This study presents a ML-based multi-satellite fusion framework, successfully generating a high-resolution, all-weather TPW dataset for the Tibetan Plateau, which represents a significant contribution to the field. However, several methodological limitations and potential avenues for improvement warrant discussion.

1）A primary concern is the framework's heavy reliance on the Himawari-8/9 (H8/9) clear-sky TPW product as the foundational reference for both bias correction and spatial downscaling. While this strategy effectively mitigates inter-sensor biases, it inherently transfers the uncertainties and potential systematic errors of the H8/9 retrievals into the final fused product. Furthermore, the adaptive correction method for cloudy conditions, which extrapolates biases from clear-cloudy boundaries, may see its efficacy diminish in regions of extensive, persistent cloud cover where valid H8/9 reference pixels are distant. Future work could enhance robustness by incorporating cloud physical properties (e.g., from microwave sounders) or assimilating short-term numerical weather prediction fields to guide corrections in areas with minimal clear-sky information. 

2）It is not clear why PWV data from polar imagers such as MODIS or MERSI are not included in the analysis. These valuable data should provide a critical reference or add new information to the fused  product. 

3) Regarding validation, while the use of 44 GNSS stations is valuable, their sparse and uneven distribution, particularly over western and northern TP, limits the ability to comprehensively assess the product's accuracy across all topographic and meteorological regimes. The validation might not fully capture errors in the most data-scarce regions. Supplementing the evaluation with data from intensive field campaigns (for  example，Scientific Expedition on the Tibetan Plateau）, additional independent satellite retrievals, or a cross-validation study during periods with varied cloud cover would strengthen the confidence in the product's performance.