This paper presents a novel algorithm for deriving below-ground urban stormwater networks using graph theory concepts. The paper is well written, well-articulated, and presents application to both urban hydrologic modeling and broad Earth system modeling. Although the manuscript is heavy on the method description, the applicability of the algorithm is clearly illustrated with 4 case studies. I think some shifting of paragraphs are needed to make the paper flow better. I also have some minor comments for the authors to consider.

1. It might be helpful to explain what “edge” and “node” mean at the beginning. It was clear later that edge means pipe and node means users but providing some contexts at the beginning would be helpful.
2. On page 4, at the beginning, I was confused about line 92-94. It seems counter-intuitive that higher weight results lower weighted BC value. Also, “a higher weight suggests a larger resistance to water flow and thus a lower flow rate.” This part also confuses me. If the pipe is made out of the same materials, how will a higher weight lead to larger resistance to water flow? Is it because the length of the pipe is longer? If that’s the case, then the pipe should be included because of its importance, right? I think the weighting process needs to be better explained up front to avoid confusion.
3. On page 8, bullet point 8, “those components that are unreachable after converting the network to an undirected graph by ignoring edge directions”, did you check if these pipes are actually not important edges? How rigorous is this approach?
4. I got a lot of questions when I saw Figure 5. For example, I wonder how pipes sizes are assigned based on BC and permissible min/max diameter. I found out that these were explained later in the manuscript. I think some indications directing the readers to the section where these are explained would be helpful.
5. On page 11, line 204, “assigning two lanes to those road types not listed in Table 2”, why?
6. On page 12, line 216, maximum discharge for a pipe when depth of water is flowing at 94% of diameter, not full!
7. In table 3, what does the LULC number mean?
8. On page 20, in table 6, the last column appeared for the first time without giving any context. It was explained later in the manuscript, but some explanations are needed when it first appeared.
9. On page 25, I don’t see any drainage pipes captured for the center of the city. Why is that the case?
10. The first sentence in Introduction does not flow well. Please revise.
11. Is Figure 1 an original creation or is it obtained from other sources? Please ensure that IP is not infringed.

This article addresses a highly needing yet challenging problem, deriving the topology of urban drainage networks from land surface data. A novel algorithm was developed and when applying to four various urban areas the accuracy (60-75%) is acceptable, especially given the complexity of the problem and uncertainties of the input data.

Specific comments are as follows:

1. The term "Below-ground Urban Stormwater Networks (BUSNs)" seems created by authors? Why not more commonly used term, such as "Urban Drainage Networks"?

2. Although not explicitly said, Figure 1 and line #23 seem indicating that authors focused on separate sewer systems (i.e., not combined sewer systems) and only stormwater drainage networks (i.e., not sewer networks)? Noting there are hundreds of cities in the US that have combined sewer systems, how well would this algorithm apply to those systems?

3. Validation was performed using a metric for coverage as the goal seems to be deriving the "topology". I'm curious if authors considered and compared slope and size of pipes? How would slope and size be implemented in large-scale urban hydrologic modeling?

4. Line 15: "urban population will grow from half to more than two-thirds of the total population by 2050." I'd suggest to delete "from half", or add "from half by 2008".

5. Line 24: "most urban modules in existing hydrological models..." provide references and/or give examples.

6. Line 159: "60% of a pipe length from the real BUSN is within this buffer zone, the pipe is considered “covered”." Did authors consider other values as the criteria? I'm curious how sensitive this criteria would be.

7. Line 381 vs. line 9: 59-76% vs. 60-75%. Which one is correct?

Comments

The manuscript entitled “An Algorithm for Deriving the Topology of Below-ground Urban Stormwater Networks” proposes a novel algorithm for estimating Below-ground Urban Stormwater Networks (BUSNs) from existing data based on the Graph theory concepts. The paper is interesting. However, the manuscript has some shortcomings which need to be improved prior to its publication. The recommendation is that the article needs Major Revisions before it can be considered for publication.

The following suggestions must be revised:

1. The abstract should be carefully rewritten as English expression needs improving and the structure is not as clear as the main part of the paper. The novel algorithm needs more explanation.
2. Now the approximate computation method of drainage capacity for urban flood modeling is a common method in the area where the BUSN data are sparse, this should be mentioned in the introduction section.
3. There are many drainage catchments in urban city, and the drainage pipe network is generally laid out according to the catchments. How to consider this in the algorithm?
4. The article only describes the pipes without mentioning the rainwater nodes and inlets, which also play a great role in the urban flooding process.
5. Validation section is weakly written. It is verified by the “covered” of the distribution of the pipe network, which is relatively rough, and there is no comparison of key parameters such as pipe diameter, slope, and flow direction.
6. The author should check the whole manuscript carefully, there are some errors in the interpretation of the diagrams.