IBM demonstrates Deep Thunder
in Hawaii
The capabilities developed and utilized for a number of past experiments
were demonstrated at two locations in Hawaii in March 1999. The first
was part of the International Workshop on Next Generation Climate Models
for Advanced High Performance Computing Facilities, March 1-3, 1999, organized
by the Research Organization for
Information Science and Technology (RIST) of Japan. The second
was part of a presentation at the Maui
High Performance Computing Center (MHPCC) in Kihei on March 4, 1999
among a series of seminars on weather
forecasting. The system was adapted to Hawaii as shown below.
For this and any of the subsequent images, you can view a higher-resolution
version by simply clicking on it. You can also interact with this
map via a scene in PanoramIX or simplified VRML.
Two mesoscale forecasts were produced at 6.5 km resolution in a region
roughly 650x650 km in extent, one for each of the aforementioned venues.
In both cases, raw observations and access to data assimilation via LAPS
for the pre-processor step were unavailable. Further, useful data
from the ETA synoptic scale model from NCEP,
which are computed at 32 km resolution, but sampled at 80 km for public
availability, did not exist. Hence, the results of a lower-resolution
model, AVN, which covers Hawaii were used for both initial and boundary
conditions.
The first forecast was done for March 2 and a sample result is shown
below. One goal was to investigate what really could be done with
limited resources. In this case, the input data was downloaded on
a laptop (IBM Thinkpad 600) via modem connection from the hotel where the
workshop was taking place. The laptop was connected via an inexpensive
ethernet hub to two IBM RS/6000 43P-260 workstation. Both of these
workstation have an SMP configuration with two, 200 MHz 630 processors.
One of the machines was used for running RAMS, the other was used for I/O
and doing interactive visualization. There was no opportunity to
optimize performance for this configuration, yet an 18-hour forecast was
completed in roughly four hours. Output from RAMS every 10 minutes
of forecast time were provided for browsing visualization, such as the
image below or via animation produced during
the workshop. The 109-frame animation
shows typical convective effects and wind patterns in Hawaii. (The
animation
can also be viewed at higher resolution, but the file is three times bigger.)
The image shows a terrain map, pseudo-colored by predicted surface humidity
overlaid with coastline maps with some major cities marked for 2 PM local
time on March 2. Predicted winds are illustrated by streamlines with
directional arrows, colored by speed. In the animation, one can observe
how the lee side of the islands, particularly the big island of Hawaii,
is wetter. One can also observe vortex shedding in the winds off
the coast of that island. This particular time step can also be examined
via a flyover animation, simplified VRML
geometry and a PanoramIX scene. (The flyover
animation can also be viewed at higher resolution, but the file is
about two times bigger.)
The second experiment on March 4, at MHPCC,
utilized sixteen 160 MHz P2SC thin nodes for computation on their very
large SP. An additional 160 MHz P2SC thin node was used for I/O.
The aforementioned workstations were available for interaction. Output
from RAMS every 10 minutes of forecast time were provided for browsing
visualization and tracking while the simulation was running. In addition,
interactive visualization applications for the analysis
of post-processed model results were available operationally, which
are illustrated below.
Output from RAMS every 10 minutes of forecast time were provided for
browsing visualization, such as the image below or via animation
produced during the workshop. The 145-frame animation
shows typical convective effects and prevaling winds in Hawaii. (The
animation
can also be viewed at higher resolution, but the file is three times bigger.)
The image shows a terrain map, pseudo-colored by color-filled contour bands
of predicted temperature overlaid with coastline maps for 7 PM local time
on March 4. Predicted winds are illustrated by arrows, colored by
speed. Total cloud water density is illustrated by white, translucent
isosurfaces and predicted reflectivities by cyan, translucent isosurfaces.
In the animation, one can observe how the lee side of the islands, particularly
the big island of Hawaii, is wetter. One can also observe vortex
shedding in the winds off the coast of that island. This particular
time step can also be examined via a flyover
animation, simplified VRML geometry and
a PanoramIX scene. (The flyover
animation can also be viewed at higher resolution, but the file is
about two times bigger.)
After each RAMS execution, all of the results
are collected and reorganized into a form that can be used by standard
meteorological analysis tools as provided by NWS (e.g., AWIPS), FSL and
others. These post-processed data were made available for two interactive
applications. This includes all of computed variables from the model,
but at hourly resolution unlike the browser application that worked with
a subset of variables but at six times the temporal resolution. Here
is a sample image and animation created
with this application for the previously discussed forecast run at MHPCC
initiated on Thursday, March 4 at 12Z UTC (2 AM HST).
A surface variable (precipitable water) has been selected for display
as pseudo-colored filled contour bands, which are overlaid on a topographic
map. Coastlines (black) are draped on the surface. An upper
air variable (relative humidity) has been selected for display via surface
extraction. The surface at 90% is requested in translucent white,
which corresponds roughly to a cloud boundary. Another field (vertical
wind speed) has been selected to show as a vertical slice, which is pseudo-color
contoured. Any of the three-dimensional fields available from the
model can be visualized with either of these methods. The upper air
wind data can be seen along two vertical profiles, which are specified
interactively, and via streamribbons. The direction of the model
wind field along these "virtual sounding" is shown via vector arrows.
Both the arrows and ribbons are pseudo-colored by horizontal wind speed.
The length of the arrows also corresponds to the horizontal speed.
Points along the profile are used as seeds for the streamribbon integration.
Each profile is realized as a pseudo-colored tube, which is contoured by
the variable selected for isosurface realization (i.e., relative humidity).
The visualization for the profile marked 2
can help illustrate the vortex shedding described earlier.
The other application is a RAMS slicer, which provides two- and 2-1/2-dimensional
interaction with surface and upper layers of the model data. Here
is a sample image and animation for the
same forecast.
Five different surface variables have been selected in a combined visualization.
Precipatable water is shown as pseudo-color. Wind velocity is illustrated
as vector arrows, colored by speed. Colored line contours of maximum
reflectivity in increments of 3 dbZ are shown. These planar representations
are deformed vertically by relative humidity to create a shaded surface.
A coastline map (black) is draped on the surface. Finally, temperature
values at discrete locations are also shown by value on the surface.
Any of the surface and upper air fields available from the model can be
visualized with any of these methods. One can see the distinctions
between the leeward and windward sides of the islands in this representation,
particularly for the large island of Hawaii.
To evaluate these model results, it is useful to compare
them to actual observations, which are
discussed on the next page.
lloydt@watson.ibm.com
