Multiscale Models for the
Tropics: A Systematic Route for Improving Theory, Computational, and
Predictive Strategies
Andrew Majda
One of the unexplained striking features of tropical convection is the
observed statistical self-similarity, in clusters, superclusters, and
intraseasonal oscillations through complex multi-scale processes
ranging from the mesoscales to the equatorial synoptic scales to the
intraseasonal/planetary scales. On the other hand, the accurate
parameterization of moist convection presents a major challenge for
accurate prediction of weather and climate through numerical
models. After a brief survey of the observational record, this
lecture summarizes recent work giving insight into these complex issues
through the paradigm of modern applied mathematics done by the lecturer
with various collaborators. This part begins with new
multi-spatial scale, multi-time scale, simplified asymptotic models
derived systematically from the equatorial equations on the range of
scales from mesoscale to equatorial synoptic to planetary/intraseasonal
(Majda 2006.) All these simplified models show systematically that the
main nonlinear interactions across scales are quasi-linear where eddy
flux divergences of momentum and temperature from nonlinear advection
from the smaller scale spatio-temporal flows as well as mean source
effects accumulate in time and drive the waves on the successively
larger spatio-temporal scales. Furthermore, these processes which
transfer energy to the next larger, longer, spatio-temporal scales are
self-similar in a suitable sense. The lecture continues with a
brief summary of the multi-scale MJO models (Biello-Majda) and recent
multi-cloud models (Khouider -Majda) for superclusters and their
fidelity with key features of the observational record.
Superparameterization is a promising recent alternative strategy for
including the effects of moist convection through explicit turbulent
fluxes calculated from a cloud resolving model. Basic scales for
cloud resolving modeling are the microscales of order 10km in space on
time scales of order fifteen minutes where vertical and horizontal
motions are comparable and moist processes are strongly
nonlinear. Systematic multi-scale asymptotic analysis (Majda
2006) is utilized to develop simplified microscale mesoscale dynamic
(MMD) models for interaction between the microscales and
spatio-temporal mesoscales on the order of 100km and 2.5 hours.
The new MMD models lead to a systematic framework for
superparameterization for numerical weather prediction generalizing the
traditional column modeling framework.
Finally this lecture ends with a new use of the multi-scale cloud
models in the intraseasonal regime to produce realistic looking MJO
analogue waves with intermittently propagating smaller scale eastward
convection embedded in a planetary scale envelope moving at 5-7 ms-1
for flows above the equator. In the model, there are accurate
predictions of the phase speed from linear theory and transitions from
weak regular MJO analogues to more realistic strong multi-scale MJO
analogue waves as climatological parameters vary. With all of
this structure in a simplified context, these models should be useful
for MJO predictability issues in a fashion akin to the Lorenz 96 model
for predictability issues in the midlatitude atmosphere. This last work
is joint with the lecturer, his Ph.D. student Sam Stechmann, and
Boualem Khouider. Most of the papers in this research program can
be found at Majda's faculty website: http://www.math.nyu.edu/faculty/majda/.