Mick E. Mouse

I have two main duties as an Associate Scientist with the Geophysical Statistics Project (GSP) and the Data Assimilation Research Section (DAReS) at the National Center for Atmospheric Research. Essentially, my job is to ensure the GSP visitors, post-docs, and collaborators have everything they need, be it data, code, hardware, methodologies, programming, programming advice, home repair ... AND that anyone, anywhere, on any architecture, can download, install and run the Data Assimilation Research Testbed (DART) software.

Select Projects

Data Assimilation Research Testbed

The Data Assimilation Research Testbed (DART) is an effort led by Jeff Anderson to develop a suite of software to explore data assimilation methodologies instead of data assimilation programming. Our intent is to develop a very modular suite of models and observations to facilitate research in data assimilation and forecasting. My presentation at the "Workshop on Ensemble Methods" in Exeter (Oct 2004) was titled: An Operational-Quality Ensemble Assimilation System for NCAR's CAM Climate Model" [pdf]

Stochastic Multiresolution Models for Turbulence

This project is a collaboration with Brandon Whitcher, Jeff Weiss (University of Colorado, Boulder), Thomas Lee (CSU), Doug Nychka and myself. We are attempting to quantitatively describe the coherent structures (eddies, vortices) in turbulent fluids using stochastic multiresolution models. The source of the "data" is the result of a high-resolution numerical integration of the Navier-Stokes equation. The method will result in the ability to automatically detect and separate "target" features (ones that resemble some generic template) from a 2D image, allowing examination of either the background or the target features.

Gibbs Sampling/Wind Blending

This is work with Chris Wikle (U of Missouri), Ralph Milliff (CoRA), Doug Nychka (GSP), and Mark Berliner (Ohio State). We combine the information from satellite observations of surface winds over the oceans with the best-guess estimate of the surface winds from the operational weather centers. The result is a suite of spatially- and temporally-complete winds that have realistic variance at all spatial scales. Our domain is the entire Tropical Pacific at 6hourly and half-degree resolution. The winds are modeled with a Bayesian Heirarchical model which incorporates propagation of large-scale modes according to the laws of physics. The models are estimated with a gigantic Gibbs Sampler that runs in parallel on NCAR's 1600-processor SMP IBM RS/6000 (as well as many other) platform(s).

Publications (reverse chronologically, sort-of)

Hoar, T.J., R.F. Milliff, D. Nychka, C.K. Wikle, and L. Mark Berliner, 2003

Winds from a Bayesian Heirarchical Model: Computation for Atmosphere-Ocean Research. Journal of Computational and Graphical Statistics, Volume 4, Number 12, 781-807

Whitcher, B., Weiss, J., Nychka, D. and Hoar, T.J., 2002

A multiresolution census algorithm for calculating vortex statistics. manuscript.

Milliff, R.F., Hoar, T.J., and Madden, R.A., 1998:

Fast, eastward moving disturbances in the surface winds of the equatorial Pacific. Tellus, 50A, 26-41.

Jones, R.H., Y. Zhang, and T.J.Hoar, 1997:

Constructing surface wind space-time maps from scatterometer data. manuscript

Trenberth, K.E., and Hoar, T.J., 1997:

El Niño and climate change. Geophys. Res. Lett., 24, No. 23, 3057-3060.

Various useful (to me) links.

• My CVS cheat-sheet
• Instructions for blending winds with the Gibbs Sampler
• My Matlab instructional demo