This is my first review of the manuscript, "How many models do we need to simulate hydrologic processes across large geographical domains?" I appreciate the authors’ work, which is both relevant and holds significant technical implications for the hydrological community. Sampling uncertainty is a critical issue, well-known to some but often overlooked by many. This study effectively highlights its importance within the context of large-sample hydrology, which, thanks to the widespread availability of the CAMELS dataset, is rapidly gaining traction.

The introduction is technically strong, well-written, and cites relevant literature. However, I suggest that the authors define “model” within the paper as a specific model configuration rather than a “modeling framework.” Clarifying this distinction early on would prevent confusion, ensuring readers understand that the paper focuses on specific model structures rather than broader frameworks.

The methods section, however, falls below the expected standard for this type of paper. I found it challenging to follow the results due to the methods being insufficiently described. Some concepts became clear only later, in the limitations paragraphs of the discussion section. To enhance clarity, I recommend adding a schematic of the procedure and clearly defining key terms like “model equivalent” and “best model,” which only become fully understood in the results section. The methods currently feel hastily written; a clearer presentation would significantly enhance the paper’s accessibility and relevance (I have included specific comments in the annotated PDF). Additionally, the problem statement is vague in places (notably in the initial statement), and it would be beneficial to restate the specific answers to these questions in the conclusion, effectively closing the problem statement.

The results are sound, and the discussion is well-articulated and engaging.

Based on these points, I recommend major revisions prior to publication.

See attachment!