Home | About Us | Contacts | Help | Feedback | SEARCH Global | Australia | NSW | Vic. | Qld | WA | SA | Tas. | ACT | NT | Ant. | Weather & Warnings | Hydrology | Climate | Numerical Prediction | About Services | Learn About Meteorology | Registered User Services |

Previous Chpater | Back to Index | Next Chapter

METEOROLOGICAL AND RELATED
DATA AND PRODUCTS

Resource Use

Performance

Observational Data

Processed Data and Products

Planned Outcome: Satisfaction of present and future needs for continuous reliable data and information on Australian weather and climate.

Meteorological and Related Data and Products is the major output of the basic observation, communications and data processing systems provided to maintain a round-the-clock nationwide weather watch and to meet present and future national and international needs for raw and processed meteorological data. It involves two component outputs:
· observational data; and
· processed data and products.

The basic observational, communications and data processing systems that deliver this major output also provide the common foundation on which virtually all the research, services and international outputs of the Bureau depend.

The observation system includes a Bureau-staffed surface and upper-air network of 50 stations, a surface network of over 800 synoptic observation stations, over 6000 volunteer rainfall observing stations and a range of other specialised networks and facilities such as weather watch radar, flood warning, lightning detection, drifting buoys, solar and terrestrial radiation, ozone and satellite data reception. The communications system consists of an integrated network of satellite, radio, facsimile and computer facilities for data collection and forecast and warning dissemination. The major analysis and prediction centres are the National Meteorological Operations Centre in Melbourne and the seven Regional Forecasting Centres, one in each State and the Northern Territory. Engineering, workshop and computing facilities that constitute an integral part of the observation, data collection and processing systems in the Bureau Head Office and the Regions are also key contributors to the output. These facilities include the joint Bureau-CSIRO High Performance Computing and Communications Centre (HPCCC).

National planning, management and coordination of the individual systems and activities, together with a number of central operations functions, resides with the Observations and Engineering Branch and the Central Operations and Systems Branch in the Bureau Head Office. The remaining operational activities are the responsibility of the Regional Offices and Field Meteorological Offices in each State and the Northern Territory.

To ensure the effective output of meteorological and related data and products, particular attention is given to regular monitoring and review of the Bureau's basic technical systems, the replacement of obsolescent systems and the introduction of new technologies via an ongoing re-equipment program, and regular liaison with users of the data and products of the basic systems to ensure their continued appropriateness and effectiveness.

Resource Use

The resources committed to Meteorological and Related Data and Products in 1999-2000 are summarised in Table 3 and given in more detail in Table 4.

Performance

Performance during 1999-2000 was assessed at two levels in terms of the:
· contribution to the achievement of the planned outcome; and
· quality, quantity and price of the outputs directed to the achievement of the planned outcome relative to agreed target levels.
The measures used were as published in the Portfolio Budget Statements 1999-2000 for the Environment and Heritage Portfolio (Budget Related Paper No 1.7).

Performance indicators relating to the achievement of the planned outcome for 1999-2000 are given in Table 5.

4. Meteorological and Related Data and Products expenses and revenue ($'000) and staff level for 1999-2000 compared with reconstructed estimates for 1998-99 and with the 1999-2000 Budget Estimates and Budget plus Additional Estimates appropriations.

	ESTIMATED	BUDGET 2	BUDGET &	ACTUAL
	ACTUAL		ADD. EST.
	1998-99	1999-2000	1999-2000	1999-2000
	($'000)	($'000)	($'000)	($'000)
FINANCIAL
EXPENSES
Employee Expenses (Appropriation)	47,573	47,567	46,778	48,650
Employee Expenses (Section 31)	293	291	870	871
Supply of Goods and Services (Appropriation)	22,554	25,480	26,544	29,439
Supply of Goods and Services (Section 31)	994	1,000	1,234	1,012
Operating Lease Rentals	7,862	7,321	7,368	6,882
Depreciation	33,167	25,410	25,375	27,753
Other Goods and Services Expenses (WMO Contribution)	0	0	0	0
Capital Use Charge	14,098	14,680	14,661	12,585
TOTAL PRICE OF OUTPUT	126,541	121,749	122,829	127,192
REVENUE
Appropriations	125,254	120,458	120,685	120,685
Sale of Goods and Services	1,288	1,291	2,103	2,103
Miscellaneous - other	0	0	41	41
TOTAL REVENUE	126,541	121,749	122,829	122,829
STAFFING
Staff Years (actual)
- Funded from Employee Expenses (Appropriation)	765.6	698.4	698.4	690.1
- Funded from Supplier Expenses (Appropriation)	5.9	4.6	12.2	12.2
- Funded from Section 31 Receipts	1.5	2.9	6.9	6.9
- Funded from Capitalised Salaries (Asset Replacement)	0.0	62.5	62.5	62.5
TOTAL	773.0	768.4	780.0	771.7
2. Includes adjustments made to initial output allocations (as published in the Portfolio Budget Statement 1999-2000) in the light of further analysis of the overall resource situation following the Budget.

Table 5. Indicators of achievement of the planned outcome: Satisfaction of present and future needs for continuous reliable data and information on Australian weather and climate.

The extent to which:· the density, representativeness, accuracy, homogeneity, continuity and reliability of the national meteorological observation network are sufficient to: - meet essential future national and international needs for Australian climate data; - provide the basis for routine nationwide weather watch and numerical prediction operations; and - provide a common foundation for the provision of basic and special weather services; · the meteorological data from the observational network are transmitted to the National Meteorological Operations Centre (NMOC) and Regional Forecasting Centres (RFC) error-free and within cut-off times and meteorological information, forecasts and warnings are communicated promptly and accurately to users; · the central and regional computer systems and computing advisory and consultative services meet the defined requirements of all Bureau programs and contribute to overall efficiency and productivity gains; · equipment installations satisfy the requirements of Bureau programs, are carried out within time and cost estimates and the performance and reliability of operational services are maximised; · the NMOC provides reliable, timely analysis and forecast guidance products that impact positively on the quality of services; and · the NMOC and the RFCs provide, in combination, a sufficiently comprehensive and responsive nationwide and regional scale weather watch and monitoring operation to detect and react immediately to the first evidence of developing dangerous weather and provide a foundation for the provision of routine basic and special weather services.

A summary of the 1999-2000 performance results for Meteorological and Related Data and Products against planned targets for quality, quantity and price is given in Table 6. Detailed discussion of these results follows under the component output headings.

Table 6. Summary of 1999-2000 performance in terms of the quality, quantity and price of Meteorological and Related Data and Products.

	PERFORMANCE MEASURE	TARGET f	ACTUAL
QUALITY	Percentage of scheduled observations from the regular surface, upper air and space-based networks received on time and within prescribed limits	85-95%	80-98%
	Duration of significant outages of major items of field equipment	7-22 days	7days (Av.)
	User satisfaction with the functional capacity and availability of mission critical communication system components	TBA	90%
	User satisfaction with the functional capacity and availability of mission critical computing system components	TBA	78%
	Accuracy of centralised analyses and forecast guidance as measured by statistical evaluation procedures - 72, 120 and 168-hour Anomaly Correlation - Gain in skill of model forecasts over persistence - Error in Model Output Forecast Maximum Temperatures - Error in Model Output Forecast Minimum Temperatures - Error in 24-hour sea-state forecasts	TBA	80,55,40% 30 3°C 2.5°C 0.7m
	User assessment of the value of forecast guidance product components	TBA	88%
QUANTITY	Numbers of fully operational stations: - Upper air	50	50
	- Synoptic cooperative	452	439
	- Climate cooperative	66	25
	- Automatic	380	396
	- Rainfall	7,460	7,337
	- Drifting buoys	17	13
	- Voluntary observing ships	88	91
	- River height	740	1,532g
	- Weather watch radars	49	50
	- Satellite ground	11	9
	- Solar radiation	14	14
	- Total ozone	5	5
	Number of inter-office communication links operational	97	84
	Number of automated data collection links operational	903	903
	Number of international communications circuits operational	9	9
	Number of Regional Forecasting Centres	7	7
	Number of AIFS-equipped offices	29	30
PRICE	Cost of operation of surface, upper-air and space-based observing networks	$68.252m	$77.344mh
	Cost of installation and maintenance of operational monitoring and prediction systems	$15.218m	$12.136m
	Cost of operation of national and international meteorological communications networks	$11.033m	$9.071m
	Cost of operation of central and distributed data processing systems	$22.724m	$22.659m
	Cost of operation of numerical and manual operational analysis and prediction systems	$6.907m	$5.982m
f Targets published in the 1999-2000 Environment and Heritage Portfolio Budget Statement (PBS). TBA - targets for these measures were not established at the time the PBS was published. g Change of definition - see discussion under Observations below h Price variations generally are due to (i) a re-valuation of assets post-budget which affected depreciation expenses and Capital Use Charges; and (ii) changes to the formula for calculating operating lease rentals (OLR) which led to a re-balancing of OLR across all Major Outputs. In the case of operating observing networks, variation (i) resulted in depreciation expenses increasing by approximately $7m and Capital Use Charges by approximately $3m.

Observational Data

Observations

The Bureau operates an efficient, integrated observations program that is designed to meet the data requirements of its services and research outputs, and its national and international commitments. Particular emphasis is placed on ensuring that the quality of data is maintained to the exacting levels required for the national climate record.

The network extends throughout the Australian region including territories in Antarctica and the Indian and Pacific Oceans, using observational systems operated by Bureau staff, volunteers and contractors. Observations from this network are supplemented by data from automatic weather stations, drifting buoys, aircraft, ships, meteorological satellites and weather watch radars. In addition to the conventional meteorological variables such as wind, temperature, humidity, rainfall and pressure, specialised data on atmospheric ozone, solar radiation, chemical constituents (such as carbon dioxide, hydrofluorocarbons and aerosols) and ocean swells and waves are also collected. The main components of the observation network are summarised in Table 7.

The program is planned and coordinated from within the Bureau's Head Office Observations and Engineering Branch, but the majority of operational activities are implemented through its Regional Offices. Specialised activities such as the satellite, ozone and solar radiation programs, major equipment installations and some aspects of the marine program are undertaken from within the Head Office. The funding and management of the Baseline Air Pollution Station at Cape Grim in Tasmania is also coordinated from within the Observations and Engineering Branch, but the station operates quasi-autonomously.

Table 7. The main components of the Bureau's observation network as at 30 June in each year from 1997 to 2000.

Data quality procedures within the observations program were increasingly automated, in line with the increasing proportion of data collected in digital format. The condition of instrumentation, however, and the competence of observers continued to be critical to ensuring data quality overall. Substantial resources are dedicated to inspection visits by specialised Regional staff to stations and facilities, and to regular training and refresher programs, in order to maintain the necessary observing standards. The performance and calibration standards of instruments, maintained by the Bureau's Physics Laboratory, also play a critical role in ensuring the quality of the data is maintained.

As a baseline, the Bureau seeks to benchmark its observation performance against the criteria set by the WMO concerning the appropriate standards and methods of obtaining meteorological observations. Unfortunately, the large uninhabited area within the Australian observing region, including the surrounding oceans, prevents the Bureau from achieving the network density specified by the WMO. However, the standards for quality and frequency of data can be achieved. In recent years, the Global Climate Observing System (GCOS) has been defined at a density and frequency sufficient to enable climate variability and climate change to be detected and monitored. The observing requirements of the Australian component of GCOS are also taken into account when establishing performance targets for the Bureau's Observational Data Output.

This year saw a general increase in the percentage of scheduled observations performed on time and within prescribed accuracy limits within the surface, upper air and space-based networks. Performance targets of 85-95% of scheduled observations were met or exceeded for the upper air wind measuring network (88%), for the surface synoptic network (88%) and for the spaced-based network (98%). Though marked improvement in the quality of the upper air temperature and humidity (radiosonde) observations was evident again this year, they remained below target at 80% of scheduled observations due, inter alia, to difficulties experienced in staffing remote observing stations. An example of the trends in the scheduled observations performed for 12UTC (Coordinated Universal Time, equivalent to Greenwich Mean Time) at Bureau-staffed surface and upper air stations, relative to the performance level in 1987, is given in Figure 16. The figure is indicative of the historical trend for thisparticular subset of stations and times.

Figure 16. Trends in the percentage of scheduled observations performed at 12 UTC at Bureau-staffed surface and upper air observation stations, relative to the performance level achieved in 1987. Missed observations are due to staff and consumables shortages.

The number of synoptic and climate cooperative stations fell below target this year though the potential data losses were largely offset by the accelerated installation of automatic weather stations (AWS), with 21 being installed during 1999-2000. Some AWS were also fitted with automatic cloud-base and cloud-amount detectors, in the form of laser ceilometers, to fulfil some of the cloud observation functions of the observing station. The cooperative network, however, continues to be very important for ensuring a wide geographic coverage of the continent and valuable visual observations.

Implementation of the `electronic field book', a laptop system which replaces the traditional handwritten observation entries, was completed during 1999-2000. This innovation has proved to be very reliable and greatly reduced the training required for cooperative observers. In addition, its criteria-based quality control greatly improved the quality of data being entered in real time into the climate data bank. The software, which was an in-house development, has generated the interest of overseas meteorological organisations.

Of the 18 drifting buoys deployed by Australia during 1999-2000, including 10 which are Bureau-owned, 13 were fully operational for the year, which is four below target. The number of drifting buoys that are fully operational at any given time varies as new ones are deployed and old ones cease transmission. As long as there are at least 12 reporting within the Australian region, the data needs of the analysis and prediction systems are not seriously compromised.

The number of river height stations providing data for hydrological services continued to grow this year in response to user demand for new and improved service coverage. This demand, which is negotiated through State-based Flood Warning Consultative Committees, sets the target for the size of the network each year. The large discrepancy in Table 6 betweenthe target value for this year (740) and the reported result (1,532) is due to a recent change in the definition of the network. In previous years only a "core" set of stations where the Bureau owned and/or operated a significant component of the data collection and communication equipment was identified as the flood warning reference network. However, with the introduction of the Australian Integrated Forecast System (AIFS) and the subsequent integration into AIFS of river height data from all sources, including from stations owned and operated by other agencies, the definition of the river height network has been expanded to include all sources of river height observations which are available to the flood forecasting system.

In order to manage the quality, maintenance and calibration requirements of the Bureau's observation networks more efficiently, the relational data base of observation network metadata, SitesDb, was developed further this year, with a Web based version providing easier access by all staff and full implementation throughout the network providing a sound management tool for many programs.

Significant use was made of remotely sensed satellite data to supplement the data available through surface-based networks. The Bureau continued to benefit substantially from the operational meteorological satellite programs of other countries (such as Japan, China, the US and the European Union) and contributed to these programs through the operation of satellite ground stations. The Bureau maintained nine fully operational satellite ground stations during 1999-2000, two less than the nominated target. The difference arose firstly through the early decommissioning of the Chinese geostationary satellite FY-2A, ahead of schedule, with the effect that the Bureau Turn Around Ranging Station (TARS) for this satellite was only partly utilised this year. China launched a replacement satellite, FY-2B, in late June and the Sydney ground station is now operational. Secondly, there were delays in receiving approval for the installation of satellite reception facilities at the New South Wales Regional Office for service to the Sydney Olympic Games.

The Japanese geostationary meteorological satellite GMS-5, which at year's end was the Bureau's source of hourly imagery over the Australian region, was due to be replaced during 1999-2000 by the Multifunctional Transport Satellite (MTSAT). Unfortunately the $US94msatellite was destroyed in November during a launch vehicle failure. Until the launch of a replacement in early 2003, GMS-5 is now programmed to remain in operation. However, since its nominal five-year design life will soon expire, the Bureau has developed contingency plans to cover the possibility of GMS-5 failure over the next few years.

A major upgrade of the satellite data archive was completed this year, with significant improvements to quality control and monitoring systems, as well as a Web-based facility for the selection and ordering of satellite data and products (www.bom.gov.au/climate/satellite).

Engineering

The Bureau has a substantial inventory of complex technical facilities distributed across Australia, including some 57 meteorological offices, many in regional Australia and the Territories. Most of these sites contain specialised facilities for meteorological observations and telecommunications and for the provision of information to users of Bureau services. Additionally, there are weather watch radars, meteorological satellite reception stations, automatic weather stations and other installations, often in remote and inaccessible sites. The engineering support function has responsibility for maintaining this equipment inventory and for installing new and replacement facilities when required. A combination of engineers and technical officers working from the Bureau's Head Office and Regional Offices, in partnership with external contractors, perform this work.

The specialised skills and knowledge of the Bureau engineering staff were also used to great effect in support of aid programs to other National Meteorological Services in the region, particularly in the South-West Pacific and South-East Asia, and in commercial projects, often overseas, through the Bureau's Special Services Unit.

Major engineering achievements during 1999-2000 included:
· the installation of 21 new automatic weather stations;
· the relocation of the Brisbane Airport weather radar;
· the relocation of a Doppler weather radar to the Sydney area in support of the 2000 Olympic Games forecasting effort;
· the installation of automatic weather balloon facilities (autosondes) at Eucla and Charleville;
· the installation of an electrolytic hydrogen generator at Charleville;
· the upgrade of the weather radar at Hall's Creek;
· the construction and installation of equipment at the new Meteorological Office at Brisbane Airport; and
· the provision of specialist engineering support and engineering advice to National Meteorological Services in Papua New Guinea and Fiji.

The new Meteorological Office at Brisbane Airport was opened during the year.

The effective operation of the Bureau's extensive observation networks relies upon the satisfactory installation and maintenance of observation equipment and facilities within time and cost constraints. Following a rigorous annual planning exercise, schedules and budgets were established for all new facilities, strategic upgrades and ongoing maintenance. The performance of the engineering support function is measured by the extent to which planned installations and replacements are achieved within schedule and budget, and by the effectiveness of maintenance activities in ensuring the reliable operation of equipment. Statistics of equipment faults and maintenance performance were maintained through a simple management information system. Substantial progress was made on improving the operational management of engineering activities with the implementation of a comprehensive interactive computer-based information system, which allows staff to maintain records of all aspects of their engineering work. The resulting database allows monitoring of faults and rectification procedures, providing important performance information, and leading to more effective management of equipment maintenance. In this way, comprehensive statistics are built up to monitor procedural efficiencies and effectiveness and to benchmark these against previous years and against international standards.

In 1999-2000, all equipment installations that were not delayed by external influences were completed on time, within budget and to user requirements. Where influences external to the Bureau forced delays, plans were adjusted within established budgetary constraints. All major equipment faults were repaired according to the Bureau's equipment maintenancestrategy, which establishes a priority for repair, based on the criticality of the site for the successful delivery of services in the short-term (days to weeks). Equipment outages at high priority sites, including those critical for monitoring severe weather events such as tropical cyclones and for supporting aviation operations, were kept to a minimum this year, and the average duration of significant outages of all major items of field equipment was about seven days. Figure 17 shows the average duration of outage per equipment fault from 1990 to 1999 for automatic weather stations and weather watch radars. The main influences on outage time were the availability of staff and/or spare parts and the location of the equipment.

Figure 17. Average duration of outage per equipment fault from 1990 to 1999 for (top) automatic weather stations and (bottom) weather watch radars.

The number of faults per automatic weather station per year decreased this year, to an average of 2.8, as the benefit of the equipment replacement program, funded through the Government response to the 1996 Review of the Operation of the Bureau of Meteorology, began to take effect. The incidence of faults within the weather watch radar network, however, remained high, averaging about 4.5 faults per radar per year, owing to the increasing age profile of the equipment and the fact that many are in use 24 hours a day.

Processed Data and Products

Communications

Telecommunications systems and services are a vital part of the Bureau's integrated operational infrastructure. The Bureau operates and manages its own specialised telecommunications systems making use of services leased from telecommunications carriers. The collection of meteorological observations (often from remote locations), the exchange of data and products between Bureau offices, and the dissemination of the Bureau's services such as forecasts, warnings and specialised products, all require telecommunications facilities.

The Bureau's telecommunications network, Weathernet, connects the National Meteorological Operations Centre with each of the seven capital city Regional Forecasting Centres, the Canberra and Townsville Meteorological Offices, the Sydney Airport Meteorological Unit, the Antarctic Meteorological Centre at Casey, eight Weather Service Offices, 38 other Field Meteorological Offices and many radar sites. In 1999-2000, Weathernet comprised 84 inter-office communication links. The published target of 97 for this measure was inappropriate and will be revised downward.

Weathernet is implemented on Telstra's Accelerated Frame Relay service. The Transmission Control Protocol/Internet Protocol (TCP/IP) supports the various communications applications carried on Weathernet, for exchange of operational data and products, as well as providing organisation-wide access to a large range of information from the Bureau's internal 'Intranet', and the global Internet. During 1999-2000, plans for the next generation of Weathernet were developed and a trial upgrade to the main Weathernet link between Melbourne and Sydney to Asynchronous Transfer Mode was implemented. The high bandwidth connection (4 megabits per second) will allow improved data distribution mechanisms to be tested.

An automatic text-to-speech (TTS) system was implemented in the Queensland and South Australian Regional Offices. TTS allows for automatic generation of voice content for recorded telephone weather services. The system adopted makes use of concatenated human voice. Thousands of speech elements, recorded by a voice artist, are computer assembled corresponding to the text content of actual forecasts and warnings produced in forecasting offices. The composite voice file is then available for download to recorded telephone service providers. The first TTS outputs in Queensland were public weather forecasts including marine forecasts for the 1196 Dial-it service in Rockhampton, Townsville and Cairns. The TTS system is expected to be implemented in all remaining Bureau Regional Offices by the end of 2000.
The Australian Meteorological Data and Information Service System (AMDISS) progressed substantially during 1999-2000. A comprehensive suite of information is now available on the Bureau's Web server (http://www.bom.gov.au) for free access as part of the basic serviceand by accredited user access under cost recovery arrangements. The average weekly hit rate for 1999-2000 was 2 million, with peak periods coinciding with tropical cyclones. Monthly hits exceeded the 17 million mark in March 2000 (Figure 18). The Bureau Web site consistently rated in the top three Federal Government sites in Australia for the year.

Figure 18. Monthly hits (millions) to the Bureau's Web site from March 1996 to June 2000 with notable tropical cyclone events marked.

Melbourne is one of the three World Meteorological Centres (WMC) of the WMO Worl Weather Watch (WWW). Along with Moscow and Washington, Melbourne plays a pivotal role in the communication of WWW data and products to the world meteorological community. Nine international links are currently operational from Melbourne, to Tokyo, Bracknell (UK), Jakarta, Wellington, Noumea, Nadi, Port Moresby, Honiara, and Singapore. The number has been at that level since 1995, with one addition (Singapore) and one line being discontinued (New Delhi) in the same year, 1998-99. The major elements of this Global Telecommunications System (GTS) are shown in Figure 19.

Figure 19. The major telecommunications links of the Global Telecommunications System of the World Meteorological Organization World Weather Watch , showing the interregional and regional circuits (thin lines) and the Main Telecommunications Network (bold lines).

During the year several improvements were made to the Bureau's international communication links including:
· upgrading of links to Indonesia and Bracknell (UK) to Frame Relay;
· upgrading of links to Japan and New Zealand to use TCP/IP, with an ISDN backup being provided for New Zealand; and
· connecting the Solomon Islands via the Internet.

The Computer Message Switching Systems (CMSS), which is at the centre of the Bureau's telecommunications, receives and forwards meteorological data (reports) and processed products both domestically and internationally. In recognition of the increased productivity and improved services made possible through CMSS, the Technology in GovernmentCommittee awarded the Bureau a 2000 Government Technology Productivity Silver Award. CMSS shared this award with SILO, the Internet Web site established in 1997 to provide a specialised meteorological information service for rural communities, decision makers, researchers and educationalists. This was the fourth such award presented to the Bureau in the last 13 years. The Bureau previously received a Gold Award for its Automated Regional Operations System (AROS) in 1988 and a Silver Award for the satellite control ground station at Crib Point (Victoria) in 1994 and for the Australian Integrated Forecast System in 1998.

It is difficult for users to make informed assessments of the performance of individual communications components, owing to the tightly coupled nature of the Bureau's computing and communications infrastructure, and to the move to client server computing and the use of the World Wide Web. User-perceived slow communications, for example, can be the end result of computing infrastructure or computer applications design problems or problems associated with external communications carriers. The Bureau's mission critical communications systems are Weathernet, Local Area Networks supporting operational areas, the Computer Message Switching System, Digital Facsimile System, Australian Meteorological Data and Information Service System and connections to the Global Telecommunications System.

No systematic surveys of the overall user satisfaction of mission critical communications systems have been conducted. However user surveys looking at specific components of the communications infrastructure have been conducted, particularly following the implementation of new systems. User surveys following the recent implementation of the CMSS within the Australian Integrated Forecast System indicated that 90% of users were satisfied with the ease of use, functionality and availability of the CMSS component.

The timely and accurate transmission of meteorological observations, exchange of data and graphical information between Bureau offices, and dissemination of the Bureau's services such as forecasts, warnings and specialised products, are dependent on the effective and efficient operation of the Bureau's communication systems. During 1999-2000, 99% of surface data and 95% of upper air data were received at the National MeteorologicalOperations Centre before the nominated cut-off times for input into the Bureau's analysis and prediction systems and 95% of output products from these systems were delivered to the Regional Forecasting Centres before the scheduled deadlines for dissemination. These performance levels were more than sufficient for the effective communication of meteorological information, forecasts and warnings to users.

Computing

The Bureau's computing infrastructure includes the central computing systems, which are an integral part of the operations of the National Meteorological Operations Centre and which support the large scale numerical modelling research in the Bureau of Meteorology Research Centre, distributed computing systems as part of other specialised facilities and programs, and the computing systems which support the Bureau's regional operations.

The joint Bureau of Meteorology/CSIRO High Performance Computing and Communications Centre (HPCCC) continued to provide reliable and efficient high performance computing services in support of the operational and research needs of both organisations. A significant development during the last quarter of 1999 was the adoption of a supercomputer strategic plan to scale operational and research systems through until 2003. The strategy included the replacement of the current NEC SX-4 supercomputer with an SX-5 supercomputer by August 2000 and the addition of a second supercomputer by November 2000.

The NEC SX-5 supercomputer for the High Performance Computing and Communications Centre was delivered during the year to replace the previous NEC SX-4.

An extensive upgrade to the Hierarchical Archive Mass Storage system was commenced in 1999 to prepare for the upgrade to the NEC SX-5 supercomputer, and to satisfy the growing demand for data storage of numerical weather prediction (NWP) products. The upgrade of the mass storage system increased the capacity of the Bureau's mass storage silo from one Terabyte to 104 Terabyte with a possible further expansion to 1000 Terabyte. The huge volume of data generated by the supercomputer is a significant issue for the Bureau with some 400 to 500 gigabytes of new NWP data generated each week.

During the year, the Australian Integrated Forecast System (AIFS) was made operational in Western Australia, South Australia and Tasmania/Antarctica. In addition, a special version of AIFS was implemented at the Navy's Fleet Weather and Oceanographic Centre (FWOC) and the Defence Meteorological Support Unit (DMSU). A temporary AIFS-equipped forecasting office was also set up in September 1999 at Rushcutters Bay on Sydney Harbour for the Pre-Olympic Sailing Regatta, to test operational systems for the 2000 Sydney Olympic Games. The final AIFS installation completed the modernisation of the computing infrastructure that commenced with the Victorian Regional Office installation in 1996.

User surveys of mission critical computing systems commenced in 2000, starting with detailed user surveys following the implementation of the AIFS in Regional Offices. These indicated that 78% of users were satisfied with the speed, ease of use, functionality and availability of AIFS applications.

Analysis of outage times for mission critical computing systems showed that the computing systems were extremely reliable during the year with an overall availability of around 99.74%. The reliability of AIFS continued to improve in 2000 with an average availability across all systems of 99.84%.

The Bureau's transition to the Year 2000 and through the `additional' leap-day on 29 February, passed with no impact on the Bureau's operations and hence on the Bureau's services to the community. This was the result of an extensive Year 2000 Compliance Project which aimed to ensure that all essential Information Technology (IT) and embedded systems, on which the Bureau is heavily dependent, performed correctly during the transition to the Year 2000 (from 31 December 1999 to 1 January 2000) and during 29 February 2000.

The Bureau's National Meteorological Operations Centre, as one of three WMO World Meteorological Centres (WMC), acted as a Year 2000 Situation Centre along with WMC Washington, WMC Moscow and the UK Meteorological Office in Bracknell. Melbourne was allocated the responsibility for monitoring, reporting and coordinating actions for the Regional Telecommunications Hubs (RTH) in Region V, Antarctica and part of Region II. These included RTHs in China, India, Iran, Japan, Saudi Arabia, Thailand, New Zealand andAustralia. No significant international data losses were reported during the Year 2000 transition.

A substantial legacy of the Year 2000 Project was the development of a comprehensive risk management strategy for operational systems and the consolidation of Bureau wide contingency plans. These plans have been used subsequently during disruptions to power and telecommunications in Darwin and Perth.

Analysis and prediction

The analysis and prediction function embraces the basic meteorological analysis and prediction operations needed to support the provision of weather and climate services and to fulfil Australia's international obligations under the Convention of the World Meteorological Organization. The National Meteorological Operations Centre (NMOC) in Melbourne, the seven Regional Forecasting Centres (RFCs) in the capital cities, the Townsville and Canberra Meteorological Offices and the Antarctic Meteorological Centre at Casey function as an integrated national network to produce a range of manual and automated guidance products which support the nationwide operational forecast and warning services provided by the Bureau of Meteorology.

Each RFC is responsible for detailed meteorological and related analysis and prediction for its State or Territory and adjacent ocean areas. The Darwin RFC has a special responsibility for the preparation of analysis and prediction products for the tropical region between the central Indian and the central Pacific Oceans. NMOC serves as the central operational hub, combining the roles of operational communications and computing with meteorological and oceanographic analysis and prediction functions. The computing resources provide the processing capacity for operation of the centralised numerical weather and oceanographic analysis and prediction systems, while the communications system is used for national and international data exchange and to disseminate generated products.

Computer models are used within NMOC to prepare products showing the current orpredicted conditions in the atmosphere or ocean. During the year, several operational models were used to provide products with different emphases:
· the Global Assimilation and Prediction System (GASP) for predictions to seven days;
· the Limited Area Prediction System (LAPS) for more detailed 1 to 2 day predictions over the Australian region;
· a version of the LAPS system especially adapted for tropical regions (TLAPS), including the capability to generate a fine mesh for the prediction of tropical cyclone movement and development;
· a fine scale version of LAPS (MESOLAPS) over Australia to provide more detail on the flow patterns; and
· an atmospheric transport model to provide predictions of the movement of atmospheric pollutants or volcanic ash.

Routine operational oceanographic products include:
· analyses of the sea surface and sub-surface temperature;
· sea-state predictions over three domains (global, regional and southeast Australia); and
· long-range predictions of the sea surface temperature of the central Pacific Ocean to assist in the preparation of the Bureau's seasonal outlooks.

The performance of the Bureau's analysis and prediction systems during 1999-2000 was assessed on the basis of a range of statistical measures of the accuracy of the centralised analysis and forecast guidance products.

For the global prediction system (GASP) anomaly correlation targets of 75%, 55% and 40% were set for the 72, 120 and 168-hour Mean Sea Level Pressure (MSLP) predictions respectively over the latitude bands from 20°S to 60°S. The anomaly correlation is the time-dependent correlation between the observed anomaly of a field (in this case MSLP) and the forecast of the anomaly, where the anomaly is defined with respect to the model's mean field, and is a common measure of skill in numerical weather prediction. The higher the correlation the better the predictive skill of the model. Figure 20 shows the time history of the anomalycorrelations of the global prediction system for the three forecast periods from early 1998 to June 2000. This sequence follows a significant upgrade to the GASP model in 1996 and a long term trend has therefore not yet been established. The variation over the last two years has been quite small although the trend has been slightly downward. Improvements in model resolution and model physics are expected to increase the anomaly correlation over the next 12 months.

Figure 20. Anomaly correlation of 72 hour (top), 120 hour (middle) and 168 hour (bottom) MSLP predictions for 20° to 60°S from the global prediction system.

For the limited-area prediction system (LAPS), the performance measure identified was the gain over persistence in the 12 month running mean skill score for the 24-hour prediction of MSLP, for which a target of 30 points was set. The persistence prediction, which assumes that today's pattern will continue unchanged, acts as a threshold for predictive skill. Thus the gain in the skill of the model over the skill of persistence is a measure of the improving accuracy of the model's predictions. Figure 21 shows the performance of the limited-area system as described by this measure from 1972 to June 2000. Starting from a difference of around 8 skill points between persistence and the limited-area model in 1972 the difference has steadily increased to around 30 for this year. The large increase in skill since 1996 reflects a significant improvement in model resolution and physics made possible following an upgrade in supercomputing facilities. Enhancements to the model implemented this year, including a doubling in the horizontal resolution, bringing the computational grid down to 37.5 km, and an extra ten levels in the vertical, had a positive impact on the accuracy of the predictions of the upper layers of the atmosphere but little impact on the MSLP field.

Figure 21. Values of the S1 skill score, a measure of the errors in prediction, for 24 hour forecasts of mean sea level pressure from operational and persistence prognoses over the Australian region. The values shown are 12-month running means. The persistence predictions, based on the assumption that today's pattern will continue unchanged tomorrow, show relatively large errors and no long-term improvement. The operational predictions from the Australian region analysis and prediction system have shown general improvementover time, with a strong improvement associated with the latest upgrade. The original base analysis used for verification purposes has been discontinued; results during the overlap period show a slight shift in the measure of skill, but the trends are similar.

A system of forecast guidance, called Model Output Forecasts (MOFs), is also derived from the LAPS model. This system, which is based on a linear regression of historical predictions and near-current weather observations, provides three-hourly predictions of weather elements out to 48 hours for 447 sites throughout Australia. These predictions are used as guidance by the RFCs in their provision of detailed weather forecasts. The statistical performance measure identified for this system was the annual average root mean square error of the 24-hour MOFs of maximum and minimum temperatures averaged over all States. Targets for this measure of 3°C and 2.5 °C for maximum and minimum temperatures respectively, were set for 1999-2000. Figure 22 shows the annual average root mean square of the 24-hour MOFs of maximum and minimum temperature for all capital cities and the national average for the period 1995 to 1999. Averaged over all States, the error in maximum temperatures has generally improved since 1995, dropping to 2°C in 1999-2000. The error in minimum temperatures has decreased slightly over the last five years but, minimum temperatures being more susceptible to local effects, there has been no steady trend. Averaged over all States, the error in minimum temperatures for 1999-2000 was 2.1°C.

An extension of the Model Output Forecasts was introduced in November 1999 to provide forecast guidance out to seven days. This version, which was derived from the GASP model, produced 7-day forecast guidance for 689 locations throughout Australia.

Figure 22. Annual average root mean square (RMS) error in the 24-hour Model Output Forecast (MOF) maximum (top) and minimum (bottom) temperatures for capital cities and the national average, from 1995 to the present.

For oceanographic products the performance measure identified was the annual average root mean square (rms) error in the 24-hour predictions of sea state, from the Australian Region Wave Model, averaged over all available Australian instrumented wave observations. Thetarget set for 1999-2000 was 0.7 metres. Figure 23 shows the trend in both the number of available observations from wave-rider buoys and the errors in the 24-hour sea state forecasts, from January 1999 to the present. For 1999-2000, the annual average rms error was at the target level. The introduction of new satellite wave height data to the starting analysis of the wave model is expected to lead to improvements in the forecasts of waves and swell, which should show up as a decrease in this error measure.

Figure 23. Verification statistics for the Australian Region Wave Model showing the number of wave-rider buoy observations and the annual average root mean square error in the 24-hour forecasts of sea state.

Other model enhancements implemented during 1999-2000 included: a new version of the MESOLAPS model with a horizontal resolution of 12.5 km over the entire Australian domain and a special 5 km resolution version specifically for the Sydney region to assist weather forecasters provide services during the Sydney 2000 Olympic Games; increased resolution and better use of satellite data to define cloud and moisture in the TLAPS model; and an improved high resolution model for the prediction of tropical cyclones.

User surveys to assess the value of forecast guidance products commenced in 1999 as part of the Review of the Analysis and Prediction Program. The overall performance of the NMOC, the Regional Specialised Meteorological Centre in Darwin and the forecasting office at Casey (Antarctica) was assessed and the percentage of users who rated them as above average (rating of 4 or 5 on a 5 point scale) was 88%. Over 95% of users surveyed stated that the quality of the output had improved over the last five years.

During the year, NMOC continued to exceed service levels for timely delivery of forecast guidance, with 98% of forecasts received before the scheduled deadline for dissemination. Despite an increase in volume, the distribution of analysis and forecast guidance products improved during the year, due to the high level of automation of the dissemination systems and improved computing and communications systems.

Other significant activities undertaken by the NMOC during the year included the provision of specialist forecast support to the Dick Smith/John Wallington Trans-Tasman balloon flight, and weather advice during the single-handed round-the-world voyage of the paraplegic sailor Vince Lauwers.

As part of international efforts to enhance global data exchange, the NMOC commenced the distribution of standard-hour weather observations and expanded the range of analysis and forecast products available to countries in South-East Asia and the South-West Pacific.

Since 1995, the NMOC has served as a WMO Regional Specialised Meteorological Centre (RSMC) for Environmental Emergency Response (EER), with responsibility for providing advice on the atmospheric transport of pollutants resulting from nuclear disasters, volcanic eruptions, forest fires, chemical incidents and other causes. The atmospheric transport prediction system continued to be maintained in a state of readiness so that requests for advice could be satisfied quickly. The system was used this year to provide advice on the transport of radioactive particles following the accident at the Tokaimura nuclear fuel fabrication facility in Japan in September 1999. The RSMC also took part in a joint trial of the system with the Comprehensive Nuclear Test Ban Treaty Organisation from 3-5 May and a full global trial involving meteorological services, national atomic energy agencies and the International Atomic Energy Agency on 27 June 2000.

Previous Chapter | Back to Index | Next Chapter