Jeff Feren (to laps_feedback)

During the Spring-early Summer period (at least), there have been
several incursions of vertically deep, anomalously cold air over the
Victorian region that have not been well handled by LAPS. In particular,
the progged 500hPa temperatures associated with the cold pool have often
been a few degrees higher than observed. In contrast, the U.S. Aviation
Model (NCEP) has tended to be highly accurate in its depiction of the
upper level cold pool ^Ö if there has been any bias with this model, it
has sometimes slightly overdone the cooling. There is, however, a
tendency for later model runs of LAPS to play ^Ñcatch-up^Ò to NCEP. The
deficiency in LAPS has, amongst other things, impacted on the quality of
forecasting of ^Ñcold air mass^Ò thunderstorms over the State. Thus, some
forecasts (based largely on LAPS) have indicated that thunderstorm
development would be no more than a marginal prospect near the coast,
whereas thunderstorms did, in fact, develop and extended as far north as
the northern Victoria/southern N.S.W. region. The most recent example of
such an occurrence was on 2/1/02. On my nightshift, during the period of
amplification of the upper cold pool prior to the thunderstorm
development, I noted that an 18Z AMDAR trace indicated a 500hPa
temperature at Melbourne already 3 degrees lower than that on the +6
hour LAPS prog.

Is it possible to rectify this problem?

A possibly related matter is the question of exactly what temp/wind
flight data is actually incorporated into the new model run of LAPS,
particularly during the daylight savings period. Thus, it seems that the
00Z and 12Z models are more or less run at the same local time (in
Melbourne), regardless of the time of year, yet the balloons are
released 1 hour later (for States observing DST) during the summer
period. For the 12Z model run, in particular, there have been several
times when we have been receiving the new model output whilst upper
sonde data is till being received. I understand that a web site was to
be set up which would enable forecasters to know exactly what (upper
wind/temp) data was included in the latest initialisation of LAPS
(GASP). Is there such a site?

Geoff Feren
Vic Regional Office

Jim Fraser

Geoff,

Regarding the operational aspects of your query, the LAPS model run is
kicked off in NMOC at ~01:00Z and ~12:55Z (+/-10min) each day. This time
has been chosen to enable timely product generation while still ensuring
that most of the local Australian observations are in and available for
use in the data assimilation. There may occasionally be a few stragglers
but generally all the sonde data is in by this time.

As far as finding out exactly what data was included in the latest
initialisation of LAPS (GASP), something like this is available on the
NMOC webpage for LAPS at
  http://nmocrtop.ho.bom.gov.au/pks/system_component_times_laps_pt375.html
and for GASP at
  http://vortex.ho.bom.gov.au/gasp/obs_multi.html
The LAPS charts show the observational data distribution derived directly
from the box files that are used by the analysis component of the
assimilation. The GASP charts show the global data available in the
operational realtime database for the 6hr time window around the 11z/23z
assimilation basetime.

There are some caveats with these plots though:
* although they show distribution of available data they don't indicate
  whether the data was actually used in the assimilation - some data may
  be rejected for whatever reason.
* they do not indicate the actual weight paid to the data inside the
  the assimilation scheme, data weights vary depending on time of
  measurement, data type, density of surrounding data points etc.

regards
Jim Fraser, NMOC