In this manuscript, Lin et al., evaluate changes in terrestrial organic matter input into the Arctic ocean over the last 18,000 years. The chosen study sites drain vast quantities of permafrost, such that enhanced terrestrial OC input might indicate enhanced permafrost thaw. The authors identify several pulses of terrestrial OC burial, some of which may correlate to known climatic events (e.g, meltwater pulse 1a).  There is remarkably low terrOC input during the Holocene but lots of variability in terrOC delivery between 10 to 16 kyr. Each peak shows different compositional characteristics, suggesting distinct terrOM sources derived from different mechanisms. However, it would be interesting to try and unravel this further using your existing data (see comments below). Overall, the manuscript is well written, the figures are clear and the captions offer an appropriate level of detail. A few comments are included below that would be worth exploring further...

Comment 1: it is hard to tell whether the pulse in terrestrial OC is due to permafrost thaw or enhanced delivery of plant/soil OC. This could be assessed by measuring 14C values in different lipids, but I realize is beyond the scope of this paper. However, there is existing literature that could be helpful - for example, Feng et al. (2013; PNAS) explored how the 14C signature of different compound classes varied across the Pan-Arctic region and lignin phenols appear to mostly derive from recent carbon, whereas n-alkanes are derived from older carbon sources (e.g,. permafrost). Although you don’t have 14C measurements, it could imply distinct sources for lignin vs n-alkanes in your samples. 

...but another way to tackle this could be to look at other n-alkane indices such as the carbon preference index. This is frequently used to assess changes in OC maturity and may provide further insights into the type of terrOC that is being delivered into the marine realm, especially during the three pulses. If it was older OC, it may yield slightly lower CPI values. It may also tell you whether you are reworking old petrogenic OC into the marine realm too (which could be a CO2 source if it was oxidised; see work by Sparkes 2016 the Cryosphere, but also work by Bob Hilton/Valier Galy etc)

Comment 2: Export vs preservation

It is important to confirm that the increase in leaf wax mass accumulation rates is not due to enhanced preservation but is reflecting enhanced terrOC export. This could be explored by calculating MARs of of short-chain alkanes (algal-derived), mid-chain alkanes (moss or macrophyte derived), and long-chain alkanes (vascular plant) n-alkanes during the three “pulses”. If all three increase, it might suggest that OC preservation is the main driver. But if only the mid- and long-chain alkane MARs increase, it would imply enhanced terrOC input. 

Comment 3: a note on pAq

The authors use the pAq index (the ratio between mid vs long chain n-alkanes) to infer changes in wetland input, but I would note that’s its more complex than this. For example, both Sphagnum moss and aquatic macrophytes are characterized by similar lipid distributions (Baas et al., 2000; Ficken et al., 2000), so without knowledge of the local vegetation, its challenging to say whether the pAq ratio is due to changes in moss input or macrophyte input. Perhaps you could narrow this down by drawing upon predicted vegetation patterns during the Holocene/LGM etc.  There is also great work by Jorien Vonk, Bart van Dongen, Orjan Gustafsson etc in similar pan-arctic regions that might be useful. For example, Vonk 2009 paper in Org. Geochem shows that “…the C₂₅/(C₂₅ + C₂₉) n-alkane ratio is most suitable for terrestrial OM source apportionment in these coastal regions”. This might be worth exploring alongside the pAq (although I suspect you will get similar results!).

Minor comments:

L63: v/v – the v’s should be italicised – change throughout

L205: correct that its used for macrophytes – but also for sphagnum mosses.

L252: is this statistically significant?

L365: dialkyl

L366: and also found in peats – and lakes - and marine sediments! 

L412: subscript CO2

Summary

This paper presents organic geochemical proxies used to distinguish the source of organic matter (OM) in two 17-18 kyr records from the Laptev sea. The data is integrated with pan-Arctic records in an attempt to investigate local versus global drivers of enhanced OM flux during deglaciation. It is a very nice study and paper that fits well within the scope of Climate of the Past. However, I also think there is some room for improvement in he presentation and discussion of the data. Specifically, 1) a clearer presentation of sedimentation rate changes in the studied cores (and how they impact the mass accumulation rate estimates) is important, 2) a more in-depth discussion about the timing and link to OM flux/source changes associated with previously identified meltwater events in the Laptev Sea between MWP-1A and MWP1B. These are discussed further in the points below:

Specific comments

• Line 25: “Additional terrOM MAR peaks coincided with periods of enhanced inland warming, prolonged ice-free conditions, and freshwater flooding, which varied between regions.”. I wonder if this sentence could be more specific, for instance, be specific about what periods evidence for coastal erosion in response to sea-level rise are identified, and at what periods are inland warming and freshwater flooding seen? Also - maybe the last sentence could be re-written so that the expression ‘regional terrOM fluxes’ only occurs once.
• Could the authors add a bit more information on how a dR value of -95 +/- 65 years for the Marine20 calibration curve was derived from Bauch et al, 2001? It is good to describe the conversion of old dR values when making them compatible with the Marine20 calibration curve, it is good bookkeeping.
• Result, Section 4.1 chronology: “The mass accumulation rates (MARs) of all biomarkers were largely affected by the pronounced sedimentation rate changes and thus, showed similar temporal changes in all terrestrial biomarkers, including HMW n-alkanes, HMW fatty acids, and lignin phenols (Fig S3, contents of each biomarker in Fig S4). Fig 2a and Fig 5 i, j show the mass accumulation rate of HMW fatty acids as a representation.”. Mass accumulation rates play a very important role in the environmental interpretations presented in this paper. In a number of areas the author’s highlight how important the number of age-depth control points, and more generally changes in sedimentation rate, can impact these calculations. Furthermore, one of the most general questions about how offshore sedimentary processes responded to regional and global climate changes, concerns how the mass accumulation of sediments (sedimentation rate) changed. If there is an influx of material from land, one would expect that there would be enhanced deposition of sediments offshore. One question not answered in the current paper is - do we see this? I think it would be great if the authors added a ‘sedimentation rate’, or ‘mass accumulation rate of sediments’ panel in Figures 2 and potentially 5. The age depth models in the supplementary material do not really show the changes in sedimentation rate through time (not in enough detail), and it would be nice to clearly see how this impacting the calculated MAR’s of other components (like TerrOM).
• Line 253: “remained rather constant despite the periods of peak MAR (Fig S5).” I think you should specify what MARs you are discussing.
• Line 375: “Before the Bering Strait opened at around 11 kyr BP (Jakobsson et al., 2017), the coastlines of the Beaufort Sea and the Chukchi Sea were connected, allowing the potential westward transport of terrOM from North America. Therefore, we consider the record before 11 kyr BP from the Chukchi Sea (4-PC1) as a representation of terrOM signal from the North American Arctic.”. I am not sure that I buy this argument. On one hand this paper is trying to disentangle regional from global climate drivers for TerrOM delivery to the Arctic, but then wants to use a record from the Herald Canyon in the Chukchi Sea as a proxy for deglacial processes operating across Arctic North America? Even when we look at the basic sedimentology, the deglacial records that have been published from the Canadian Beaufort Sea have a very high detrital carbonate content, which is not mirrored in the Chukchi Sea records. Maybe there is something I have misinterpreted, in which case it would be good to clarify what is meant in this sentence.
• Line 385: “Age-depth models for these records were recalibrated against the Marine20 calibration curve (Heaton et al., 2020) or a combination of Intcal20 (Reimer et al., 2020) and Marine20 curves, depending on the original studies to achieve congruent age control across all records. Reservoir ages were taken from the original publications.” Is this accurate? or were reservoir corrections taken from the original publications and updated to fit with the Marine20 calibration curve by . . . and then specify how this was done.
• Line 407-409:”The rapid global sea-level rise during meltwater pulse 1A (mwp-1A) was an important process in terrOM mobilization across the pan-Arctic region. TerrOM MAR peaks during this period are observed widely in records from the Eurasian Arctic and the Bering Sea”. The title of this section is ‘Pan-Arctic factor: sea-level rise’ but no mention is made here of the Canadian Arctic/Beaufort Sea. I think it is hard to argue this without data from the Canadian Arctic, and it does not seem like that data exists (i.e in Fig 5, the Beaufort Sea records do not extend that far back in time). I can imagine that there may be a difference in glaciated versus non-glaciated margins etc. I at least think that this needs to be discussed in the text, as I am not at all convinced that the Chukchi Sea record is representative of North America. A core from the Herald canyon cannot tell us about all the processes operating across the northern coast of Canada.
• Lines 431-435: “In the North American Arctic, terrOM MAR peaks appeared during the interval between mwp-1A and mwp-1B (Fig 5d, e). Inland warming in North American began at approximately 13.5 kyr BP, while the Eurasian Arctic remained cold (Brosius et al., 2021). This regional temperature discrepancy possibly explains the exclusive terrOM MAR peaks observed in the North American Arctic during the interval between mwp-1A and mwp-1B (Fig 5c, g)”. I think one of the most important observations that is not picked up in this paper is the link between TerrOM fluxes in the Beaufort and Chukchi seas and the d18O excursion reported by Spielhagen et al., 2005 in the outer Laptev Sea. All of these events appear to occur between MWP-1A and MWP-1B and there seems to be a coincidence in timing, even with some of the HMV Mar’s in the Laptev Sea (PS51/154) and Fram Strait. However, this is hardly discussed in the paper and I wonder if it deserves more attention (see next comment)
• Lines 465-470: “However, freshwater events were less likely to be the cause for terrOM MAR in the western Laptev Sea. While freshwater flooding events were recorded in an icedammed lake upstream of the Lena River (14.9 ± 2.0 kyr BP) and in a sediment record from the Laptev Sea (PS2458, at 12.7 ky BP) (Spielhagen et al., 2005; Margold et al., 2018), the timing of these events did not 470 correspond with any of the terrOM MAR peaks in cores PS51/154 and PS51/159. This temporal mismatch suggests that Siberian freshwater pulses had little impact on the increase in terrestrial biomarker MAR in the western Laptev Sea.”. The d18O peak in foraminfera from PS2458 (Fig 7 in Spielhagen et al 2005) does seem to overlap or is very close to some of the increased terrigenous biomarkers MAR’s shown in figure 5 of this paper. I think it would add a lot to recalibrate the ages of PS2458, and plot this data on one of the summary figures. It seems to be an extremely complementary dataset to the goals of this study (looking at processes impacting terrigenous OC mobilization to the Arctic). The current arguments that it is not correlated to any of the documented periods of enhanced TerrOM flux is not really supported by the current presentation of data. It may be true – but can it be shown more clearly?