Remote Sensing for Ocean Circulation and Prediction Greg Jacobs Naval Research Lab Oceanography As Reported by Bryan Carbol On April 21 2004 Greg Jacobs presented Remote Sensing for Ocean Circulation and Prediction to ASEN 6210 Remote Sensing Seminar Ocean assimilation is an effort to construct the best estimate of the current ocean state This is done using currents thermodynamics and all other properties that help define the state of the ocean This is a difficult proposition due to the inability to observe everything and computational limitations to modeling everything This presentation gave an overview of the Naval Research Lab s effort to incorporate all knowledge of the ocean in order to present the best estimate in one place Ocean assimilation serves many purposes By using the melding of observed data and complex modeling it is possible to better predict non deterministic features While models will always eventually diverge the use of ocean assimilation can aid in lengthening predictive times Ocean assimilation can also aid in the correction of both forcing fields and offers inaccurate dynamics It is important to recognize that this is not a replacement of dynamics a synchronization of both dynamics and data is necessary NPOESS Jason 1 TOPEX Poseidon Most studies of the oceans are based upon satellite altimetry While such satellites instruments as TOPEX Poseidon NPOESS and Jason 1 have provided excellent data the problem lies in trying to correlate the entire ocean body to the properties sensed only at the surface By using historical in situ observations it has been found that sea surface height can be related to subsurface temperature and salinity When using only sea surface height data the current model converges very slowly to these observations By combining all information gained a correction can be applied to the model for more accurate results As stated in the slide show model dynamics are corrected using an error in forcing What assimilation does is estimates the forcing error It does this by minimizing a cost function so that the model state matches the observations The problem in this process is that the covariance functions and matrices are too large It becomes too computationally expensive to complete Correction of Ocean State through Assimilation Computational resources available C Running a non assimilative model takes time T Assimilation adds to T by factors of 5 up to 1000X For high resolution models proper dynamical representation may require that T be a majority of C In such a case only simplified assimilation may be possible C t Cos T No Assim Nudging OI Kalman Filter 4D Var 18 As can be seen in the above chart assimilation becomes too computationally expensive for the NRL s resources at the optimal interpolation OI method of minimizing the cost function 4D Var is a direct way of minimizing the cost function but is not currently computationally possible With the extremely expensive cost to using assimilation why is it valuable to the study of oceans The types of properties and features that can be measured by various satellites are dependent upon time and spatial scales The following two graphs depict the types of properties and features that can sensed without and with the addition of assimilation Due to the spatial and temporal resolution of TOPEX Poseidon it can only sense things in the shaded area But with the use of assimilation the capabilities can be extended to include eddies and fronts as well as coastal upwelling Currently the NRL is taking sea surface height and temperature measurements via satellite and using this observable data in conjunction with its models The result on a global scale is a representation of deep ocean eddies as well as vertical thermal structure By limiting the coverage to regional higher resolution models can be used in coastal regions Ocean assimilation has been used effectively for many different applications Current modeling for oil spill clean up in Spain and contamination dumping in the Gulf of Mexico Fishery services have requested surface chlorophyll and SSH products for tuna fishing in the Gulf of Mexico The Tuna Research and Conservation Center requested SSH analysis for their effort in tagging tuna Many oil drilling companies and organizations have used products provided by the NRL for eddy and current monitoring and The NRL provides a multitude of operational product to the US Navy to aid the warfighter in keeping him safe and effective Products ranging from tidal analyses to water clarity assessments have been used during actual operations The NRL has many plans to expand the amount and type of data that is being used in their ocean assimilation There are currently investigations in the use of shore based RADAR drifters moored current monitors unmanned vehicles both underwater and airborne as well as ship and aircraft gathered data The use of optical data from space is being examined as is the ability to enhance even further the nesting ability of the application As evidenced in the application of the NRL s current products ocean assimilation is an extremely effective tool With plans for further research and enhancements the services of the NRL will continue to be an invaluable tool to anyone with oceanic interests