Kovacek and Weijs have created a comprehensive database for ungauged sub-basins in British Columbia, utilizing various other databases to aid in different hydrological purposes. Overall, the article is very interesting, but there are inconsistencies that make it difficult to follow in certain places. Some minor changes are required before publishing this paper.

Major comments:

• A definition of the basin considered in this study is needed. Basin is a term that is interchangeable with catchment and watershed, but it typically refers to the entire drainage area of a river. In this article, ‘basin’ represents the local watershed of each river-reach. The term ‘sub-catchment’ or ‘sub-basin’ is more appropriate here.
• To understand the process more easily, a flowchart showing different steps of BCUB database development in the methodology section would be helpful.
• The reason for using the HydroBASINS watersheds (level-5 and 6) to subdivide the study region is understandable. However, the underlaying hydrography data in the HydorBASINS and BCUB databases are different. So, there is a chance of missing a part of the sub-catchments located near the regional boundary in BCUB database. For example, a part of the sub-catchment of the PCR region, located near the boundary between PCR and FRA, may overshoot to the FRA region due to hydrographic data inconsistencies. How was this issue addressed during the development of this database?

Specific comments:

• It is sometimes difficult to follow the article due to inconsistencies in the statements. For example, the line 76 in the motivation section, “The accuracy of stream network delineation improves with increasing DEM resolution.” The transition from the previous lines to this one is not smooth.
• Another example of inconsistency is in line 134, where the delineation of the stream network is discussed after the description of the pour point selection process from the stream network. It would be more appropriate to discuss the stream network delineation process before selecting pour points.
• Line 103: Please provide the minimum drainage area threshold used to delineate the stream network from USGS 3DEP.
• Figure 7: This is a nice figure to show the impact of using DEM with different resolutions. The plot with colored density would be more helpful to understand the figure.
• When using QGIS version 3.28 to open the dataset, it displays the pour point location instead of the sub-basin polygon. Has the delineated sub-basin geometry been excluded from the database?

The paper introduced a new large sample dataset BCUB for British Columbia Canada. This data includes more than 1 million basins over the region, which is far more than any of the current available dataset. It for sure will support water resources research and hydrological modeling as presented by the authors in their abstract. I support to publish it with minor revisions.

It would be nice if authors can convince us the necessity to have 1.2 million basins (sub-basins or sub-catchment), it’s very difficult to find the one you are interested and very difficult to use all of them in a regional scale. If possible, authors can introduce some specific implementations of these large-sample basins. It is also very difficult to find real observation of river discharge to support further analysis.

Please check the unit of precipitation, e.g., 2028mm/day for gauge 1269663 must be wrong?

What is the difference in the number of detected sub-basins when using the two spatial resolution?