The authors have made many changes and additions in detail in the course of the revision. Unfortunately, this has not improved the quality of the manuscript; rather, the fundamental deficiencies have become even more apparent.
In the introduction and in the discussion (lines 89-100 and 718-724, respectively) it is claimed that with the help of a model for simulating potential peat mineralization it is possible to determine the extent of C losses due to peat decomposition or the resulting CO2 source function of drained peatlands more precisely than possible with the methods available so far under field conditions according to the IPPC guidelines. However, due to a serious factual error, the authors could not provide the evidence for this, contrary to their statements. Namely, when designing their model, the authors wrongly assumed that changes in the C stock of the studied peatlands under real conditions in the field are determined solely by the microbial degradation of the peat. However, this is only true for ecosystems without growing plants. As soon as these occur, as in the case of the studied meadows, the changes in the C-supply result in a first approximation from the resultant of the opposing C-fluxes C-input of plant assimilates and C-output by microbial decomposition (respiration) of the old peat. The importance of current C-input may be relatively small on drained peatlands compared to C-output in the form of heterotrophic respiration, but it still exists in the form of shoot litter, roots, and rhizodeposition. Unlike the authors' modeling approach, common methods for measuring C supply, such as examination of the entire peatland soil profile and the C budget approach based on CO2 flux measurements (i.e., determination of NECB differences), which the authors also used, implicitly or explicitly account for this fact. Thus, failure to account for C input when modeling C dynamics inevitably leads to findings that, a priori, cannot correspond to actual changes in the C stock of the peatland ecosystem, nor are they comparable to results obtained using the other methods mentioned above.
While it is not shown directly by the authors, it is clear from converting the respiration values shown in Table 4 from d m yr-1 to kg CO2 m-2 (line 482) that the changes in C stocks determined using the two methods do not in fact match. Namely, the C losses determined using the modeling approach are always lower than the NECB differences determined by the C budget approach. The authors do not address this issue at all. Although the consistency of the C losses determined by both methods is the crucial prerequisite for using the modeling approach instead of the very costly C budget approach, which is currently considered a standard method according to IPCC guidelines. Which in itself would be very welcome because of the many advantages that the modeling approach brings.
This dilemma is not helped by the close correlative relationship between the modeled potential respiration and the summary NECB ifferences and their individual C budget components, which is undoubtedly determined by the dominance of hydrological conditions. This is due to the above-mentioned fact that changes in C stocks are always determined simultaneously by C input and C output, and both C pathways can in principle change independently of each other. This would have become apparent very quickly if, in addition to ryegrass, a fen meadow with more water-tolerant grass species such as reed canarygrass had also been studied. Very likely, the effect of increased groundwater levels on plant-relevant CO2 fluxes and C export would have been weaker here, ultimately leading to a shift in the correlative relationships between modeled potential respiration and measured NECB difference and C budget components. And that one would always have to measure the CO2 fluxes and the NECB differences in order to determine the changed correlative relationships would not be at all in the sense of the matter.
In view of this, must again recommend rejection of the manuscript.
However, this assessment does not mean that the authors' effort to complement or replace the methods used so far to determine C losses from peatlands with a modeling approach that couples C and water dynamics is worthless or meaningless. On the contrary, such an approach has great potential to address the weaknesses of the standard methods. This concerns, not least, the precise capture of the spatiotemporal dynamics of the processes and the ability to reliably capture C losses at the field and landscape scales. However, one will only obtain an accurate and generalizable approach if all processes relevant for C dynamics and their regulation are integrated in it. Interesting suggestions in this sense can be found in van Huissteden et al. (2006), Premrov et al. (2021, Sci. Total Environ. 754, 14243), Pohl et al. (2015, Biogeosciences, 12, 2737-2752), and Tiemeyer et al. (2016). The last paper also already pointed out the particular large influence of groundwater level in summer on the climate impact of peatlands. |