**2. Materials and Methods**

#### *2.1. Field Operations and Sampling*

Sampling took place in the CENTS long-term rotation and tillage experiment at the department of Agroecology, Aarhus University Flakkebjerg (55◦19 N, 11◦23 E) [18]. The soil is a sandy loam with 14.7% clay, 13.7% silt and 69.6% sand contents, average bulk density of 1.53 Mg m−<sup>3</sup> and an average organic matter content of 2% [19]. Soil pH (water suspension) at the time of sampling ranged between 7.4 and 8.3, with a mean value of 7.8. The long-term experiment consists of a randomized split-plot design with four crop rotations (R1–R4) as main plots and four levels of tillage (direct sowing, shallow harrowing, deep harrowing and inversion tillage) as sub-plots, arranged in four replicate blocks [20]. For this study, sampling was restricted to the R2 rotation, which consists of 3 years of winter barley (*Hordeum vulgare*) followed by 1 year winter rape (*Brassica napus*) and 2 years winter wheat (*Triticum aestivum*), all with straw retention after harvest [18]. The last crop before sampling was substituted with winter rye (*Secale cereale*) followed by fodder radish (*Raphanus sativus*) as a winter cover crop. Sampling was also restricted to direct sowing (NT) and inversion tillage (AuT). Finally, only three of the four experimental blocks in the long-term experiment were used for sampling and are considered here as replicate samples.

The NT treatment consisted of sowing by direct drilling using a single-disk drill (2002 to spring 2006) and later a single chisel coulter drill (autumn 2006–present), and straw retention at harvest. The AuT treatment consisted of seeding with a drag coulter seed drill, straw retention at harvest and inversion ploughing followed by rolling before sowing [20]. On the year of sampling, inversion tillage in the AuT sub-plots was carried out in late October.

In late August, glyphosate (1 L ha−<sup>1</sup> Roundup Bio) was applied to the NT and AuT sub-plots in the R2 rotation, killing the established mixture of cover crops, volunteers and weeds. This allowed the center portion of each sub-plot to remain bare (B), while the edges of the sub-plots were re-established exclusively with winter rye volunteers (V). Clippings carried out in the V portions of AuT showed that a growth of approximately 600 kg ha−<sup>1</sup> DW in aboveground winter rye biomass at the time of tillage in October.

In early December, two intact topsoil cores (Ø = 20 cm, h = 20 cm) were extracted side by side from the center (B) and edge (V) of each AuT and NT subplot. Core extraction was carried out by slowly pressing a steel cylinder into the soil with a hydraulic press, and subsequently digging around the buried cylinder with spades to manually retrieve the cores. This resulted in two separate sets of twelve samples covering four treatment combinations (AuT/NT × B/V) per block. One set of cores was designated as reference for field conditions (Reference), while the second set of cores was designated as experimental set to undergo simulated precipitation in the laboratory (lab-rain). All soil cores were transported back from the field and put in storage at 2–4 ◦C on the same day as extracted, awaiting sample preparation.

Additionally, composite topsoil samples (0–20 cm depth) were taken from the core sampling areas of all tillage and cover crop treatment combinations. Representative subsamples of loose soil were then used for total soil C and total N content analysis in a Vario Max Cube organic elemental analyzer (Elementar, Germany).
