Applying Time-Expended Sampling to Ensemble Assimilation of Remote-Sensing Data for Short-Term Predictions of Thunderstorms

Round 1

Reviewer 1 Report

Please see attached pdf "Review_TES_WoFS.pdf"

Comments for author File: Comments.pdf

Author Response

See uploaded pdf file "Response-1".

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments:

 

This article applied the time-expanded sampling (TES) method to the convection-allowing ensemble-based Warn-on-Forecast System (WoFS) for four severe storm events to reduce the computational cost constrained by the real-time assimilation of remote sensing data from radars and geostationary satellite GOES-16. Forecasting experiments tripling a smaller ensemble size are conducted and compared with the forecasting experiment using the original ensemble size. Results show that forecasts using the TES method have the same capability and quality as the forecast using original ensemble size in predicting hourly rainfalls and probabilities of tornados and damaging winds, and the predictive capability and quality are basically not sensitive to the expanded time. This interesting finding suggests that TES is useful for cost-saving real-time applications of WoFS in assimilating remote sensing data and generating short-term severe weather forecasts. In my opinion, it could be potentially accepted after it undergoes some minor revisions.

 

Minor concerns:

1. Section 2: the time resolutions of the conventional, radar and satellite data, as well as the spatial resolution of satellite data, have not been given in the description. Additional description about EnKF should be given before the introduction of Ensemble spread and horizontal and vertical localization.

2. Why Figures 3 to 5 only give the statistics about one event? What about the results of other events?

3. A few abbreviations did not give the full name when they are first used. Some indicators also lack of formulas. For examples, CR, ETS, POD, SR, CSI, FB, etc. Please check throughout the manuscript and make corresponding modifications.

4. Line 214: “from observing system simulation experiments”should be“from observing system to simulation experiments”?

5. Line 263: RMSE is a conventional estimation index. Why use innovation to replace error to create a new name RMSI?

6. Tables 3 and 4 should be combined as one table.

7. Line 329: “Since ym and Hm(x) represent the same truth”, using“represent” here seems not proper as these two values are just approximations, please rewords it.

8. Are the results of figures 6 and 7 averaged for the four events? If so, please give additional statement in the captions of these figures.

9. Figures 10 and 14 are unclear, please revise them.

Author Response

See uploaded pdf file "Response-2".

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

I appreciate the author's detailed responses, they have adequately addressed  my concerns of the previous manuscript. I only have a few minor edits to add that should be addressed, after which I recommend this manuscript for acceptance. I do not need to see another revised version.

Minor Edits:

Line 124 (line 113 in original submission): Authors responded this was corrected to 5 km, but the revised version I see still has 15 km.

Table 1 : Please also change TB to BT to be consistent with main text.

Figure 6: Authors mentioned replotting this figure for color consistency of other figures, but the revision I have still shows the original figure (e.g. E36 is black, but should be red to be consistent with Fig 4, 7)

Line 473: RS should be SR.

Figs 8, 11: I think I now understand, all panels are valid at the same time, but columns 2 and 3 are initialized from analyses further back in time.  The column headers say "1-hour, 3-hour, 6-hour" which to me indicates different forecast hours from the same forecast IC. I think more can be done to avoid confusion to future readers: perhaps revise this to also say, e.g., the IC time on each header?  Or, someplace such as the caption or main text (i.e. lines 509-511) it should be clarified such as, "...for forecasts valid at 0200 UTC 29 April but initialized from different analyses 1-, 3-, and 6-hours prior." 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf