Algorithm for Improved QPE over Complex Terrain Using Cloud-to-Ground Lightning Occurrences

Round 1

Reviewer 1 Report

I made some minor comments and revisions in the enclosed PDF.

Other questions/concerns:

It is no doubt tricky using lightning as a means to quantify precipitation amounts. CG lightning does have a good relationship with precipitation but not always.

The methodology behind using the Kalman filter is sound but keep in mind the variety of dynamic/thermodynamic conditions that affect deep convective development as well as aerosol content that can impact the cloud microphysics. All these can affect the electrification of storms which may give your methodology some issues.

Why was lightning polarity not addressed in this study?

Comments for author File: Comments.pdf

Author Response

Answers to Reviewer 1

It is no doubt tricky using lightning as a means to quantify precipitation amounts. CG lightning does have a good relationship with precipitation but not always. 

·      Thank you for the question. As the authors of this paper and other authors (Petersen and Rutledge (1998) had mentioned, the relationship depends on many factors like meteorological conditions, geographical location, meteorological system, among others. There is not unique mathematical relationship between convective precipitation and lightning but even with this non-unique relationship these are two phenomena that occurs most of the time in the same location at the same time, and with the proper calibration the authors consider that can be useful in the estimation of precipitation, specially where other methods fail.

The methodology behind using the Kalman filter is sound but keep in mind the variety of dynamic/thermodynamic conditions that affect deep convective development as well as aerosol content that can impact the cloud microphysics. All these can affect the electrification of storms which may give your methodology some issues. 

·      Thanks you for asking this.  As we mentioned in the previous question, lightning-precipitation relationship (LPR) depends on many variables, and this makes a non-unique mathematical relationship. We agree with the reviewer about that the cloud microphysics can affect LPR but the purpose of the research is not explaining the physics behind LPR, the purpose of this paper in the estimation of precipitation, the Kalman filter is a good option because it changes the model precisely when LPR is changing and it does not still fixed while the the relationship is changing like the other studies do.

Why was lightning polarity not addressed in this study? 

·      Very interesting question, thank you for asking it. No more than 3% of the CG for the analyzed storms were found to be positive for our space domain and our studied period, hence we only  consider the total number of CG flashes, regardless of the polarity.

NOTE. All the speeling-typos-writting corrections sent by the reviewer were fixed as it can be seen in the paper´s new version. 

Reviewer 2 Report

p { margin-bottom: 0.1in; line-height: 120%; }

Rain is indeed one of the most important climatic parameters that we need to measure accurately, nevertheless due to its high spatio-temporal variability it is among the most challenging ones for monitoring. Thus a methodology that estimates rain amounts based on lightning measurements is desirable.

This paper is part of a series and presents the algorithm used in the other papers. It is important and useful as it will allow other researchers to use this method.

One major issue that needs clarification is how does the algorithm translate the link between lightning and precipitation from an individual storm scale into a seasonal scale?  As stated in section 1.1: “Research findings show that the lightning-precipitation relationship varies over a wide range of scales; temporal (minutes to seasonal) and spatial (storm scale, e.g., [34]) to complete regions (about 104 km2; e.g. [33]). “ So regarding this statement it should be explained how the algorithm solves this basic issue.

Another issue is to what type of storms is this algorithm relevant? Please explain what type of adaptations should be done for using this algorithm to other types of rain systems in other locations around the globe?

Author Response

Answers to Reviewer 2

Rain is indeed one of the most important climatic parameters that we need to measure accurately, nevertheless due to its high spatio-temporal variability it is among the most challenging ones for monitoring. Thus a methodology that estimates rain amounts based on lightning measurements is desirable.

This paper is part of a series and presents the algorithm used in the other papers. It is important and useful as it will allow other researchers to use this method.

·       We thanks the reviewer these comments, we also consider that this methodology can contribute to better precipitation estimation specially where other techniques present problems and thus, this can be a good complement to the others.

One major issue that needs clarification is how does the algorithm translate the link between lightning and precipitation from an individual storm scale into a seasonal scale?  As stated in section 1.1: “Research findings show that the lightning-precipitation relationship varies over a wide range of scales; temporal (minutes to seasonal) and spatial (storm scale, e.g., [34]) to complete regions (about 104 km2; e.g. [33]). “ So regarding this statement it should be explained how the algorithm solves this basic issue.

·      Very good question, thanks. Our algorithm operates in the other way around the reviewer is thinking (it goes from seasonal scale to 5 minutes time scale). We initiate the Kalman with the STI model  (seasonal), which in general is the space, and time climatological average. So, we consider this mode is the best to initiate the Kalman filter, the Kalman filter for sure, tracks better the time changes that occur at every storm, but it can be happen that some storm behaves as the average so, for this storm the STI will be a very good estimator. However, if the weather system that produced lightning activity, moves away from our domain, or if it simply dissipates, there will be less lightning data available to improve precipitation estimation. This is not an intrinsic weakness of the employed system; it is merely the way a Kalman filter (or any other sequential data assimilation methodology) works. If there are less data available for assimilation/optimal estimation, their impact will be less significant, and the estimation of a quantity will be weighted more towards the employed model itself. In  our paper published in the Journal of Applied Meteorology and Climatology we are proposing as a future work a storm tracking methodology to address this issue, so then, this work is on progress and it will be published in a future contribution.

Another issue is to what type of storms is this algorithm relevant? Please explain what type of adaptations should be done for using this algorithm to other types of rain systems in other locations around the globe?

·      Interesting question, thanks for asking. Regarding the type of rain systems, as long we have cold convective storms that produce lightning, this method can be applicable to any system with these features.