Reply on CC2

Community commetns 4: Irina Kurina: The manuscript “Holocene wildfire regimes in forested peatlands in western Siberia: interaction between peatland moisture conditions and the composition of plant functional types” by Angelica Feurdean et coauthors is of great interest and raises a very relevant topic for scientific research. This research is based on a variety of evidence, it is used a multi-proxy approach to the detailed palaeoenvironmental reconstruction. Age-depth models for the studied peat cores as a necessary basis for work looks very strong. I think also that the study area related to the southern part of Western Siberia, on the one hand, is poorly studied and, on the other hand, this is the one of the most suitable places in Eurasia for such research because of the widespread occurrence of peatlands. Although the design of the manuscript (so called paperwork) leaves much to be desired. And the interpretation of the data obtained raises many questions. Therefore below I provide my comments (questions) to the manuscript.

R: We agree that the link between the water table and climate needs to be further demonstrated, but the evidence provided by our study is sufficient to suggest that dry peatland conditions are likely to burn often and /or severely. In revising our paper, we are extending the explanation on testate amoeba hydrological reconstruction slightly, but the full record of testate amoeba (TA) assemblages as well as the hydrological reconstruction and interpretation at the four sites will be published extensively in a future publication (Diaconu et al., in prep). TA assemblages record onsite (very local) changes in peatland hydrology whereas macrocharcoal in peatlands record local fire occurrence, generally strongly influenced by the local peat moisture (Blyakharchuk and Kurina, 2021;Magnan et al., 2012;Feurdean et al., 2020). Thus, although not identical, the spatial scale of the two proxies is rather local. The water table in peatlands is driven by the interplay between precipitation and evapotranspiration and that a substantial amount of water is lost via evapotranspiration, which is largely controlled by air temperature (Swindles et al., 2019;Diaconu et al., 2018: Pleskot et al., 2021. If the peatland type is oligotrophic or mezzooligotrophic (isolated or partially isolated from other water sources except for precipitations) then peatland moisture is a direct reflection of the hydroclimate conditions. Even when a peatland is dependent on the discharge of a water source, the quantity of the discharge is still dependent on the climatological conditions, which is indirectly reflected in the DWT. The good agreement between the high fire activity, low water table, and regional high summer temperature reconstructions, makes us suggest the existence of a link between fire, peatland hydrology, and summer temperatures (see similar findings from literature (Turetski et al., 2014;Kettrige et al., 2015, However, in  R: These are predictions for potential future forest-fire interactions as a response to future climate changes based on what we have learned from our palaeoecological data. We do not consider that this information is unfounded, but rather it follows logical findings from our long-term records. We, therefore, decided to keep these here. R: Thank you, we have added a sentence acknowledging this. It reads: ''Despite that Siberia contains a large extent of forested peatlands, particularly its western part (Vompersky et al., 1994;Liss et al. 2001;Kirpotin et al., 2021) no studies have explicitly explored the interactions between peatland moisture, vegetation composition, and fire regime in this region''. R: We started the introduction by acknowledging the role of fire in boreal forests generally (l.60-75). We moved to recognise that a lot of boreal forests grow on peatlands and that the relationship between such forests and fire is less know (l. 76-89). We then introduced the usefulness of palaeoecological research to capture past changes in fire regime in forest ecosystems (90-91). It is not relevant at this point whether a forest occurs or not on peatland, but that an understanding of the dynamic of a forest ecosystem, which contains species that live decades to centuries, needs long-term records. Finally, the primary aim of our study is to look at forested peatlands. We believe that the introduction follows a logical path, however, in revising the paper we will make it clear throughout the other parts of the paper that we accurately use the terms forests and forested peatlands.

Sorry, it is difficult for me to understand the key idea of your research. So I read the introduction of your manuscript. Lines 60-75 -you write about wildfires in forests.
Line 170 -you write "To determine the regional changes in forest composition, we created composite records of PFTs". Could you explain, how do you determine the regional changes in forested peatland composition? How do you separate the composition of forests and forested peatlands? Based on my individual experience, I cannot imagine that tree composition is the same in a forested peatland and in forests, which surround it. In most cases they are different. As confirmation, let us look at the description of modern conditions in the studied peatlands   Fig 1) so that in the end we obtained a regional picture of forest composition and dynamics. Each peatland we cored is forested and composed of tree species mentioned at 2.1 (please note a slight modification to accommodate Loyko's suggestion). The vegetation at the coring points was indeed slightly open as we have chosen this approach to make our coring easier, but otherwise, these peatlands are forested. The principle of pollen analysis is that it records species composition on a scale ranging from local to regional (Bennet and Willis, 2001). Given the large extent of forested peatlands we cored, we assume that the bulk of our pollen diagram indicates the tree composition on these forested peatlands. It is also highly probable that an unknown proportion of the pollen rain comes from outside peatlands. Thus, although it is not possible to separate the pollen originating from forested peatlands from pollen coming outside of the peatlands, in revising this manuscript, we will be more careful in describing the past forest composition and dynamics regarding the location of these forests and in linking the local variability in peatland hydrology (local) with forest composition (representative of the larger area likely with more variable moisture conditions).
Line 320 -you used Ti concentration as "possible indicator of water influx". Could you provide any reference to the researches that confirm this idea? I am surprised to see such interpretation of the Titanium peaks. As far as I can consider from different papers (as example, Kempter and Frenzel 2008 in Science of the Total Environment; Margalef et al. 2014 in Palaeogeography, Palaeoclimatology, Palaeoecology;Hutchinson et al. 2016 in Regional Environmental Change -you cited the last reference) Ti is mainly precipitated from atmosphere. Its increasing peaks in a peat core (or lake sediments) can be caused by wind or soil erosion, by enchanced precipitation, or by increased production of the ecosystem. There are many reasons for positive peaks of Ti, but I never heard about river flood as the reason. If we look in your manuscript in Line 320 you write "The detrital element Ti, a possible indicator of water influx, was high in the bottom profiles that were rich in minerogenic material.". Then, in Lines 326-327 you write "Proxy records from Siberia attest to warmer and drier-than-present climate conditions between 9 to 6 ka … (Groisman et al. 2012)". The age of bottom profiles in the studied mires with high peaks of Ti is about 8.5-7.0 ka (I take it from place Rybnaya mire, Fig. S5). So it coincided with period of drier climate conditions in Siberia. I think that flood events should be happen if precipitation increases, but precipitation was reduced at the period. How can you explain this discrepancy? I can imagine that this period of drier climate conditions might contribute to frequent fires, deforestation and enchanced soil erosion. This is just my opinion, but I think this is more reliable explanation for the Ti peaks, than river flood, that you suggested. R: Thank you for this comment which indicates that we need to clarify our explanation of the geochemical features of the peat profile. The higher Ti content in the basal, minerogenic portion of Rybanya and, in fact, also at UC is not connected to floods but reflects the minerogenic substrate at the sites and therefore pre-dates the inception of the peat. We refer to subsequent (higher in the profile) Ti fluctuations as indicating possible flood events i.e., low frequency but high magnitude events which may not reflect the overall climatic trend at the time. As a lithogenic (or geogenic) indicator Ti can be seen as an indicator of detrital input reflecting the mineralogical content (in comparison to the highly organic nature most peat profiles). The source of this material will reflect both the type of mire and the events leading to this input. In an ombrotrophic context the input will be aeolian. Here the landscape position of the site means that (at this stage in the mire) a fluvial input (reflecting possible flooding or channel position change) is feasible as a transport mechanism for the delivery of such material and associated lithogenic signal (also seen in other lithogenic indicators but Ti has been selected as indicative).
Line 181 -you used the transfer function developed for the pan-European region (Amesbury et al. 2016) to derive the water depth from the studied peat cores. This transfer function was developed mainly (or even especially) for ombrotrophic and oligotrophic peatlands, but you applied this to the mesotrophic mires. I think this might increase the incorrectness of the reconstructed values of DWT in your study. Why did you not used the transfer function developed for Asian peatlands (Qin et al. 2021 in QSR), because the studied mires are located in Asia, but not in Europe? Apart from that, I can say that the transfer function developed for Asian peatlands includes more places with higher values of pH and therefore, I guess, it might be more suitable for reconstruction of DWT in the studied mires. Also I suggest adding testate amoeba diagrams from the studied peat cores to the Supplementary Materials of your manuscript. It is very important and interesting data. Furthermore, it would be very helpful to show the efficiency of the transfer function in your peat cores. There are standard statistic indexes (the chi square distance of fossil testate amoeba assemblage to the closest modern analogue from transfer function training set; goodness-of-fit statistic; the number of rare taxa and the number of absent taxa) indicating to what extent the fossil testate amoeba complexes in the cores correspond to the testate amoeba complexes embedded in the transfer function. We should avoid the situations when the half of taxa from fossil testate amoeba assemblages are absent in the taxon list of the transfer function and really do not contribute to the water depth reconstruction R: The study of Qui et al. (2021) was not available at the time of running our quantitative DTW reconstructions and assembling our datasets for this paper. However, the pan European transfer function (TF) is not limited to European sites but includes peatlands from Siberia and the sites having a large PH variation. Although, Qui et al. (2021) states that the Asian TF performs similarly to other large-scale TF, a grouping of some taxa were less representative for our records than in Amesbury et al (2016). A full comparison of the Siberian profiles using both training sets, representative diagrams for each site, and climate-based reconstruction are in prep for publication Diaconu et al. In the revision manuscript, we will therefore retain our reconstructions but add a few sentences on the numerical performance of transfer function.
Lines 328-329 -you write "Warm summer temperatures likely enchanced evapotranspiration and consequently lowered peatland water levels, leading to drier surface conditions". You explain this for the period of "a temperature and moisture optimum between 6 and 4.5 ka BP (Groisman et al. 2012)" in Siberia. Although, if we look at the pollen diagrams from the studied mires (especially at Rybnaya mire Fig. S4a), we can see the increase of conifer pollen (Pinus sylvestris, P. sibirica) at this period and the decrease of Betula pollen. We can consider that conifers spread when precipitation exceeds evaporation (a prerequisite for the existence of the taiga). How can you explain this discrepancy when likely conifer trees indicate increased moisture (precipitation exceeds evaporation), but testate amoeba based DWT in mire indicate low levels (evaporation exceeds precipitation)? In general, I can conclude it looks very strange that reconstructed water levels in your peat cores are not coincided directly to the climate changes, which you take from the monograph by Groisman et al. (2012) for Siberia. I think this is an additional argument that the reconstructed DWT values from the studied peat cores do not reflect hydroclimate changes of the study area.
R: The temporal succession in the main tree taxa at our sites is in good agreement with other pollen records from western Siberia. There is also a good agreement between our and the few other charcoal records from the region in indicating a high fire activity between 8-4.5/ 5 ka. Wildfires occur predominantly during the growing season (sprint to autumn) but are most severe in summer associated with dry soils or peatland conditions. Peatland conditions must have been dry at this time (8-4.5 ka), to allow such high fires to occur, and lower local moisture conditions are also what TA show. In revising this paper, we will look more into agreements/disagreements between vegetation composition, TA hydrological conditions, and regional Siberia climatic conditions.