2.2. Experimental design and soil incubation
The soil from the OF treatment (15 year-old) represents active SOC. The soils from the BF treatment for 15 years and BF treatment for 23 years plus an extra 815 days of laboratory incubation (BF+, the additional incubation is to further consume active organic carbon) represent relatively resistant SOC (Zhang et al., 2017). Therefore, we refer to the OF SOC as active C, the BF and the BF+ SOC as relatively resistant C in this study. Although all three soils may still contain various components with different turnover times and protection mechanisms, here we operationally used the three soils to represent active and relatively resistant SOC (Conant et al., 2008; Townsend et al., 1997), and compared the relative vulnerability of these SOC pools to the direct and legacy effects of drying-wetting cycles for two reasons. First, we measured the proportion of active C (dissolved organic carbon/total carbon, DOC/TC) and found that the proportion of active C in the OF soil is the highest, and that in the BF+ soil is the lowest (Table 1). Second, we measured the chemical composition of SOC of these three soils and found that the molecular index of SOC stability (or resistance) in the OF soil is the lowest, and that in the BF+ soil is the highest (Fig. 2). In addition, Barré et al. (2010) constrained a three-pool-model using observed soil organic C data in long-term BF soils and found that most of the organic C are stable C (turnover time of several centuries or more). Basic properties of these three soils can be found in Table 1 and Zhang et al. (2017).
The incubation experiment consisted of a nine 10-d DWC period and a 28-d extended period, with a 10-d stabilization period between them (Fig. 1). In the DWC period (0-90 d), soils were subjected to three water regimes, including one constant-moisture control (mean moisture content: 60% WHC), one mild DWC treatment (80-40% WHC), and one strong DWC treatment (100-20% WHC), each with three replicates. After the 90-d DWC period, all treatments were adjusted to constant moisture of 60% WHC for ten days to a consistent state (90-100 d). All soils were kept at 60% WHC during the subsequent 28-d extended period (100-128 d).
For each soil, a subsample of 60 g (dry weight basis) was placed in a 1-L incubation jar at 60% WHC with deionized water, and the soils were pre-incubated at 22°C for 10 days in the dark. To achieve soil drying, 45 g and 20 g silica gel were placed in the incubation jars for 100-20% WHC and 80-40% WHC treatments, respectively (Fig. S1). These two weights were chosen because our pre-experimental data showed that these two weights of silica gel could accelerate the loss of soil moisture to the expected levels (i.e. 20% and 40% WHC) in 10 days. Inside each jar, 10 ml 0.5 M NaOH solution was used to trap CO2 from SOC decomposition. At the end of each drying period, the NaOH solution and silica gel were replaced, and the soil samples were rapidly rewetted by adding deionized water with a syringe. The trapped CO2-C in the NaOH solutions were immediately measured in the form of total inorganic C using a Lotix combustion TOC analyzer (Teledyne Tekmar, USA), and the silica gel was dried in an oven at 105°C overnight to regenerate it. After replacing the silica gel and NaOH solution, the jars were sealed immediately with gas-tight lids and kept at 22°C in the dark. A previous study proved that silica gel does not absorb CO2(Harrison-Kirk et al., 2013). Three blanks without soil were incubated at the same condition to quantify the amount of C in the air and initial NaOH solution. Prior to measuring the total inorganic C, the NaOH solutions in the DWC treatments were adjusted back to 10 ml with deionized water.
After the 90-d DWC period, all soils were adjusted to a constant moisture content of 60% WHC and maintained for 10 days (stabilization period, 90-100 d). Note that we did not measure respiration for the stabilization period. On day 100, we did a destructive soil sampling. Specifically, 8 g subsamples were used to determine dissolved organic carbon (DOC) and microbial biomass carbon (MBC), 10 g subsamples were stored at -80°C until the measurement of microbial biomass, and 10 g subsamples were incubated for another 28 d (extended period, 100-128 d) in which the CO2-C in the NaOH solutions were measured on day 103, 107, 114, 121 and 128.
In total, we incubated 30 jars ((3 × 3 + 1) × 3) consisting of three soils (OF, BF and BF+ soils), two DWC treatments (100-20% and 80-40% WHC) and one constant-moisture control treatment (60% WHC), one blank, and three replicates for each treatment.