However, model construction may be severely influenced by tillage operations. The impact of tillage on
the temperature dependence of SR was studied to reveal the temporal patterns of model quality of
temperature-driven SR-models. To obtain SR, CO2 fluxes were measured with a dynamic chamber
technique in treatments of an energy crop rotation amended with biogas residues (BR) and mineral
fertiliser (MF). Measurementswere performed with short intervals during the first three days after tillage
operations, then with extending intervals between measurements up to 35 days after tillage.
Additionally, soil concentrations of hot-water extractable organic carbon (HWC) were determined
before and during the experiment. Overall, in all treatments individual CO2 fluxes were affected by the
extent of soil disturbance and fertiliser treatment. The highest tillage-induced fluxes where observed
after disking in MF treatment. Tillage also induced an immediate increase of HWC, indicating additional
labile C and a fast response of microbial activity. However, the change of HWC lasted only one day and
approximated the pre-tillage values within a week. Even though BR soil had a higher HWC content, the
increased C mineralisation in one repetition of MF suggests that buried plant residues might have a
higher influence on SR after tillage than the type of fertiliser. Directly after soil disturbance by tillage it
was impossible to construct temperature-driven models for SR in all treatments. Assuming that the
coefficient of determination is appropriate with R20.5 and the model quality is good with
NRMSE0.15, the qualities of the models increased continuously with time, but were unsatisfying
for at least two weeks. During this time, SR showed a high sensitivity to changing environmental
influences like precipitation and soil moisture or available C for microbial turnover, rather than
temperature. The treatment BR showed a less sensitive pattern, which might be attributed to an altered
soil structure and microbial activity of soil after long-term application of an organic fertiliser like BR.
Therefore, temperature-driven models for the prediction of soil derived CO2 emissions should be applied
carefully for the days and weeks after tillage and verification by measurements in shorter intervals is
advisable.