Interactive comment on “Improved ELMv1-ECA Simulations of Zero-Curtain Periods and Cold-season CH4 and CO2 Emissions at Alaskan Arctic Tundra Sites” by Jing Tao et al

Line 16-17 "simulated cold-season emissions at three tundra sites were improved by 84% and 81%" it is not clear what metric the 84% and 81% refer to, is this the mean absolute error? Please specify. Line 17-19 "...zero-curtain period in Arctic tundra, accounted for more than 50% of the total emissions" This statement is slightly misleading. This is the case in the model, but the study showed that this part was overestimated compared to the observations. I would therefore add something like "in the model, compared with around 45% (30-60%) in the observations" Introduction Lines 60-62 " However, current land models tend to significantly underestimate soil temperature during the cold season over permafrost regions (Dankers et al., 2011; Tao et al., 2017; Nicolsky et al., 2007; Yang et al., 2018b). One possible reason is that many land models fail to appropriately account for the latent heat released during soil water freezing" It is true that many land surface models did underestimate soil temperatures but, more recently, improved snow schemes have removed a lot of this problem. For example, your first reference Dankers et al (2011) has a followup study Burke et al (2013) which includes a multilayered snow scheme and removes the majority of the winter cold bias although a small cold bias remains. I highly recommend adding some discussion of snow here to make it clear that the latent heat is not the only (or even the biggest) factor. Most recent LSM’s (e.g. in CMIP6) do represent latent heat, if not particularly well, I suggest clarifying that to "One possible reason is that while