|
Model soundings from the BUFR data (raw model data, no interpolations) for the 12 UTC 26 July 2001 operational Eta run are shown for the grid columns nearest four stations. The RAP and PHP soundings are for the day-one forecast event while the SLN and P#G soundings are for the day-two forecast event. Dry adiabats are in brown, moist adiabats in red, and mixing ratio lines in blue.
|
 |
Most of the soundings show the effects of both the convective parameterization
and the grid-scale scheme. First you see the daytime boundary layer or evening
residual layer. Then, the BMJ convective scheme smooths out the sounding but
strong low-level advection maintains instability. For instance, CAPE remained
2000 J/kg at RAP and SLN during a three- to four-hour period throughout which
the convective parameterization was triggering, dribbling out a steady light
rain totalling around one-third of an inch. Meanwhile, a grid-scale thunderstorm
starts to build, and after a few hours it generates heavy precipitation, cooling
the low levels with evaporation. CAPE from the lowest 500 meters finally drops
below 1000 J/kg at three of the four locations. The moisture lofted in the grid-scale
updraft keeps the BMJ convective scheme active, though most of the precipitation
is falling from the unrealistic, broad grid-scale updraft core. The continued
activation of the BMJ scheme smooths the sounding, masking some of the signature
of the grid-scale thunderstorm.
 |
Here is one that the convective parameterization didn't affect very much. This 03 UTC sounding at P#G shows the effects of the grid-scale condensation scheme. It looks strikingly like cloud-resolving simulations (such as with 1 km grid spacing) of a small, growing cumulus tower. A cloud-topped mixed layer is evident above 800 hPa. As the tower rises, the ambient air immediately above it is forced to rise and cools dry-adiabatically. This creates a steep lapse rate on top of the cumulus cloud, promoting turbulent mixing in the model which may cool the top of the cloud. A stable lapse rate results between the disturbed ambient air and the undisturbed air above. |
The sequence from RAP shows the build-up of the cumulus tower leading to the grid-scale storm. The grid-scale cloud, supported by strong grid-scale vertical motion, shows up as subtle inflections in the sounding. Cloud base and cloud top are marked by reductions in the lapse rate, and the lapse rate gradually steepens upward in the cloud like the way a moist adiabat is curved. Profiles of cloud water will be shown on the next page.
Also, the upward transport of high-dewpoint air and low-level replenishment from below increases the column precipitable water (PW), leading to a PW maximum in the heavy rain area.
These soundings have a moist absolutely unstable layer (MAUL). You would think this would create instant convective overturning — that's exactly what the grid-scale motions are trying to do! Bryan and Fritsch (2000) show examples of MAULs from observations and high-resolution model simulations of a squall line. They present a convincing argument that MAULs can result from deep layer ascent and are caused by vigorous mesoscale ascent into a squall system. We might expect to see more of these pathological-looking soundings as we move to finer resolution, and they might even be a realistic representation of a certain part of the squall line environment! However, it is unrealistic to see them for hours on end at one location, and, at 22 km grid spacing, the model cannot respond realistically.
