Institute for Mathematics Applied to Geosciences (IMAGe)


IMAGe Theme-of-the-Year (T-O-Y)

The Theme-of-the-Year is a program to focus on specific areas of research that will benefit from intense collaborative effort. The topics will be selected by the IMAGe external advisory panel and will be coordinated by a Visiting Co-director.

Theme for 2007: Statistics for Numerical Models

Numerical models are vital to simulate geophysical, chemical and ecological processes and to understand the relationship among components in the Earth system. As models have become larger and more complex, their construction, validation and analysis are no longer amenable to simple approaches and statistical summaries. Statistical science in the past 20 years has advanced to handle the interpretation of complicated multivariate, spatial and temporal data sets and it is well suited to tackle the massive outputs from numerical experiments that are now the norm in the geosciences. This theme is undertaken with the goal of matching cutting edge statistical methods to the needs of geophysical model development and to make statisticial scientists aware of the particular scientific issues and research in the geophysical modeling community.

The T-O-Y is pursued in partnership with the Statistical and Applied Mathematical Sciences Institute (SAMSI), located in Research Triangle Park, NC and the Mathematical Sciences Research Instititute (MSRI), Berkeley, CA. Both SAMSI and MSRI are NSF Mathematics Institutes with an international stature in the mathematics community. The main acitivites will be a series of three workshops and a summer graduate workshop, all held at NCAR, that dovetail with SAMSI activities on its programs on random matrices and on computer models and with the MSRI summer school program. IMAGe participation will be valuable in representing a suite of geophysical models within the SAMSI program. Derek Bingham, a faculty member at Simon Fraser University, British Columbia, CA, and an expert in the design of computer model experiments will serve as T-O-Y co-director and will visit NCAR through the year with an extended stay in May. Montse Fuentes, a faculty member at North Carolina State University and a SAMSI Fellow, will be the main liason with the SAMSI programs.

The main planned workshops/meetings are:

  1. Geophysical Models at NCAR: Scoping and Synthesis. 13-14 November 2006.
  2. Application of Random Matrices Theory and Methods. 7-9 May 2007
  3. Application of Statistics to Numerical Models: New Methods and Case Studies. 21-23 May 2007
  4. Summer Graduate Workshop on Data Assimilation for the Carbon Cycle. 8-13 July 2007
The first workshop is intended as a scoping and brainstorming meeting where four NCAR modeling groups will interact with a large group of statisticians interested in the design and analysis of computer experiments. The NCAR geophysical models/groups are
  • Upper atmosphere model (TIEGCM) (HAO: Maura Hagan, Ray Roble, Art Richmond)
  • Single column boundary layer model (RAL/MMM: Josh Hacker)
  • Two dimensional turbulence in Navier Stokes flows (IMAGe: Annick Pouquet)
  • Land component of the NCAR climate model (CGD: Gordon Bonan; U Kansas: Johannes Feddema)
The intent is that concrete problems will be identified that will help structure the statistical working group activity at SAMSI. For each modeling group a statistical researcher will serve as a liason to guide collaboration among the modeling group and the statistical working groups.

Workshops II and III will be more traditional conferences but will include a blend of tutorial and research talks. In each workshop, ample time will be reserved for discussion and also for presentations on progress on the specific modeling project initiated in the first workshop.

The summer graduate workshop will be in partnership with MSRI and focuses on the mathematical tools such as inverse methods and data assimilation to estimate the surface sources of carbon dioxide. The determination of sources of carbon in the Earths' atmosphere is an important area of biogeochemistry and crucial in quantifying human emissions of greenhouse gases. The summer graduate workshop will build off of the successful model that was held July 2006 at MSRI and will feature morning tutorial lectures with reinforcing afternoon computational exercises and projects. The Data Assimilation Research Testbed (DART) will be used as a software framework for the mathematical and statistical methods.