Review: "Optimizing Precipitation Parameterizations in Regional Climate Model (RegCM5): A Case Study of the Upper Blue Nile Basin (UBNB)"

A set of 7 numerical experiments using Regional Climate Model version 5 (RegCM5) was conducted to understand the better precipitation parametrization to reproduce the monthly and daily precipitation over the upper blue Nile basin (UBNB). The numerical experiments combine different cumulus convection schemes (over the land the Kain-Fritsch, or Grell, or Emanuel, or Tiedke, or Emanuel  are used, combined with Emanuel over the sea) and large-scale precipitation schemes (SUBEX and NoTo). Simulations were driven by ERA5 reanalysis and PP7 precipitation data was used as observational reference. For the comparison between simulated and observed precipitation it is used some statistical indices. The main conclusion is that predominates a general overestimation of the simulated rainfall in most seasons of the year, with an acceptable  performance being obtained by the combination of Emanuel for cumulus convection and NoTo for large-scale precipitation schemes. The analysis of simulation performance in the region is a relevant topic with large potential to add new  knowledge related to models' physical parameterizations behavior in the tropics. However,  as highlighted in the major points, there are some statements/interpretations that need to be clarified/better founded before the acceptance of the manuscript. 

 Major points

1. The title does not correspond to the analysis presented in the manuscript. The authors did not conduct any “optimization” in the model since they used the default code and parameters.  The manuscript only presents a comparison of different schemes to represent the moist deep convection and large scale precipitation. One suggestion of title is: 

“Comparison of precipitation parametrizations in Regional Climate Model (RegCM5): A case study for Upper Blue Nile Basin (UBNB)”

1. In L226-240 (L287-289 of conclusions) the authors attributed the simulated precipitation errors to the difficulty of the simulations to reproduce important features of large scale circulation in the region. However, no figures/tables and their interpretations are presented to justify the claims which are only based on the previous literature. The authors need to improve this part of the manuscript by presenting/analyzing how the experiments simulate the mentioned circulation features. 
2. (L264-271) After identifying a considerable overestimation of rainfall by the simulations, the authors present several arguments that could have provided more realistic simulations. My question is: Why didn't the authors use these recommendations in their numerical experiments? In particular, the use of a hydrostatic dynamics at 10 km resolution, which is considered the interface (or limit) of validity for this approximation, is highly questionable. Why did not the authors use the MOLOCH that is non-hydrostatic and makes the RegCM5 code very fast?  The authors need to clearly justify their decision to use the hydrostatic approximation core.
3. Figure 4 and 5 are redundant since both are showing the same information of the biases (underestimation or overestimation)  of  the simulations in relation to the observed precipitation.  In addition, Figure 4 has different scales in each season  and it is not easy to compare them. In this way, it should be better to combine the top panels of Figure 4 with the relative biases in Figure 5 to have only one Figure showing the biases. 
4. The statements  in L19-20 do not correspond to what is presented in the manuscript. In the actual form, no sensitivity numerical tests (by changing the core from hydrostatic to non-hydrostatic or increasing/reducing the number of vertical levels) or the circulation pattern were presented in the manuscript. I am suggesting: 1)  to remove "Sensitivity analysis … precipitation outputs”; 2) to include in the manuscript an analysis of the circulation patterns (point 2) to justify the affirmation in the end of phrase. 
5. To improve understanding, the authors could reorganize the introduction. Suggestion:

Starting with L24-277 and following with L37-45. After that, L27-37 would be followed by L46-58. 

I noted that the main objective of the text is not clearly established at the end of the introduction. There are some indications of the objectives in L64-67. I suggest that the authors reorganize the objectives clearly at the end of introduction. 

Minor points

In many parts of the text (L31; L46; L142-150; L195; L265), the citations are not correct. For example: L108 - in (Giorgi et al, 2023b) should be in Giorgi et al. (2023b); L148 should be “by Holslag et al. (1990)”; L149 should be “by Zeng et al. (1998)” and many others in the text. Please, check all the text. 

L30-31 - should be “... interannual) controlled by the Global …. Abtew et al. (2019) found … ”

L37 - should be “ … rainfall since it is the main rain …"L39; 

L109-111 - should be “are driven by atmospheric variables and SST from ERA5 reanalysis data from ECMWF (Hersbach et al., 2020) with 0.25o x 0.25o of horizontal resolution for the period 2000-2009. For evaluation, observed … for the period 2001-2009” since SST was already defined in L31. 

L118 - should be “The domain has a 10 km horizontal … longitude (Fig.2 ), involving …””

L121 - The 18 vertical levels is a very small number for a 10km horizontal resolution. Why only 18 vertical levels?

L127 -  should be “... in the different numerical experiments …” since the authors used the code as it is, i.e., they do not change any parameter or physical parametrization. 

L139-140 - should be “Hence, a new set of simulations was conducted by using Nogherotto-Tompkins (NoTo; Nogherotto et al., 2016) microphysic scheme, which treats the mixed …. Over East Africa, Godoshava and Semazzi (2019) revealed …. In addition, Kalmár et al. (2021) …”

L150 - In many parts of the text, the authors used Emanuel_NoTo, Tiedke_Noto and Kain-Fritsch_NoTo.  This information should be  in Table 2. 

L157 - remove “calibration”; it is more common to use “relative bias” instead “percent bias”.

L158 - What is the interpretation of the RSR index? What is the range of acceptance of RSR? It was calculated considering a  time series or for the mean spatial pattern in UBNB? Please, clarify  it in the text and also include a reference for the index. 

L163 - remove “estimated”

L165 - remove “to check spatiotemporal distribution” since this information is already in the beginning of the phrase.

L180 - should be “ … for the period 2001-2009 since the year 2000 was considered as …”

L190 - What is the meaning of  “reduces the significance between CCs”? Please, clarify in the text. 

L194 - should be “ are presented in Fig. 3b”.

L183 - should be “ … UBNB the Fig. 3 and Table 3 show the evaluation of the simulated precipitation for the seven numerical experiments””

L202 - Suggestion “The performance of the experiments S5, S6 and S7, which  use the NoTo microphysics scheme, is evaluated by analyzing the spatial pattern  …” 

L206 - should be “The model also overestimated rainfall in eastern …”

L218 - should be “need an improvement”

L221 - should be “exhibit a low correlation”

L242-243  - should be “ characteristics simulated in the experiments S5, S6 and S7 and observed by PP7 during the FMAM, JJAS and ONDJ (Fig. 7). “

L245 - should be “... (S6) slightly better represents the observed distribution of the daily precipitation during the FMAM (Fig. 7a)”

L278 - I think that should be better “non-convective precipitation and reduces the biases in the simulated total precipitation.”

L296 - What is the meaning of “refine the physical parametrizations”?

Figures: 1) The labels in most of the figures need to be improved since they are very small, which makes it difficult for readers to interpret the figures. 2) Please, highlight the UBNB basin (shown in Fig. 1) in panel (a) of Figure 2.

This manuscript investigates the performance of the RegCM5 model with a hydrostatic dynamical core in simulating precipitation over the complex terrain of the Upper Blue Nile Basin (UBNB). The study leverages high-resolution ERA5 reanalysis data to evaluate various convective and microphysical parameterization schemes, ultimately identifying the Emanuel convection scheme coupled with the Nogherotto-Tompkins (NoTo) microphysics scheme as the optimal configuration. The following points should be addressed to improve the clarity, robustness, and overall quality of the manuscript:

Major points:

1. The statement that even the best parameterization shows a 66% bias is significant. This level of error raises concerns about the model's credibility for the region. The authors should provide a more comprehensive discussion on how to interpret this result. Specifically, they should:  a) Clearly state what level of bias is considered acceptable for the intended applications of the model in this study. b) Discuss whether the model, despite the bias, successfully captures the dominant spatial and temporal patterns of climate variability, which can sometimes be more important than the absolute values for certain applications.
2. The methodology does not mention a threshold for defining a "wet day" (e.g., >0.1 mm/day). Please clarify if all precipitation values, including trace amounts, were used in the analysis. The low frequency of zero precipitation values shown in Figure 7 suggests the model (and possibly the observations) rarely simulates completely dry days. Confirming the wet-day threshold is important for interpreting intensity and frequency results.
3. The manuscript compares different parameterizations but lacks a detailed discussion on the physical reasons why the optimal scheme (Emanuel + NoTo) performed best. The authors should elaborate on the key physical differences between the schemes (e.g., in their triggering of convection, closure assumptions, or ice-phase processes) and provide a hypothesis for why this specific combination is more suitable for representing the dominant atmospheric processes in the UBNB.

Minor points:

1. The 10km resolution may be insufficient to capture local processes in the UBNB's complex terrain. Justification for this choice (e.g., computational limits or sensitivity tests) should be provided, along with discussion on how resolution impacts simulation accuracy.
2. Figures 1 and 2 should be merged to clearly show the UBNB within the larger model domain, improving readability.
3. L158 - Although monthly evaluation is common, adding daily-scale metrics would strengthen the assessment of rainfall distribution and intensity.
4. The spatial discontinuities in simulated precipitation (Fig. 4), especially in the northeast, require physical explanation—g., topography representation, land-atmosphere feedbacks, or convective processes.
5. The selection of Emanuel+NoTo needs clarification, as Kain-Fritsch+NoTo performed well in explained variance for key seasons (Fig. 6). The trade-offs and reasons for the final choice should be explicitly explained.