 |
| Titel |
Towards Compressed Super-Parameterization: Test of NAM-SCA under Single-Column GCM Configurations |
| VerfasserIn |
J.-I. Yano, S. Kumar, G. Roff |
| Konferenz |
EGU General Assembly 2012
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 14 (2012) |
| Datensatznummer |
250061101
|
|
|
|
|
|
| Zusammenfassung |
| The NAM–SCA (nonhydrostatic anelastic model with segmentally–constant approximation)
is a type of cloud–resolving model (CRM) with a highly–inhomogeneous distribution of
finite volumes of highly varying sizes in a horizontal direction under a two–dimensional
geometrical configuration. These finite–volume positions and sizes are dynamically adapted
with time so that a best performance is achieved with a relatively limited number of finite
volumes. The concept can be compared with “compression” techniques used for digital
images. The physics are also limited only to a minimum microphysics.
The present paper reports implementation of NAM–SCA into a single column version of
global atmospheric models as a compressed super–parameterization, and discusses
various implementations. ECHAM and ACCESS are chosen as the host models. A
stand–alone single–column model (SCM) is also developed, in which the other
physics are prescribed by observations. A simple radiation scheme is also added as
required.
Overall, it is found that, under this configuration, the lowest–resolution (Δx = 16Â km)
NAM–SCA with the smallest domain size (L = 32Â km) often works the best. Neither
increase of resolution or domain size does not lead to a better performance until both
cross thresholds (Δx -¤ 1Â km, L -¥ 256Â km). The result suggest that even a simple
parameterization (with a small degree of freedom) can perform in a reliable manner as
long as it is constructed in physically consistent manner. Especially, the mesoscale
organization is not necessarily a key for successful parameterization as long as only the
domain–averaged outputs are in concern as the case for the standard parameterizations. |
|
|
|
|
|
|