 |
General: 663 |
P4: Reduction of Temporal Discretization Error in an Atmospheric General Circulation Model (AGCM)
Author: Daisuke Hotta , Advisor: Eugenia Kalnay (AOSC)
Problem Statement Presentation
Project Proposal
Abstract
An atmospheric general circulation model (AGCM) is a computer program which numerically integrates the system of partial di erential equations which describes the physical laws governing the ow of the Earth's atmosphere. Integration of AGCM is a central part of weather predictions and climate projections. Currently, due to restrictions on computational costs, most AGCMs adopt relatively low-order tempo- ral discretizations, under the premise that errors arising from temporal discretizations are negligible compared to those associated with spa- tial discretizations or physical parametrizations. However, recent study shows that temporal discretizations is likely to be a signi cant source of model errors. The goal of this project is to reduce the temporal discretization error. Two di erent approaches will be examined: the rst approach is to test a new time integration scheme, called Lorenz N-cycle scheme, which requires the same computational costs as the conventional scheme but yet can yield forecasts of higher-order accu- racy. The second approach is to empirically estimate and correct model errors using the methodology presented by Danforth et al. (2007).
MidYear Progress Report and
Presentation
Final Presentation ,
Final Report
