Understanding processes controlling near-surface salinity in the tropical ocean using multi-scale coupled modeling and analysis

PIs: Carol Anne Clayson and James Edson (WHOI), Eric Skyllingstad (OSU)


This project uses a combination of data analysis and high resolution models to understand how precipitation, evaporation, diurnal variability, temperature stratification, and barrier layers all impact the upper ocean structure and mixing processes, and then in return how this variability affects both the air-sea exchanges of heat, moisture, and momentum and atmospheric convection and precipitation.  The goal of this project is to provide understanding of the complex processes that control the distribution of salinity in the ocean boundary layer over a wide range of spatial and temporal scales, including both oceanic and atmospheric processes, using observations and a hierarchy of modeling approaches.

Location of SPURS-2 mooring displayed on the ocean precipitation field from GPCP satellite data. The SPURS-2 mooring and the meteorological instruments. Releasing the rawinsondes that make the atmospheric boundary layer measurements during one of the cruises.

Location of SPURS-2 mooring displayed on the ocean precipitation field from GPCP satellite data. The SPURS-2 mooring and the meteorological instruments. Releasing the rawinsondes that make the atmospheric boundary layer measurements during one of the cruises.

Our investigation uses existing data sets taken in the eastern tropical Pacific Ocean during the SPURS-2 field campaign from both ships and buoys. Data from the SPURS buoy is one of the largest collections of direct observations of ocean evaporation ever made, and can be used to help improve the way that we estimate these air-sea exchanges in models, from satellites, and from other buoy and ship data.  Simultaneous atmospheric measurements using radiosondes from the ship are also available to help with the coupled model simulations.

We are using very high vertical and horizontal resolution models, in order to investigate how important many of these fairly small scale features are to the larger coupled ocean and atmosphere.  The ultimate goal of the project is to understand what types of measurements are required to shed light on these processes at the frontal region of the warm/fresh pool, and what measurements would be needed to improve the way we parameterize these processes in coarser-resolution models.

This project was funded by NOAA’s Climate Variability and Predictability program in support of NOAA’s contribution to TPOS 2020.

The SPURS-2 mooring and the meteorological instruments

The SPURS-2 mooring and the meteorological instruments.

Releasing the rawinsondes that make the atmospheric boundary layer measurements during one of the cruises.

Releasing the rawinsondes that make the atmospheric boundary layer measurements during one of the cruises.


We have completed the analysis of the SPURS-2 data, and this data has been combined with previously available data to produce a new version of the COARE algorithm, with improved bulk transfer coefficients for moisture and heat fluxes. Since the COARE algorithm is heavily used by scientists around the world for studies from individual storms to global energy and water cycle variability, an improved algorithm is a valuable asset to the community.

Using both the oceanographic data from the buoy and the atmospheric profile soundings from the cruise, we have used a combination of very high resolution ocean-only and coupled models to investigate the importance of small-scale variability on coupling within this region. Using a one-dimensional ocean model, we investigated the impact of resolving the variability in the upper meters of the ocean on the upper ocean and sea surface temperature. At lower resolutions, even at one meter resolution, the stabilizing impact of the radiation was spread out over a deeper depth than in reality, and the thermocline was too deep, leading to a cooler sea surface temperature. This was true also of rain events, as the freshwater was less mixed throughout the upper column in the high resolution model, and the pycnocline depths were greater in the lower resolution model, again leading to a cooler sea surface temperature.

This impact was also seen in our coupled WRF-ROMS atmosphere-ocean model. In this model, WRF was triple nested and the inner horizontal resolution was 4 km, where the ROMS ocean model was coupled. Comparing simulations between the vertical resolutions of the ocean model at 1 m or at 5 m, we demonstrated that the higher resolution is much more capable of simulating the diurnal warming of the sea surface temperature, as well as being able to resolve the fresh lens that occur under raining conditions. We have also investigated whether being able to resolve these shallow ocean features affect clouds and rainfall. In our ten-day simulations, we demonstrated a difference in the location of the peaks of rainfall, with the peaks of the rainfall happening closer to the equator in the higher resolution ocean model, demonstrating the coupled impact and importance of resolving the diurnal warming and the fresh lenses in this region.

Lessons Learned

Our focus in this project was providing guidance on the types of data that would be important for further process studies in the tropical Pacific region. There are three main types of data that would be important to have to help us further understand the atmosphere-ocean coupling and how accurately our models are representing this:

1. Direct flux observations are required, particularly in stable and low-wind and high-wind convective regimes, to clarify parameterization differences. The SPURS-2 dataset did not have many observations at these extremes, which is exactly where the largest differences among the flux paramaterizations occur.

2. Ocean observations of temperature and salinity with higher vertical resolution (<0.25m) in the upper few meters of the ocean. As shown by our model simulations, representing the depths where the heat and freshwater is has a strong impact on how stable the upper ocean is, and so how deep these properties get mixed down from the surface, which provides the link between the surface and the deeper ocean.

3. Atmospheric boundary layer profiles at high vertical (several meters) and horizontal resolution to understand surface-convection coupling, small-scale events such as cold pools. For understanding the ocean-cloud links, it is necessary to have a large number of events so that statistics of these events can be gathered and compared with the models.


Clayson, C. A., 2020: Small-scale ocean variability and air-sea interactions. NASA GSFC invited seminar, 13 October 2020 (virtual).

C. A. Clayson, 2020: The importance of radiometric/skin SST on air-sea fluxes. FRM4SST ISFRN Workshop, 18 September 2020 (virtual conference based in UK).

Skyllingstad, E., 2020: Modeling the effects of freshwater rain flux on convective coupling in the eastern tropical Pacific. AGU Fall Meeting, 11 December 2020 (virtual conference based in US).

Edson, J. B. 2020: Recent advances in marine platforms and sensors for air-sea interaction studies. AMS Annual Meeting, 15 January, 2020, Boston, MA.

Edson, J. B. 2019: Observations of air-sea interactions over the tropical oceans. CLIVAR workshop, 2 May 2019, Boulder, CO.

Clayson, C. A., 2019: Coupled ocean-atmosphere boundary layer measurements (with an emphasis on fluxes). 2019 US CLIVAR Summit, 7 August 2019, Long Beach, CA.


SPURS-2 data is available at the NASA PODAAC.

One-dimensional ocean model time series showing the impact of varying vertical resolutions. Upper panel: sea surface temperature from three model simulations, with “High” vertical resolution a telescoping resolution starting at under 0.1m near the surface. Second panel: thermocline difference between the three model simulations. Third panel: halocline difference between the three model simulations. Bottom panel: mixing depth difference between the three model simulations. In all cases, the lower the resolution, the deeper the solar heating and the freshwater gets deposited in the ocean, reducing the amount of stabilization that the upper ocean exhibits, impacting the surface temperature.

The data from the SPURS-2 experiment, demonstrating the large number of additional datapoints that are being used for an updated COARE parameterization, and their impact on the transfer coefficients.

Accumulated 10-day total rain amounts between the higher resolution ocean model (E15) and the lower resolution model (G2), with the shift away from the equator in the peaks of precipitation in the lower resolution model.