the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 18 Oct 2024
ReefTEMPS: The Pacific Islands Coastal Temperature Network
Abstract. While the rise in global ocean temperature continues its course, reaching 1.45+/- 0.12 °C above pre-industrial level according to the World Meteorological Organization in 2023, marine heat waves frequencies and intensities increase. Consequently, coral reef ecosystems which are among the most vulnerable environments are strongly impacted with dystrophic events and corals experiencing increasing frequencies of bleaching events. That has devastating consequences for the Pacific Island Countries and Territories (PICTS) that strongly rely on these ecosystems. In-situ observation remains the best alternative for providing accurate characterization of long-term trends and extremes in these shallow environments. This paper presents the coastal temperature dataset of the ReefTEMPS monitoring network in which moored stations are implemented over a number of PICTS over a wide region in the Western and Central South Pacific from New Caledonia to French Polynesia. These in situ temperature time series are unique in several ways: in the length of some historical stations dating back to 1958 for the oldest, thus providing more than 65 years of daily data; in the number of countries sampled (16 PICTS) ; and in the variety of coral ecosystems monitored (from atolls to high islands and from barrier reef’s external slopes to shallow and narrow lagoons). Measurement devices have evolved over the years to provide increasingly precise and frequent observations so that the ReefTEMPS network was endorsed as a French National Observation Service in 2020, a label ensuring quality controlled and open access data of long-term observations. All stations are publicly available in ASCII or formatted NetCDF files, either on the ReefTEMPS dedicated Information System which also allows quick visualisation of time series, or in the SEANOE marine data platform. All links and accesses to these temperature time series are provided herein. The quality control and longevity of these temperature time series allows diagnosing long-term trends, highlighting the influence of multiple processes on temperature dynamics (e.g., internal waves, cyclones, seasonal and climate modes) and documenting the time evolution of extreme events. All files are made publicly available in dedicated SEANOE repositories (DOI provided herein).
- Preprint
(7523 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on essd-2024-394', Anonymous Referee #1, 27 Nov 2024
Review of the manuscript:
ReefTEMPS: 1 The Pacific Islands Coastal Temperature Network
By Romain Le Gendre, David Varillon, Sylvie Fiat, Régis Hocdé, Antoine de Ramon N’Yeurt, Jérôme Aucan, Sophie Cravatte, Maxime Duphil, Alexandre Ganachaud, Baptiste Gaudron, Elodie Kestenare, Vetea Liao, Bernard Pelletier, Alexandre Peltier, Anne-Lou Schaefer, Thomas Trophime, Simon Van Wynsberge, Yves Dandonneau, Michel Allenbach and Christophe Menkes
This paper presents coastal temperature data from the ReefTEMPS network of moored stations at a number of PICTS in a broad region of the western and central South Pacific, from New Caledonia to French Polynesia. The in situ temperature time series are considered unique in several respects, both in terms of the longevity of some historical stations - the oldest dating back to 1958 and providing daily data for more than 65 years - and in terms of the number of countries surveyed (16 PICTS) and the diversity of coral ecosystems monitored (from atolls to high islands and from outer reef slopes to narrow, shallow lagoons). All data from the stations are publicly accessible via the dedicated ReefTEMPS information system, which also enables rapid visualization of the time series, or via the marine data platform SEANOE. The longevity of these temperature time series makes it possible to diagnose long-term trends, show the influence of various processes on temperature dynamics and document the temporal evolution of extreme events.
General Comments
The manuscript (ms) is well structured and written and demonstrates the importance of continuous measurements, especially on moored systems, to determine the physical and biochemical processes that influence biodiversity. However, some improvements should be made before publication; adding some information that I think is missing could improve the text. I found that the authors are vague in some parts of the text and use a very discursive approach without being precise. In particular, in section 3, the authors show the evolution of the sensors used without specifying where (stations, areas) the changes took place. It is difficult to say whether the improvements or changes have taken place at all stations in the network or only in certain areas. See my specific comments.
I appreciate the inclusion of examples of key applications in Chapter 6. Since these are processes determined by long-term measurements on various time scales, it would be interesting to relate these changes, especially those in temperature, to ENSO. This topic is already briefly discussed on the website, but it would be an added value to discuss it in the handbook.
Specific comments
In the abstract:
Line 50: 16 PICTS are indicated, while only 14 are mentioned on the ReefTemps website; which number is correct?
Line 57: Quality control is mandatory before sharing data and showing possible trends, while the length of a time series is the added value and the key factor for its determination. Please delete quality control from the sentence.
2.1 History:
Figure 1 is missing information mentioned in this chapter as the start of AusAID support in 2011, which led to the deployment of several sensors on different islands. In 2012, the station on Wallis and Futuna was taken over by NC University, so it became the network?
In line 57 a dot is missing after GOPS
2.2 The current ReefTEMPS network
Line 85: again 16 or 14 PICTS? Please enter the number of actual stations or monitored sites here (active, inactive and dismissed). Later in line 92 you indicate that the New Caledonian components of the backbone component of the ReefTEMPS network include 43 stations, so it would be interesting to indicate the total number of stations belonging to this network right at the beginning of the ms.
3.1 Oceanographic bucket
Line 22: Please move “The nominal acquisition time for both stations was 7 am local time and the targeted depth using the bucket was 0.5 m" in line 20 after “…nearly 47 years” .
3.2 Compact autonomous loggers
Line 27: ".... Were used to monitor coastal temperatures"... please specify where?
Facilities: It is difficult for the reader to tell at first glance which of the platforms listed in Table 1 are still active. I suggest dividing them into active and inactive or color-coding the two/three cases.
4. Processing and quality control
The overall strategy is well described for data measured after 2010; however, I cannot see a clear description of procedures when you recovered the instrument. Have CTD casts been taken each time you recover and redeploy the sensors? How often do you replace the instrument? It seems like you are only removing outliers, what about sensor drift? How do you compensate for any discrepancy (align mismatch) between the end of the time series and the new time series?
4.2 The procedure for obtaining the homogenized monthly long-term data is very interesting and impressive. It could be of additional value to show how the data was corrected. I propose to include a time series as an example of a raw and a QC-corrected time series (Anse Vata or Phare Amédée).
6.2 Characterize physical processes at various timescales
Line 9-10 please rephrase the sentence and highlight the importance of the parameters time series in identifying physical processes at various time scales.
Appendix A
Table A1.
- Not everyone knows the area in which the stations are located, and it is not immediately obvious from the “nom_station”. Please add a first column with the area/region (example Cook Island.
- Add the last sensor type that was used in the table heading after the positions.
- See my previous comment on the "active" column
- Latest type sensor column: explain why you use full names for the sensor types such as “Thermistor" and “Multiparameter" and then abbreviations such as TSG, MG and SEAU (I could not find what this means in Table 3). I suggest standardizing the information in the column and giving the full name for all instruments.
Citation: https://doi.org/10.5194/essd-2024-394-RC1 -
RC2: 'Comment on essd-2024-394', Anonymous Referee #2, 17 Dec 2024
Evaluation of the MS “ReefTEMPS: The Pacific Islands Coastal Temperature Network” by Le Gendre et al.
General comments:
The manuscript reports on the ReefTEMPS observation network and focuses on coastal high-frequency seawater temperature time series acquired in 16 Pacific Island Countries and Territories since 1997 and historical daily series back to 1958 for the older. The MS is well written, and embraces a wide scope, which is very much appreciated: from historical context to network and database description, key applications and opening development and perspectives. Section 2 on network description, and section 4 on data processing and quality control are sometimes rather vague or too general to provide a direct and fair vision on current/past sampling effort and data quality.
The data set is impressive and unique by the spatial (mostly intertropical, from roughly 140°E to -178°E) and temporal extent covered (pluri-annual to multi-decadal HF measurements and daily historical time series). Acknowledging the very nice collaborative effort and important work achieved for data aggregation, quality check and dissemination, I also find the dataset of very high relevance. Examples of key scientific applications are well developed (section 6). Such in situ temperature time series are indeed required to accurately characterize local conditions and marine heatwave stress in highly diverse and sensitive coral reef ecosystems, as well as to analyze fine scale coastal processes and dynamics poorly captured by satellite SST. Warming trend analysis carried out at climatic time scale after homogenization of the long-term series is another asset of the MS.
However, some information is missing or lacking precision in the presentation of the network/dataset and regarding the quality check. By adding precise information on total and yearly sampling effort and major breakdown of data series (in particular the proportion still active), authors could greatly ease the understanding and potential reuse of dataset.
The data are available by different means and in different formats (csv, NetCDF), with a dedicated visualization service. Some data series from the NetCDF archive are not shown on the ReefTEMPS portal and the reason why is not clear to me. This should be quantified and explained. Also, the archive comprises 481 nc files in total + one ascii file. For the variable TEMP, there are 185 files with both raw and validated data. I find the presence of “duplicated” data series with different quality levels in the same directory confusing.
Exploring the dataset itself, 95 files follow the naming convention indicated in section 4.1 (with depth indication), among which 21 raw data series, 57 visually checked (0C) and 17 reduced data sets (3A or 3B). Plotting the visually QC’ed series, I found obviously bad values remaining and flat Quality Flag values at 0. These bad data may fragilize the confidence in the entire data series and impede direct use of the data set, e.g. for satellite data validation. The database needs to be systematically and carefully double-checked and updated by the authors to remove all spurious values and achieve the highest possible number of series visually checked, which to my opinion, and according to Section 4.1. and Figure 3, should be a minimal requirement for dissemination and for publication in ESSD.
Specific comments:
Section 2. The description of the data set is sometimes too vague or general. Complementary information should be provided on the total number of series, equivalent in total year of observation, on the proportion (and quantity) of active vs interrupted data series and major breakdown by origin (sensor type, e.g. tide gauge vs benthic loggers), depth (please consider indicating 0-10 m available as ground-truth for satellite SST), single vs. multiple depth (verticals).
Figure 1. Consider showing the number of series available for each year on the timeline.
Appendix A, Table A1. Please consider presenting the table differently, starting with active series and followed by past/interrupted ones. A supplementary summary figure showing data availability by site since 1997 would ideally complement Table A1.
Page 6 line 87. Indicate the amount/proportion of stations/series stopped.
Page 6 line 93 “and the longest time series” please indicate the number of series with a minimum duration of 10 years
Figure 2. Consider using different symbols on the map for ongoing vs. interrupted series (e.g. circles vs. squares).
Section 4 & 5. These sections describe the data life cycle, quality control, management and dissemination in a very general way. The total number of data files and major breakdown by QC level should be quantified and explained somewhere, potentially in section 4.1 or in section 5.
Line 73. Indicate the time interval at which maintenance and recalibration were performed, either systematically or in general with some exceptions. If possible, provide feed-back on typical stability and results from intercomparison.
Section 6.1. Marine heatwaves refer to discrete events with significant deviation to a baseline or climatology. Please consider showing some pluriannual or climatological mean in Figure 5(b,c) and indicate anomaly/intensity in the text in order to figure out how extreme were those events.
Page 14 line 66. “Played a key role…”. Replace by something more explicit like “negatively impacted the health of ecosystems…”.
Page 14 line 90. “large biases (more than 2°C)…”. The text on satellite SST vs. in situ comparison is a bit short and should be extended from analysis of Figure 5.
Technical corrections:
Page 8 line 21. Repetition of “1977”, the sentence should start by “At the Amédée …”
Figure 5. Consider indicating color legend above panel “a” for New Caledonia, Fiji and French Polynesia, as in Figure 6. Indicate sampling depth directly in panels.
Figure 6. Indicate depth in panels. The Y-label Temperature (°C) is missing. Consider inserting the legend for the black curves “elevation” and “wind speed” in panels “a” and “c” directly. Satellite SST data in panel “d” are hardly visible.
Figure 7. Legend on the trend on yearly warmest months could be removed as it is not showed.
Citation: https://doi.org/10.5194/essd-2024-394-RC2
Data sets
ReefTEMPS: The Pacific Island coastal ocean observation network D. Varillon et al. https://doi.org/10.17882/55128
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 229 | 52 | 8 | 289 | 10 | 7 |
- HTML: 229
- PDF: 52
- XML: 8
- Total: 289
- BibTeX: 10
- EndNote: 7
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1