PI and organization: I. Karagali (DTU Wind Energy)

Co-Is: J. She (DMI), J. Hoyer (DMI), J. Murawski (DMI)

Sea Surface Temperature (SST) is considered an essential variable by the Global Climate Observing System (GCOS) due to its important role in climate related, bio-chemical and oceanic CO2 processes. The SST diurnal variability (Figures 1 & 2), driven by the coincident occurrence of low wind and solar heating, is not properly resolved in CMEMS models and products. Not resolving SST variability results in biases of the total heat budget estimates, thus, demised model accuracies. In addition, diurnal SST variability complicates merging of SSTs from different satellite sensors which impacts efforts to create climate datasets and L4 SST products. Not resolving the diurnal variability of the upper ocean temperature may additionally result in large errors when modelling harmful algal blooms.





Figure 1. Vertical distribution of SST during night or under strong wind conditions (red line) and during the day under light winds (black line). From Minnett P. & Kaiser-Weiss A. (2012)








Figure 2. Example of diurnal signals in SEVIRI SST retrievals, on June 28 2018.

DIVOST-COM aimed at resolving the centimeter-scale vertical temperature structure of the upper water column by representing diurnal variability of the Sea Surface Temperature (SST), through developing and using the existing 1-dimensional General Ocean Turbulence Model (GOTM). The project rationale was to “develop and integrate a diurnal variability model with HBM, i.e. the Baltic MFC 3-D physical-biological model, and the Level 4 SST, i.e. one of the SST TAC products for the region, thus improving the CMEMS modelling and satellite products for the Baltic Sea”.  This was achieved by using GOTM and assessing its ability to represent:

  1. diurnal variability of the upper ocean as observed from infra-red SST retrievals (SEVIRI).
  2. night-time upper ocean temperature similar to L4 SST products
  3. bulk SST as simulated by HBM.

In DIVOST-COM, sensitivity tests performed to assess the diurnal variability seen from hourly satellite SST retrievals from SEVIRI and simulated by GOTM and HBM showed that mean SEVIRI diurnal variability could reach or exceed 1°C at test locations and that was better represented by GOTM compared to HBM (Figure 3).


Figure 3. Statistics of mean diurnal variability (left) and its variance (right) from 4 months of SEVIRI SST (red), GOTM (blue) and HBM (yellow) simulations at selected test sites.







Sensitivity tests were also performed to assess the optimal GOTM set-up, especially focusing on the parameterizations for the absorption of light within the water column. It was found that an added 9-band parameterization could be more representative than the standard 2-band option included in GOTM (Figure 4).


Figure 4. PDF of the differences of hourly SEVIRI SST minus GOTM (blue) and SEVIRI minus HBM (pink) for all grid points with diurnal variability of 1°C and more, during the day-time period from 07:00 to 19:00 for four dates with strong diurnal signals. In the left panel, GOTM simulations use the standard 2-band parameterization for light absorption, in the right panel the added 9-band parameterization. 



GOTM was used to simulate the Baltic Sea temperature of the upper 25 m, from June to August 2018. Mean temperature anomalies, i.e. deviations from the night-time well mixed conditions, from GOTM matched SEVIRI in spatial and temporal extend (see example in Figure 4). DIVOST-COM has also provided some insight on the representative depth of the L4 SST analysis and SEVIRI SST, through direct comparisons with GOTM, indicating lower biases within the upper 5 meters and the upper 20 cm, respectively.

Figure 5. Example of mean temperature anomalies from the corresponding night-time foundation values, for GOTM (left), SEVIRI (middle) and HBM (right) for June 01, 2018.

GOTM and its supporting infrastructure developed for DIVOST-COM in a semi-operational mode can be directly applicable to CMEMS MFC and relevant TAC systems, thus allowing for usage by developers who are interesting in performing similar investigations for other CMEMS relevant areas, e.g. the Mediterranean and Black Seas. Results and infrastructure from DIVOST-COM can be applied for producing and delivering a new product for SST diurnal variability or provide an improvement, as an added variable, to existing products.


Minnett, P.J., and Kaiser-Weiss, A.K. (2012). Near-Surface Oceanic Temperature Gradients. Discussion Document, Group for High Resolution Sea-Surface Temperature. 7 pp. https://www.ghrsst.org/wp-content/uploads/2016/10/SSTDefinitionsDiscussion.pdf