Impact of East Australian Current observations Tasman Sea eddies in an ocean model
Description of Work:
Can observations of the East Australian Current using a HF ocean surface radar improve model forecast skill of meso-scale eddies in the Tasman Sea? The study will use observations at Coffs Harbour (30S, 153E) which extend approximately 100 km east across the East Australian Current (EAC) and perform assimilation impact studies on a domain encompassing upstream of Coffs Harbour, the EAC separation (at approximately Smoky Cape, 31 S), and the Tasman front (across to New Zealand), with a particular emphasis on meso-scale eddies.
The HF OSR measures surface currents in the top few tens of centimetres of the ocean, on a few km resolution with a range of around 100km over 10 minute time scales. The OSR is part of the IMOS ACORN facility and is planned to commence operation in February 2012. Routine data delivery could be expected by mid 2012. Observations show the EAC is largely barotropic, so OSR should be representative of the depth-integrated current.
The Ocean Model CLAM? Assimiliation
The HF OSR provides currents in regions where the two radars overlap (and the subtended angles of the ray are greater than ~20 Outside of this region there is another equally extensive area where there is only one useful current vector component resolved. While not suited to visual interpretation, single current vector components can be assimilated into ocean models. The model already assimilates altimetry, SST and temperature and salinity profiles, so any skill improvement will be in excess of this. The assimilation of HF OSR observations may also be useful in the situation where altimetry is degraded (due to loss of satellites or other problems). It would be useful to quantify the impact of assimilating OSR currents in the absence (or reduction) of altimetry. Possible candidate data-sets for skill evaluation are feature tracking, surface drifters (these are probably drogue to a few metres depth), or synTS. The first two sources will probably generate sparse data-sets. Maybe the evaluation will be achieved by looking at the increments in SSH?
Links to other Work
We have previously looked at the impact of observations on models using the error estimates in the data assimilation system (Oke et al., 2009). It would be instructive to see how data withholding experiments compare to the observation network design study tool.
Oke, P. R., Sakov, P. & Schulz, E.W., 2009, A comparison of shelf observation platforms for assimilation in an eddy-resolving ocean model, Dynamics of Atmospheres and Oceans, 48, 121-142, doi:10.1016/j.dynatmoce.2009.04.002.