In this manuscript, the authors present the integration of a new version of the Noah-MP land surface model (Noah-MPv5.0) into the Land Information System (LIS). They investigate the performance of the new LIS/Noah-MPv5.0 system against the previous version of the system (LIS/Noah-MPv4.0.1), focusing on a number of hydrological states and fluxes. They find slightly degraded performance of the new system for most of the variables investigated.

I think that this manuscript will eventually provide a useful contribution to the community and constitute a valuable resource for user of LIS. However, I would suggest a few changes before the manuscript is accepted for publication.

Major comments:

In its current form, the manuscript demonstrates the differences between LIS/Noah-MPv5.0 and LIS/Noah-MPv4.0.1 in detail, but in my opinion falls short when it comes to the interpretation of these differences. From the perspective of LIS users, I think that an understanding of the model changes that lead to the observed differences would be valuable, especially in cases where the model performance is degraded. The ‘Discussion’ section could be a good place to include more of an interpretation, but as it stands, that section focuses mostly on future model development without having established that the suggested future changes would address the cause of the differences observed currently.

For example, the authors note the somewhat contradictory increase in surface soil moisture and simultaneous decrease in latent heat flux. I would have like to see a more in-depth discussion for the cause of this behavior. Are there changes to the model processes or parameters that would explain this? Given that the radiative forcing (and wind forcing) is the same in both experiments, would it make sense to check whether there has been a change in the soil temperature? Or look at a map of LAI to investigate whether this is caused by the bug fix to the vegetation fraction scaling that is mentioned.

Given the above discrepancy between the soil moisture and latent heat changes, I would also suggest including an additional ET product in the evaluation. While the GLEAM product is certainly a good choice, it is somewhat dependent on the soil moisture assumptions that it makes. So, I would suggest including a soil-moisture independent product like ALEXI to further investigate the conflicting SM and LH responses.

It is a bit unclear how the variables that are evaluated were chosen. From a LIS user perspective, I am wondering whether it would be useful to include an additional figure that would show the results from a more comprehensive and standardized benchmarking framework, like ILAMB, as this would provide a high-level overview of the changes across additional model variables.

Minor comments:

Generally, a lot of the Figures are small, making it hard to see details. Since several panels often share a color bar, maybe you can just have the two color bars at the side, which might allow you to increase the subpanel size.

l.24: Here and elsewhere in the paper I would suggest avoiding formulations like ‘negative bias’ and instead say ‘under-/overestimates soil moisture’, as this is more intuitive.

l. 104 – 105: If a ‘scorecard’ type evaluation is the intention, then I don’t understand why the authors opted not to use one of the existing comprehensive benchmarking tools, like ILAMB (see major comment) or include a scorecard type figure in the paper. Also, the quotation marks are mismatched.

l. 147: “…control the soil process timestep.”

l. 159: “…of benchmark simulations with LIS coupled with Noah-MP”. Rephrasing to avoid ‘coupled simulations’ since it implies coupled to a GCM.

l. 174 – 175: Was STATSGO used over the US and FAO elsewhere?

l. 218 – 219: There are several areas where the precipitation bias does not align with the soil moisture bias (for example Northern Canada or Southern Brazil). What is the suspected cause for the soil moisture differences there?

Figure 2: Here and in all other bias figures, I think it would be helpful to include the mean absolute bias values as well, to get a sense for the model changes without the impact of compensating errors.

l. 314 – 315: “…uses snowpack physics consistent with other land snowpacks…”

l. 329: “…despite the overestimation of soil moisture…”

l. 418 – 419: What do the authors think is the reason that these new parameters are only effective at reducing the bias in some regions?

This paper presents a detailed evaluation of an upgrade to the NASA LIS. The paper is well written and should be useful to LIS users.

When summarizing the results, I suggest adding a table, in the form of a scorecard (with a stratification per variable, season and domains), that summarizes the magnitude and significance of the changes in the results obtained for the two LIS versions that are evaluated in the paper.

I would also suggest that more details be provided on the use of the Github submodule mechanism to streamline synchronization, or at least a reference on how this works and helps keeping versions in sync.

In many figures, grids with statistically significant differences are shown with gray dots. However, the technique used to assess whether the differences are statistically significant is not explained in the text. Please provide more details on the method used.

Finally, I have read the comments made by Anonymous Referee #1 and agree that the discrepancy between model differences for soil moisture and LH needs to be investigated further. For the CONUS domain, I suggest looking at each season separately, as well as separating the evaporation and transpiration components of evapotranspiration if this is possible in LIS. Even if a comparison to observations of the two components is not possible, it could provide useful information as to the origin of the differences.