In this manuscript, Wang et al compiled and post-processed half-hourly flux data from these 34 EC flux sites in the Heihe River Basin to create a continuous, homogenized time series dataset. This work filled the gaps in carbon flux data and auxiliary meteorological data of the Heihe carbon flux network, and generated a carbon flux dataset. Then the half-hourly NEE measurements were partitioned into GPP and Reco. The diurnal, seasonal, and inter-annual variabilities of carbon flux across diverse ecosystems in the Heihe River Basin were explored based on the gap-filled and partitioned dataset. This post-processed carbon flux dataset with a total of 34 EC sites in the Heihe River Basin is very valuable and important.

Overall, I find the paper compelling and fit for publication after minor revision. I only have a few comments as below:

1. ERA5-Land dataset is a reanalysis dataset. It would be great if some evaluation for variables from ERA5-land vs observations can be added.
2. If possible, comparing the MDS with RF results will be interesting.
3. The RF model was used for meteorological and carbon-flux data post-processing. The readers will be curious to know the construction and parameter selection of the RF model.
4. Please give some explanation why only Rg, Ta, and VPD are selected as input of the RF model to predict NEE.