Sydney Olympic WWRP
Forecast Demonstration Project
Minutes and Action
Items from WWRP FDP Meeting Sydney May31-June 1, 1999
Participants
WWRP: Browning (WWRP SSC Nowcast Representative), T. Keenan (Chair
of S2K FDP Working Group), C. Adams (WWRP SSC Impacts Advisor), C. Collier (U. Salford,
GANDOLF), B. Golding ( UKMO, NIMROD), J. Wilson ( NCAR, Autonowcaster), P. Joe (EC,
CRPDS), M. Eilts( NSSL, WDSS, by telephone)
BOM representatives: E. Spark (Olympic Project Manager), B. Buckley
(NSWRO), G. Furler (RDNSW), K. O'Loughlin( RDVIC), P.Purdam (BMRC), A. Treloar( NSWRO),
P.May (BMRC), J. Colquhoun (NSWRO), G. Evenhuis (NSWRO), J. Gill (SPB), G. Mills (BMRC),
R. Hicks (SPB), J. Chattaway (NSWRO), J. Bally (BMRC), I. Bell (BMTC),M. Hassett (COSB),
C. Ryan (COSB), G. Holland (BMRC)
List of email addresses as Attachment 1.
1. SUMMARY OF STATUS and MAJOR ISSUES
Keenan, BMRC
Major NWP Groups at FSL and CAPS have encountered problems in
support funding and unlikely to participate. The FSL Option is still being pursued but
doubtful.
No funding support for inclusion of CPOL radar at Badgery=s Ck.
Kurnell radar is now major radar source and as will be discussed later has some major
clutter and potentially clear air signal problems.
BOM particularly concerned with impact studies hidden costs and
resource requirements. Need to accept that BOM is official forecast, FDP will enhance
official forecast, BOM has final say on forecast policy and that interaction with media
must be sensitively handled and reflect the joint nature of FDP.
Little progress so far with verification sub-group. Impacts advisory
group has been formed.
WMO Workshop planned as part of FDP for 2 weeks in November 2000.
Participants required to support this aspect of program. WMO support will be available for
these activities.( Contact Zhaochong Lei after WMO Congress in May)
Logistical issues (customs, travel dates/numbers) require attention
by participants.
Browning, WWRP SSC
Provided outline of history of FDP and emphasized that readily
identified measures of success for each system are necessary for it to be included in the
FDP. Impact of overall project dilution by the withdrawal of the NWP groups had to be
recognized. Very much now focused on 0-6 h time frame with observationally based
state-of-the-art-systems. To be useful systems must lead to a response by users,
particularly the forecasters. There needs to be criteria established for
"successful" implementation of the systems.
2. SYDNEY OLYMPIC OPERATIONS
Extensive summary provided by E. Spark, BOM Olympics Program
Manager. SOCOG highly centralized operations for weather impact at major venues. Actual
venues will not have final say on weather related operations. BOM briefers will be
available at SOCOG and at sailing (Rushcutter=s Bay) with access to web. INFO2000 is an
hourly text based system not well suited for enhanced weather products. Final decision
making process for main venues centrally controlled and requirement is for categorical Y/N
based on fax and oral briefing. No real history of interacting with weather information
and less than desirable for FDP in that they (SOCOG) represent a customer without a known
history (or skill) in this area. Expect sailing to be interested.
During discussion of BOM thunderstorm warning process that will
continue during the FDP the prospect of using other more traditional BOM customers for
impact studies was raised and thought valuable e.g. aviation raised for consideration. See
later considerations at end of meeting to change the focus to more "traditional and
experienced" users.
Purdam reported on how communications and system integration was to
occur in Sydney. BMRC to provide mirror AIFS server from which WWRP will access BOM data.
No major impact on BOM communications although its expected it will take 15 min to NIMROD
NWP information from Bragg in Melbourne. Current Melbourne Sydney communications seem
adequate although this needs to be end-end tested. Approach is outlined in Attachment 2.
Action: Keenan to test Melb-Sydney communication
of LAPS 15km resolution NIMROD data subset and composite radar data.
3. DATA
The document outlining the implications of the WWRP on BOM comms needs
to be completed (Hassett-STCM).
AWS sites in place but no decoder for 1, 10 min data. There are no
AWS=s off the E-W line, but 6 BMRC stations still could be used for filling holes which
could be important if radar clear air compromised by environment.
Action Keenan to investigate if decoder will
make AWS data available on AIFS and arrange transfer of 6 AWS's from Darwin AWS research
network for Olympic period. Buckley and Keenan to coordinate integration within NSW
operations.
Issue of validation was discussed at AWS sites. Quality of
non-Bureau sites was raised as an issue.
Action: Spark, Keenan to report on best AWS
sites for validation.
Profilers: time lines - after Sept 99 trial. It may be preferable
for the Nowcaster to use AMDAR data for low-mid level winds.
Aerosonde: Holland reported up to 8-10 aerosondes would be available
and could be employed in data impact studies. Indicated possible Nowra site for operations
with command at NSWRO. This is a potentially unique aspect to the FDP.
Action: Short 1-2 p proposal (Buckley, May, Mills) to be
prepared and forwarded to Holland. Holland to indicate requirements for operations centre
at NSWRO.
Rawinsondes: Keenan reported the RDNSW had agreed to provide 6 h
rawinsonde ascents from Sydney during FDP.
NWP Status (Mills, BMRC)
Change in LAPS/meso-LAPS resolution (to 37, 12.5 km respectively) in
~ 2 mo. Running in research mode at 5km and expect this to be available by September 99.
Work to continue on 6h assimilation cycle at 12.5 km resolution with 3h cycle by Olympics.
15 km resolution data in GRIB format to be provided by BMRC for
NIMROD. Test dataset has been provided to all WWRP participants for test purposes.
NIMROD (Golding, UKMO)
Some problems in UK in moving HPC110 system with regard to different
LAN's
Action: Purdam to investigate if there is a problem with BOM
LAN.
Salford and UKMO to have one person sharing duties for September 99.
Funding OK for 1999, but need some additional support for IOP.
Output gridded accumulation/rate fields and with text output.
Production of gif images no major problem and require background and projection.
Projection issued sorted out at meeting.
15 min cycle (see products) 0-6 h forecasts.
No formal training material is available.
Interest in employing satellite data.
Appropriate Z-R for Sydney. MP to be employed as first guess.
Action: Papers on NIMROD to be provided by Golding to Bell.
Keenan to provide details of MCIDAS data structure and example to Golding. Keenan to
obtain DSD data for Sydney Z-R relation.
GANDOLF(Collier, U. Salford)
Funding OK for 1999 and funds being sought to support 2000
operations. Two people to come for 1999 (Fox and Ph.D student).
10 min forecasts to t+3 h. Based on 4 cloud class conceptual model
and can include satellite data if used for NIMROD.Text messages and alerts available.
OK, tuning of thresholds for this and NIMROD?
Note that NIMROD and GANDOLF are combined into a HPC110 single
system.
NIMROD HPC110 to accompany personnel to Sydney.
AUTONOWCASTER (Wilson, NCAR)
Forecasts storms > 35 dBZ based on combination of data including
analysis of wind provided by adjoint model at 2 km resolution. Requires good clear air
sensitivity or emphasis on mesonet. Hence AWS mesonet enhancement desirable. Manual
analysis of gust fronts possible if clear signature down.
Operated at New Mexico, Sterling, Redstone Alabama.
Funded for 1999 and 2000 although help required on housing of staff.
Want 15 min 30, 60 120 min forecasts? Probably not 120 as not
proven.
Forecaster training material OK and web based product production OK.
Lots to do, especially data ingest, clutter issues, growth/decay
module.
There are significant concerns regarding clear air detectability on
radar as noted above.
Autonowcaster now being installed at BMRC prior to shipment to
Sydney.
CPRS (Joe, Ac)
System HPC180 or PC with Linux.
Funding OK for this year. Will arrive 1 September with systems
person for setup. Consists of science, graphics and distribution module. Can produce
meteogram and gif image products.
Mesocyclone algorithm with folding algorithm now operating. BWER
identifying SWX
McGill rainfall accummulation algorithm. Autocorrelation approach
giving 0-90 min pt location forecasts.
Tuning of algorithms required (Paul to interact with Andrew Treloar
on trade offs between FAR and POD for the detection algorithms). As for other systems,
urgently need test data.
Will consider purchase of PC in Australia with setup to occur whilst
at BMRC.
Action: P. Joe to contact P. Purdam/Keenan for order details. AC
to reimburse BMRC.
WDSS
Open WDSS may be ready mid September. Funding probably OK, but is
not guaranteed though. Ken Howard to lead WDSS S2K involvement.
Simple web output OK, but this to get forecaster to go and do full
session on host if of interest for max. benefit (applies to other systems also).
Include training material. 1 person coming at a time during Sydney
Olympic period.
Again radar QC issues to be addressed.
VERIFICATION Sub-Group (Reoprt by Browning based on
Bubnova input).
No formal progress in sub-group establishment at this stage.
What products to be available?
Need common formats.
There may be a need to verify both time series and map type
products.
Action: Bally to provide Bubnova with product information
(Attachment 3).
IMPACT ASSESSMENT (C. Adams, CSU)
Establishment of sub-group (Pielke, Adams, Jay, Macauly, USA). Keen
for Australian involvement particularly for responses after significant events. . Will
have up to US5K from USWRP proposal to NSF to support Australian student involvement.
Keenan reported that Bally, Gill BOM to be involved and probably group from Macquarie (R.
Blong). Should know of success in NSF proposal by October 99.
Action: Keenan to contact James Cook University, Townsville for
additional Australian involvement.
There are still significant questions regarding the methodology. But
the approach is to assess the enhancement to BOM services with preference to focus on1-2
long term BOM forecast users e.g. SES, aviation etc. See discussion later.
K. O'Loughlin (RDVIC) reported on link to WMO Public Wx program
How to quantify improvements with lack of ground truth.
Action: Adams to prepare survey instruments etc before
September.
RADAR QC Committee and ISSUES (Keenan, BMRC)
Test Kurnell data sent to QC committee(some problems with UF
conversion)
DSP Ugrade required for Kurnell to keep up with RT Communications.
Communications upgrade required from Kurnell to NSWRO (256 kbs)
Test SE Australia radar composite sent to NIMROD group.
Sea and Land clutter major problem. Compounded by sidelobes. Lack of
clear air echoes on test data although radar sensitivity is good. Affects autonowcaster
although severe Wx not affected. Longer term investigation required.
There was a suggestion to focus on the NCAR clutter rejection, but
at the moment the consensus was for each group to do there own QC in short term, at least
through start of 1999 trial.
BMRC is to investigate modified scans for better clutter
suppression.
5 min updates highly required.
Action Items:We decided that it is best to do our own QC for
pilot program, and ongoing testing. This may lead to the central supply of post QC data at
later time. Purdam to supply radar projection codes (radar-lat/long-radar). BMRC to sort
way to get 5 min scans from Kurnell. May to undertake further test scans on Kurnell.
Keenan to sort purchase of additional RVP6 DSP board.
5. PRODUCTS
Need to define web-based standard products. Need gif outputs for the
September trial for an ensemble type web site.
WWRP groups to choose basic web output for their system(using
guidelines).
Standard maps/areas for rainfall etc. 100km *100km? Displays to
include optional range rings?
Web based automatic 6/8 panel display + point click
meteograms/meteogram ensemble.
General agreement to conform to indicated approah.
Action: Bally and Purdam to co-ordinate products including
provision of map background information (including verification/forecast site specifics-in
consultation with Spark, NSWRO)
6. FDP EMPHASIS
Now that the high-resolution NWP components of the FDP have
withdrawn, the FDP is focused even more sharply than before on the observationally-based
0-6 hour period, with particular reference to rain and severe weather. Browning was
disappointed, therefore by the shortfall in BOM observational resources compared to those
indicated in Tom Keenan's
proposal which was endorsed by the November 1998 meeting of the WWRP
SSC. On the other hand there is continuing enthusiasm for the FDP from those attending the
Sydney planning meeting.
Browning also concerned about the rather minimal opportunity for
impact studies relating to use of the forecasts by the Olympic officials. It was clear
that there will be greater opportunities for provision, and assessment of impact, of the
FDP-enhanced products sent via the Sydney regional office to OTHER SPECIALISED USERS.
Accordingly, Browning recommends that there should be more emphasis on the enhanced
products being demonstrated to a wider range of specialist users, without of course
loosing sight of the important profile provided by the Olympic requirements. This will
have the added advantage of providing a continuing focus for the FDP as it continues
beyond the end of the Olympics through November, during which time a larger sample of
weather events will be obtained. This would require more leg-work by BOM staff
in involving end users in the experiment but the potential benefits
will be correspondingly greater for the impacts research and, I think, for the Bureau too.
Suggested Keenan incorporate these ideas into the planning of the September 1999 trial and
into the updated FDP planning documentation which he will be submitting to WWRP for review
in October 1999.
All at the meeting endorsed these views.
Action, Browning to write formally to BOM suggesting above
requirement. Keenan and others to discuss with BOM for official BOM endorsement and to
finalise a plan to be submitted to BOM executive for approval. Appropriate NSW and HO BoM
staff to discuss and liaise with BoM users in terms of potential involvement in the form
of a demonstration project. Treloar to identify and possible product service in above.
Approach would still be to enhance official BOM forecasts.
7. TRAINING
3 BOM forecaster groups will be training in September. The 1st
of these is for sailing. The focus for WWRP will be on the 2nd and 3rd
groups (general and severe weather groups) training on the 14-17th and 20-23rd
respectively as these groups will be located at the RFC rather than at the yachting venue.
Training on WWRP systems is required as forecasters will have good access to systems and
system experts (advocates, formerly referred to as champions).
Action Items for each Participant: Training Centre needs a
summary of :
Who is coming and when,
Documentation
Training/presentation material (electronic form preferred)
BMRC to supply data for Sydney hail storm and a more typical
convective day for testing/training development as well as for system development/testing.
(Treloar to provide dates)
WMO workshop- This will cover theory, background techniques, case
studies, real time application. The participants will be from centres with a reasonable
prospect of being able to use these types of system in the near future.
8. OTHER ISSUES
There is a recommended modification of the aims see above.
Some concerns were given regarding the ongoing support and remote
access to the systems in the non-IOP periods.
Action: BMRC is committed to provide some computer support and
trouble shooting for the continued operations of the systems (in conjunction with NSW RO).
Archiving of data and products. This needs to be sorted on both
original data set and products. Again gets back to what are required data sets for running
the different systems. All the systems have some level of own archive, but data storage
remains a question over the 12 month period. Need RO support and more discussion via email
on archive process. May/Bally to coordinate.
9. TIMELINES
Bubnova to prepare verification plans ASAP include a small working
sub-group.
IMPACTS
Buckley to coordinate "sound out" meeting with SES et al,
prepare/explain project to improve Bureau products, have examples of displays from groups
within 2 weeks (3-4 OHT=s from each group) - n.b have to have background maps out end of
the week.
May/John Chattaway/Jim Elliot (SRFW) to liase with Hydrology, Sue
Lawrence for Aviation.
Adams to review status in mid June.
Methodology, Bureau plan and impact plan by end of October.
Data sets for development /testing - End of June
Treloar, Purdam to coordinate - radar, satellite images (MCIDAS),
sounding data and try and get .AXF format AWS data.
Bally, Treloar to interact with Joe re. Tuning for sensitivity of
mesoscyclone detection etc.
ALL PARTICIPANTS: ASAP contact David Jasper (d.jasper@bom.gov.au)
regarding any importation of equipment so that arrangements can be made for duty free
import. This may not be too much of an issue as NCAR have provided funds for pc=s etc to
be purchased in Australia, BMRC has workstations for WDSS and the UK team plan on bring
their machine with them when they come for the initial installation. Paul Joe will send
specs to BMRC to purchase PC for CPRDS (Env. Can. to fund).
Purdam to contact/liaise regarding lack of IP addresses and other
network issues with Geert and Mike Hassett.
