Science & Math Colloquium
October 8, 2013 | 11:45 AM to 12:35 PM
Dr. Matthew Hoffman, Assistant Professor of Mathematics, Rochester Institute of Technology
Numerical Weather and Ocean Prediction and Data Assimilation
Presented by Dr. Matthew Hoffman, Assistant Professor of Mathematics, Rochester Institute of Technology
Abstract: An integral part of numerical weather prediction for years, the objective of data assimilation is to provide close approximations to an unknown trajectory of a dynamical system through the input of a time series of observational data. For a planetary atmosphere or ocean, this amounts to estimating the past and current values, as well as the future evolution, of physical variables, such as temperatures, winds, currents, salinities, and dust propagation using both in situ and satellite observations. I will speak about data assimilation, specifically ensemble based assimilation methods, and how these methods quantify uncertainty and adjust the state estimates from short term numerical simulations to provide a new set of initial conditions for the next model runs without linearizing the system. I will discuss the methods themselves and then results from my work implementing an assimilation system for the Chesapeake Bay among other applications.