Data from: Large subsurface carbon stocks in a long-term no-tillage site are vulnerable to potential mineralization

This dataset includes laboratory incubations of soils from the R.J. Cook Agronomy Farm Long-term Agroecosystem Research site in Pullman, Washington. The purpose of this dataset was to evaluate depth distributions of carbon pools following 17 years of no-till management. Soil cores were sampled to 153 cm depth at 25 locations representing 5 distinct soil series, a range in landscape positions, and a range in profile organic carbon contents. The soils were divided into 10 cm depth increments from 0 to 30 cm depth, and by genetic horizon from 30 - 153 cm depth. Air-dried, sieved soils were wetted to 50% water-filled pore space to initiate the incubation, and held at 21 degrees C for 350 days. CO₂ release was measured using alkaline traps that were replaced twice weekly for the first 2 weeks, once weekly from weeks 2-6, every 2 weeks from weeks 6-19, and every 4 weeks thereafter through day 350. Ancillary measurements to aid interpretation included C:N ratios, d¹³C of organic C, inorganic C content, soil pH, and texture.

Alternative compartmental model structures were evaluated to fit the incubation time series. Alternative fitting methods were also considered, including fitting a single set of transit times (k-values) to all samples, versus fitting a unique set of turnover times for each samples. Model comparisons supported the use of a 3-pool linear decay model with a single set of k-values optimized for all samples collectively, and no requirement to allocate total soil organic carbon (SOC) across the pools. This model therefore represents 4 actual kinetic pools of soil carbon, with 3 pools explicitly modeled, and the remaining carbon not estimated by the model comprising a fourth, resistant carbon pool. The transit times represented by this model were approximately 5 days, 2 months, and 2 years.

At the conclusion of the initial incubation measurements, we also initiated a new 60-day incubation on a subset of samples to evaluate changes in microbial community composition over time. A subset of 40 samples were selected by random stratification across horizons and across the range in total SOC. Dry samples were wetted to 50% water-filled pore space to initiate the incubation, and were held for 60 days at 21°C.  Phospholipid fatty acids (PLFA) samples were collected prior to incubation, and after 5 and 60 days, timepoints corresponding with the transit times of the fast and intermediate kinetic pools.

Results showed that SOC concentrations declined with depth, but the fraction that was mineralized in 350-day laboratory incubations increased with depth. We found 51% larger stocks of readily mineralized SOC stocks for the 30-153 cm depth (18.3 ± 4 Mg C ha^-1) compared to 0-30 cm (12.1 ± 1.6 Mg C ha^-1), when summing the kinetic pools that had transit times of up to 2 years. Although deep SOC had typical chemical fingerprints of extensive microbial processing, including narrow C:N ratios and ¹³C enrichment, the processing and incubation of soils reactivated decomposition of these substrates. Over the first 60 days of incubation, ratios of gram + to gram – bacteria decreased in deep soil and increased in near-surface soil. This suggested that bacteria favoring high-quality substrates were reactivated in deep soil when exposed to optimal decomposition conditions. These findings of large stocks of vulnerable SOC in subsoil show the importance of extending SOC monitoring below 30 cm depth.