In the manuscript entitled “Representation of a two-way coupled irrigation system in the Common Land Model”, the authors implemented several new features regarding irrigation water applications and withdrawals in a land system model and evaluated the model with new features based on observation-based datasets. Although these new features are mainly adopted from other models, it is impressive that the authors implemented them in a different land model. I congratulate the authors for this huge amount of work and believe that it is qualified to be published on EGU HESS. However, there is still some room for improvement in the manuscript, which are listed below.

Major Comments

1. Materials and Methods:
   • Before introducing the crop module, it would be better to first describe the sub-grid cell level structure of the model CoLM. How are the calculations conducted over different land-use tiles? How are rainfed and irrigated cropland differently treated? Etc.
2. Results:
   • Evaluation:
     • The authors adopted the crop model used in the Community Land Model version 5 (CLM5) and activated the crop model, which means that the crop phenology is simulated. I think it is important to also compare the simulated and observed LAI at some single points to show how the model reproduces the phenology of different crop types.
     • I do not think comparing the simulated and observed energy fluxes averaged over all irrigated regions in the USA is enough. I would compare them over single grid cells with intense irrigation, or at least, average across the irrigated regions in different climate zones.
     • Similarly, I would do some evaluations of streamflow over some bigger catchments (which can consist of several sub-catchments that the authors used) with intense water withdrawal. In the supplementary Figure S7, it would be nice to also plot the boundaries of catchments. A map showing the difference between simulated and observed streamflow will be better than the bar plots in Figure 8.
     • For the crop yield, I would also like to see a map showing the difference between simulated and observed crop yield in administrative regions, rather than the bar plot in Figure 9.
     • It is always recommended to add maps showing difference between simulations and observations in Figure 3 and 4.
   • It would be interesting to add some results showing the new features of CoLM irrigation module. For example, showing the simulated water withdrawal from different sources, showing simulated terrestrial water storage, showing how lakes or reservoirs change with irrigation water withdrawal (if they do), etc. This will help readers to know what can they use the model to do.
3. Discussion

3.1 I have to admit that I am a bit disappointed that all the evaluation work was done solely in the USA. I understand that in the USA there are generally more abundant observations, but it should at least be discussed in the discussion that some efforts are needed if other researchers want to use this model for studies over other regions.

Minor comments

L193 What is this crop growth stage? When crop LAI > 0?

L205-207 Is this coefficient identical in all grid cells? Or can we assign different values in different regions?

L221-222 A table showing the main difference among them would be helpful.

Figure 6 I would change the tick of y-axis because it is not clear to see the absolute number of irrigation water withdrawal.