As many of you may know, we have been working on ROMS adjoint-based algorithms for several years. Several papers have been already published about these algorithms. Over the last two years Andy Moore and I with the help of Gregoire Broquet and Brian Powell have completely reworked all the 4D-Var data assimilation algorithms in ROMS. The data assimilation algorithms include the primal formulation of incremental 4D-Var (I4D-Var) in model space, and the two dual formulation algorithms (4D-PSAS and R4D-Var) in observation space that can be configured in strong constraint (model is perfect) or weak constraint (model has errors) modes. In addition, we built the adjoint in the entire data assimilation system for 4D-PSAS and R4D-Var to quantify the observation sensitivity and flow dependent error covariances via ensemble 4D-Var. There are also algorithms for the observations impact, error covariance modeling, error hypothesis testing, posterior error estimates, eigenvector analysis of posterior error covariance, array modes, degrees of freedom, preconditioning, and several diagnostic tools. We have prepared several tutorials for how to use these algorithms in WikiROMS.
We just submitted three papers describing these algorithms to Progress in Oceanography that probably will be published in 2011, after going to customary review process. The papers are comprehensive and with increased degree of complexity. Andy did an excellent job describing these algorithms. The simulations carried out are unique and monumental Perhaps, these are the must read papers for everybody working on ocean 4D-Var data assimilation:
- Moore, A. M., H. G. Arango, G. Broquet, B. S. Powell, A. T. Weaver, and J. Zavala-Garay, 2011a: The Regional Ocean Modeling System (ROMS) 4-dimensional variational data assimilation systems, Part I: System overview and formulation, Prog. Oceanogr., 91, 34-49, doi:10.1016/j.pocean.2011.05.004. PDF
- Moore, A. M., H. G. Arango, G. Broquet, C. A. Edwards, M. Veneziani, B. S. Powell, D. Foley, J. Doyle, D. Costa, and P. Robinson, 2011b: The Regional Ocean Modeling System (ROMS) 4-dimensional variational data assimilation systems, Part II: Performance and application to the California Current System, Prog. Oceanogr., 91, 50-73, doi:10.1016/j.pocean.2011.05.003. PDF
- Moore, A. M., H. G. Arango, G. Broquet, C. A. Edwards, M. Veneziani, B. S. Powell, D. Foley, J. Doyle, D. Costa, and P. Robinson, 2011c: The Regional Ocean Modeling System (ROMS) 4-dimensional variational data assimilation systems, Part III: Observation impact and observation sensitivity in the California Current System, Prog. Oceanogr, 91, 74-94, doi:10.1016/j.pocean.2011.05.005. PDF
You are welcome to download these papers in their final form as submitted to the journal. These are the primary papers to reference in the future in your publications with ROMS 4D-Var algorithms. In some cases, you need to read these papers several times to grasp the concept and the overwhelming details that are needed when running these algorithms.
To our knowledge ROMS is the first and only ocean framework to offer these three 4D-Var data assimilation algorithms and distributed freely to the ocean modeling community. It is also the first time that a systematic comparison is offered between I4D-Var, 4D-PSAS, and representer-based R4D-Var. We also offered all these algorithms to the user community at the same time that they were developed and released an example for the California Current System, WC13, in the svn repository https://myroms.org/svn/src/test/WC13. We call these papers affectionally as the ROMS 4D-Var Trilogy.
Good luck using these algorithms