changing microbial P concentrations, suggests an accelerated turnover of the microbial P 533 pool when available soil Pi is low. Rapid microbial P turnover may help to provide P to plants 534 under low Pi conditions, by promoting a rapid uptake and also release of P from the microbial P 535 pools. Gross Po mineralization was shown to release significant amounts of P, accounting for on 536 average 32% of total isotopically exchangeable P at the end of the 8 d incubation experiment. By 537 better understanding Po contributions to plant-available P, more accurate soil P fertility 538 recommendations can be made, thus promoting overall greater P-use efficiencies in agriculture. 539 540 24 541 Acknowledgements 542 The landowners that agreed to let us sample their fields and provide farm management history 543 data are kindly acknowledged for their participation. This project was funded by and supported 544 by an NSERC CGS Doctoral Scholarship, the Swiss National Science Foundation, Ontario 545 Ministry of Agriculture, Food and Rural Affairs, and the Canada Research Chairs program. An 546 NSERC Michael Smith Foreign Study Supplement allowed for the study abroad opportunity. 547 The authors declare that the funding sources had no role in the study design, collection, analysis 548 and interpretation of the data. Two anonymous reviewers are also thanked for their time and 549 contributions to an improved manuscript. 550 551 25 552 References 553 Achat, D.L., Bakker, M.R., Morel, C., 2009a. Process-based assessment of phosphorus availability in a 554 low phosphorus sorbing forest soil using isotopic dilution methods. Soil Science Society of America 555 Journal 73, 2131-2142. 556 Achat, D.L., Bakker, M.R., Augusto, L., Saur, E., Dousseron, L., Morel, C., 2009b. Evaluation of the 557 phosphorus status of P-deficient podzols in temperate pine stands: Combining isotopic dilution and 558 extraction methods. Biogeochemistry 92, 183-200. 559 Agriculture and Agri-Food Canada (AAFC), 1998.The Canadian System of Soil Classification.3rd edition. 560 NRC Research Press, Ottawa. 561 Alef, K., 1995. Soil respiration. In: Alef, K., Nannipieri, P. (Eds.), Methods in applied soil microbiology 562 and biochemistry. Academic Press, London, pp. 214-219. 563 Analytical Methods Committee, 1987. Recommendations for the definition, estimation and use of the 564 detection limit. Analyst 112, 119-204. 565 Arcand, M. and Schneider, K. 2006.Plant- and microbial-based mechanisms to improve the agronomic 566 effectiveness of phosphate rock: a review. Annals of the Brazilian Academy of Sciences 78, 791-807. 567 Bowley, S.R., 2008. A Hitchhiker's Guide to Statistics in Plant Biology.2nd Edition. Any Old Subject 568 Books, Guelph, ON, Canada. 569 Brookes, P.C., Powlsen, D.S., Jenkinson, D.S., 1982. Measurement of microbial biomass phosphorus in 570 soil. Soil Biology and Biochemistry 14, 319-329. 26 571 Bünemann, E.K., 2015. Assessment of gross and net mineralization rates of soil organic phosphorus – a 572 review. Soil Biology and Biochemistry 89, 82-98. 573 Bünemann, E.K., Marschner, P., McNeill, A.M., McLaughlin, M.J., 2007. Measuring rates of gross and 574 net mineralization of organic phosphorus in soils.Soil Biology and Biochemistry 39, 900-913. 575 Bünemann, E.K., Oberson, A., Liebisch, F., Keller, F., Annaheim, K.E., Huguenin-Elie, O., Frossard, E., 576 2012. Rapid microbial phosphorus immobilization dominates gross phosphorus fluxes in a grassland soil 577 with low inorganic phosphorus availability. Soil Biology and Biochemistry 51, 84-95. 578 Bünemann, E.K., Steinebrunner, F., Smithson, P.C., Frossard, E., Oberson, A., 2004. Phosphorus 579 dynamics in a highly weathered soil as revealed by isotopic labeling techniques. Soil Science Society of 580 America Journal 68, 1645-1655. 581 Cade-Menun, B.J., Preston, C.M. 1996. A comparison of soil extraction procedures for 31P NMR 582 spectroscopy.Soil Science 161, 770-785. 583 Condron, L.M., 2004. Phosphorus - Surplus and deficiency. In: Schjønning, P., Elmholt, S., Christensen, 584 B.T. (Eds.), Managing Soil Quality: Challenges in Modern Agriculture. CAB International, Cambridge, 585 MA, USA, pp. 69-84. 586 Condron, L.M., Tiessen, H., 2005. Interactions of organic phosphorus in terrestrial ecosystems. In: 587 Turner, B.L., Frossard, E., Baldwin, D.S. (Eds.), Organic Phosphorus in the Environment. CABI 588 Publishing, Cambridge, MA, USA, pp. 295-308. 589 Efron, B., 1978. Regression and ANOVA with zero-one data: Measures of residual variation. Journal of 590 the American Statistical Association 73, 113-121. 27 591 Entz, M.H., Guilford, R., Gulden, R., 2001. Crop yield and soil nutrient status on 14 organic farms in the 592 eastern portion of the northern Great Plains. Canadian Journal of Plant Science 81, 351-354. 593 Fardeau, J.C., Guiraud, G., Marol, C., 1996. The role of isotopic techniques on the evaluation of the 594 agronomic effectiveness of P fertilizers. Fertilizer Research 45, 101-109. 595 Fardeau, J.C., 1993. Available soil phosphate – its representation by a functional multiple compartment 596 model.Agronomie 13, 317-331. 597 Fardeau, J.C., Morel, C., Jappe, J., 1985. Exchange kinetics of phosphate ions in the soil-solution system - 598 experimental verification of theoretical equation. Comptes Rendus De L'Academie Des Sciences Serie III 599 - Sciences De La Vie 300 300, 371-376. 600 Fernandez, G.C.J., 1992. Residual analysis and data transformations: Important tools in statistical 601 analysis. Horticultural Science 27, 297-300. 602 Food and Agriculture Organization of the United Nations (FAO).2015. World reference base for soil 603 resources 2014.FAO, Rome. Available at: http://www.fao.org/3/a-i3794e.pdf (accessed 10 Nov 2016). 604 Fortune, S., Robinson, J.S., Watson, C.A., Phillips, L., Conway, J.S., Stockdale, E.A., 2005. Response of 605 organically managed grassland to available phosphorus and potassium in the soil and supplementary 606 fertilization: field trials using grass-clover leys cut for silage. Soil Use and Management 21, 370-376. 607 Fraser T., Lynch D. H., Entz M. H., Dunfield K. E. 2015. Linking alkaline phosphatase activity 608 with bacterial phoD gene abundance in soil from a long-term management trial.Geoderma 257- 609 258:115-122 28 610 Frossard, E., Achat, D.L., Bernasconi, S.M., Bünemann, E.K., Fardeau, J.C., Jansa, J., Morel, C., 611 Rabeharisoa, L., Randriamanantsoa, L., Sinaj, S., Tamburini, F.,Oberson, A., 2011. The use of tracers to 612 investigate phosphate cycling in soil-plant systems. In: Bünemann, E.K., Oberson, A., Frossard, E. (Eds.), 613 Phosphorus in Action. Biological Processes in Soil Phosphorus Cycling. Springer, Heidelberg, pp. 59-91. 614 Frossard, E., Fardeau, J.C., Brossard, M., Morel, J.L., 1994.Soil isotopically exchangeable phosphorus: A 615 comparison between E and L values. Soil Science Society of America Journal 58, 846-851. 616 Frossard, E., López-Hernández, D., Brossard, M., 1996. Can isotopic exchange kinetics give valuable 617 information on the rate of mineralization of organic phosphorus in soils? Soil Biology and Biochemistry 618 28, 857-864. 619 Frossard E, Condron, L.M., Oberson, A., Sinaj, S., Fardeau, J.C. 2000. Processes governing phosphorus 620 availability in temperate soils. Journal of Environmental Quality 29, 15-23. 621 Gallet, A., Flisch, R., Ryser, J.P., Frossard, E., Sinaj, S., 2003.Effect of phosphate fertilization on crop 622 yield and soil phosphorus status. Journal of Plant Nutrition and Soil Science 166, 568-578. 623 Gosling, P., Shepherd, M., 2005.Long-term changes in soil fertility in organic arable farming systems in 624 England, with particular reference to phosphorus and potassium. Agriculture Ecosystems and 625 Environment 105, 425-432. 626 Hoffman, D.W., Richards, N.R., 1954. Soil Survey of Bruce County, Ontario. Report No. 15 of the 627 Ontario Soil Survey. Environmental Farm Service, Canada Department of Agriculture and Ontario 628 Agricultural College, Guelph, ON, Canada. 29 629 Hoffman, D.W., Richards, N.R., 1952. Soil Survey of Perth County, Ontario. Report No. 15 of the 630 Ontario Soil Survey. Environmental Farm Service, Canada Department of Agriculture and Ontario 631 Agricultural College, Guelph, ON, Canada. 632 Knight, J.D., Buhler, R., Leeson, J.Y., Shirtliffe, S.J., 2010.Classification and fertility status of 633 organically managed fields across Saskatchewan, Canada. Canadian Journal of Soil Science 90, 667-678. 634 Kuono, K., Yuchiya, Y., Ando, T., 1995.Measurement of soil microbial biomass phosphorus by an anion 635 exchange membrane method. Soil Biology and Biochemistry 27, 1353-1357. 636 LECO Corporation. LECO instruction/operations manual for the FP-428 Nitrogen and Protein 637 Determinator version 2.4.LECO Corporation, St. Joseph, MI, USA. 638 639 López-Hernández, D., Brossard, M., Frossard, E., 1998. P-isotopic exchange values in a relation to Po 640 mineralization in soils with very low P-sorbing capacities. Soil Biology and Biochemistry 30, 1663-1670. 641 Macklon, A.E.S., Grayston, S.J., Shand, C.A., Sim, A., Sellars, S., Ord, B.G., 1997. Uptake and transport 642 of phosphorus by Argostiscapillaris seedlings from rapidly hydrolysed organic sources extracted from 643 32 644 McGill, W.B., Cole, C.V., 1981. Comparative aspects of cycling of organic C, N, S and P through soil 645 organic matter.Geoderma 26, 267-286 646 McLaughlin, M.J., Alston, A.M., Martin, J.K., 1988. Phosphorus cycling in wheat-pasture rotations 647 II.The role of microbial biomass in phosphorus cycling. Australian Journal of Soil Research 26, 333-342. 648 Miller, Robert O., Kotuby-Amacher, Janice, Rodriguez, Juan B. 1997. The measurement of soil pH in a 649 soil-water saturated paste. OMAFRA accredited method. In Western States Laboratory Proficiency 650 Testing Program Soil and Plant Analytical Methods.Version 4.00.p 15. P-labeled bacterial cultures. Plant and Soil 190, 163-167. 30 651 Morel, C. Plenchette, C., Fardeau, J.C. 1992.The management of phosphate fertilization in wheat 652 crops.Agronomie 12, 565-579. 653 Morel, C., Tiessen, H., Stewart, W.B., 1996.Correction for P-sorption in the measurement of soil 654 microbial biomass P by CHCl3 fumigation. Soil Biology and Biochemistry 28, 1699-1706. 655 Nash, D.M., Haygarth, P.M., Turner, B.L., Condron, L.M., McDowell, R.W., Richardson, A.E., Watkins, 656 M., Heaven, M.W., 2014.Using organic phosphorus to sustain pasture productivity: A 657 perspective.Geoderma 221-222, 11-19. 658 Oberson, A., Friesen, D.K., Rao, I.M., Bühler, R., Frossard, E., 2001. Phosphorus transformations in an 659 oxisol under contrasting land-use systems: The role of soil microbial biomass. Plant and Soil 237, 197- 660 210. 661 Oberson, A., Frossard, E., 2005. Phosphorus management for organic agriculture. In: Sims, J.T., 662 Sharpley, A.N. (Eds.), Phosphorus: Agriculture and the Environment. ASA, CSSA, and SSSA, Madison, 663 WI, USA, pp. 761-779. 664 Oberson, A., Joner, E.J., 2005. Microbial turnover of phosphorus in soil. In: Turner, B.L., Frossard, E., 665 Baldwin, D. (Eds.), Organic Phosphorus in the Environment. CABI, Cambridge, MA, USA, pp. 133-164. 666 Oberson, A., Pypers, P., Bünemann, E.K., 2011. Management impacts onbiological phosphorus cycling in 667 cropped soils. In: Bünemann, E.K., Oberson, A., Frossard, E. (Eds.), Phosphorus in Action. Biological 668 Processes in Soil Phosphorus Cycling.Springer, Heidelberg, pp. 431-457. 669 Oehl, F., Frossard, E., Fliessbach, A., Dubois, D., Oberson, A., 2004.Basal organic phosphorus 670 mineralization in soils under different farming systems. Soil Biology and Biochemistry 36, 667-675. 31 671 Oehl, F., Oberson, A., Sinaj, S., Frossard, E., 2001a. Organic phosphorus mineralization studies using 672 isotopic dilution techniques. Soil Science Society of America Journal 65, 780-787. 673 Oehl, F., Oberson, A., Probst, M., Fliessbach, A., Roth, H.R., Frossard, E., 2001b.Kinetics of microbial 674 phosphorus uptake in cultivated soils. Biology and Fertility of Soils 34, 31-41. 675 Ohno, R., Zibilske, L.M., 1991. Determination of low concentrations of phosphorus in soil extracts using 676 malachite green. Soil Science Society of America Journal 55, 892-895. 677 Olsen, S.R., Cole, C.V., Watanabe, F.S., Dean, L.A., 1954. Estimation of available phosphorus in soils by 678 extraction with sodium bicarbonate. United States Department of Agriculture Circular 939, 1-19. 679 Ontario Ministry of Agriculture and Food (OMAFRA), 2009.OMAFRA publication 811: Agronomy 680 guide for field crops.Available at: 681 http://www.omafra.gov.on.ca/english/crops/pub811/3fertility.htm#table3-7 (accessed 2 Feb 2014). 682 Parkinson, J.A., Allen, S.E., 1975. A wet oxidation procedure suitable for the determination of nitrogen 683 and mineral nutrients in biological material. Communications in Soil Science and Plant Analysis 6,1-11. 684 Quiquampoix, H., Mousain, D., 2005. Enzymatic hydrolysis of organic phosphorus. In: Frossard, E., 685 Turner, B.L., Baldwin, D.S. (Eds.), Organic Phosphorus in the Environment. CABI Publishing, 686 Cambridge, MA, USA, pp. 89-112. 687 Ramnarine, R., Voroney, R.P., Wagner-Riddle, C. Dunfield, K.E., 2011. Carbonate removal by acid 688 fumigation for measuring the d13C of soil organic carbon. Canadian Journal of Soil Science 91, 247-250. 689 Randriamanantsoa, L., Morel, C., Rabeharisoa, L., Douzet, J.M., Jansa, J., Frossard, E., 2013. Can the 690 isotopic exchange kinetic method be used in soils with a very low water extractable phosphate content 691 and a high sorbing capacity for phosphate ions? Geoderma 200, 120-129. 32 692 Randriamanantsoa, L., Frossard, E., Oberson, A., Bünemann, E.K., 2015. Gross organic phosphorus 693 mineralization rates can be assessed in a Ferralsol using an isotopic dilution method. Geoderma 257-258, 694 86-93. 695 Roberts, C.J., Lynch, D.H., Voroney, R.P., Martin, R.C., Juurlink, S.D., 2008. Nutrient budgets on 696 Ontario organic dairy farms. Canadian Journal of Soil Science 88, 107-114. 697 Rutherford, P.M., McGill, W.B., Figueiredo, C.T., Arocena, J.M. 2007. Total Nitrogen. In M.R. Carter 698 and E.G. Gregorich (Eds), Soil Sampling and Methods of Analysis 2nd ed. CRC Press Boca Raton, FL, pp 699 239-250. 700 Saleh-Lakha, S., Miller, M., Campbell, R.G., Schneider, K., Elahimanesh, P., Hart, M.M., Trevors, J.T., 701 2005.Microbial gene expression in soil: Methods, applications and challenges. Journal of Microbiological 702 Methods 63, 1-19. 703 SAS Institute, 2008. SAS Version 9.2 for Windows, Cary, North Carolina. 704 Saxton, K.E., Willey, P.H. 2006.The SPAW model for agricultural field and pond hydrologic 705 simulation.In V.P.Singh and D.K. Frevert (Eds), Watershed models. CRC Press, Boca Raton, FL., pp. 706 401–435. 707 Schneider, K.D. 2014.Understanding biological contributions to phosphorus availability in organic dairy 708 farm soils.PhD Thesis.University of Guelph, Guelph, ON, Canada. 709 Sheldrick, B.H., Wang, C., 1993. Particle size distribution. In: Carter, M.R. (Ed.), Soil Sampling and 710 Methods of Analysis. CRC Press, Boca Raton, FL, USA, pp. 499-507. 33 711 Sen Tran, T., Fardeau, J.C., Giroux, M., 1988.Effects of soil properties on plant-available phosphorus 712 determined by the isotopic dilution phosphorus-32 method. Soil Science Society of America Journal 52, 713 1383-1390. 714 Soil Survey Staff, 2014.Keys to Soil Taxonomy, 12th ed. USDA-Natural Resources Conservation 715 Service, Washington, DC. 716 Voroney, R.P., Brookes, P.C., Beyaert, R.P., 2008. Soil microbial biomass. In: Carter, M.R., Gregorich, 717 E.G. (Eds.), Soil Sampling and Methods of Analysis. 2nd Edition. CRC Press, Boca Raton, FL, USA, pp. 718 637-651. 719 Yevdokimov, I., Larionova, A., Blagodatskaya, E., 2016. Microbial Immobilisation of phosphorus in soils 720 exposed to drying-rewetting and freeze-thawing cycles. Biology and Fertility of Soils 52:685-696. 34 Table 1:Select soil properties and productivity indicators for the four field soils used in the study.Mean (standard deviation) of 4 subplots per field. Soil no. a Soil Texturea Total Organic Cb Total Inorganic Cc Total Nd (g N kg-1) Olsen STPf (mg P kg ) Total P (mg P kg-1 ) pHe -1 (g C kg-1) (g C kg-1) g Microbial Ci Organic Ph (mg P kg-1) (mg C kg-1) 1 silt loam/loam 20.7 (1.1) 4.1 (0.4) 2.2 (0.1) 7.4 (0.05) 2.8 (0.6) 684 (61) 216(32) 539 (30) 2 loam 27.3 (3.5) 4.5 (1.5) 2.6 (0.3) 7.2 (0.33) 5.5 (1.6) 781 (109) 222 (51) 744 (60) 3 silt loam 24.3 (1.2) 6.0 (1.0) 2.5 (0.1) 7.3 (0.17) 6.8 (3.3) 809 (81) 273 (23) 813 (137) 4 silt loam 23.8 (3.0) 6.9 (2.9) 2.1 (0.2) 7.5 (0.52) 11.2 (1.8) 729 (135) 145 (45) 522 (54) Particle size distribution was measured by the pipette method (Sheldrick and Wang, 1993) b Totalorganic C was measured using a Costech ECS4010 elemental analyzer (Costech Analytical, Valencia California) at the University of Saskatchewan Stable Isotope Facilities after the removal of inorganic carbonates by HCl fumigation for 48 hrs (Ramnarine et al., 2011) c Total inorganic C was calculatedas the difference between total C measured on bulk soil using a Costech ECS4010 elemental analyzer at the University of Saskatchewan Stable Isotope Facilities and total organic C measured as described above. d Total N was measured by dry combustion using an automated Dumas method according to LECO FP428 protocol (LECO Corporation; Rutherford et al. 2007) 35 e f pH was measured as a soil-water saturated paste (Miller et al. 1997) Soil test phosphorus as indicated by NaHCO3 extractable P (Olsen STP) g Total P was determined according to the wet digestion method of Parkinson and Allen (1975)using H2SO4 and H2O2 h NaOH-EDTA extractable organic P (Cade-Menun and Preston, 1996) i Microbial biomass C was determined according to the chloroform fumigation method using a kC factor of 0.35 (Voroney et al., 2008) 36 Table 2:Parameters determined from the isotopic exchange kinetics experiment,along with soil respiration rates measured during the 7-day incubation experiment for each of the four soils.Pw = water-extractable Pi(mg kg-1), where Pw = 10 (L kg-1) *Cp (mg L-1); m and n arecoefficients accounting for the immediate and slow physico-chemical reactions, respectively that describe the rate of decrease of radioactivity in the soil solution with time; E1min = amount of isotopically exchangeable P during the first minute of the experiment.Mean (standard error) of n=4 subsamples. Within columns, means followed by the same letter are not significantly different (p < 0.05). E1min Pw Soil no. n CO2 release rate m (mg P kg-1) (mg P kg-1) (mg C kg-1 soil d-1) 1 0.03a (0.03) 0.38a (0.01) 0.09a(0.01) 0.40a (0.15) 15.3a (2.0) 2 0.07a (0.03) 0.37a (0.01) 0.13b (0.01) 0.55a (0.15) 13.1a (2.0) 3 0.10a (0.03) 0.36a (0.01) 0.12ab (0.01) 0.78a (0.15) 15.9a (2.0) 4 0.24b (0.03) 0.32b (0.01) 0.14b (0.01) 1.62b (0.15) 17.1a (2.0) 37 Table 3: Specific activity (SA), which is the ratio of 33Pi: 31Pi of the water-extractable Pi pool (Pw) and the microbial P pool (indicated by hexanol-labile P) (Phex)and the ratio of SAPhex to SAPw for the four field soils at 1 and 8 days after soil labeling. Mean and standard deviation (n=4); within columns, means followed by the same letter are not significantly different (p <0.05). Soil no. SAPw SAPhex SAPhex : SAPw d1a d8 d1a d8 d1a d8a 1 2.3b (1.3) 0.9a (0.2) 0.4b(0.2) 0.6b (0.1) 0.2b (0.1) 0.8b (0.3) 2 1.2ab (0.3) 0.8a (0.2) 0.03a (0.03) 0.2a (0.1) 0.03a (0.02) 0.3a (0.1) 3 1.3ab (1.0) 0.7a (0.3) 0.1a (0.03) 0.3a (0.2) 0.1ab (.03) 0.4a(0.1) 4 0.7a (0.2) 0.5a (0.1) 0.1a (0.04) 0.2a (0.1) 0.1ab (.03) 0.3a (0.1) a statistical analysis was performed using log-transformed data. 38 Table 4: Extrapolated and measured E-values, gross Po mineralization (P min), Pi immobilization (P imm) and net Po mineralization determined for four field soils at the end of an 8-day incubation experiment. Daily P fluxes are also shown. Mean and standard deviation (n=4); within columns, means followed by the same letter are not significantly different (p < 0.05). Soil no. Extrapolated E-value Measured E-value Gross P min - Gross P imm mg P kg-1 soil - Net P min Daily gross Daily gross P min P imm - mg P kg-1 d-1 - Daily net P min 1 13.8a (4.5) 23.7a (4.6) 9.9a (2.1) 8.6a (3.1) 1.3a (3.6) 1.20a (0.26) 1.10a (0.39) 0.16a (0.44) 2 17.8a (5.5) 24.7a (7.3) 6.9a (5.4) 6.3a (1.7) 0.6a (6.2) 0.86a (0.68) 0.78a (0.21) 0.08a (0.77) 3 20.5ab (5.3) 30.0a (8.8) 9.5a (4.2) 6.6a (1.3) 2.9a (5.0) 1.17a (0.52) 0.81a (0.16) 0.36a (0.61) 4 29.8b (5.7) 42.6b (9.2) 12.8a (7.3) 5.7a (2.1) 7.0a (9.2) 1.58a (0.90) 0.71a (0.26) 0.87a (1.13) a Values for all 16 soil samples, including those after 1 day of incubation can be found in supplemental information (Table D1). 39 Figure Captions Figure 1:Recovery of 33P in the water-extractable Pi pool (Pw) (top) and the hexanol-labile P pool (Phex) (bottom) versus water-extractable Piconcentrations (mg kg-1) at 1d (left) and 8 d (right) after soil labeling. If the relationship between the variables was found to be significant, the linear regression equation is shown.The shapes indicate the different soils in the study: soil 1 ( ), soil 2 ( ), soil 3 (O), soil 4 (+). Figure 2:Specific Activity (ratio of 33 Pi:33Pi) of the hexanol-labile P pool (SA Phex) versus water- extractable Pi concentrations (mg kg-1) at 1d (top) and 8 d (bottom) after soil labeling. The relationship between the variables was found to be significant for day 8 and the regression equation is shown. The shapes indicate the different soils in the study: soil 1 ( ), soil 2 ( ), soil 3 (O), soil 4 (+). Figure 3: Gross Po mineralization (top), Pi immobilization (middle), and net Po mineralization (bottom) in relation to water-extractable Pi after 1(left) and 8 (right) days of soil labeling with 33P. If the relationship between the variables was found to be significant, the linear regression equation is shown. The shapes indicate the different soils in the study: soil 1 ( ), soil 2 ( ), soil 3 (O), soil 4 (+). 40 Figure 1 Figure 2 Figure 3