Soil carbon and nitrogen mineralization after the initial flush of CO2

Soil health evaluation with biological activity requires standardization for greater understanding across environments. Soil‐test biological activity (STBA) may be an important indicator of soil N availability, but how it relates to long‐term soil N mineralization (NMIN) has not been documented. This study evaluated short‐ and longer‐term C and N mineralization in five soils from Georgia and North Carolina to validate associations between STBA and net NMIN. Although mathematical descriptions of cumulative C mineralization (CMIN) were logical and consistent among soil types, descriptions of net NMIN were complicated by the need to fit to single and double exponential models for different soil types. Rather than relying on exponential model fitting, emphasis on associations between linked processes of CMIN and NMIN resulted in simple, logical, and relatable interpretations. Soil‐test biological activity is a simple, rapid, and robust indicator that shows strong association with soil NMIN for up to 150 d.


INTRODUCTION
Management to achieve healthy, functioning soil is needed to meet production and environmental goals of sustainable agricultural systems (NRC, 2010). Soil biological activity and its role in nitrogen mineralization (NMIN) are key components of soil health evaluation (Doran et al., 1994;Franzluebbers, 2016). Soil biological activity is most often determined from aerobic incubation under standardized laboratory conditions (Anderson, 1982;Sainju, Stevens, Evans, & Iversen, 2013;Waksman & Starkey, 1924). Net NMIN, on the other hand, can be determined under either aerobic or anaerobic incubation conditions (Nyiraneza, N'Dayegamiye, Chantigny, & Laverdiere, 2009;Schomberg et al., 2015). The substrate leaching-incubation method has been a standard to determine pool sizes and activities of NMIN (Stanford & Smith, 1972). The method is time and resource consuming, and therefore, more rapid methods including anaerobic incubation have been used (i.e. 7 or 14 d) (Curtin et al., 2017;Reussi Calvo, Wyngaard, Orcellet, Sainz Rozas, & Echeverria, 2018). Other relatively simple and rapid procedures correlated with net NMIN have been explored, including cold-and hot-water extractants, weak bicarbonate and high-salt extractants, mild caustic base extractants, oxidizing extractants, and short-term carbon mineralization (CMIN) and NMIN assays (Curtin et al., 2017;Nyiraneza et al., 2009;Schomberg et al., 2009). Chemical extraction approaches have value, but they essentially sidestep the issue that NMIN is a biological process.
The objective of this study was to evaluate how soil-test biological activity (STBA) in its proposed form of aerobic incubation during 3 d following rewetting of dried soil (Franzluebbers, 2016) relates to longer-term soil CMIN and NMIN (i.e. up to 150 d). The hypothesis was that STBA would have strong association with net NMIN for longer periods of incubation than simply 0-24 d, as evaluated previously. If true, STBA could be useful for rapidly estimating potential NMIN under a diversity of environmental conditions.  Table 1. After representative field sampling, soil was dried in a forcedair oven (55 • C) for ≥3 d followed by gently crushing and passing through a screen with 4.75-mm openings.

MATERIALS AND METHODS
Soil was thoroughly mixed in a bucket and aliquots of 50 g and 74 ml were scooped into handling bottles (variable mass and volume were due to objectives in an ancillary study). Ten replicates were processed for each soil and extraction date (i.e. 10, 24, 60, and 150 d) (n = 190). The Mecklenburg CL did not have soil incubated and extracted at 150 d due to limited quantity available. Soil water content and bulk density were optimized for each soil during incubation, ranging from 145 to 308 g kg −1 and 1.01 to 1.51 Mg m −3 , respectively.
Soil was wetted to 50% water-filled pore space with a pipette delivering water to the top of each sample. Soil was incubated at 25 • C in a sealed 1-L canning jar in the presence of an alkali trap (10 ml of 1 mol L −1 NaOH). At the end of 3 d, alkali traps were removed and replaced again at 10, 24, 60, and 150 d, sequentially. Determination of evolved CO 2 -C was by acid titration with vigorous stirring to neutralize remaining T A B L E 1 Soil characteristics [soil-test biological activity (STBA), mg CO 2 -C kg -1 soil 3 d -1 , mean ± standard deviation; soil clay, sand, total organic C (TOC), and total soil N (TSN) concentrations (g kg -1 soil); and soil pH] from five soils evaluated alkali to a phenolphthalein endpoint following precipitation of carbonate with excess 1.5 mol L −1 BaCl 2 . Soil at the end of each incubation period was oven dried (55 • C, 3 d), sieved through a screen with 2-mm openings to homogenize, and then a 10-g subsample extracted with 2 mol L −1 KCl, filtered, and frozen. Inorganic N (NO 3 -N + NO 2 -N + NH 4 -N) was determined from a thawed aliquot on segmented flow chemistry (Bran-Luebbe, Autoanalyzer III) using hydrazine and salicylate-nitroprusside methods. Initial inorganic N was determined in a similar manner from four subsamples of each soil.
Statistical analyses of soil CMIN and NMIN data were conducted with the general linear model (SAS Institute, 2015) using a completely randomized design with factorial arrangement of 5 soil types × 4 extraction periods × 10 replications. Significance was declared at p ≤ .05. Cumulative CMIN and net NMIN data were fitted with nonlinear regressions using SigmaPlot v. 14 (Systat Software, San Jose, CA). Simple correlation and linear regression were used to assess strengths of association.

RESULTS AND DISCUSSION
Cumulative CMIN data were fitted to a double exponential model for all five soils. The first pool represented a readily mineralizable pool of organic matter, and the second pool represented a slowly available pool characterized by basal soil respiration. Coefficient of variation among replicates of different soils and incubation times was 10 ± 6%. This relatively small variation expressed the repeatable nature of the laboratory incubation procedure. Model regression coefficients for cumulative CMIN were vastly different among soils, reflecting differences in pool size and activity of the two main components ( Table 2). Multiplication of the pool size (a) and activity constant (k 1 ) for the small, readily mineralizable pool of C resulted in 1-d CMIN values of 254.7, 133.4, 66.3, 23.8, and 24.4 mg CO 2 -C kg −1 soil for the Mecklenburg CL, Georgeville SiL, Appling cSL, Norfolk LS, and Wedowee SL, respectively. Multiplication of the pool size (b) and activity constant (k 2 ) for the larger pool of C mineralized at a T A B L E 2 Non-linear regression coefficients for the description of cumulative C mineralization and net N mineralization from five soils evaluated. General forms of regressions were: C = a ⋅ (1 − e -k1⋅d ) + b ⋅ (1 − e -k2⋅d ) and N = N i + p ⋅ (1 − e -k3⋅t ) + q ⋅ (1 − e -k4⋅t )

Cumulative C mineralization (mg C kg -1 soil)
Net N mineralization (mg N kg -1 soil) Note. a, pool size for the small, readily mineralizable pool of C; k 1 , activity constant for a; b, pool size for the larger pool of C mineralized at a steady-state level; k 2 , activity constant for b; Ni, initial inorganic N; ps, pool size for the small, readily mineralizable pool of N; k 3 , activity constant for ps; q, pool size for the larger pool of N mineralized at a steady-state level; k 4 , activity constant for q; CL, clay loam; cSL, coarse sandy loam; LS, loamy sand, SL, sandy loam; SiL, silt loam; ND, not determined; NA, not applicable. steady-state level (i.e. basal soil respiration) was 75.6, 10.4, 9.5, 2.6, and 2.1 mg CO 2 -C kg −1 soil d −1 for the Mecklenburg CL, Georgeville SiL, Appling cSL, Norfolk LS, and Wedowee SL, respectively. The 1-d flush of CO 2 and basal soil respiration rate in these five soils were closely associated (r 2 = .87, p = .02). Both were also closely associated with soil organic C and total soil N (p ≤ .03). Mineralization of N also followed a nonlinear pattern (Figure 1). In all soils, there was an initial release of NH 4 -N during the first 10 d. Ammonium-N returned to a basal concentration level in all soils, except the Norfolk LS. Soil NH 4 -N concentration remained the dominant inorganic N form throughout F I G U R E 1 Inorganic N accumulation for five soils during incubation for up to 150 d. Least significant difference bars are for NH 4 -N, NO 3 -N, and summation of the two components as total inorganic N for each soil. Total inorganic N accumulation was fitted to a nonlinear regression equation. Note difference in scale between top and bottom panels the 150-d incubation period in the Norfolk LS. Although soil NO 3 -N eventually became the dominant form of inorganic N in the Wedowee SL, NH 4 -N retained significant concentration. Soil NO 3 -N accumulation was initially delayed but rapidly became the dominant inorganic N form in many of the soils.
A shift in dominance from NH 4 -N during the first 10 d to NO 3 -N afterward suggests that nitrifying microorganisms were present but delayed in their activity. Drying of soil is known to affect nitrifying microorganisms (Franzluebbers, Weaver, Juo, & Franzluebbers, 1994). Low and slow NO 3 -N accumulation over time in the Norfolk LS and Wedowee SL suggests that nitrifying microorganisms may have been inhibited in these two soils, perhaps due to presence of switchgrass for >10 yr. There is growing evidence that nitrifying microorganisms may be inhibited when exposed to vigorous root systems of warm-season grasses (Byrnes et al., 2017;Subbarao et al., 2015).
Inorganic N accumulation data were best fit to either oneor two-pool nonlinear regression models ( Table 2). The diversity of regression models made it more difficult to interpret coefficients among different soils. Taking the product of the pool size (ps) and its rate constant (k 3 ), instantaneous NMIN was 19.3, 27.4, 2.1, 0.3, and 2.6 mg N kg −1 soil for the Mecklenburg CL, Georgeville SiL, Appling cSL, Norfolk LS, and Wedowee SL, respectively. Instantaneous NMIN was not related to total organic C and N, unlike the close association observed between instantaneous CMIN and total organic C and N. Coefficient of variation in inorganic N accumulation among soils and incubation times was 12 ± 8%. Coefficient of variation was greater for individual inorganic species, with variation of 31 ± 28% for NH 4 -N and 39 ± 40% for NO 3 -N.
As determined and defined previously (Franzluebbers & Stuedemann, 2001;Franzluebbers, Pershing, Crozier, Osmond, & Schroeder-Moreno, 2018;Schomberg et al., 2009), net NMIN during 0-24 d was set as a standard. Compared with that at 0-24 d, net NMIN at 0-10 d was 0.8 ± 0.1 times greater, at 0-60 d was 1.5 ± 0.2 times greater, and at 0-150 d was 2.1 ± 0.5 times greater. As expected and shown in Figure 1, rate of soil NMIN declined with increasing length of time. This was likely a function of microbial incorporation of N into newly formed biomass and complex byproducts but more important, due to exhaustion of easily mineralized organic substrates with progressive time under incubation (Feng & Simpson, 2009). Long-term mineralization assays are conceptually useful but do not reflect changing conditions in the field, particularly in systems with continuous C inputs from growing plants as would be expected in perennial pastures and conservation cropping systems with a goal of achieving continuous living root systems. Difficulties in fitting inorganic N accumulation data to a consistent nonlinear regression model may have been one of the reasons why earlier scientific investigations failed to sustain momentum in predicting soil N availability from mineralization assays (Sierra, 1990;Smith, Schnabel, McNeal, & Campbell, 1980). Earlier focus may have been too much on mathematical quantification of model parameters describing the NMIN process. I suggest here that differences among soils in organic C and N resources of the active fraction may be a more practical target for defining soil N availability through organic material mineralization. This latter view can be evidenced by strong association of STBA with declining need for exogenous N in crop (Franzluebbers, 2018b;Yost et al., 2018) and pasture systems (Franzluebbers, Pehim-Limbu, & Poore, 2018).
In conclusion, STBA offers a simple, rapid, and robust indication (not direct estimation) of soil NMIN for up to 150 d. Greater utilization of this soil-testing tool will inform land managers of the important role that biologically active organic matter plays in nutrient cycling. Abundant data exist to relate STBA with net NMIN during 24 d of incubation (Franzluebbers, 2018a;and references therein). This study clearly demonstrated that STBA was also highly related to net NMIN during 150 d of incubation. It appears that once more readily available pools of C and N are exhausted, length of incubation becomes less important to overall NMIN estimation. Therefore, a simple yet practical means of estimating soil NMIN gives growers a reliable way of predicting N fertilizer requirements and avoiding unnecessary N inputs (Franzluebbers, 2018a(Franzluebbers, , 2020. This understanding is important so that (a) agricultural systems can become more efficient in N utilization and (b) agricultural advisers can help growers remain profitable and avoid unnecessary losses of N to the environment.

ACKNOWLEDGMENTS
Erin Silva, Ashley Turner, and Ellen Leonard provided excellent technical support. Funding was provided by USDA Agricultural Research Service.