Leaching of soils during laboratory incubations does not affect soil organic carbon mineralisation but solubilisation

Laboratory soil incubations provide controlled conditions to investigate carbon and nutrient dynamics; however, they are not free of artefacts. As carbon and nitrogen cycles are tightly linked, we aimed at investigating whether the incubation-induced accumulation of mineral nitrogen (Nmin) biases soil organic carbon (SOC) mineralisation. For this, we selected two soils representative of the C:N ratio values found in European temperate forests, and applied two incubation systems: ‘closed’ beakers and ‘open’ microlysimeters. The latter allowed leaching the soil samples during the incubation. By the end of the 121-day experiment, the low C:N soil significantly accumulated more Nmin in beakers (5.12 g kg-1 OC) than in microlysimeters (3.00 g kg-1 OC) but there was not a significant difference in SOC mineralisation at any point of the experiment. On the other hand, Nmin did not accumulate in the high C:N soil but, by the end of the experiment, leaching had promoted 33.9% more SOC solubilisation than beakers. Therefore, we did not find evidence that incubation experiments introduce a bias on SOC mineralisation. This outcome strengthens results from soil incubation studies.


Introduction
Laboratory soil incubations have been extensively applied in various areas of research [1,2] and are commonly used to investigate the potential mineralisation of soil organic carbon (SOC) under optimum conditions for microbial activity [3]. Compared to field experiments, laboratory incubations have the advantages of standardising soil samples (e.g. particle size, bulk density) and controlling environmental drivers of mineralisation (e.g. temperature, moisture content). During the decomposition of soil organic matter, ammonification transforms organic nitrogen into NH 4 + , which is transformed into NO 2 -and NO 3 through nitrification.
Ammonification is part of the decomposition process of soil organic matter and leads to a net production of CO 2 , even though the chemoautotrophic oxidation of NH 4 + requires CO 2 . In a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 the field, mineral nitrogen (N min ) is susceptible to being leached down the profile. However, in beakers, the 'closed' and most frequently used incubation system [4], this removal does not occur. As a result, N min may accumulate in the soil sample potentially affecting SOC mineralisation. In an attempt to minimise this incubation-induced accumulation of N min , in this study we applied an 'open' incubation system, referred to as microlysimeter [5][6][7], that permits the leaching of soil samples during the experiment. Most incubation studies with N-amended soils apply the treatment at the beginning of the experiment and have shown that increases in N min concentrations typically decrease native SOC mineralisation [8,9]. Potential reasons for this decrease are the decline in N-mining by soil microbial communities [10] and the suppression of extracellular enzymes [11]. However, there are also incubation studies that have shown no effect [12] or a positive effect of N min on SOC mineralisation in cases of soils with severe N limitation [13]. As the availability of N min is influential on substrate decomposition [14], bulk soil C:N ratios can contribute to explain the dynamics of these two elements in soils.
In this study, we compared 'closed' (i.e beakers) and 'open' (i.e. microlysimeters) incubation systems and tested whether N min accumulation occurred and whether it had an effect on SOC mineralisation. We hypothesise that the effect of the incubation-induced accumulation of N min on SOC mineralisation depends on the C:N ratio of the soil, with suppressed and stimulated mineralisation in low and high C:N soils respectively. This hypothesis is of particular relevance to laboratory incubation studies. If proved true, C mineralisation measured in these experiments could be differently biased depending on the C:N ratio of the incubated sample.

Ethics statement
Sites from where soils for this study were sampled are part of the Swiss Federal Institute of Forest, Snow and Landscape Research network. All necessary permits were obtained before sampling. This study did not involve endangered or protected species.

Selection, sampling and characterization of study soils
To test our hypothesis, we selected two soils (Table 1) from the database of the Swiss Federal Institute for Forest, Snow and Landscape Research (WSL) [15]. In May 2014, this database contained data on 1,050 soil profiles spread across Switzerland. The C:N ratio of the two soils selected (12.51 ± 0.03 and 17.43 ± 0.17, mean ± s.e.m.) was representative of the values most commonly found in European temperate forest mineral soils [16]. But at the same time, these two soils were different within the range of probability distribution of C:N values in Swiss forest soils (2 nd and 7 th decile); therefore, representing soils of contrasting N min dynamics.
In August 2014, we collected three soil composites within a 40 × 40 m 2 plot at each site. Each composite was the product of mixing eight 0-20 cm depth soil cores collected from one of three non-overlapping areas of the plot. This sampling strategy enabled us to account for spatial variability. Each composite constituted an experimental replicate. Soil samples were collected with a 5 cm diameter Humax corer. After collection, samples were transported in portable fridges to the lab where they were freshly sieved by hand ( 2 mm) and stored at 3.5˚C until the beginning of the experiment in March 2015.
We measured soil pH on 40˚C dried composite subsamples. Part of these dried soil subsamples were also milled and fumigated with HCl to measure total organic carbon and total nitrogen with an Elemental Analyser (vario MICRO cube, Elementar, Germany). The rest of data in Table 1 is part of the WSL database [15,17].

Incubation experiment
We incubated fresh soil (sieved to 2 mm; 40 g equivalent dry mass; adjusted to 0.8 g cm -3 bulk density) in sterilised glass beakers and microlysimeters. Soil samples in both incubation systems were placed separately into 2 litre airtight glass jars that contained 20 ml of distilled water to ensure the headspace was moist. Lasting 121 days, the experiment was conducted under aerobic conditions, in the dark and at 25˚C. The experiment started after a 10-day preincubation which was run under the same conditions as in the incubation assay.
During the course of the experiment, we repeatedly measured SOC mineralisation and the concentration of N min and organic carbon (OC) in two types of soil water extracts (SWE). The CO 2 -C product of the mineralisation of SOC was captured in 20 ml NaOH (1M) traps placed into the 2 liter glass jars. Subsequently, the amount of C trapped was determined by the change of conductivity of the NaOH [18]. Mineral N (NO 3 and NO 2 -) in SWE was quantified with a continuous flow analyser (San++, Skalar, Breda, the Netherlands). We also measured NH 4 + but the concentrations were below the detection limit (1 ppm) and are not reported. Finally, OC in SWE was determined with a TOC analyser (DIMA TOC-2000 Dimatec, Essen, Germany). Since the beginning of the experiment, NaOH traps were taken out and replaced for fresh ones on days 4, 13, 30, 63 and 121. On these five days, microlysimeters were leached (SWE l ) and beakers and microlysimeters were destructively extracted with water (SWE e ); hence we obtained two types of SWE samples. To produce SWE l samples, 30 ml of nutrient solution [19] without N or P were added to each microlysimeter. Subsequently, after equilibration for 30 minutes, the systems were leached by applying a suction of -20 kPa for 25 minutes. To obtain SWE e samples, three replicates of each combination of soil and incubation system (n = 12) were extracted with distilled water (1:5 soil:water) after shaking for 30 minutes at 100 rpm. SWE l and SWE e samples were filtered (1.6 μm MGA glass microfibre, Sartorius) before analysis of N min and OC.
Until the time of a given leaching date, the total quantity of N min or OC solubilised ([SWE T ] t ) for samples in microlysimeters, was the sum of the quantity of N min or OC found in the SWE of the destructed sample ([SWE e ] t ), plus the cumulative quantity of N min or OC removed in previous leaching cycles ( P tÀ 1 i¼1 ½SWE l i ) (Eq 1).
Eq 1 does not apply to samples incubated in beakers. As 'closed' systems were not flushed, their total quantity of N min or OC solubilised is equal to SWE e .

Data analysis
Statistical analysis was performed with the R software (version 3.3.2) [20]. Incubation data (i.e. SWE e -N min , SWE e -OC, OC mineralised) were normalised relative to the total OC content of the bulk soil at the beginning of the experiment. Incubation systems and soil types were compared by Student's t-tests. Results can be found in the Supplementary Material section. Errors given in the text, tables and graphs are standard errors of the mean. All data produced in this study is open access [21].

Soluble mineral nitrogen
Results show that the concentration of soluble N min in the incubated samples related to the C: N ratio of the soil (Fig 1A, S1 Table). In both incubation systems, the low C:N soil accumulated N min non-linearly over the incubation period, but by day 121 this accumulation was significantly smaller in microlysimeters compared to beakers (p < 0.05, t-test). In the high C:N soil, N min remained constant in both incubation systems (Fig 1A). When examining the total N min produced (Fig 1B, S1 Table), there was no significant difference between beakers and microlysimeters for any of the two soils.

Soluble organic carbon
Soluble OC did not accumulate in any soil or incubation system (Fig 2A, S2 Table). However, when considering the total soluble OC produced over the incubation period (Fig 2B, S2  Table), we observe that by day 121, the high C:N soil released more OC in microlysimeters than in beakers (p = 0.07, t-test). Additionally, the cumulative OC leached from microlysimeters represented about 0.96 ± 0.13% and 1.79 ± 0.10% of the cumulative CO 2 -C released, in the low and high C:N soil respectively.

Soil organic carbon mineralisation
By day 121, there was no significant difference in the cumulative C mineralised between soils or incubation systems (Fig 3, S3 Table). Quantitatively, the C mineralised by the low C:N soil was 90.52 ± 4.00 g kg -1 OC in beakers and 92.23 ± 5.89 g kg -1 OC in microlysimeters. The high C:N soil mineralised 87.30 ± 2.62 g kg -1 OC in beakers and 90.39 ± 1.80 g kg -1 OC in microlysimeters.

Discussion
'Closed' systems promoted the accumulation of N min in the low C:N soil, but this accumulation did not significantly affect SOC mineralisation Although we measured an increase of 2 and 3 times the initial N min concentration in microlysimeters and beakers respectively by the end of the incubation of the low C:N soil (Fig 1A), we  Soil incubation leaching and nitrogen accumulation did not observe a significant effect of these two levels of N min on SOC mineralisation (Fig 3). Therefore, this result contradicts our initial hypothesis, even considering that the increase of N min in this soil in beakers (140.27 ± 3.70 μg N min g -1 soil) was the same magnitude as than treatments applied in N-amended incubations and where a negative effect on SOC mineralisation was found [9]. One explanation for this result could be that the concentrations of N min were not high enough to inhibit decomposition enzymes [22]. A second explanation could be that N-amended incubations are commonly short-term assays (i.e. hours, days) that apply a unique treatment at the beginning of the experiment [9,12,13,23]. Contrary to these experiments, the accumulation of N min in this study was gradual and hence, soil microorganisms may have progressively adapted to the increasing N min concentrations [23].
Mineral N did not accumulate in the high C:N soil and leaching did not induce N limitation In the high C:N soil in beakers, we observed a constant concentration of N min (Fig 1A). This was probably due to a steady state between the production and the immobilisation of N min by soil microorganisms [24,25]. Leaching did not induce N min limitation in the high C:N soil. This conclusion is supported by the fact that the total N min produced over the incubation period ( Fig 1B) was not significantly larger in microlysimeter than in beakers.

Leaching promoted the solubilisation of SOM in the high C:N soil
Leached OC represented on average only~1% of the SOC mineralised as CO 2 ; therefore, it is unlikely that the depletion of labile C from leaching had a significant effect on the mineralisation of SOC. We did not observe an accumulation of soluble OC over the incubation period in any soil or incubation system, but a stable concentration within the range of 2.8-4.5 g kg -1 OC (Fig 2A). This dynamic reflects an equilibrium between the solid and the aqueous phases of soil organic matter [24][25][26][27][28]. This explanation is also supported by the higher solubilisation of SOC in the leached high C:N soil compared to the un-leached system (Fig 2B).

Conclusions
In this study, we aimed at investigating whether the incubation-induced accumulation of N min biases SOC mineralisation. We selected two soils that (i) were representative of the C:N ratio values found in European temperate forests, and (ii) differed on their net nitrogen mineralisation. Results demonstrated that the progressive accumulation of N min , which only occurred in the low C:N soil, did not have a significant effect on the mineralisation of SOC. In parallel, N min did not accumulate in the high C:N soil, but leaching promoted higher solubilisation of SOC. Our results are based on two representative European temperate forest soils, but incubations are applied to a wider range of soil types. Therefore, to test whether the results of this study hold independently of the characteristics of the incubated samples (e.g. pH, texture), future work should be undertaken with a broader range of soils, also including strongly N-limited ones.
Supporting information S1 Table. Results of Student's t-tests to compare two incubation systems (i.e. leached microlysimeters and un-leached beakers) in two soils (i.e. high and low C:N ratios). Data tested are soluble mineral nitrogen (NO 3 --N and NO 2 --N) relative to total soil organic carbon (Fig 1A and 1B) over the 121-day incubation (g kg -1 OC). SWE e -N min was measured in 1:5 soil water extracts (Fig 1A) and SWE T -N min was calculated (Eq 1) for the total soil water extract as the sum of extracted and leached N min (Fig 1B). Cell values: Significance code based on p-values (''1, '.'0.1, ' Ã '0.05, ' ÃÃ '0.01, ' ÃÃÃ '0.001), t-value, p = p-value. (DOCX) S2 Table. Results of Student's t-tests to compare two incubation systems (i.e. leached microlysimeters and un-leached beakers) in two soils (i.e. high and low C:N ratios). Data tested are soluble organic carbon (OC) relative to total soil organic carbon (Fig 1A and 1B) over the 121-day incubation (g kg -1 OC). SWE e -OC was measured in 1:5 soil water extracts (Fig 2A) and SWE T -OC was calculated (Eq 1) for the total soil water extract as the sum of extracted and leached OC (Fig 2B). Cell values: Significance code based on p-values (''1, '.'0.1, ' Ã '0.05, ' ÃÃ '0.01, ' ÃÃÃ '0.001), t-value, p = p-value. (DOCX) S3 Table. Results of Student's t-tests to compare: in row-1, two incubation systems (i.e. leached microlysimeters and un-leached beakers) in two soils (i.e. high and low C:N ratios) and in row-2, two soils when incubated in two incubation systems. Data tested are cumulative carbon mineralised relative to total soil organic carbon (Fig 3)