Review of "XBT data collected along the Southern Ocean “chokepoint” between New Zealand and Antarctica, 1994-2024", Giuseppe Aulicino et al.

Summary: The paper nicely describes a valuable dataset of XBT observations across the Southern Ocean from New Zealand to the Ross Sea, Antarctica. These data cover 1994 to 2024 along the PX36 Ship of Opportunity (SOOP) line. The quality of the data is described and suitable QC procedures followed to improve the dataset for end users. Depth corrections have been applied appropriately.

The authors have done an excellent job in summarising the data collected and presenting examples of the use with eddy and front identification using complimentary datasets such as satellite altimetry and Argo datasets.

I recommend publication after some corrections to improve the manuscript (see below) and access to the data could be improved given the current situation post-hurricane in the USA where the NCEI servers are difficult to access. Similarly, publication of some code to access and display the data would be useful to go along with the manuscript. 

Manuscript comments:

Introduction: I recommend including reference to the GOOS Ship of Opportunity program (SOOP) and the SOOPIP (Ship of Opportunity Implementation Panel) in the manuscript. Certainly the authors are part of these programs and the data makes a significant contribution to the SOOP program.

Line 109, suggest the reference should be Cowley et al 2013 as it refers to depth and temperature variations over time.

Table 2 contains Hanawa 1995 FRE values for the T7 probe, not manufacturer FRE as stated in the title. If the Hanawa FRE is used, please include in reference. Also confusion over which FRE the non-corrected Depth values use in the dataset (see comments in the data section below).

Line 155. Suggest including Cowley & Krummel with Parks et al reference.

Line 187, 202, 206: Suggest 'log sheets' or similar in place of 'launch clipboards'.

Line 208. Grammar: 'resulted the most correspondent' perhaps should be ' corresponded'

Line 230. Grammar: 'unusally' should be 'unusual'

Line 276-277. The reference for nodc has an out-of-date table. I think the authors have used the most recent update for the Cheng correction, but these tables do not contain that update. Perhaps the authors can ask Cheng to provide a link to the latest tables. Also, the second link to iap data is to the global ocean dataset, not the XBT correction tables.

Line 302. What is 'SBdy'?

Line 319. Grammar: please check the sentence as it could be improved to make clearer the author's meaning.

Section 2. I think you need a subsection here describing how you created the section plots. What algorithms did you use to make these? Are the section data available for the user to download? Can you include the code in a python notebook or similar so the user can create them?

Data comments: 

I am unable to download the data via https, but was able to access files via ftp (although this is a bit slow). There is some difficulty accessing the servers at NCEI. Can you provide an alternative route to the data? Perhaps a Zenodo link or another DOI to the full dataset (one link) would be useful in addition to the individual NCEI links.

I reviewed the xbt_araon_XXXIV_all_qc.txt file.

1. Fall Rate Coefficients listed in the file (manufacturer T7 coefficients) do not match the ones in the paper (Hanawa 1995 coefficients). Please check which coefficients were used as this will make a difference to the Cheng 2014 corrections applied.

2. Label 'Bot. Depth [m]' is not the bottom depth, but the XBT maximum depth reached.

3. I loaded up the data and plotted it using Python and found the quality flags are well applied, however there are some bad data remaining (hit bottoms and some leakage) in a few of the profiles. Generally, the QC looks acceptable for the one file I reviewed.

4. It would be very helpful and enhance the paper if you could provide some code to read and plot the data for the users. A python notebook would be very helpful. Below is the basic code I used to load and look at the data. You could expand on this. 

import pandas as pd

import matplotlib.pyplot as plt

import os

# Specify the file path

file_path = '/path/to/file/xbt_araon_XXXIV_all_qc.txt'

# Preprocess the file to remove lines starting with //

with open(file_path, 'r') as file:

    lines = file.readlines()

# Filter out lines starting with //

filtered_lines = [line for line in lines if not line.startswith('//')]

# Write the filtered lines to a temporary file

temp_file_path = '/Users/cow074/Downloads/temp_file.txt'

with open(temp_file_path, 'w') as temp_file:

    temp_file.writelines(filtered_lines)

# Read the space-delimited text file into a DataFrame, ignoring the header rows

df = pd.read_csv(temp_file_path, sep='\t', skiprows=0)

# remove the temporary file

os.remove(temp_file_path)

# add column names

df.columns = ['Cruise','ProfileNumber','Type','mm/dd/yyyy','hh:mm','longitude','Latitude',

              'XBT_depth','Depth','Depth_corr','Temperature','Temperature_corr','QF']

# Display the DataFrame

print(df)

# Select the data where QF is < 3

df_filtered = df[df['QF'] < 3]

# plot depth_corr vs Temperature_corr

plt.scatter( df_filtered['Temperature_corr'], df_filtered['Depth_corr'])

# flip the y-axis

plt.gca().invert_yaxis()

plt.ylabel('Depth')

plt.xlabel('Temperature')

plt.title('Depth vs Temperature')

plt.show()

plt.pause(1)

plt.close()

# loop through the data in groups of 10 profileNumbers and plot df_filtered data coloured by ProfileNumber

for i in range(0, max(df_filtered['ProfileNumber']), 10):

    # plot the data where ProfileNumber is in the range i to i+10

    df_temp = df_filtered[(df_filtered['ProfileNumber'] >= i) & (df_filtered['ProfileNumber'] < i+10)]

    plt.scatter(df_temp['Temperature_corr'], df_temp['Depth_corr'], c=df_temp['ProfileNumber'])

    plt.gca().invert_yaxis()

    plt.ylabel('Depth')

    plt.xlabel('Temperature')

    plt.title('Depth vs Temperature')

    # include a legend with ProfileNumber

    plt.colorbar(label='ProfileNumber')

    plt.show()

    plt.pause(1)

    plt.close()

Very useful paper presenting XBT data collected in under sampled region of the South Ocean over 30 years of fields expeditions.

Please see my remarks to the main text of the manuscript in the supplemented file.

My concerns are about section plots presented. I recommend to  review the interpolation parameters used for section plots presented in  the 'Supporting Information' document, in order to avoid some artefacts, which I believe are caused by too small search radius set during section creation. It is also important to use same interpolation setting for all section and provide this information (i.e. horizontal search radius and vertical search radius used for interpolation in ODV) in the text.

It also will be helpful to add a profiles markers on the section plots as well as section boundaries set in ODV for each section (this will allow to access spatial limits during interpolation).

It will be useful to have plots in  the 'Supporting Information' document numbered individually (like S1-1, ..., S1-35), - it will make reference to some individual plot more convenient. As for now all 36 section plots are referred to as Fig S1, therefore, below, I  use page numbers to refer to some particular plots.

Remarks to the section plots in the 'Supporting Information' document: 

I believe some of the features on the plots appears as result of using too short horizontal search radius, so interpolation algorithm is only using data from one profile not taking into account its neighbors . Provided that distance between profiles (n) is about ~20 km, I would recommend to employ 3n value, i.e. ~60km.

I only mention few pages/section plots below, but these remarks are related to all section plots - there are many gaps in the stations position which could be the cause for the interpolation artefacts now existing everywhere.

Page 4: sharp gradient at 60°S most likely caused by the short search radius and the gap in the station position along the section line.

Page 5: sharp negative gradients at 55°S and 60°S and positive at 57° and 62°

Page 6:  sharp negative gradients at 55°S and positive at ~57° (this one is possibly caused by the gap in the station position along the section line - but it is difficult to be sure because data points are not plotted)

Page 7: sharp negative gradients at 54, 56, 57°S and positive in between as well as at 62, 63, & 64°S could be possibly caused by the gaps in the station position along the section line - but it is difficult to be sure because data points are not plotted.