Background & Summary

The PIRATA (Prediction and Research Moored Array in the Tropical Atlantic) program has been conducting annual oceanographic cruises in the Tropical Atlantic Ocean since 1997 to study the ocean-atmosphere dynamics and their impact on climate1,2,3. PIRATA is driven not only by fundamental scientific questions, but also by the societal need to improve the predictability of temporal and climatic variability and their associated events. This program is maintained by a close collaboration between institutions from the United States (National Oceanic and Atmospheric Administration [NOAA]), Brazil (Instituto Nacional de Pesquisas Espaciais [INPE], with contribution from Diretoria de Hidrografia e Navegacão [DHN]) and France (French Institut de Recherche pour le développement [IRD] and Météo-France). Since the beginning of the program, an IRD team in France, which became the IMAGO team http://imago.ird.fr/ in 2007 has managed the technical operations in the fields of physics and marine chemistry. At now, the French PIRATA team manages 6 meteo-oceanic buoys located in the eastern tropical Atlantic and the Gulf of Guinea, along with 3 equatorial current moorings. Figure 1 presents the buoys network managed by the Brazil, France and USA. Physical oceanographic and meteorological parameters are mainly collected by the continuous systems implemented on the buoys and through CTDO2 (Conductivity, Temperature, Depth, Dissolved Oxygen) vertical profiles. Oceanographic and meteorological measurements from the buoys are transmitted in real time by the Argos or Iridium systems (depending on the type of buoys) and made available by the Global Telecommunication System, the PIRATA NOAA’s Pacific Marine Environmental Laboratory (PMEL) website, and via anonymous ftp (https://www.pmel.noaa.gov/gtmba/pirata). Since 2004, during the annual PIRATA French cruises, scientists have seized the opportunity to increase the number of biogeochemical parameters and their spatial coverage. Hydrographic stations along specific longitudes (principally 10∘W) have been set up on a recurring basis each year. Over the years, we have increased the number of parameters measured. This has been achieved by increasing the number of technical personnel and scientific equipment that can be taken on board these missions. The measured parameters are salinity, dissolved oxygen, nutrients (since 2004), phytoplankton pigments (since 2011), marine pH (since 2021) and total alkalinity (since 2023). Our aim with these samplings is to provide high quality data for studying the spatial and temporal variability of these parameters, to gain a better understanding of the underlying processes that drive the ocean’s biogeochemical cycles4 and to help identifying the impacts of climate change on ocean acidification in the Tropical Atlantic5. These biogeochemical data are recorded, stored and made freely available to the public. They are accessible through the open scientific data repository SEANOE (SEA scieNtific Open data Edition)6. This paper is a presentation and a-state-of-the-art of this biogeochemical dataset7.

Fig. 1
figure 1

The PIRATA buoys network: red marks (French), white marks (US) and green marks (Brazilian).

Full size image

Methods

Sample Collection and Analysis

Two categories of sample collection are occurred. Firstly, surface waters are collected from the thermosalinograph system, at regular intervals (every GPS degrees in latitude or longitude) throughout the ship’s journey. Secondly, seawater samples are collected on CTD-O2 casts, particularly along the 10∘W section. A number of other specific profiles are also produced. Figure 2 shows the distribution of sampling stations from the beginning. Seawater samples are taken using NISKIN 8 L bottles mounted on a rosette. In accordance with standard operating procedures, we have established a systematic approach to closing the Niskin bottles. Furthermore, we adjust these sampling levels based on the results obtained from the descending CTD profile. The CTD unit is a SeaBird SBE911+ model since the beginning of the sampling. The number of profiles and samples has evolved and increased over the years. This number is highly dependent on the available vessel time. Priority is given to mooring operations, and problems can occur, as vandalism2,3. The Table 1 summarizes the French PIRATA cruises since the beginning and the evolution of the parameters sampled. Tables 2, 3, 4 and 5 are descriptions of parameters with the vocabulary used in the MEDATLAS database. Water masses share common characteristics and properties. These include both conservative (potential temperature and salinity) and non-conservative (oxygen, silicate, phosphate and nitrate) variables. These six key properties are based on observational data and allow to identify the water masses8. Figure 3a,b,c,d,e show the vertical distribution of these variables, allowing, for example, the characterisation of the Eastern South Atlantic Central Water (ESACW) at 500 m depth. Figure 3f,g show the vertical distribution of the carbon parameters in relation to the other parameters, which provide information on the uptake of the carbon parameters.

  • Salinity. We use the method recommended by the World Ocean Circulation Experiment (WOCE) Hydrographic Programme (WHP). Since 1978, oceanographers have employed the Practical Salinity Scale (PSS78)9, which defines salinity as the ratio of the electrical conductivity of a seawater sample to the conductivity of a standard potassium chloride solution in which the mass fraction of potassium chloride is 32.4356  × (10−3)10. We use a Guidline 8410A Portasal salinometer, installed in an air conditionned lab onboard. The salinometer water bath is adjusted to be in the same range than the room temperature (+/−1 °C). The samples bottles are stored 24h before the analysis in the air-conditioned lab11. To collect of the salinity samples, we use glass bottles with disposable plastic insert and bakelite screw cap11. We use reference material to calibrate and validate our salinity measurements: IAPSO (International Association for the Physical Sciences of the Oceans) standard seawater, distributed by Ocean Scientific International (OSIL). Since 2021, the implementation of an automation recording for the salinity measurements represents a significant advancement in terms of enhancing data accuracy and efficiency. This approach eliminates the need for manual data entry, ensuring consistency, reliability in the collection of values, avoids copying errors and saves time on calculations requiring salinity values (D.O., pHt and total alkalinity). The acquisition is automatically recorded using a Python program developed by our colleagues at the Laboratoire d’Océanographie Physique et Spatiale (LOPS) at Brest. Figure 3a shows the vertical distribution of the salinity during the PIRATA FR34 mission in 2024. Technical validations are described in a following chapter and Tables 6 and 7.

  • Dissolved Oxygen. Dissolved oxygen is collected and analyzed on board using the method described by Culberson et al12. This is a potentiometric measurement method based on the classic Winkler iodometric method13. A solution of manganous chloride (MnCl2) is added to a known quantity of seawater, and this is immediately followed by the addition of an alkaline sodium hydroxide-sodium iodide solution (NaOH/NaI). The principle of the reaction is as follows: the addition of manganous hydroxide (Mn(OH)2) precipitates and reacts with the dissolved oxygen in the water, forming a hydrated tetravalent oxide of manganese (MnO(OH)2). Upon acidification, the manganese hydroxides dissolve. In the acid solution, the tetravalent manganese in MnO(OH)2 acts as an oxidizing agent and liberates iodine (I2) from iodide ions (I−). The requisite amount of thiosulfate for the titration of each mole of I2 is two moles. Since two moles of I2 are formed for each mole of O2, the resulting stoichiometric equation is four moles of thiosulfate equaling one mole of O2. By determining the concentration of the thiosulfate solution and the volume required to titrate the liberated iodine, the amount of dissolved oxygen in the seawater sample can be readily calculated. All along the years, we have used 2 Metrohm titrator models with a 10 mL buret and employing a platinum electrode. The Metrohm titrator automatically and very precisely delivers a volume of thiosulfate by dynamically detecting the equivalence point. Over time, the systems have evolved from the 798 to the 848 model (see table 7). The typical measurement deviation of the 10 mL cylinder is +/− 20 μL. The principle of sampling is based on the recommendations made by Go-Ship (experts reports and guidlines)14 and Culberson12. We use glass bottles with long ground glass stoppers that are precisely calibrated in volume (about 120 mL). The primary precaution to be observed is the prevention of air bubbles becoming entrapped within the glass bottles during the process of collecting the sample. Prior to embarking on the cruise, the thiosulfate solution and also the potassium iodate solution (Supelco, potassium iodate volumetric standard, secondary reference material for iodometry, traceable to NIST SRM Certipur©) are prepared. This solution is employed to calibrate the sodium thiosulfate solution used in the titration process. The thiosulfate concentration is monitored on a daily basis. This calibration also encompasses a control with a 0.02 N standard potassium iodate solution provided by OSIL. Figure 3b shows the vertical distribution of the dissolved oxygen during the PIRATA FR34 mission in 2024. Before the cruise, an intercomparison exercise with other labs is performed to qualify the reagent and the method. Technical validations and performances are described in a following chapter and Tables 6 and 7.

  • Dissolved inorganic nutrients. Dissolved inorganic nutrients (nitrite, nitrate, phosphate and silicate) are collected and pasteurized according to the methods described by Aminot et al.15 and Daniel et al.16. Important precautions must be taken during sampling to avoid contamination or damage. In the case of the nutrients, it may be advisable to use gloves when collecting. For example, powder free vinyl gloves could be recommended for sample collection at sea. After three rinses, seawater is sampled in 30 mL narrow-mouth polypropylene bottles. The stopper is tightened with pliers before being pasteurized in an oven at 80°C for 2.5 hours. Then, samples are stored at room temperature protected from light in a dedicated crate for this use. The long-term preservation of nutrient samples is improved by protecting them from the possible photolysis of the organic matter present.

    The nutrients analysis are realized in our lab on shore. Analyses have been performed using the Seal Analytical AA3 HR system from 2004 to 2023. Currently, analyses are being conducted with the new Continuous Segmented Flow Analyzer (CFA) AA500. The analyses of the various nutrients are described in detail in the book by A. Aminot and R. Kerouel17. Another reference document for these analyses is the experts report written by Becker et al.18 from the GO-SHIP repeat hydrography manual. The baseline solution employed is ultrapure water. The standards solutions are prepared by weighing ultrapure salts of the nutrients, and the secondary standards for calibration are prepared in low-nutrient seawater (LNSW), which is sourced from the cruise. Approximately 200 litres of LNSW are collected near the equator. This nutrient-poor matrix is subsequently utilised in the laboratory to prepare the water employed in the creation of working standards, which are utilised to generate calibration curves during the course of analysis. It is imperative to employ a matrix that is as proximate as possible to the analysed seawater in order to circumvent the influence of optical effects on the continuous flux analysis. Aminot and Kerouel17 set out the detection limits of the nutrients as follows: nitrate, 4 nmol/L; nitrite, 2 nmol/L; phosphate, 3 nmol/L; and silicate, 0.04 μmol/L. Figure 3c,d,e show the distribution of the nutrients during the PIRATA FR34 mission in 2024. Technical validations and performances are described in a following chapter and Tables 6 and 7.

  • Phytoplankton pigments. The sampling of seawater for the determination of phytoplankton pigments have started in 2011. Six levels of depth are collected from the surface to a depth of one hundred meters. One liter of seawater is filtered onto a 25 mm GFF filter. Pigments are very sensitive to light. The filters are preserved in aluminium foil and must be rapidly frozen and stored at  − 80°C to prevent degradation during and after the cruise until analysis. A total of 23 distinct pigments are analyzed, with the list provided below. Chlorophyll c3, Chlorophyll (c1,c2), Peridinin, 19’Butanoyloxyfucoxanthin, Fucoxanthin, Neoxanthin, Prasinoxanthin, Violaxanthin, 19’hexanoyloxyfucoxanthin, Astaxanthin, Diadinoxanthin, Alloxanthin, Diatoxanthin, Zeaxanthin, Lutein, Divinyl chlorophyll-b, Chlorophyll-b, Alpha-carotene, Beta-carotene, Pheophorbide a, Divinyl chlorophyll-a, Chlorophyll-a, Pheophytin-a. The analytical procedures were conducted using a 1200 HPLC Agilent Technologie system with a diode-array detector, in accordance with the methodology described in the paper by Van Heukelem L. and Thomas C.19, as well as the technical chapter by Ras J. and Claustre H. in the NASA Technical publication20. Pigments are extracted in “spiked” methanol. This refers to methanol containing vitamin E acetate (3 mL of vitamin E acetate stock solution at approximately 2.5g.L-1 in 1 liter of methanol) which is employed as an internal standard to calculate the extraction yield. The extraction yield is calculated from the area of the internal standard peak in the absence of samples (three runs at the commencement of the analysis sequence) and the area of the internal standard peak with sample. The internal standard, i.e. the vitamin E acetate peak, is detected at 222 nm, with no interference from phytoplankton. The pigments are then separated on a C8 column, 3 mm in diameter, 150 mm long and with a porosity of 3.5 μm. The column used was Agilent Technologie’s ZORBAX Eclipse XDB-C8. The software, Chemstation©from Agilent, generates a chromatogram for each sample. The peaks corresponding to the pigments are integrated, and the pigments are identified based on their retention times and spectra, with quantification performed by calculating the area of each peak. The concentration is obtained from the calibration lines for each pigment, which are certified pure pigments marketed by the DHI Group (Lab products) in Denmark and accompanied by a certificate of analysis for each pigment. Figure 4 shows the distribution of the chlorophyll a (as example of a pigment) during the PIRATA FR34 mission in 2024 along the 10W section. Technical validations are described in a following chapter.

  • pH (total scale). Measuring the pH of seawater is a challenge because the conventional method of measurement, by electrochemistry, can be distorted because of the ionic strength of seawater, implying an uncertainty in the measurement of  ±0.02 pH units21. A UV-visible spectrophotometric method was proposed by in 199322, requiring the use of meta-cresol purple, a coloured indicator. The colour of this indicator varies over the pH range of seawater (7.4-8.4) and provides a determination with an uncertainty of ±0.003 pH units. Since 2021, we have been measuring pH in the marine environment (total scale pHt) by spectrophotometry using a dye. The measurements are carried out in accordance with the descriptions and measurement protocols set out in various scientific publications that are recognized by our field, eg. Clayton et al.22, Dickson et al.23, Liu et al24. The samples are collected following the recommendations from the reference papers. From the niskin bottle, a Tygon tube is introduced into the bottom of a 500 mL glass bottle. Once the bottle is full, the sample water is allowed to overflow during several seconds before the tube is removed and the bottle is capped without the presence of trapped air bubbles. Then, the samples are analyzed as soon as possible. The reading cells are filled with the seawater and placed in a Julabo CORIO CO water bath, which is set to a temperature of 25°C. We use m-cresol purple at a concentration of 2 mM as the color indicator. This indicator was prepared before the cruise in the laboratory. The cresol is supplied by the University of South Florida and diluted to a concentration of 2 mM. A Shimadzu spectrophotometer fitted with a 10 cm cell holder is employed for the measurements. The temperature of the measuring cells (25°C) is maintained by the Peltier system of the Northwest Quantum©cell holder. Figure 3f shows the distribution of the pHt during the PIRATA FR34 mission in 2024. Technical validations and performances are described in a following chapter and Tables 6 and 7.

  • Total alkalinity. In order to determine the total alkalinity, the same bottle sampling method that used for pH measurement is employed. Potentiometric measurements of total alkalinity are conducted onboard during the PIRATA FR cruises from 2023. During the 2023 cruise, the analyses were conducted in accordance with the standard operating procedure SOP3b25, which outlines the methodology for determining total alkalinity in seawater via open-cell titration. Each sample was analyzed two or three times to ensure the highest possible degree of accuracy. The analysis time was approximately 10 minutes, which limits the number of samples that could be analyzed on board. In 2024, a decision was made to change the analytical methodology employed in order to reduce the time required for analysis and to enable the analysis of a greater number of samples on board, while maintaining equivalent precision. The analyses were thus conducted in accordance with the two-end point determination described by the protocol published in the article by Fiz Perez et al. (Marine Chemistry, 1987)26. The analysis time is approximately four minutes, with each sample analyzed twice. The resulting data was then averaged to obtain the most accurate value possible. An 888 Metrohm titrator is employed, controlled by OMNIS software, to perform the analysis. The temperature of the samples is monitored by a Pt 100 temperature probe. The pH metric determination is conducted using a hydrochloric acid solution with a concentration of 0.03 N, which is tested with Dickson CRM (batch 203). The volume of the sample is measured using a Knudsen pipette, which had previously been calibrated gravimetrically at the Brest laboratory (approximately 50 ml determined to 1/100th of a ml). The sample is aspirated using a KNF pump. Figure 3g shows the distribution of the total alkalinity during the PIRATA FR34 mission in 2024. Technical validations and performances are described in a following chapter and table Tables 6 and 7.

Fig. 2
figure 2

Distribution of sampling stations (profiles and surface).

Full size image
Table 1 List of PIRATA cruises and biogeochemical samples collected since 1997.
Full size table
Table 2 Descriptions of the nutrients parameters following the information provided in the library P09 of the “MEDATLAS Parameter Usage Vocabulary” hosted and managed by the British Oceanographic Data Centre (BODC) by means of the NERC Vocabulary Server (NVS2.0).
Full size table
Table 3 Descriptions of the dissolved oxygen and salinity following the information provided in the library P09 of the “MEDATLAS Parameter Usage Vocabulary” hosted and managed by the British Oceanographic Data Centre (BODC) by means of the NERC Vocabulary Server (NVS2.0).
Full size table
Table 4 Descriptions of the carbon cycle following the information provided in the library P09 of the “MEDATLAS Parameter Usage Vocabulary” hosted and managed by the British Oceanographic Data Centre (BODC) by means of the NERC Vocabulary Server (NVS2.0).
Full size table
Table 5 Examples of few descriptions of the phytoplankton pigments following the information provided in the library P09 of the “MEDATLAS Parameter Usage Vocabulary” hosted and managed by the British Oceanographic Data Centre (BODC) by means of the NERC Vocabulary Server (NVS2.0).
Full size table
Fig. 3
figure 3

Vertical distributions of biogeochemical parameters measured during PIRATA FR34: salinity (a), dissolved oxygen (b), NOx (c), Phosphate (d), Silicate (e), pH (f) and total alkalinity (g).

Full size image

Data Records

The chemical analysis dataset from the PIRATA French cruises is available at the SEANOE repository: https://www.seanoe.org/data/00470/58141/7. The data is available in a single zip file containing 3 folders: Surface (sea surface samplings), Reports (chemistry cruise report) and BTL (water column discrete samplings). The description of the dataset is summarised as follows: “Bottle” (during the CTD-O2 upcasts) and Surface data (discrete samples of salinity, oxygen, pigments, nutrients, pH and alkalinity) collected in the Gulf of Guinea during cruises carried out in the framework of the programs ≪Prediction and Research morred Array in the Tropical Atlantic≫2. The files provide a substantial amount of information. At the beginning, the columns start by the metadata. A description of the metadata is outlined below:

Cruise : name of the cruise.

Station : number of the sampling station.

Type : B for bottles station.

yyyy-mm-ddThh:mm:ss : date.

Longitude [degrees_east], Latitude [degrees_north] : geographical position.

Table 6 List of the Certified Materials (CRM) and standards used for the analysis.
Full size table

Bot. Depth [m] : bathymetric depth .

Table 7 Use of the analytical instruments in the marine chemistry laboratory of the IMAGO team.
Full size table

DEPTH [m] : depth of the sampling bottle.

Fig. 4
figure 4

Vertical distribution of the Chlorophyll a (CPHL) along the 10W section, PIRATA FR34.

Full size image

Bottle nb : number of the sampling bottle on the rosette.

The following columns concern the parameters collected and measured. The descriptions of the column headings are given in Tables 2, 3, 4, 5. Parameters descriptions are provided in the library P09 of the ≪ MEDATLAS Parameter Usage Vocabulary ≫ hosted and managed by the British Oceanographic Data Centre (BODC) by means of the NERC Vocabulary Server (NVS2.0). More informations could be found through the SeaDataNet services https://vocab.seadatanet.org/search. To go further, the biogeochemical data obtained from the PIRATA French cruises described above are transmitted to the SEANOE (SEA Scientific Open Data Edition) services (https://www.seanoe.org/html/about.htm). As illustrated in Table 2, the following nutrients parameters are described in accordance with the information provided in the library P09 of the “MEDATLAS Parameter Usage Vocabulary”. For each parameter, the acronym used in the database, the units, and a brief description are provided. As illustrated in Table 3, the information regarding salinity and dissolved oxygen are described by the same way than in the Table 2. Correspondingly, Table 4 provides data on carbon parameters, including pH and alkalinity, while Table 5 presents information on phytoplankton pigments. These tables represent the data collection and the way they are saved in the dataset repository SeaDataNet services https://vocab.seadatanet.org/search.

Technical Validation

The IMAGO marine chemistry laboratory adheres to rigorous standards of validation in all aspects of biogeochemical determinations. Our measurements of accuracy, trueness, and precision align with the definitions set forth in the publications of the JCGM (Joint Committee for Guides in Metrology)27, and our team is ISO 9001 certified since 2009. This certification attests to our commitment to excellence in laboratory practices. We build control charts to easily control the precision and accuracy of our measurements. We follow procedures described in reference documents25,28. During all French PIRATA cruises, for each CTD cast, we take two samples at the same pressure and therefore have a duplicate of the same water mass. This duplicate changes with each cast to test the watertightness of the Niskin bottles. This test is done for each biogeochemical parameters except for the pigments. Another procedure conducted on board is the CTD cast test, whereby all Niskin bottles are closed in order to sample a single water mass. Replicates of the parameters are collected in order to determine the precision of the measurements. Figure 5 presents an example control chart for the deviation of salinity doublets measured during the PIRATA FR34 mission. The observed outlier values are primarily attributable to deficiencies in the Niskin bottles closure during the cast. The validation of the analytical measurements is achieved through the use of certified reference materials (CRM) (Table 7). The purpose of the CRM is to act as a calibration standard for analytical systems, as well as to validate the accuracy of the measurements. The CRM are employed for salinity11, nutrients18,29, pH and total alkalinity30 measurements. Figures 6 and 7 depict examples of control charts for the nutrients (Nitrate, Phosphate and Silicate) during the PIRATA FR34 mission. Following the distribution of the measurements, we employ various lots of certified reference materials (CRMs) provided by Kanso Technos Co. (Table 6). No certified reference material is available for determining the dissolved oxygen in seawater. The thiosulfate calibration employs a standard iodate solution provided by OSIL, which serves as a reference for comparing the routine calibration with the iodate solution prepared in-house, thereby ensuring the reliability of the calibration. The analytical performance criteria for the phytoplankton pigments determination are evaluated by the use of internal and external standards. The internal standard, vitamin E-doped methanol, is injected at the beginning of the sequence, with the aim of calculating the extraction yield. The recommended external standard, in each series of analyses, is a DHI-certified sample containing several pigments (MIX pigments). A calibration curve is established for each of the 23 pigments. The pure pigments standards are supplied by DHI, accompanied by a certified sheet indicating the purity and concentration. Our team contributes to experts working groups. We organize and participate to yearly workshops on the best working practices for dissolved oxygen or carbon parameters (Total alkalinity and pHt) https://www.odatis-ocean.fr/en/activities/technical-workshops. We participated also to interlaboratory exercices29 for nutrients analysis https://www.jamstec.go.jp/scor/calibration.html. All these activities allow us to maintain contact with our colleagues in the different laboratories to improve our practices. Pasteurization15,16 offers a means of conserving nutrients. A collaborative study was conducted with colleagues from the Ifremer DYNECO Laboratory. The objective is to ascertain whether the methodology employed for storage of the nutrients is a viable and effective solution. They participated to the International Nutrient Inter-comparison Voyage (https://wp.csiro.au/iniv/) in 2023 hosted by CSIRO NCMI Hydrochemistry. The objective of this project was to assess the inter-comparability of seawater nutrient measurements and to identify the most effective shipboard methodologies for reducing error between different international laboratories. Samples were also collected and pasteurized for analysis in France several weeks later. The storage method yielded satisfactory results, which will be published in due course.

Fig. 5
figure 5

Deviation of salinity doublets measured during the PIRATA FR34 mission.

Full size image
Fig. 6
figure 6

Control charts for the nutrients using the KANSO RMNS (lot CR). The solid lines are the targets and the dashed lines are the warning limits (targets  ± 2σ).

Full size image
Fig. 7
figure 7

Control charts for the nutrients using the KANSO RMNS (lot CH). The solid lines are the targets and the dashed lines are the warning limits (targets  ± 2σ).

Full size image