Treating strong adjoint sensitivities in tropical eddy-permitting variational data assimilation
نویسندگان
چکیده
A variational data assimilation system has been implemented for the tropical Pacific Ocean using an eddypermitting regional implementation of the MITgcm. The adjoint assimilation system was developed by the Estimation of the Circulation and the Climate of the Ocean consortium, and has been extended to deal with open boundaries. This system is used to adjust the model to match observations in the tropical Pacific region using control parameters which include initial conditions, open boundaries and time-dependent surface fluxes. This paper focuses on problems related to strong adjoint sensitivities that may impede the model fit to the observations. A decomposition of the velocities at the open boundaries into barotropic and baroclinic modes is introduced to deal with very strong sensitivities of the model sea surface height to the barotropic component of the inflow. Increased viscosity and diffusivity terms are used in the adjoint model to reduce exponentially growing sensitivities in the backward run associated with nonlinearity of the forward model. Simplified experiments in which the model was constrained with Levitus temperature and salinity data, Reynolds sea surface temperature data and TOPEX/POSEIDON altimeter data were performed to demonstrate the controllability of this assimilation system and to study its sensitivity to the starting guesses for forcing and initial conditions.
منابع مشابه
Strategies for nested and eddy-permitting state estimation
[1] Both ocean process and prediction studies increasingly rely on state estimation (i.e., data assimilation) to provide the most complete representation of how the ocean circulates. This study applies the formalism and methodology of state estimation, recently developed for the global, coarse-resolution problem, to eddy-permitting state estimation in regional domains. Two major challenges exis...
متن کاملOn Sensitivity Analysis within the 4DVAR Framework*
The sensitivity of model forecasts to uncertainties in control variables is evaluated using the adjoint technique and the ensemble generated by the reduced-order four-dimensional variational data assimilation (R4DVAR) algorithm within the framework of twin-data experiments with a quasigeostrophic model. To simulate real applications where the true state is unknown, the sensitivities were estima...
متن کاملA Method of Successive Corrections of the Control Subspace in the Reduced-Order Variational Data Assimilation*
A version of the reduced control space four-dimensional variational method (R4DVAR) of data assimilation into numerical models is proposed. In contrast to the conventional 4DVAR schemes, the method does not require development of the tangent linear and adjoint codes for implementation. The proposed R4DVAR technique is based on minimization of the cost function in a sequence of low-dimensional s...
متن کاملEfficient gradient computation for dynamical models
Data assimilation is a fundamental issue that arises across many scales in neuroscience - ranging from the study of single neurons using single electrode recordings to the interaction of thousands of neurons using fMRI. Data assimilation involves inverting a generative model that can not only explain observed data but also generate predictions. Typically, the model is inverted or fitted using c...
متن کاملAdjoint Implementation of RosenbrockMethods Applied to Variational DataAssimilation Problems
In the past decade the variational method has been successfully applied in data assimilation problems for atmospheric chemistry models [9, 10, 11]. In 4D-var data assimilation a minimization algorithm is used to find the set of control variables which minimizes the weighted least squares distance between model predictions and observations over the assimilation window. Using the adjoint method, ...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 2006