Analysis of Lateral Boundary Effects on Inner Domain of COAMPS
Abstract
One of the major problems with a Limited Area Model is the introduction of error from the lateral boundaries. The boundary condition provides a source of forcing to the interior of the model. This forcing typically originates from a global model such as NOGAPS. The transition at the boundary from one model to another invariably produces errors. Traditionally, the way to minimize boundary error is to move the boundary as far away from the area of interest as possible. In this way, the errors do not have time to infest the LAM with bad information. Moving the boundary far away from the area of interest increases the computational forecast load and decreases its timeliness. This study looks at how close the lateral boundary can be to minimize computational time and still maintain a forecast that is useful. It was found that when the entire inner COAMPS nest was analyzed, the differences between the control forecast and the test forecast where within the natural variability of the control grid. It was also found that there where localized areas within the model domain that differed between the control domain and the test domain by up to 20 mb for the sea level pressure after a six day forecast.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 2004
- Accession Number
- ADA427223
Entities
People
- Brad G. Harris
Organizations
- Naval Postgraduate School