Large-time uniqueness in a data assimilation problem for Burgers' equation

TL;DR

This paper proves that for Burgers' equation, collecting enough observational data in a 4D-Var data assimilation scheme guarantees the uniqueness of the optimal initial condition, addressing nonconvexity issues.

Contribution

It establishes a theoretical result showing that sufficient data collection ensures uniqueness of the minimizer in 4D-Var data assimilation for Burgers' equation.

Findings

Uniqueness of minimizer is guaranteed with enough data

Addresses nonconvexity in the data assimilation cost function

Provides theoretical foundation for data assimilation reliability

Abstract

There is currently a great deal of interest in the 4D-Var data assimilation scheme, in which one uses observational data to find the optimal initial condition for a differential equation by minimizing a cost function over the set of all possible initial states. For nonlinear models this cost function can be nonconvex, and so the uniqueness of minimizers is not guaranteed. In this paper we apply 4D-Var to Burgers' equation and prove that, once a sufficient amount of data has been collected, there can be at most one physically reasonable minimizer to the variational problem.