NCAR Visitor Center - One-Day Closure - Saturday 10/25 more info>
- UCAR Home
- About Us
- For Staff
Sue Ellen Haupt
Date and Time: Wednesday, March 19, 2014 1:30 PM - 3:00 PM
Location: FL2-1022 Large Auditorium
Although atmospheric models can provide a best estimate of the future state of the atmosphere, due to sensitivity to initial conditions, it is intractable to predict the precise future state. In particular, predicting a particular realization of an evolving flow field requires knowledge of the current state of that field and assimilation of observations into the model. An example is modeling atmospheric transport and dispersion of a contaminant when the observation is of the transported contaminant, a problem that exemplifies the issue of uncertain turbulent flow. We will discuss the inner vs. the outer variability and how both can be recovered with judicious use of the observations. In this case, the problem is compounded by the fact that the field observed is a tracer that is advected and mixed by the flow field, but does not directly alter the flow field. This one-way coupled system presents a challenge: one must first infer the changes in the flow field from observations of the contaminant, then assimilate that data to recover both the advecting flow and information on the subgrid processes that provide the mixing.
This work demonstrates using a genetic algorithm to optimize the match between the observed flow and the model. Given contaminant sensor measurements and a transport and dispersion model, one can back-calculate unknown source and meteorological parameters. In this case, we demonstrate the dynamic recovery of unknown meteorological variables, including the transport variables that comprise the "outer variability" (wind speed and wind direction) and the dispersion variables that comprise the "inner variability" (contaminant spread). The optimization problem is set up in an Eulerian grid space, where the comparison of the concentration field variable between the predictions and the observations forms the cost function. The transport and dispersion parameters, which are determined from the optimization, are in Lagrangian space.
We then discuss the broader applicability of this general approach, specifically blending physical modelling, ground truth observations, and artificial intelligence methods to optimize the match between the two and argue that this approach can be advantageous for discovering information about dynamical/physical processes.
This seminar will be Webcast - Webcast link