Atlantic Multi-decadal Oscillation Modulates the Relationship Between El Ni&ntilde;o-Southern Oscillation and Fire Weather in Australia

Liu, Guanyu; Li, Jing; Ying, Tong

doi:https://doi.org/10.5194/acp-2022-858

Preprints

https://doi.org/10.5194/acp-2022-858

Preprints

29 Mar 2023

| 29 Mar 2023

Status: a revised version of this preprint is currently under review for the journal ACP.

Atlantic Multi-decadal Oscillation Modulates the Relationship Between El Niño-Southern Oscillation and Fire Weather in Australia

Guanyu Liu, Jing Li, and Tong Ying

Abstract. El Niño–Southern Oscillation (ENSO) is an important driver of fire weather in Australia. The correlation between ENSO and Australian fire weather has strengthened over the recent two decades. However, the causes for this change have not been well investigated. Here, using reanalysis datasets and numerical model simulations, we show that the Atlantic Multi-decadal Oscillation (AMO) could potentially modulate the ENSO-Australian fire weather relationship. The correlation between ENSO and Australia fire weather index (FWI) increases from 0.17 to 0.70 when AMO shifts from its negative phase to its positive phase. This strengthening effect can be explained by the atmospheric teleconnection mechanisms. Specifically, the positive AMO phase, manifested as a warming North and Tropical Atlantic, generates Rossby wave trains and results in high pressures over Australia, which increases local temperature and wind speed but decreases precipitation. This signal superimposes ENSO and thus serves to enhance the ENSO effect on Australian fire weather.

Received: 24 Dec 2022 – Discussion started: 29 Mar 2023

Download & links

Preprint (PDF, 1272 KB)

Supplement (1172 KB)

Guanyu Liu et al.

Status: final response (author comments only)

RC1:
'Comment on acp-2022-858', Anonymous Referee #2, 06 Apr 2023

This study states that the ENSO-Australian fire relationship can be modulated by the phase changes of AMO. Specially, Atlantic warming may induce warmer temperature and less precipitation, which serves to enhance wildfires when combined with positive ENSO phase. This result is useful in understanding the recent shift in the ENSO-Australian fire relationship and in future wildfire projection in Australia. However, to further improve this manuscript, there are a few issues and questions that need to be addressed:
1. The authors mainly looked at the AMO effect on positive ENSO-Australian fire relationship, i.e., the modulation of Australian fire weather during El Nino conditions. Could they also examine La Nina conditions? Will responses of Australian fire weather also be strengthened during La Nina?

2. While it is plausible that AMO modulates the ENSO-Australian fire relationship, another major decadal climate variability, PDO, also shifted its phase around 2000. PDO may exert an even stronger impact on Australian weather. How could the authors exclude the effect of PDO?

3. The authors only used fire weather to represent fire activities. This may not be exactly equal to the actual fire counts or emissions. It is suggested to validate the correlation between ENSO and Australian fires using other proxies such as burned area, fire counts, etc.

4. Figure S8 did not give any statistical significance test of SST anomaly, which should be presented for clarity.

Citation: https://doi.org/10.5194/acp-2022-858-RC1
- AC1: 'Reply on RC1', Prof Li, 02 May 2023
  
  We thank this reviewer for the helpful suggestions, which have greatly helped us improve the manuscript. We have studied the comments carefully and made revisions, which we hope will meet the journal standards. Our detailed responses to each of the comments and suggestions are in the supplement. The referee’s original comments are shown in blue. Our replies are shown in black. The corresponding changes in the manuscript are shown in italic black.
  
  Citation: https://doi.org/10.5194/acp-2022-858-AC1
RC2:
'Comment on acp-2022-858', Anonymous Referee #5, 07 Apr 2023

Review of “Atlantic multi-decadal oscillation modulates the relationship between the El Nino-southern oscillation and wildfires in Australia” by Liu et al. (MS ID#acp-2022-858)
The authors attempted to explore the underlying mechanism of AMO shift affecting the relationship of ENSO with wildfires in Australia using several reanalysis data and numerical sensitivity simulations. They showed the correlation between ENSO and Australia fire weather index increased from 0.17 to 0.70 when AMO shifted from its negative phase to positive phase. The specifical impacting process was that the positive AMO can generate Rossby wave trains and result in the high pressure in Australia and then increased the local temperature and wind speed but decreased precipitation. This paper is well writing and the finding is valuable. I think it can be published after several corrections.

1. This work mainly explored the effect of AMO shift on the wildfires in Australia. It is ok. What about the role of global warming? How to distinguish its role?

2. The authors highlighted the roles of SST anomalies over the tropical Atlantic and northern Atlantic. The SST effect over tropical Atlantic is ok and the signals are significant. But it is not for north Atlantic, and the correlations between North Atlantic SST and local factors are much weak, in which most of grids have not passed the significant test. So, the main discussions of this paper are suggested to focus on the tropical Atlantic SST anomaly.

3. To explain the effects of SST anomalies over tropical and north Atlantic on wildfires in Australia, the authors mentioned that the SST anomalies can generate Rossby wave and then affecting the local atmospheric circulations. However, it should be noted that this process involves both hemispheres, and whether the Rossby wave can spread to the southern hemisphere due to trade winds is a matter of debate. The authors can do further analyses to validate this process, such as EP flux, streamfunction, etc.

4. Line 55-: Is the FWI calculated from these local factors? or it obtained from the other source?

5. Line 75: How to choose the positive and negative phase years of Nino 3.4?

6. Section 2.2: In the OBE, the monthly SST variability is added to the climate mean. However, the early sentence mentioned that the trends were added to the model. Which is right?

7. For the OBE, the authors have run eight members. Is there any difference for the initial condition?

8. Line 100: Both the words of “rainfall” and “precipitation” are used in this paper. Strictly, there are some differences between them. Please keep the same expression across this paper.

9. For the orders of Figures in this paper, it seemed disorder. For example, Figure 2 are mainly explained first in the main document, rather than Figure 1.

10. Line 137: This sentence confused me and it not agreed well with the display in Figure 2f and i.

11. Line 168: I have not found the results of OBE. To increase readability, please showing the parallel panels of OBE with reanalysis.

Citation: https://doi.org/10.5194/acp-2022-858-RC2
- AC2: 'Reply on RC2', Prof Li, 02 May 2023
  
  We thank this reviewer for the helpful suggestions, which have greatly helped us improve the manuscript. We have studied the comments carefully and made revisions, which we hope will meet the journal standards. Our detailed responses to each of the comments and suggestions are in the supplement. The referee’s original comments are shown in blue. Our replies are shown in black. The corresponding changes in the manuscript are shown in italic black.
  
  Citation: https://doi.org/10.5194/acp-2022-858-AC2
RC3:
'Comment on acp-2022-858', Anonymous Referee #4, 11 Apr 2023

The manuscript "Atlantic Multi-decadal Oscillation Modulates the Relationship Between El Niño-Southern Oscillation and Fire Weather in Australia" studies the relationship between AMO and ENSO-Australian fire weather. This work advances our understanding of the interactions between different climatic phenomena. After carefully revising several issues, I consider this work interesting, relevant, and worth publication in this journal.
In general, the manuscript could be enhanced by extending the discussions about the atmospheric teleconnection mechanisms by means of Rossby wave trains. Although the authors do a great job explaining the Rossby wave trains argument, I think the discussion about the Rossby wave source anomalies in Figure 4 (c-d) could be improved to clarify this point to the readers.
Minor comments:
LN 61-62. I suggest rewriting as: "WND10 is calculated using the zonal (U10) and meridional (V10) components of the wind vector to represent the intensity of the 10 m wind."
LN 62. The authors compared their results with other NCEP-NCAR and MERRA-2 reanalyses "In order to verify the robustness of our results." I am not sure that this comparison ensures the robustness of their results. I would suggest saying instead: "In order to compare our results with other reanalysis datasets"
LN 63. Please indicate the resolution of the NCEP-NCAR and MERRA-2 reanalysis in the text. Do you use the data in the original grids, or do you regrid to use the same grid as ERA5?
Figure S2. Caption: "...and (e,j) Kaplan Extended SST v2 datasets"
LN 115-117. The authors could also mention that the correlations are strengthened even with the analysis of detrended time series. If Global warming alone were the cause, the correlations would disappear.
LN 140-145. Please improve the explanation of Figure 2.
Figure 2 Lowercase i in Caption ".... and (c,f, i, l) Sea Level Pressure ..."
Figure 2. It is not clear how you do the significance test on these figures. I recommend enhancing the manuscript (or in the supplementary material) on how you do your significance test.

Citation: https://doi.org/10.5194/acp-2022-858-RC3
- AC3: 'Reply on RC3', Prof Li, 02 May 2023
  
  We thank this reviewer for the helpful suggestions, which have greatly helped us improve the manuscript. We have studied the comments carefully and made revisions, which we hope will meet the journal standards. Our detailed responses to each of the comments and suggestions are in the supplement. The referee’s original comments are shown in blue. Our replies are shown in black. The corresponding changes in the manuscript are shown in italic black.
  
  Citation: https://doi.org/10.5194/acp-2022-858-AC3
RC4:
'Comment on acp-2022-858', Anonymous Referee #1, 11 Apr 2023

This manuscript explored the shifting correlations between fire weather indicators and ENSO when AMO was in its positive and negative phases, and has shown interesting results about the combined effects of AMO and ENSO when they are in-phase and out-phase. The authors have tried to interpret the underlying mechanisms teleconnecting Atlantic SST, Australian temperature and precipitation, etc. I am gladly noticing the authors have tested several datasets for the same climate parameters in order to get robust results. This study has provided new information on the impacts of ENSO on fire weather in Australia, the method and analysis are generally solid, the writing is OK, and I suggest this manuscript should be accepted after a proper revision.
My major comments:
1) Statistically, individual AMO+ and ENSO+, and the combination of them, are in favor of a fire-prone climate (high temperature and low precipitation, Fig.2 S5 S6 S7), and this part is OK. But more studies are needed when interpreting the teleconnections: AMO was usually defined as the mean SST of the entire North Atlantic Ocean (including both North and Tropical Atlantic, as you defined in this manuscript). However, it’s interesting to notice that the responses to North Atlantic Ocean SST are much weaker than those of the Tropical, and sometimes even insignificant (Fig.3 4a 4b). Could the authors also show the SSTs time series or patterns for both North and Tropical Atlantic in the supplementary, so the readers could get an idea about at least their differences? I assume the close connection with Tropical Atlantic SST could arise from their closeness in latitudes, so the Rossby wave anomaly arising from a higher Tropical Atlantic SST is easier reaching Australia? which was indicated in Fig.4a? Anyway, more interpretations are needed in this section.
2) Did you remove the global SST warming trend from AMO records? Obviously, most AMO indies are simply the regional means of SST, which combine signals of both global SST warming and multidecadal SST variations. I strongly recommend subtracting the 60°N-60°S mean SST from AMO as Trenberth and Shea (2005) did, or subtracting SSTs of the same latitude from North and Tropical Atlantic SST when you were conducting OBE. Reviewer#2 has already noticed this problem and questioned the role of global warming as well.

Reference:
Trenberth, Kevin; Dennis J. Shea (2005). "Atlantic hurricanes and natural variability in 2005". Geophysical Research Letters. 33 (12): L12704.

Citation: https://doi.org/10.5194/acp-2022-858-RC4
- AC4: 'Reply on RC4', Prof Li, 02 May 2023
  
  We thank this reviewer for the helpful suggestions, which have greatly helped us improve the manuscript. We have studied the comments carefully and made revisions, which we hope will meet the journal standards. Our detailed responses to each of the comments and suggestions are in the supplement. The referee’s original comments are shown in blue. Our replies are shown in black. The corresponding changes in the manuscript are shown in italic black.
  
  Citation: https://doi.org/10.5194/acp-2022-858-AC4
RC5:
'Comment on acp-2022-858', Anonymous Referee #6, 18 Apr 2023

This study examines the effect of Atlantic multidecadal variability on the ENSO-fire weather relationship. A strong modulation is identified using observational data and this is explored further using model simulations.
Overall, the study provides important insights but I think would benefit from some further sensitivity testing to ensure results are well understood and robust.
The analysis is limited by data availability back to only 1981 which makes assessing the effects of longer-timescale climate variability, like the AMO, challenging. The authors do an admirable job though (including use of other, longer datasets), but could more explicitly acknowledge the limitation of data length.
There is surprisingly little discussion of the effect of AMO on ENSO itself. Elaboration on the relationship and mechanisms would be useful (e.g. Levine et al., 2017).
The analysis is primarily based on linearly detrended data over 1981-2019. Given the high decadal variability in Australian precipitation, for example, I would be slightly concerned that detrending could accidentally reduce natural variability in the data as well. Sensitivity tests where precipitation is not detrended may be useful.
Observational datasets are available for some of the variables studied here (e.g. Australian gridded climate dataset (AGCD)). It would be worth comparing ERA-5 against AGCD to ensure ERA-5 is performing adequately.
L22: Suggest changing “climate” to “global”
L32-33: This sentence seems redundant.
L38-39: Better as “shifted from negative to positive phase” I think.
L104-105: Worth adding a qualifier in here such as “typically”.
References
Levine, A. F. Z., McPhaden, M. J., & Frierson, D. M. W. (2017). The impact of the AMO on multidecadal ENSO variability. Geophysical Research Letters, 44(8), 3877–3886. https://doi.org/10.1002/2017GL072524

Citation: https://doi.org/10.5194/acp-2022-858-RC5
- AC5: 'Reply on RC5', Prof Li, 02 May 2023
  
  We thank this reviewer for the helpful suggestions, which have greatly helped us improve the manuscript. We have studied the comments carefully and made revisions, which we hope will meet the journal standards. Our detailed responses to each of the comments and suggestions are in the supplement. The referee’s original comments are shown in blue. Our replies are shown in black. The corresponding changes in the manuscript are shown in italic black.
  
  Citation: https://doi.org/10.5194/acp-2022-858-AC5
- AC7: 'Reply on RC5', Prof Li, 02 May 2023
  
  We are sorry that we made a mistake in the attachment in the previous version of the response. Please refer to the reply in this version.
  We thank this reviewer for the helpful suggestions, which have greatly helped us improve the manuscript. We have studied the comments carefully and made revisions, which we hope will meet the journal standards. Our detailed responses to each of the comments and suggestions are in the supplement. The referee’s original comments are shown in blue. Our replies are shown in black. The corresponding changes in the manuscript are shown in italic black.
  
  Citation: https://doi.org/10.5194/acp-2022-858-AC7
RC6:
'Comment on acp-2022-858', Anonymous Referee #3, 26 Apr 2023

General Comments:

Liu et al. investigated the role of Atlantic Multi-decadal Oscillation (AMO) in influencing the relationship between El Niño-Southern Oscillation (ENSO) and Australian fire weather, using both reanalysis data and ocean basin experiments (OBE). Their findings suggest that the positive AMO enhances the ENSO-Australian fire weather relationship, potentially through atmospheric teleconnection mechanisms. These results are important for advancing our understanding of fire-climate relationships. However, the approach and data used in the study may be difficult to follow, as it involves analyses from both reanalysis and OBE. It’ll also be helpful to provide additional analysis regarding the contributions of individual meteorological factors, such as increased temperature versus wind speed, to the AMO-strengthened ENSO-Australian fire weather relationship.

1. The authors analyzed both the reanalysis and OBE model results, but when each dataset is used and why they are used are not very clear. It would be helpful to have an overview at the end of the Introduction and/or in Methodology to remind the readers of these.

2. Fig. 1a shows the stronger correlation between ENSO and FWI during positive AMO, and Fig. 2 presents increased temperature, reduced precipitation, stronger high pressure, and larger winds. It’ll be interesting to show which meteorological factor contributes most to the strengthened ENSO-FWI relationship under positive AMO.

Specific comments:

1. line 40: I think the approach is not clear at this point (and after Data and Methods section). Maybe it's better to have an overview of the approaches to help readers understand the plan for investigating the Atlantic impact on Australian fire weather.

2. line 50: Please specify the meteorological factors considered in the FWI (e.g., wind, temperature, etc.)

3. line 52: "...fuel availability (drought conditions)" is confusing. Do you mean drought conditions affect fuel availability? Please explain it in detail.

4. line 63: How do you match and compare products with different spatial resolutions?

5. line 97: Why do you separate into the two regions? As mentioned by reviewer #1, AMO is usually defined by SST over the North Atlantic. It'll be helpful to elaborate on why you separated the two regions.

6. lines 104-105: Have you analyzed the effects of the ENSO or AMO leading the peak fire season on fire weather? For example, the effect of ENSO 3 or more months prior to peak fire season on fire weather in peak fire season. Chen et al. (2016) showed that NINO4 with a lead time of 10 months has a high correlation with burned areas in Australia.

7. lines 114-115: How does global warming affect the ENSO-FWI relationship? I know the analyses focus on the effect of natural variability, but it’ll be interesting to understand how global warming plays a role in modulating the ENSO-fire weather relationship, along with the contribution from AMO. Several reviewers also mentioned this.

8. lines 146-151: Figures 2j-l show the composite of AMO, while the following sentences mainly describe positive AMO. It would be helpful to include the composite analysis for positive AMO only and negative AMO only so that there would be evidence supporting the description of positive AMO.

9. lines 227-229: I wonder whether any large fire event is associated with ENSO and positive AMO in history. Or how would the results or conclusions change by replacing fire weather with burned areas or fire emissions? The authors demonstrate the effects of ENSO and AMO on fire weather while lacking actual fire events.

Technical comments:
1. lines 81-82: “In addition, simple stochastic variability…..behaviors.” This sentence is not clear. Please revise it.

2. line 96: Please specify SON in the main text.

3. lines 130-131: Broken sentence. Please revise it.

4. lines 183-185: The sentence is confusing (“…., we examine the responses of the 500 hPa geopotential height (GPH) responses …). There are two responses in the sentence.

Citation: https://doi.org/10.5194/acp-2022-858-RC6
- AC6: 'Reply on RC6', Prof Li, 02 May 2023
  
  We thank this reviewer for the helpful suggestions, which have greatly helped us improve the manuscript. We have studied the comments carefully and made revisions, which we hope will meet the journal standards. Our detailed responses to each of the comments and suggestions are in the supplement. The referee’s original comments are shown in blue. Our replies are shown in black. The corresponding changes in the manuscript are shown in italic black.
  
  Citation: https://doi.org/10.5194/acp-2022-858-AC6

Guanyu Liu et al.

Supplement

https://doi.org/10.5194/acp-2022-858-supplement

Guanyu Liu et al.

Viewed

Total article views: 565 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
447	87	31	565	39	1	4

HTML: 447
PDF: 87
XML: 31
Total: 565
Supplement: 39
BibTeX: 1
EndNote: 4

Views and downloads (calculated since 29 Mar 2023)

Month	HTML	PDF	XML	Total
Mar 2023	102	21	3	126
Apr 2023	261	46	13	320
May 2023	84	20	15	119

Cumulative views and downloads (calculated since 29 Mar 2023)

Month	HTML	PDF	XML	Total
Mar 2023	102	21	3	126
Apr 2023	261	46	13	320
May 2023	84	20	15	119

Viewed (geographical distribution)

Total article views: 562 (including HTML, PDF, and XML) Thereof 562 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 23 May 2023

Download

Preprint (1272 KB)
Metadata XML

Short summary

Fire weather in Australia is positively correlated with El Niño-Southern Oscillation (ENSO). However, their correlation is strengthened in the last two decades. Here, we show the correlation between ENSO and Australia fire weather index (FWI) increases from 0.17 to 0.70 when AMO shifts from its negative to positive phase. This is explained by the teleconnection effect. Warmer Atlantic generates Rossby wave trains and results in high pressures and a weather condition conducive to wildfires.


Total:	0
HTML:	0
PDF:	0
XML:	0