Collaborative Research: CMG: Multi-Scaled Dependent, Heavy Tailed Distributions in Geophysical Flow: Physical Mechanisms and Data Assimilation

University Of North Carolina At Chapel Hill, Chapel Hill NC

Investigators

Abstract

This project involves two linked problems. Starting from the observation that the spatio-temporal distributions of tracers in the atmosphere and ocean are non-Gaussian, the first problem is to try and attach mechanisms to the different types of non-Gaussian behavior by looking at models of various physical processes and seeing what tracer distributions result. The second problem is to try and develop ways of performing data assimilation that take account of the non-Gaussian nature of data "errors" and of situations in which uncertainty combines multiplicatively rather than additively. One of the impacts of this work may be on the estimation and prediction of oceanic flows, a topic that is becoming increasingly important for a range of applications including coastal zone management and fisheries. Data from passive Lagrangian platforms, one of the methods used to observe the ocean, are frequently characterized by non-Gaussian statistics. Other potential areas of application include the forecasting of the distributions of chemical tracers in both the atmosphere and the ocean.