OCCAM's portrayal of ENSO
Prune STROZZI, Graham QUARTLY, Meric SROKOSZ, Andrew COWARD & Beverly DE CUEVAS
1. Introduction
OCCAM is a high-resolution global model of the ocean, run with realistic forcing (here 6-hourly reanalyses from NCEP). Ocean-atmosphere heat transfer is controlled by bulk formulae, involving the difference between OCCAM sea surface temperature (SST) and the reanalysis air temperature, but there is no direct assimilation of ocean observations. OCCAM provides a good representation of the mean state of the ocean, and has an active eddy field. We first consider the mean seasonal cycle and then look at interannual variations.
Figure 1 : Illustration of sea surface temperature field from OCCAM. Details concerning the model grid, forcing, mixing etc. can be found here.
2. Mean SST fields
First we verify that OCCAM provides a good representation of the mean sea surface temperatures (SST). For this validation we use data from the satellite sensor TMI, which has covered the tropical region 38°S-38°N for Dec. 1997 onwards.Both satellite and model data show temperatures of 28-30°C at the equator, and ~12°C at 38°S and 38°N, with similar seasonal variations. One difference is that the cold tongue in the eastern equatorial Pacific is weaker (less cold) in the model than the observations. This is most noticeable during June-November, and may relate to excessive vertical diffusion in the model.
Figure 2 : Mean SST fields for model and TMI for months during period 1998-2004.
3. Interannual changes in SST
We look at changes from the mean climatology for January. During 1998, at the peak of a major El Niño, upwelling is suppressed, so the eastern equatorial Pacific has temperatures up to 2°C warmer than normal.The La Niña phase in January 2000 shows an enhanced cold tongue in both model and satellite data. Careful inspection of the time series of anomalies shows that the features in the model lag those in satellite data by a month, and those in the model are broader, signifying ...
Figure 3 : SST anomalies relative to mean for January
4. Interanual changes in temperature profile
It is important also to consider variations with depth. For this we compare OCCAM with TAO/TRITON moorings along the equatorial Pacific. The two phases of ENSO are here illustrated by Jan 1998 and Jan 1999. During the extreme El Niño phase the warm waters (>22°C) are spread uniformly across the longitudinal range 146°E-235°E, with the warmest surface waters for 190°E-215°E. However, the hottest waters (>28°C) extend deeper in OCCAM than in the real world. The moorings also show the isotherms around 18°C dipping at 225°E (based on one mooring), whereas the deepening of those isotherms occurs 20° further east in OCCAM.During La Niña the warmest waters are at the western end of the section shown, with the 22°C isotherm 50% deeper than during El Niño. On the eastern side the extreme upwelling of La Niña leads to the 22°C isotherm reaching the surface.
Figure 4 : Temperature profiles along the equator from TAO/TRUTON moorings and in OCCAM. Temperature range plotted is 10-30°C, with 22°C highlighted.
5. Interanual changes in SSH
There are changes in sea surface height (SSH) due to the expansion o the warmest waters (steric effect) and changes in circulation. Thus comparison of SSH provides information on depth-integrated changes associated with ENSO. We use the gridded altimetry product, DUACS. The large-scale SSH changes in the model have a similar distribution to the satellite product, but the model anomalies are only half the magnitude. This suggests that the effect of interannual changes in forcing are not penetrating deep enough in the model.
Figure 5 : SSH anomalies relative to mean for January
