Capacity Building in Physical Chemistry for Oceanography
New survey highlights emerging skills gaps in the physical chemistry of natural waters

Earlier this year we conducted an online survey and consultation with the ocean science community to assess what we perceive as emerging skills gaps in basic physical chemistry training and expertise in several areas of chemical oceanography, especially (but not exclusively) including the ocean carbonate system. In the survey, we asked for this information:

1. Expertise, applications, and professional roles
2. Opinions concerning skills gaps in physical chemistry for different areas of oceanography and needs for capacity building

We received well over 100 responses, with very many insightful observations and answers to our questions. We note a few highlights below, and invite you to read the Survey Report describing the skills gap survey results and associated community feedback on recommended paths forward. The great majority of respondents expressed their opinions as to the nature of the skills gaps, sometimes at length, and we believe the report provides important insights into the current state of chemical oceanography, its gaps and blind spots, and novel ideas for moving forward.  

Survey feedback highlights:

• Participation was representative across career stages (almost evenly split among PhD/postdoc, mid-career, and late career).
• Top participant research areas: Ocean carbonate system, including marine carbon dioxide removal (mCDR); laboratory measurements and instrumentation; trace metals; chemical oceanography/aqueous geochemistry; chemical sensors; standards and reference materials.
• Geographical spread of respondents: 42% North America; 32% Europe; 11.6% Far East (including Australia and New Zealand); 6.7% Central and South America; 8% rest of the world.

Responses across all career stages cited a lack of appropriate training (both undergraduate and PhD level), leading to inadequate research skills, with many specific examples given. There was universal support for a workshop to discuss the issues and to recommend next steps.

Join us for a virtual community discussion at OA Week in November

To follow up on this survey, we are convening an online community discussion on Tuesday 19 November at 1600-1730 GMT/1100-1230 ET as part of the Global Ocean Acidification Observing Network (GOA-ON) Ocean Acidification (OA) Week 2024. The purpose of this discussion will be to decide next steps for a community activity (most likely a Workshop) including its focus, content, participants, and outcomes to help address the emerging skills gap identified in the survey. Please register here. We encourage you to attend even if your professional interest is not ocean acidification, as this discussion and its applications are much broader than OA.

We look forward to meeting you in the discussion session. It will be important both as a source of new ideas, priorities, and the consensus that we need to convince others – funding organisations, for example – that we have an issue that is essential to address.

If you would like further information, or you represent an organisation that would like to participate in this effort, please get in touch with either Heather Benway (hbenway@whoi.edu) or Simon Clegg (s.clegg@uea.ac.uk).

MarChemSpec has been accepted as a project within the UN Ocean Decade, which runs from 2021 – 2030. Joining the UN Ocean Decade is part of our efforts to gain improved international visibility and encourage greater use of the modelling tools.

Our Ocean Decade webpage, which is under construction, can be found at https://forum.oceandecade.org/ventures/106957.

Our modelling posters for Ocean Sciences 2024 can now be viewed here: dissolved CO2 and mineral solubilities, and inorganic complexation of GEOTRACES trace metals.

Other activities: we are making improvements to the MarChemSpec trace metal model, and have begun work on manuscripts describing thermodynamic measurements that will support development of a MarChemSpec model of the Tris buffer solutions used for calibration of pH. The measurements have been made by our collaborators at the national metrological institutes of the USA, Japan, and Germany over the last several years.

MarChemSpec at Ocean Sciences 2024, and New Release of Modelling Software

MarChemSpec is the Marine Chemical Speciation Model for acid-base speciation, trace metal complexation, and saturation with respect to solid phases in natural waters containing the inorganic species present in seawater. Project leaders Simon Clegg and David Turner will be present at Ocean Sciences and, to coincide with the meeting, we have a new release of the modelling software. The improvements make it easier to use for many oceanographic applications, and we have also updated the trace metal model to our most recent work. These are the new features:

• The software is now able to automatically adjust an input solution (seawater or some other natural water) to fixed single values, or pairs, of these quantities: total dissolved inorganic carbon, total alkalinity, total or free pH, and partial pressure or fugacity of CO₂. This makes the model simpler to use for many oceanographic and mCDR (marine CO₂ removal) applications.
• Our model of trace metal complexation has been improved for Fe(III) hydrolysis, based on work described in one of our posters (see below). This model does not yet have the new capabilities noted above.

Talk to Us About MarChemSpec Applications

Simon Clegg and David Turner will present posters on MarChemSpec applications to marine carbon dioxide removal (CM24A-1140) and trace metal complexation (OB34C-0880).

Meet Simon and David at their posters, and visit the Ocean Carbon & Biogeochemistry (OCB) Exhibit Booth (BOOTH 512, OSM Exhibit Hall, MAP) to test drive the new software and discuss applications. Their availability at the OCB booth is likely to be broadly:

• Tuesday: 10 am – 6 pm
• Wednesday: 10 am – 4 pm
• Thursday: about 10 am, or 12 noon – 1 pm

To arrange a specific time to meet, please email Simon (s.clegg@uea.ac.uk) and David (david.turner@marine.gu.se), or talk to Heather Benway or Mai Maheigan at the booth.

Simon will also be attending the Workshop to Build Consensus Toward Science-Led Protocols for Ocean Alkalinity Enhancement (Monday morning, February 19), and the mCDR Networking Event (Monday evening, February 19).

Software Downloads

The new releases of the software can be downloaded (as zip files) from the links below. At this time both the seawater and trace metal models are available only as standalone Windows programs, and as a Windows MATLAB function. (After the meeting we will prepare Linux and macOS standalone models, and the Python function. The software page on this website will also be updated later.)

• Windows standalone seawater model
• Windows standalone trace metal model
• Windows MATLAB function

Unzip the archives to your computer, and follow the instructions in the \docs subdirectories. In the cases of the seawater model and MATLAB function we have provided revised versions of the original instruction manuals, plus shorter Supplements that describe how to use the new features. For the trace metal model, there is just a revised version of the original manual. If you have any problems installing or using the software, please contact us.

Chemical Speciation Modelling Software Version 1.01, and Tutorials

The MarChemSpec modelling software and its capabilities are described in our June 2023 post, and also under the new menu item ‘Software’ (see above). The models were released on June 15-16 at an event following the Ocean Carbon and Biogeochemistry Summer Workshop. Our tutorials were recorded, and can be viewed on YouTube here.

The new version 1.01 of the software can now be downloaded. This differs from our initial release in June in these respects:

1. An error in the model for trace metal complexation has been corrected. This affects only Cu²⁺ complexation at temperatures other than 25 ^oC.
2. The documentation of the MarChemSpec Python function has been amended to describe better how directory paths should be specified to the function.

Please go to the software page on this website for downloads and a link to our Zenodo archive.

There will be a further release of the software, with greatly increased capabilities, to coincide with AGU/ASLO Ocean Sciences in February 2024. Our group will be attending the meeting.

If you would like to be added to our email list for future announcements, please send a request to Heather Benway (hbenway@whoi.edu).

The third of our papers on chemical speciation modelling is now at the pre-proof stage of publication. It describes our model of acid-base equilibria in natural waters containing the solutes present in standard seawater: the major ions (Na⁺, Mg²⁺, Ca²⁺, K⁺, Sr²⁺, Cl^–, and SO₄^2-) plus H⁺, OH^–, HSO₄^–, and the components of dissolved carbonate (CO₂*, HCO₃^–, CO₃^2-), borate (B(OH)₄^– and B(OH)₃), and fluoride (F^– and HF):

• S. L. Clegg, J. F. Waters, D. R. Turner, and A. G. Dickson (2023) Chemical speciation models based upon the Pitzer activity coefficient equations, including the propagation of uncertainties. III. Seawater from the freezing point to 45 °C, including acid-base equilibria. Mar. Chem. (pre-proof), art. no. 104196, https://doi.org/10.1016/j.marchem.2022.104196

The paper, and its extensive supporting information, includes detailed comparisons of model calculations with the measured carbonate system constants K₁, K₂, and K_B and also the dissociation product of water (K_W) and saturation of seawaters with respect to calcite. Please get in touch with Simon Clegg (s.clegg@uea.ac.uk) or David Turner (david.turner@marine.gu.se) if you would like to know more.

SCOR Working Group 145 has now transitioned to the Chemical Speciation Task Group of the Joint Committee on the Properties of Seawater which is a permanent committee sponsored by the International Association for the Properties of Water and Steam (IAPWS), SCOR, and by the International Association for the Physical Sciences of the Oceans (IAPSO). Our work will continue under this new affiliation, focusing on the development and application of the models, free software to run the models, and the extension and basis in SI of the total pH scale (in collaboration with national metrology institutes of the USA, Germany, and Japan).

This website will be revised over the following months to reflect these changes, including to become a source for software downloads and online calculations.

Our first two papers on chemical speciation modelling have just been published. They are major outputs of this Working Group, and describe the scientific basis for tools that will be provided to oceanographers later this year. These open access papers are:

• M. P. Humphreys, J. F. Waters, D. R. Turner, A. G. Dickson, and S. L. Clegg (2022) Chemical speciation models based upon the Pitzer activity coefficient equations, including the propagation of uncertainties: Artificial seawater from 0 to 45 °C. Mar. Chem. 244, art. 104095. https://doi.org/10.1016/j.marchem.2022.104095

• S. L. Clegg, M. P. Humphreys, J. F. Waters, D. R. Turner, and A. G. Dickson (2022) Chemical speciation models based upon the Pitzer activity coefficient equations, including the propagation of uncertainties. II. Tris buffers in artificial seawater at 25 ^oC, and an assessment of the seawater ‘Total’ pH scale. Mar. Chem. 244, art. 104096.
https://doi.org/10.1016/j.marchem.2022.104096

The assessment of the total pH scale in the second paper includes definitions of various forms of pH, a summary of equations relating them, and an explanation of commonly used terms.

The release of software implementing both these models, and another one of solutions containing the solutes present seawater electrolyte (inc. carbonate and borate species), will be announced on this site in due course.

Our manuscript describing the model of seawater electrolyte has just been submitted to Marine Chemistry, and the title and abstract are given below. Please get in touch with Simon Clegg (s.clegg@uea.ac.uk) or David Turner (david.turner@marine.gu.se) if you would like to know more.

Chemical Speciation Models Based Upon the Pitzer Activity Coefficient Equations, Including the Propagation of Uncertainties. III. Standard Seawater from the Freezing Point to 45 °C, Including Acid-Base Equilibria

Simon L. Clegg, Jason F. Waters, David R. Turner, and Andrew G. Dickson

ABSTRACT: A quantitative understanding of pH, acid-base equilibria, and chemical speciation in natural waters including seawater is needed in applications ranging from global change to environmental and water quality management. In a previous study (Humphreys et al., 2022) we implemented a model of solutions containing the ions of artificial seawater, based upon the use of the Pitzer equations for the calculation of activity coefficients and including, for the first time, the propagation of uncertainties. This was extended (Clegg et al., 2022) to include the Tris buffer solutions that are used to calibrate the seawater total pH scale. Here we apply the same methods to develop a model of solutions containing the ions of standard reference seawater, based upon studies by Millero and co-workers. We compare the predictions of the model to literature data for: the dissociation of dissolved CO2 and bicarbonate ion; boric acid dissociation; saturation with respect to calcite, the ion product of water, and osmotic coefficients of seawater. Estimates of the uncertainty contributions of all thermodynamic equilibrium constants and Pitzer parameters to the variance of the calculated quantity are used to determine which elements of the model need improvement, with the aim of agreeing with properties noted above to within their experimental uncertainty. Further studies are recommended. Comparisons made with several datasets for carbonate system dissociation in seawater suggest which are the most reliable, and identify low salinity waters (S < 10) as a region for which dissociation constants of bicarbonate are not yet accurately known. At present, the model is likely to be most useful for the direct calculation of equilibria in natural waters of arbitrary composition, or for adjusting dissociation constants known for seawater media to values for natural waters in which the relative compositions of the major ions are different.

We have just submitted our first two modelling papers to Marine Chemistry (the abstracts are shown below).

The first manuscript concerns the model of artificial seawater, which is the core of any speciation model of seawater and related natural waters. It particularly focuses on acidified artificial seawater, because measurements of these solutions are essential to the definition of the ‘total’ pH scale for seawater, developed by WG 145 member Andrew Dickson.

The second manuscript describes the first speciation model of the Tris buffer solutions (equimolal Tris and TrisH⁺ in artificial seawater) used to define the total pH scale. We examine the assumptions inherent in the definition of the scale, the relationship of total pH to true total hydrogen ion concentration, and prospects for the extension of the scale to low salinities. There may be implications for the calculation of carbonate system equilibria from measured total pH.

Please get in touch with Simon Clegg (s.clegg@uea.ac.uk) or David Turner (david.turner@marine.gu.se) if you would like to know more.

Chemical Speciation Models Based Upon the Pitzer Activity Coefficient Equations, Including the Propagation of Uncertainties: Artificial Seawater from 0 to 45 ^oC

Matthew P. Humphreys, Jason F. Waters, David R. Turner, Andrew G. Dickson, and Simon L. Clegg

ABSTRACT: Accurate chemical speciation models of solutions containing the ions of seawater have applications in the calculation of carbonate system equilibria and trace metal speciation in natural waters, and the determination of pH. Existing models, based on the Pitzer formalism for the calculation of activity coefficients, do not yet agree with key experimental data (potentiometric determinations of H⁺ and Cl^– activity products in acidified artificial seawaters) and, critically, do not include uncertainty estimates. This hampers both applications of the models, and their further development (for which the uncertainty contributions of individual ion interactions and equilibrium constants need to be known). We have therefore implemented the models of Waters and Millero (Mar. Chem. 149, 8-22, 2013) and Clegg and Whitfield (Geochim. et Cosmochim. Acta 59, 2403-2421, 1995) for artificial seawater, within a generalised treatment of uncertainties, as a first step towards a more complete  model of standard seawater and pH buffers. This addition to the model enables both the total uncertainty of any model-calculated quantity (e.g., pH, speciation) to be estimated, and also the contributions of all interaction parameters and equilibrium constants. Both models have been fully documented (and some corrections made). Estimates of the variances and covariances of the interaction parameters were obtained by Monte Carlo simulation, with simplifying assumptions.  The models were tested against measured electromotive forces (EMFs) of cells containing acidified artificial seawaters. The mean offsets (measured – calculated) at 25 ^oC for the model of Waters and Millero are: 0.046±0.11 mV (artificial seawater without sulphate, 0.280 mol kg^-1 to 0.879 mol kg^-1 ionic strength);  and -0.199±0.070 mV (artificial seawater, salinities 5 to 45). Results are similar at other temperatures. These differences compare with an overall uncertainty in the measured EMFs of about 0.04 mV. Total uncertainties for calculated EMFs of the solutions were dominated by just a few contributions: mainly H⁺-Cl^–, Na⁺-Cl^–, and H⁺-Na⁺-Cl^– ionic interactions, and the thermodynamic dissociation constant of HSO₄^–. This makes it likely that the accuracy of the models can readily be improved, and recommendations for further work are made. It is shown that calculated standard EMFs used in the definition of the marine ‘total’ pH scale can be accurately predicted with only slight modification to the original models, suggesting that they can contribute to the extension of the scale to lower salinities.

Chemical Speciation Models Based Upon the Pitzer Activity Coefficient Equations, Including the Propagation of Uncertainties. II. Tris Buffers in Artificial Seawater at 25 ^oC, and the Marine ‘Total’ pH Scale

Simon L. Clegg, Matthew P. Humphreys, Jason F. Waters, David R. Turner, and Andrew G. Dickson

ABSTRACT: The substance Tris (or THAM, 2-amino-2-hydroxymethyl-1,3-propanediol, CAS 77-86-1), and its protonated form TrisH⁺, is used in the preparation of pH buffer solutions for applications in seawater chemistry. The development of an acid-base chemical speciation model of buffer solutions containing Tris, TrisH⁺, and the major ions of seawater is desirable so that: (i) the effects of changes in the composition and concentration of the medium on pH can be calculated; (ii) pH on the free (a measure of [H⁺]) and total (a measure of ([H⁺] + [HSO₄^–])) scales can be interconverted; (iii) approximations inherent in the definition of the total pH scale can be quantified; (iv) electrode pairs such as H⁺/Cl^– and H⁺/Na⁺ can more easily be calibrated for the measurement of pH. As a first step towards these goals we have extended the Pitzer-based speciation model of Waters and Millero (Mar. Chem. 149, 8-22, 2013) for artificial seawater extended to include Tris and TrisH⁺, at 25 ^oC. Estimates of the variances and covariances of the additional interaction parameters were obtained by Monte Carlo simulation (Humphreys et al., submitted to Mar. Chem.). This enables both the total uncertainty of any model-calculated quantity (e.g., pH, speciation) to be estimated, and also the individual contributions of all interaction parameters and equilibrium constants. This is important for model development, because it allows the key interactions to be identified. The model was used to quantify the difference between the operationally defined total pH scale and true -log₁₀([H⁺] + [HSO₄^–]) in Tris buffer solutions at 25 ^oC, for the first time. The results suggest that the total pH scale can readily be extended to low salinities using the established approach for substituting TrisH⁺ for Na⁺ in the buffer solutions, especially if the speciation model is used to quantify the effect on pH of the substitution. The relationships between electromotive force (EMF), and pH on the total scale, with buffer concentration artificial seawater at constant salinity are shown to be linear over a reasonable range of buffer molality. The pH of Tris buffers containing ratios of TrisH⁺ to Tris that vary from unity can be very simply calculated (not requiring a model). Other technical aspects of the total pH scale, such as the extrapolation of pH to zero buffer (at constant salinity), are examined. The model was tested against measured EMFs of cells containing Tris buffer in artificial seawater at 25 ^oC, and the mean deviation (measured – calculated) was found to be 0.13±0.070 mV for salinities 20 to 40, using TrisH⁺-Cl^– interaction parameters revised in this work. Total variances for calculated electromotive forces of the buffer solutions are dominated by contributions from just a few parameters, making it likely that the model can readily be improved with respect to accuracy. Recommendations for further work are made in order to extend the model to the 0 – 45 ^oC, and reduce errors to within or close to the experimental uncertainties of the data upon which the total pH scale is based.

We presented a poster at the Ocean Carbon and Biogeochemistry Summer Workshop (2021), entitled “Chemical Speciation Models Including the Propagation of Uncertainties: Application to the Marine ‘Total’ pH Scale“. This describes our project overall: our collaborators, the ‘timeline’, some key results, and future plans. Download and view the pdf below (click on the green section headers to move between pages).

Our next posts, which will appear soon, will describe the contents of the first two of the major results of the project – evaluated and tested chemical speciation models of Tris buffers in artificial seawater, and acidified artificial seawater, and implications for the marine total pH scale.

Our paper describing measurements of Tris buffer solubilities in various salt solutions, and measurements of water activities of aqueous Tris, has just been published. The citation and abstract are given further below.

Working Group members Simon Clegg, Andrew Dickson, and David Turner, and Jason Waters of NIST (USA), are currently working on a manuscript describing insights from an uncertainty analysis of the current best Pitzer speciation model of the Tris buffers used in seawater pH measurements. Also, associate member Frank Bastkowski of PTB (Germany) has begun Harned Cell measurements to characterise the thermodynamic properties of aqueous solutions containing dissolved equimolal TrisH⁺ and Tris in an NaCl medium. The results will be used to improve the speciation model of Tris buffer in artificial seawater (in which the major solute is NaCl).

P. Lodeiro, D. R. Turner, E. P. Achterberg, F. K. A. Gregson, J. P. Reid, and S. L. Clegg (2021) Solid-liquid equilibria in aqueous solutions of Tris, Tris-NaCl, Tris-TrisHCl, and Tris-(TrisH)₂SO₄ at temperatures from 5 to 45 ^oC. J. Chem. & Eng. Data 66, 437-455. (https://dx.doi.org/10.1021/acs.jced.0c00744)

Abstract: The substance Tris (or THAM, 2-amino-2-hydroxymethyl-1,3-propanediol) is used in the preparation of pH buffer solutions for applications in natural water chemistry, including seawater. The development of a chemical speciation model of buffer solutions containing Tris, TrisH⁺, and the major ions of seawater is desirable, so that the effects of changes in the composition and concentration of the medium on pH can be calculated. The Pitzer activity coefficient equations, commonly used in such speciation models, describe the thermodynamic properties of solutions in terms of interactions between dissolved ions and uncharged solute species. To determine some of these interactions, we have measured solubilities of Tris(s) in water and aqueous solutions of NaCl, TrisHCl, and (TrisH)₂SO₄ and the solubility of NaCl(s) in aqueous Tris(aq), from 5 to 45°C. We report measurements of the water activities of Tris solutions at 293.5 K to high supersaturation with respect to the solid. Using the Pitzer equations, we compare our results to literature data yielding stoichiometric dissociation constants of TrisH⁺ in aqueous NaCl, and to electromotive forces of cells containing dissolved Tris, TrisHCl, and NaCl. Values of parameters for the interactions of Tris with the ions TrisH⁺, Na⁺, and SO₄^2- at 25°C are determined.

Much has happened in the world since our last news item in March. However, we have been able to continue our work, even experiments, and have made good progress over the spring and summer.

First, a manuscript describing measurements of Tris buffer solubilities in various salt solutions, and measurements of water activities of aqueous Tris, is in the final stages of review by J. Chem. and Eng. Data. These measurements, carried out by Pablo Lodeiro at GEOMAR and Flo Gregson at the Bristol Aerosol Research Centre, provide important information about the interactions of the buffer substance Tris with other salts. The results are bring used to improve the chemical speciation model of Tris buffer solutions (for the calculation of pH) that we are developing. In a separate study, water activities of Tris solutions at various temperatures have just been measured by our collaborators Tian Xiaomeng and Prof. Chak Chan at City University of Hong Kong. A programme of Harned Cell measurements to be carried out by WG 145 associate members Regina Easley (NIST) and Frank Bastkowski (PTB, Germany) is being finalised.

Second, at the Ocean Sciences meeting in San Diego in February we showed draft software for the calculation of the pH of Tris buffers in artificial seawater, and the K* of the carbonate system in standard seawater. The novel features of these models are: (i) the ability to vary the composition of the solution medium from seawater stoichiometry; (ii) the calculation of uncertainty contributions from all the individual elements of the model. The draft software is still available to try here. Since the meeting we have refined the uncertainty treatment, and a manuscript describing the methods and results is currently in preparation. Here is a result we’d like to show you: ranked uncertainty contributions (as percentages of the total calculated variance) to a model-calculated pH of Tris buffer in artificial seawater of salinity 35.

The key finding is that, of the very many interactions represented by parameters in the chemical speciation model, only a very few matter. This is extremely helpful for model development.

What about dissolved organic matter, and trace metals? Due to its polydisperse nature, the thermodynamics of natural organic matter (NOM) cannot be treated using the Pitzer approach that we are applying to standard seawater and Tris buffers. However, NOM is important as a trace metal complexant, and also as a contributor to titration alkalinity in lower salinity waters.  Martha Gledhill and Pablo Lodeiro at GEOMAR  have begun using the NICA-Donnan model, developed for freshwater NOM, to characterise the chemistry of marine NOM. Together with David Turner at the University of Gothenburg, they are assessing the use of a novel model code that combines the Pitzer and NICA-Donnan approaches. The first application to be studied is a new data set from the Amazon estuary.

At the recent SCOR Meeting the lifetime of WG 145 was extended for at least a further year in recognition of what we have achieved so far and of our current and planned activities.