Review Reports - Streamflow Intensification Driven by the Atlantic Multidecadal Oscillation (AMO) in the Atrato River Basin, Northwestern Colombia

Round 1

Reviewer 1 Report

Please find my detailed comments and suggestions in the attached document.

Comments for author File: Comments.pdf

Author Response

Response to Reviewer 1

We would like to thank the reviewer for his/her helpful comments. Please find below our responses.

Reviewer#N Water-647437

“Streamflow intensification driven by the Atlantic Multidecadal Oscillation (AMO) in the Atrato River basin, northwestern Colombia”

General comments

The impact of the warm and cold phase of the AMO on streamflow of the Atrato River is investigated and statistically significant differences are found related to a change in the zonal gradient of sea level pressure and sea surface temperature gradient between the tropical central Pacific and tropical Atlantic Oceans. A connection between the MJ and ND Atrato River streamflow and the AMO over the 1995-2015 period is demonstrated. The paper is in need of clarification and several noteworthy editorial issues must be attended to before the paper is ready for publication. Hence, the manuscript is in need of major revision before it is ready for final publication in “water.” It is possible the revision will be minor if the methodology-related question is a misunderstanding by the reviewer and can be clarified with further specification in the updated version of the manuscript.

Thanks. All of your comments have been taken under consideration and the paper was revised according to them and the recommendations from the other reviewers.

The primary methodological concern is related to the statement “the trend of the time series at each grid point was eliminated to remove the climate change effects” (lines 184-185) in reference to the reconstructed SST dataset. What was the method used to eliminate the trend? Was this same method utilized to remove climate change effects in the gridded sea level pressure and 500 hPa vertical velocities? How can you be certain that the trend removal did not have an impact on the zonal gradient results of SST and sea level pressure between the tropical central Pacific and tropical Atlantic? At the very least, please provide clarification in your response to the reviewers and, if appropriate, add clarification to the methodology description in the manuscript.

The reconstructed SST data of the 1870-2015 period were used only to support that the SST anomaly patterns during the 1965-1994 and 1995-2015 periods indeed contain, respectively, the CAMO and WAMO features[1,2]. In this case, the linear trend for the 1870–2015 of the time series at each grid point was eliminated using the method of the least squares before the calculation of the standardized monthly anomalies.

For the composite analyses, monthly standardized anomalies of precipitation, SST, SLP and 500 hPa vertical velocity in pressure coordinate during the 1965-2015 period were obtained. In this case, the linear trends were not removed from the variables.

Once the methodological criticism has been addressed, the manuscript will be ready for a re-submission after the following minor editorial and technical issues have been corrected and/or addressed.

Minor issues – editorial and technical issues

[1] There are several sections of text within the manuscript when confusing or improper English phrasing is used;

The English of the new version has been revised by an English technical editor.

“the heighten of rainfalls” [<unknown meaning>??] (line 34),

The abstract was re-written.

“upward movements” [‘movement’] (line 42),

Done.

“in both Walker and Hadley cells” [‘in both the Walker…’] (line 67),

Done.

“still need further research” [‘…further investigation.’] (line 73),

This line was modified.

“and it is expected” [‘and are expected’] (line 99),

Done.

“has a privileged geostrategic location” [<unknown meaning>??] (line 103),

In this version the paragraph was modified.

“subject to rights” [<unknown meaning>??] (line 107),

In this version the paragraph was modified.

“one of the highest performance” [<unknown meaning>??] (line 122),

This phrase has been removed.

“until this point,” [<unknown meaning>??] (line 129),

In this version the paragraph was modified.

“were used” [‘was used’] (line 155),

Done.

“Prediction e Center” [‘Prediction Center’] (line 157),

Done.

“extremes” [‘extrema’] (line 174),

Done.

“patterns for the two” [‘patterns of the two’] (line 183),

In this version the paragraph was modified.

“assessed using Student’s t-test” [‘assessed using the Student’s t-test’] (lines 186 and 196),

Done.

“considering the monthly” [<unknown meaning>??] (line 207),

In this version the paragraph was modified. ‘’… according to AMO phases’’

“analyses for the years” [‘analyses of the years’] (line 217),

Done.

“differences composites” [‘difference composites’] (line 218),

Done.

“anomalies in the pressure” [‘anomalies in pressure’] (line 219),

Done.

“results for other” [‘resultes of other’] (line 255),

Done.

“and the trends detected” [‘and trends detected’] (line 275),

In this new version the trend analysis was eliminated. The phrase was removed.

“which showed” [‘and showed’] (line 302),

Done.

“is for each 0.2” [‘is 0.2’] (line 324),

Done.

“downward motions” [‘downward motion’] (lines 328, 329, 337, 352, …),

Done.

“upward motions” [‘upward motion’] (lines 329, 330, 333, 337, 352, 369, 440, …),

Done.

“connections between” [‘connection between’] (line 343),

Done.

“with a strengthening zonal” [‘with a strengthened zonal’] (line 367),

Done.

“extreme rainfalls” [‘extreme rainfall’] (line 383),

Done.

“changes by using” [‘changes using’] (line 384),

Done.

“with the nonsignificant differences” [<unknown meaning>??] (line 388),

In this version the paragraph was modified.

“show less (more) defined patterns” [what is less or more defined?] (line 429),

In this version the paragraph was modified.

“is well (little) structured” [what is well or little structured?] (line 438).

In this version the paragraph was modified.

[2] (line 147) Clarify how station representativeness was ensured.

The stream flow data of the 1965-2015 period at the stream gauge stations of Belén, San Antonio, and Bellavista from the IDEAM, located in upper ARB were used in preliminary analyses (Appendix A) based on the principal component analysis (PCA) and correlations among the stations. These analyses indicated that the Bellavista streamflow time series reproduced most of the variations shown in the other two streamflow time series. So, the streamflow data at Bellavista was used here as representative of the study domain.

[3] (lines 215-216 and elsewhere throughout the manuscript) the font changes between normal to boldface in the middle of a sentence, description, or table/figure caption.

Thanks for raising this, the change was made.

[4] (lines 222 and 233) redundancy in specifying 95% confidence level

We rewrote this paragraph.

[5] (lines 274-345) A discussion of Figs. 4 and 5 should note the ‘negative’ in the anomaly SST or SLP pattern between the CAMO and WAMO MJ panels (e.g., ‘red’ in CAMO is ‘blue’ in WAMO and ‘blue’ in CAMO is ‘red’ in WAMO in MJ).

Thanks for raising this, we have improved the discussion.

[6] (lines 376-380) “However,…by [87] and [98].” Won’t anthropogenic forcing change the mean of the SSTs and sea level pressure in the tropical central Pacific and tropical Atlantic, but have little influence on the zonal SST and SLP gradient between the two basins? Your results suggest it is the strength of the zonal gradient that is important. Will anthropogenic forcing really change the streamflow in the Atrato River?

Thanks for raising this. In the new version the fundamental role of the inter-basin gradient in the modulation of the Atrato streamflow and rainfall has been clarified. Future research could evaluate the role of anthropogenic forcing in the region.

Figures

Figure 1. Legend refers to “Rain gauge stations” when only a single station is plotted (‘station’).

Thanks for raising this, it was a mistake. However, in this new version, Figure 1 includes the three Rain Gauge stations used.

Figures 2, 4-7, A1. Should the contour also be added to the color bar to show the exact color corresponding to the contour level?

We agree, the change was made.

Figures 2, 4, 5 and Figures 6-7, A1. Why was the color bar reversed from blue negative anomalies to red negative anomalies? Please make them consistent.

In the case of precipitation, blue is commonly used for positive anomalies (wet conditions) and red for negative anomalies (dry conditions), however, in this version we use different color schemes to emphasize anomalies of precipitation. This change follows guidance on appropriate techniques for mapping climatic variables from Kaye et al. [3] and Retchless and Brewer [4].

A description of the contours is needed in the caption of Figure 5.

The description has been included (Now Figure 6).

References

Enfield, D.B.; Mestas-Nuñez, A.M.; Trimble, P.J. The Atlantic multidecadal oscillation and its relation to rainfall and river flows in the continental U.S. Geophys. Res. Lett. 2001, 28, 2077–2080. Enfield, D.B.; Mestas-Nuñez, A.M. Multiscale variabilities in global sea surface temperatures and their relationships with tropospheric climate patterns. J. Clim. 1999, 12, 2734–2746. Kaye, N.R.; Hartley, A.; Hemming, D. Mapping the climate: Guidance on appropriate techniques to map climate variables and their uncertainty. Geosci. Model Dev. 2012, 5, 245–256. Retchless, D.P.; Brewer, C.A. Guidance for representing uncertainty on global temperature change maps. Int. J. Climatol. 2016, 36, 1143–1159.

Author Response File: Author Response.docx

Reviewer 2 Report

Review of Streamflow Intensification Driven by the Atlantic Multidecadal Oscillation (AMO) in the Atrato River Basin, Northwestern Colombia by Cerón et al.

The topic of this paper is definetively interesting. The paper attempt demonstrate a long-term relationship between stream flow of the Atrato River measured in the Bellavista station in the period between 1965-2015 (51 years). Relationship of AMO and precipitation and circulation pattern are examined in detail.

The comparison is performed by computing the long-term averages of the relevant data during Warm AMO and Cold AMO periods and the difference between mean values is tests using t-test. This seems a rather oversimplified way to compare time series, which is most often performed using regression techniques, rank correlation, and spectral or wavelet coherence.

Figure 3 is the key figure of the paper, which illustrate the original streamflow data to be compared with AMO index. A visual comparison between these data (Fig 3) and the AMO index (not shown) would show immediately that there is a very little resemblance on the timescale of the ~60-70 yr (i.e., the AMO oscillation). Unfortunately a graph of the AMO index in not shown in the paper, but it should have been.

The major features of flow data show in Fig. 3 consist of an increasing trend, and shorter time scale variability; in my opinion none of them can be related to the ~60-70 yr AMO oscillation.
The long term trend, although increases in the same direction of the AMO index, does not show the sinusoidal path of AMO which identify an oscillation, in particular does not show the expected decreasing pattern in the first decade (1965-1975). In my opinion the constant trend in the flow data can be better interpreted as a consequence of the global climate changes.
The shorter term variability of the flow, which is a major feature of the signal, is discussed in terms of intense flow events, which are listed in table 1. These events have been related (correctly, I believe) to ENSO. In particular it occurs that extreme flows in the ND season (Table 1) correspond, with few exceptions, to ENSO negative peaks.
I understand that AMO is supposed to modulate ENSO, however, here the experimental evidence is that intense flow events are related to ENSO and not to AMO.

The paper put also a lot of emphasis and effort on the long-term-averaged atmospheric circulation and seasonal precipitation pattern. I have found this part of the paper very interesting. However, these results are very indirectly related with Atrato river flow and, most importantly, the experimental data (i.e., the flow record) do not support a connection between them.

Authors state that "The results show a connection between the AMO and the streamflow of the Atrato basin during the last five decades", I regret to say that in my opinion the main conclusion is not sufficiently supported by the data.

Author Response

Response to Reviewer 2

We would like to thank the reviewer for his/her helpful comments. Please find below our responses.

Review of Streamflow Intensification Driven by the Atlantic Multidecadal Oscillation (AMO) in the Atrato River Basin, Northwestern Colombia by Cerón et al.

The topic of this paper is definitively interesting. The paper attempt demonstrate a long-term relationship between stream flow of the Atrato River measured in the Bellavista station in the period between 1965-2015 (51 years). Relationship of AMO and precipitation and circulation pattern are examined in detail.

Thanks. All of your comments have been taken under consideration and the paper was revised according to them and the recommendations from the other reviewers.

We agree that Figure 3 is the key figure of the document. So we improved it. In the new version, Figure 3 shows the Atrato River Basin (ARB) streamflow and AMO index, both filtered with a 10-year moving mean filter. Similar graphics were constructed for the average precipitation in the ARB (Figure 4).

The major features of flow data show in Fig. 3 consist of an increasing trend, and shorter time scale variability; in my opinion none of them can be related to the ~60-70 yr AMO oscillation.

According to Figure 3, the smoothed streamflow time series follows the sinusoidal pattern of the AMO index. Furthermore, Table 1 shows the t-test comparisons of streamflow and precipitation time series before and after 1994 in MJ and ND.

The long term trend, although increases in the same direction of the AMO index, does not show the sinusoidal path of AMO which identify an oscillation, in particular does not show the expected decreasing pattern in the first decade (1965-1975). In my opinion the constant trend in the flow data can be better interpreted as a consequence of the global climate changes.

Please see the previous answer.

The shorter term variability of the flow, which is a major feature of the signal, is discussed in terms of intense flow events, which are listed in table 1. These events have been related (correctly, I believe) to ENSO. In particular it occurs that extreme flows in the ND season (Table 1) correspond, with few exceptions, to ENSO negative peaks.

I understand that AMO is supposed to modulate ENSO, however, here the experimental evidence is that intense flow events are related to ENSO and not to AMO.

We included an explanation on how the ENSO and AMO are connected after 1994.

Please refer to answer above in the first comment.

Author Response File: Author Response.docx

Reviewer 3 Report

Dear editor,

I found the paper by Ceron et al. interesting and well written. Thus I suggest its publication after some revision.

The authors study the relationship between the streamflow of Atro River and the Atlantic Multidecadal Oscillation. Their results are reasonable because of the well-known influence of the AMO in the region which is also linked to the migration of the Intertropical Convergence Zone.

I found the paper well written.

However, because of the result in figure 3, which I found interesting, I would suggest comparing such a result with the equivalent time series curves relative to the AMO, by possibly separating it into two time series for the May-June and November-December periods or periods one or two months earlier to check whether the model has some forecast potentiality.

I think that this addition could make the paper significantly more useful and interesting

Author Response

Response to Reviewer 3

We would like to thank the reviewer for his/her helpful comments. Please find below our responses.

Dear editor,

I found the paper by Ceron et al. interesting and well written. Thus, I suggest its publication after some revision.

The authors study the relationship between the streamflow of Atrato River and the Atlantic Multidecadal Oscillation. Their results are reasonable because of the well-known influence of the AMO in the region which is also linked to the migration of the Intertropical Convergence Zone.

I found the paper well written.

Thanks. All of your comments have been taken under consideration and the paper was revised according to them and the recommendations from the other reviewers.

I think that this addition could make the paper significantly more useful and interesting

We agree that Figure 3 is the key figure of the document. So we improved it. In the new version, Figure 3 shows the Atrato River Basin (ARB) streamflow and the AMO index, both filtered with a 10-year moving mean filter. Similar graphics were constructed for the average precipitation in the ARB (Figure 4).

Author Response File: Author Response.docx

Reviewer 4 Report

General comments

However, its structure should be partially reorganized, as there is a confusion between the contents of “Materials and Methods” and “Results”, as well as between primary and ancillary data analyses that are performed in the study. The results of this study should be clearly divided from the results already published in the scientific literature, whereas ancillary data analysis relevant to secondary stations (San Antonio and Belén) should be placed in the appendix, in order to make easier the reading of the manuscript.

I suggest the publication of this manuscript on WATER, after a MAJOR REVISION that should cover the following points:

The reasons of the choice of May-June and November-December as reference periods for the analysis of the interactions between oceanic and atmospheric forcings and runoff should be better explained in the Introduction. 3 lines 120-148: “Study area” section should contain only a brief description of the region of interest, based on data and papers already published in the literature and not a presentation of the data that are used in this study (lines 128-131, 143-148). PCA analysis (used to prove the highest significance of Bellavista station compared to other gauging stations) can be placed in the Results or in the Appendix, if the authors consider it of secondary importance. 4 line 152: The other sections of the “Material and Methods” should contain the description of the methods used for data analysis, but not the results obtained with these analyses. Statistical analyses of the data series of precipitation and streamflow (Figure 1b,c), the analysis of their variability (Table 1), together to all relevant paragraphs of the text in which these figures/tables are discussed, are results of the current study and they have to be placed in the “Results and Discussion”. Why the comparison among streamflows at Bellavista, San Antonio and Belén stations is shown in Table 2 and discussed in the text (Pag. 7 lines 258-273), if Bellavista station was already selected as the most representative by PCA analysis (pag. 3 lines 139-148)? This study should provide more detailed information on the relationship between streamflow and precipitation at Bellavista station, which is only briefly mentioned in the text (pag 7, lines 253-257). Can the authors show one figure with the time series of bimonthly precipitation in May-June and November-December, equivalent to the Figure 3? Why the annual cycles of precipitation and streamflow at Bellavista stations (Figure 1b,c) differ in some months? For example, the precipitation in June is the highest one, whereas the streamflow is rather low. In Atrato River basin, is there any effect of snowmelt on the river regime?

Specific comments:

1 line 33: I don’t think Atrato River has the highest mean streamflow (m³/s) in the world. Perhaps, it has a rather high specific streamflow (m³/s km²). 2 line 53: change in: … , due to the interaction of several phenomena acting at various temporal and spatial scales. 2 lines 82-87: this sentence is not very clear and it might be rephrased. Here and in the rest of the manuscript (e.g. pag. 6 lines 236-243), a frequent utilization of brackets to indicate two opposite cases/trends makes difficult the reading. 3. line 122: Change in: The basin is a high performance basin, with one of the highest specific water streamflows in the world… 4 lines 150-151: Figure 1 (a) Atrato River basins, (b) annual cycle of the precipitation and (c) annual cycle of streamflow at Bellavista gauge station in 1965-2015. 4 line 161: … 0.25° x 025°. The GMFD… 7 line 248: The patterns shown in Figure 2 are not “seasonal” as they refer to MJ and ND periods.

Comments for author File: Comments.pdf

Author Response

Response to Reviewer 4

We would like to thank the reviewer for his/her helpful comments. Please find below our responses.

General comments

The present study investigate the effects of low-frequency changes of oceanic and atmospheric characteristics on the regime of Atrato River. The manuscript is rather well written, with a good figure design and an extensive literature. However, its structure should be partially reorganized, as there is a confusion between the contents of “Materials and Methods” and “Results”, as well as between primary and ancillary data analyses that are performed in the study. The results of this study should be clearly divided from the results already published in the scientific literature, whereas ancillary data analysis relevant to secondary stations (San Antonio and Belén) should be placed in the appendix, in order to make easier the reading of the manuscript. I suggest the publication of this manuscript on WATER, after a MAJOR REVISION that should cover the following points:

All of your comments have been taken under consideration in the revised manuscript according to them and the recommendations from the other reviewers.

1) The reasons of the choice of May-June and November-December as reference periods for the analysis of the interactions between oceanic and atmospheric forcings and runoff should be better explained in the Introduction.

In the revised manuscript, we described better the reason for choosing MJ and ND (Section 2.1 and 2.2).

2) Pag. 3 lines 120-148: “Study area” section should contain only a brief description of the region of interest, based on data and papers already published in the literature and not a presentation of the data that are used in this study (lines 128-131, 143-148). PCA analysis (used to prove the highest significance of Bellavista station compared to other gauging stations) can be placed in the Results or in the Appendix, if the authors consider it of secondary importance.

We agree. The study area was rewritten to review only data from already published literature. We kept annual cycle in figure 1 for the illustration purpose of already published information [1,2]. The selection of Bellavista station was described in Appendix A.

3) Pag. 4 line 152: The other sections of the “Material and Methods” should contain the description of the methods used for data analysis, but not the results obtained with these analyses. Statistical analyses of the data series of precipitation and streamflow (Figure 1b,c), the analysis of their variability (Table 1), together to all relevant paragraphs of the text in which these figures/tables are discussed, are results of the current study and they have to be placed in the “Results and Discussion”.

We agree, we made the correction and included only the description of the data and methods in section “Material and Methods”.

4) Why the comparison among streamflows at Bellavista, San Antonio and Belén stations is shown in Table 2 and discussed in the text (Pag. 7 lines 258-273), if Bellavista station was already selected as the most representative by PCA analysis (pag. 3 lines 139-148)?

We agree with this comment. Now there is only information for the selected gauge station, Bellavista (Table 1).

5) This study should provide more detailed information on the relationship between streamflow and precipitation at Bellavista station, which is only briefly mentioned in the text (pag 7, lines 253-257). Can the authors show one figure with the time series of bimonthly precipitation in May-June and November-December, equivalent to the Figure 3? Why the annual cycles of precipitation and streamflow at Bellavista stations (Figure 1b,c) differ in some months? For example, the precipitation in June is the highest one, whereas the streamflow is rather low. In Atrato River basin, is there any effect of snowmelt on the river regime?

The differences mentioned by the reviewer between the annual cycles of the precipitation and streamflow at Bellavista station, are because only the rainfall from the Bellavista station, at the same location as the gauge station, was considered in the previous version. In the new version we used mean basin rainfall upstream the stream gauging station. Now we clarified this.

We included in section 3.2 the streamflow time-series at Bellavista and the mean precipitation (based on the three upstream stations of the basin -Buchado, Tagachi, and Bellavista). These time series were smoothed with a 10-year running mean filter. The time-series were later compared in Figure 4, and discussion of the results is presented in section 3.2.

Furthermore, according to the literature there are no glaciers in our study domain, thus the Atrato River Basin (ARB) is not affected by snowmelt on the river regime. See further details of Glacierized areas over South America in Braun et al. (2019) [3].

Specific comments:

6) Pag. 1 line 33: I don’t think Atrato River has the highest mean streamflow (m³/s) in the world. Perhaps, it has a rather high specific streamflow (m³/s km²).

You are right. We made the necessary corrections.

7) Pag. 2 line 53: change in: … , due to the interaction of several phenomena acting at various temporal and spatial scales.

We made the necessary corrections.

8) Pag. 2 lines 82-87: this sentence is not very clear and it might be rephrased. Here and in the rest of the manuscript (e.g. pag. 6 lines 236-243), a frequent utilization of brackets to indicate two opposite cases/trends makes difficult the reading.

Thanks for your suggestion. We made the necessary editing and English corrections.

9) Pag. 3. line 122: Change in: The basin is a high performance basin, with one of the highest specific water streamflows in the world…

The description of the study area was improved (section 2.1).

10) Pag. 4 lines 150-151: Figure 1 (a) Atrato River basins, (b) annual cycle of the precipitation and (c) annual cycle of streamflow at Bellavista gauge station in 1965-2015.

11) Pag. 4 line 161: … 0.25° x 025°. The GMFD…

12) Pag. 7 line 248: The patterns shown in Figure 2 are not “seasonal” as they refer to MJ and ND periods.

We agree. The change was made.

References

IDEAM 2014 - Estudio Nacional del Agua - ENA 2014; 2014; ISBN 9789588067704. Velásquez-Restrepo, M.; Poveda, G. Estimación del balance hídrico de la región Pacífica. Dyna 2019, 86, 297–306. Braun, M.H.; Malz, P.; Sommer, C.; Farías-Barahona, D.; Sauter, T.; Casassa, G.; Soruco, A.; Skvarca, P.; Seehaus, T.C. Constraining glacier elevation and mass changes in South America. Nat. Clim. Chang. 2019, 9, 130–136.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Please find my comments in the attached document.

Comments for author File: Comments.pdf

Author Response

Response to Reviewer 1

We would like to thank the reviewer for his/her helpful comments. Please find below our responses.

Reviewer#1 Water-647437

“Streamflow intensification driven by the Atlantic Multidecadal Oscillation (AMO) in the Atrato River basin, northwestern Colombia”

General comments

The impact of the warm and cold phase of the AMO on streamflow of the Atrato River is investigated and statistically significant differences are found related to a change in the zonal gradient of sea level pressure and sea surface temperature gradient between the tropical central Pacific and tropical Atlantic Oceans. A connection between the MJ and ND Atrato River streamflow and the AMO over the 1995-2015 period is demonstrated. The authors have done a good job of addressing my concerns based on the previous version of the manuscript. Hence, the manuscript is in need of minor revision before it is ready for final publication in “water.”

Minor issues – editorial and technical issues

[1] (line 179) should read “…using the method of least squares.”

Done.

[2] (lines 277, 283, 286, 305, 310, 318, 322, 348) error message ‘Error! Reference source not found.’ appears in the text.

We made the correction.

Figures

Figure 4. First line of the figure caption is indented too far to the right.

This was checked.

Author Response File: Author Response.docx

Reviewer 3 Report

Dear Editor

I think that the current paper can be accepted for publication.

The authors have improved their paper by editing it and adding the corrections that I suggested in Figures 3 and 4 that make their results clearer.

Author Response

Response to Reviewer 3

We would like to thank the reviewer for his/her helpful comments.

Reviewer 4 Report

The authors has significantly improved the manuscript in the present version, through an extended revision of the text and adding new figures. The structure of the manuscript was also reorganized, accordingly to the comments of the reviewer. For this reasons, I suggest the publication of this manuscript in the present version on WATER. Please, check only if these few minor suggestions are applicable.

Section “2.1. Study area”, line 119: the unit of measurement of specific streamflow should be: m3 s-1 km-2.

Section “3.3. Composite analysis”, line 299: “Since 1994, anomalous cooling in the tropical Eastern Pacific and warming in the central North Pacific in ND indicated the establishment of CPDO.”

Section “3.3. Composite analysis”, line 341: “Our results are in agreement with the finding that the atmospheric response to the tropical Atlantic warming can lead to an intensification of the upward branch of the Walker circulation in northern SA[96–98]. This process is accompanied ….”

Check if the format of the following references is correct: 2 (line 490), 19 (line 528), 30 (line 552), 58 (line 612), 68 (line 637), 80 (line 665), 84 (line 673), 85 (line 675), 88 (line 683) and 92 (line 693).

Author Response

Response to Reviewer 4

We would like to thank the reviewer for his/her helpful comments. Please find below our responses.

Reviewer#4 Water-647437

“Streamflow intensification driven by the Atlantic Multidecadal Oscillation (AMO) in the Atrato River basin, northwestern Colombia”

Comments and Suggestions for Authors

Section “2.1. Study area”, line 119: the unit of measurement of specific streamflow should be: m3 s-1 km-2.

Done.

Thanks for your suggestion; we made the change.

Author Response File: Author Response.docx