The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology CLIVAR WGOMD Workshop on Ocean Mesoscale Eddies, Exeter, April 2009 Trevor J McDougall CSIRO, Hobart, Australia TEOS-10, the new Thermodynamic definition of Seawater: what it is and how to use it

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology Background The existing 1980 International Equation of State (EOS-80) has served the community very well for almost 30 years. EOS-80 provides algorithms for density, sound speed, heat capacity and freezing temperature. However, it does not provide expressions for entropy, internal energy and enthalpy. All these things are best derived from a single Gibbs function from which all these thermodynamic quantities, as well as potential temperature, can be found in a consistent manner. Also, there is a little more data that has now been incorporated, making the algorithms more accurate.

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology Main Terms of Reference for WG127 To examine the results of recent research in ocean thermodynamics with a view to recommending a change to the existing internationally accepted algorithms for evaluating density and related quantities including enthalpy, entropy and potential temperature. To examine the feasibility of using simple functions of three- dimensional space to take account of the influence of composition anomalies on the determination of density in the ocean.

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology SCOR/IAPSO Working Group 127 SCOR/IAPSO WG127 on The Thermodynamics and Equation of State of Seawater Trevor J. McDougall, Chair (Australia) Rainer Feistel (Germany) Chen-Tung Arthur Chen (Taiwan) David R. Jackett (Australia) Brian A. King (UK) Giles M. Marion (USA) Frank J. Millero (USA) Petra Spitzer (Germany) Dan Wright (Canada) Associate member, Peter Tremaine (Canada)

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology Features of the new TEOS-10 Thermodynamic Equation Of Seawater SCOR/IAPSO Working Group 127 has settled on a definition of The Reference Composition of seawater. This was a necessary first step in order to define the Gibbs function at very low salinities. This Reference Composition, consisting of the major components of Standard Seawater, has been determined from earlier analytical measurements. The definition of the Reference Composition enabled the calculation of the Absolute Salinity of seawater that has this Reference Composition (making use of modern atomic weights).

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology Chemical Composition of Standard Seawater – the Reference Composition Using the available information and 2005 atomic weight estimates, mole fractions of standard seawater can be determined. The Na + contribution is determined by the requirement achieve exact charge balance. The resulting “Reference Composition” is shown to the right. Millero, F. J., R. Feistel, D. G. Wright and T. J. McDougall, 2008: The composition of Standard Seawater and the definition of the Reference- Composition Salinity Scale. Deep-Sea Research I, 55, Na + Mg 2+ Ca 2+ K+K+ Sr 2+ Cl – SO 4 2– HCO 3 – Br – CO 3 2– B(OH) 4 – F–F– OH – B(OH) 3 CO 2 Sum Solute Mole fraction Mass fraction Reference Composition

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology Reference Salinity S R is defined to provide the best available estimate of the Absolute Salinity S A of both (i) seawater of Reference Composition, (ii) of Standard Seawater (North Atlantic surface seawater). S R can be related to Practical Salinity S P (which is based on conductivity ratio) by S R = ( /35) g kg –1 x S P. The difference between the new and old salinities of ~ g kg –1 (~0.47%) is about 80 times as large as the accuracy with which we can measure S P at sea. Reference Salinity as a stepping stone to Absolute Salinity

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology Absolute Salinity Anomaly Practical Salinity S P reflects the conductivity of seawater whereas the thermodynamic properties are more accurately expressed in terms of the concentrations of all the components of sea salt. For example, non-ionic species contribute to density but not to conductivity. The Gibbs function is expressed in terms of the Absolute Salinity S A (mass fraction of dissolved material) rather than the Practical Salinity S P of seawater. S A = ( /35) g kg –1 x S P +  S A (x,y,p)

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology Determining Absolute Salinity

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology Absolute Salinity Anomaly The Absolute Salinity Anomaly  S A is determined by accurately measuring the density of a seawater sample in the laboratory using a vibrating beam densimeter. This density is compared to the density calculated from the sample’s Practical Salinity to give an estimate of  S A We have done this to date on 811 seawater samples from around the global ocean. We exploit a correlation between  S A and the slicicate concentration of seawater to arrive at a computer algorithm to estimate  S A =  S A (x,y,p). S A = ( /35) g kg –1 x S P +  S A (x,y,p)

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology Absolute Salinity Anomaly

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology Absolute Salinity Anomaly

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology Absolute Salinity Anomaly

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology Absolute Salinity Anomaly

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology Why adopt Absolute Salinity S A ? The freshwater content of seawater is (1 – 0.001S A ) not (1 – 0.001S P ), and S A and S P are known to differ by about 0.47%. There seems no good reason for continuing to ignore this known difference in ocean models. Practical Salinity expressed in the PSS-78 scale is outside the system of SI units. PSS-78 is limited to the salinity range 2 to 42. Density of seawater is a function of S A not of S P. Hence we need to use Absolute Salinity in order to accurately determine the horizontal density gradients (for use in the “thermal wind” equation).

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology What is a Gibbs Function? J. Willard Gibbs From a Gibbs function, all of the thermodynamic properties of seawater can be determined by simple differentiation and algebraic manipulation.

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology EOS-80 Density Heat Capacity Sound Speed Freezing Point Temperature: IPTS-68 Salinity: PSS-78 Seawater: TEOS-10 Temperature: ITS-90 Salinity: RCSS-08 Fluid Water: IAPWS-95 Ice Ih: IAPWS-06 New and old definitions of seawater Seawater: EOS-80

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology Percentage error in the horizontal density gradient Measurement of the horizontal gradient of density is the main way that oceanographers are able to estimate the ocean circulation (the “thermal wind” equation). These are actually differences in density at constant pressure, so what is dynamically important is

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology Percentage error in the horizontal density gradient This figure is for data from the world ocean below 10 Mpa ~ 1000 m. 60% of the data is improved by more than 2%. This improvement is mainly due to including the effects of seawater composition on the horizontal density gradient. The effect of having a better  in TEOS-10 vs EOS-80 is a factor of six smaller (the red data uses S R in place of S A ).

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology Improvements in Freezing Temperature at high pressure

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology Present practice regarding “heat” in oceanography To date oceanographers treat potential temperature  as a conservative variable. We also mix water masses on S -  diagrams as though both salinity and potential temperature are conserved on mixing. Air-sea heat fluxes result in a change in an ocean model’s potential temperature using a constant specific heat capacity. That is, we treat “heat” as being a constant times  How good are these assumptions? Can we do better?

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology The First Law of Thermodynamics in terms of  We would like the bracket here to be a total derivative, for then we would have a variable that would be advected and mixed in the ocean as a conservative variable whose surface flux is the air-sea heat flux. If we take thermodynamic reasoning leads to The First Law of Thermodynamics is written in terms of enthalpy as

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology The specific heat capacity of seawater at p = 0 dbar Specific heat capacity J kg -1 K -1

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology The concept of potential enthalpy h 0 Just as  is the temperature calculated after an adiabatic change in pressure, so potential enthalpy is the enthalpy of a fluid parcel after the same adiabatic change in pressure. Taking the material derivative of this leads to and the last two terms are very small, being no more than 0.15% of the leading term, even at a pressure of 40 MPa = 4,000 dbar. This means that the First Law of Thermodynamics can be written as

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology The difference between potential and conservative temperatures,  S

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology Error in using entropy as a heat-like variable S

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology In order to implement Conservative Temperature in an ocean model we interpret the model’s temperature variable as , and provide a polynomial for density in the form  (S A, ,p), and provide the forward and inverse algorithms  (S A,  ) and  (S A,  ). Then heat fluxes are simply fluxes of  times the constant Similarly, models should be initialized with Absolute Salinity S A and the output should be compared to observations of S A, not of Practical Salinity S P. Improving Heat Conservation in Ocean Models

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology Improving Heat Conservation in Ocean Models Conservative temperature is 100 times closer to being “heat” than is potential temperature. The algorithm for conservative temperature has been imported into the MOM4 code and it is available as an option when running the MOM4 code. The figures show the expected influence of sea-surface temperature in the annual mean, and seasonally.

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology Improving Heat Conservation in Ocean Models

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology Improving Heat Conservation in Ocean Models

The Centre for Australian Weather and Climate Research A partnership between CSIRO and the Bureau of Meteorology Recommended changes to oceanographic practice Adopt the TEOS-10 definition of the Gibbs function for seawater, requiring the use of its new algorithms for density, sound speed, enthalpy, etc, ( these algorithms are available on the TEOS-10 web site ) Adopt Absolute Salinity S A ; requiring the use of the new algorithm to go from the present conductivity-based measure of salinity, S P, to S A ( McDougall, Jackett & Millero, Ocean Science, 2009 ). Continue to report Practical Salinity S P to national data bases because (i) S P is a measured parameter and (ii) we need to maintain continuity in these data bases. Note that this treatment of salinity is similar to what we presently do for temperature; we store in situ temperature, but we publish potential temperature.