Interpreting Modeled Oceans Across CMIP Eras and the Latest Observations
Lawrence Livermore scientists in the Atmospheric, Earth, and Energy Division developed a new framework that allows ocean climate models to be most directly compared across numerous CMIP generations alongside the latest generation observations. These models have been in development for sixty years, with dramatic increases to model realism a feature of the latest generation. Alongside these developments, the observation-based Equation of Seawater (EOS) and our understanding of the properties of seawater have been evolving since the first physical oceanographic insights were published in the early 19th century, with the first EOS defined off the density measurements of Forch, along with the measurements of Knudsen and Ekman in the early 20th century. These varied insights were merged by the work of Fofonoff et al. and Sweers into the first EOS variant which served until 1980, then replaced by Joint Panel for Oceanographic Tables and Standards (JPOTS) by the International Equation of State of Seawater (EOS-80; ICES et al., 1981; for details see Millero, 2010). Flaws and inconsistencies in the thermodynamic properties of seawater that underpinned the EOS-80 definition were identified, and in 2010 these issues were resolved with the redefined Thermodynamic Equation of State of seawater (TEOS-10; IOC et al., 2010).
The EOS used in any one model configuration underpins the accurate calculation of key derived quantities, that are defined by fixed variables, equations, and coefficients and are often difficult to decipher. As numerous variants of EOS are used across the CMIP archive, approximations are most often used in practice introducing an inconsistency to model intercomparison studies. Considering the global ocean is responsible for the bulk of the 1971-2018 Earth System excess heat storage (396 ZJ, 91%) due to anthropogenic influence, even small analysis inconsistencies may lead to large (~ZJ scale) error propagation. As the role of transient ocean heat uptake plays such a large role in Earth’s energy budget, any changes to ocean operation will strongly impact realized near-surface warming – so accurate analyses are of the utmost importance.
This work outlines a framework to compare model ocean fields most directly across numerous CMIP eras and EOS variants. This facilitates the direct comparison of model fields with the latest Thermodynamic Equation of State (2010; TEOS-10) generated observational insights, which account for aspects of the real world which were ignored by all previous EOS variants. The work also advocates for modelers to define and document the fixed variables, equations, and coefficients that are required for the correct calculation and comparison of model-derived quantities to observations more clearly. In addition, it advocates for the migration toward the TEOS-10 standards of model configuration and initialization, aiming to align models and observations more tightly in future CMIP phases.
IOC, SCOR, and IAPSO (2010) The International Thermodynamic Equation of Seawater—2010: Calculation and Use of Thermodynamic Properties. Intergovernmental Oceanographic Commission, Manuals and Guides No. 56, UNESCO (English), Paris, 218 pp. Available online at: http://www.teos-10.org/pubs/TEOS-10_Manual.pdf (accessed November 8th, 2021)
Millero, F. J. (2010) History of the equation of state of seawater. Oceanography, 23 (3), pp 18-33. DOI: 10.5670/oceanog.2010.21