OPTIMAL ESTIMATION IN METEOROLOGY was presented by Gérald Desroziers during the ALATNET seminar on Data Assimilation, June 11-22, 2001.

• Introduction
  • estimation theory
    • observations are not perfect,
    • the estimate will also be approximate,
    • the estimate must be an optimal combination of observations,
    • it is better to overdetermine the problem,
    • a model can be used to give a preliminary estimate,
    • model are never perfect,
    • the estimate should fit observations within their errors.
  • 1st approach: minimisation of the estimation error variance
  • 2nd approach: Weighted Least Squares
  • 3rd approach: Maximum-likelihood
• Using a background
• Optimal Interpolation
• Structure functions
• Using a model: the Kalman Filter
  • Principle of a 3D-Var or Optimal Interpolation scheme