Statistical-Observational Analysis of Skillful Oceanic Predictors of Heavy Daily Precipitation Events in the Sahel

Round 1

Reviewer 1 Report

Review on “statistical-observational analysis of skillful oceanic predictors of heavy precipitation events in the Sahel” by Diakhate et al.

 

The manuscript discussed the impacts of SSTs over the Nino3.4 region and the Mediterranean Sea on the Sahel rainfall. The study helps us to further understand how and where SSTs over the Nino3.4 and the Mediterranean Sea to affect the Sahel rainfall on interannual timescales. It can be consider to publish after Major revision according to the following comments.

 

Major Comments:

In Sections 3.1 and 3.2, the authors compared the hindcasts and observations. However, the authors should specifically discuss how they calculate these hindcasts. I also confused why the authors discuss the lag-correlation relationship between hindcasts and observations. When selecting the predictor(s), the author may first discuss the lag-MAC analysis. After that, hindcasts can be obtained based on associated lag-MAC analysis. The authors can be referred to some references, e.g., Wang et al. (2015). Additionally, the authors should unify the usage of “forecast” or “hindcast” in the manuscript.

Wang, B., B. Xiang, J. Li, P.J. Webster, M.N. Rajeevan, J. Liu, and K.-J. Ha, 2015: Rethinking Indian monsoon rainfall prediction in the context of recent global warming. Nat. Commun., 6, 7154.

 

How to validate the prediction skill when taking the SSTs over the Nino3.4 region and the Mediterranean Sea as the predictor? I suggested the author can taking 1981-2006 as the training period, whilst the period 2007-2016 is used to test/calibrate the prediction skill for the predictors according to the obtained relationship.

 

The authors discussed SSTs over the Nino3.4 region and the Mediterranean Sea are separately used to as predictor to forecast the Sahel rainfall. Is there any connection of SST between the Nino3.4 region and the Mediterranean Sea? How about the case when taking SSTs over the Nino3.4 region and the Mediterranean Sea as combined predictor to forecast the Sahel rainfall?

 

It is much better If the authors can provide associated physical explanation in detail.

 

Specific comments:

L52: Was the reference [33] formally published? If yes, the authors should cite the formally published version. If not, why? It seems that the S4CAST model used in this study is based on this reference.

 

L82-85: With respect to the definition of heavy/extreme rainfall event, have the authors considered the duration of each event? Does the inclusion of duration of event significantly affect the analyses in this study?

 

 L92-97: Have the authors applied the filter to the SSTs? The authors should comment on this.

 

L139: Figs.1bd

 

L144: I can’t find the Fig. S1.

 

L145-146: “Indeed, ……”. Please complete this sentence.

 

L153: 60S-60 --> 60°S-60°N. Please correct it in other places.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

This manuscript explores spatial/temporal predictive skill of extreme precipitation events in the Sahel using an SST (sea surface temperature) based statistical forecast framework. The authors claim that the Mediterranean SST is a key driver for the extreme precipitation in the Sahel region, whereas the response from the ENSO forcing is influential only over the southernmost Sahel. I find the topic of the study relevant for the decision makers and sub-seasonal to seasonal forecasting community. Following are my main concerns: 

It has been reported that the land-atmosphere interaction/feedback is quite strong in the Sahel of Africa. The leading statistical values in present analysis could both be effects of a different cause. There could be persistence in the variables forced remotely or locally. I am wondering, to which extent the current results, i.e., the magnitude of the predictive skills and changes, are affected by these missing information? Would the current interpretations be changed if the local feedbacks were adequately isolated in the present analysis?

L96-97: “Throughout the whole manuscript, the random phase test….” Could the authors explain why (and superiority over others) this method was used?

Fig. 4: It shows that the number of significant correlation coeff. increases in lag-3. Can the authors add the possible reason behind this increment in the lag-3 time? Is it due to the development of the ENSO event?

Figure: RMSE maps need to be improved, especially the color scales.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The authors have answered all my concerned questions which are correspondingly well resolved. No further revision is needed. Thus, I recommend to accept it.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Thank you for the revision. I have very minor comments:

Fig. 6b title: Please explain the term 'U(-1)'

The very last figure should be named as Fig. 8 (I guess?)

L358-360: "In this paper fundamental elements were provided to develop a prompt and accurate forecasting system of heavy and extreme daily precipitation events over West Africa, as shown in the analyzed case studies (Figs. 7-8)."  Please rephrase the statement (e.g., "prompt and accurate". It currently sounds very strong. There are caveats though).

 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf