Meteorological and Related Research is delivered through three individual outputs that contribute to the achievement of the desired outcome. The developments in each of these outputs during 2003-04 and their contribution to the outcome is provided below.

PURE RESEARCH 

A small but essential component of BMRC activity is directed to pure research, which has intrinsic scientific value and is aimed at advancing the science of meteorology. This research lays the foundation for other research in the BMRC by ensuring that the work is well connected with current international advances in the science.

• Increased understanding of the substantial changes in the nature of storm development that have occurred since the mid-1970s through studies analysing the atmospheric circulation in the southern hemisphere over the past fifty years. For example, there has been a 30 per cent reduction in the intensity of storms and an eastward shift in the centre of storm activity relative to the southwest of Western Australia. These changes are consistent with the observed reduction in rainfall over southwestern Australia in recent decades (Figure 14).
• Continued development of a technique to extract the predictable component of climate variations from the unpredictable 'weather noise' of winter fronts and storms. The technique was used to demonstrate the extent to which seasonal climate variations can be predicted in the northern hemisphere.
• Continuation of research on the Madden-Julian Oscillation (MJO), which controls much of the week-to-week variation of the weather in our tropics and sub-tropics. The research has helped clarify the physical processes through which coupling between the ocean and atmosphere control the behaviours of the MJO. The work has included analysis of the influence of the MJO on the El Niño phenomenon

Figure 14. The dominant mid-latitude storm tracks, which affect southern Australia, for the periods 1949 to 1968 (upper) and 1975 to 1994 (lower). Shown here is the relative impact (contours and shading) of the storms at 300hPa as they develop and propagate eastward. The associated troughs are deepest in those regions of highest contour values and darkest shading. Analysis shows that storm development is about 30 per cent weaker in the latter period and that there has been an eastward shift in the area most affected by storms. These changes are consistent with the reduction in rainfall observed over southwest Western Australia in recent decades.

Contribution towards outcome

• Pure research in the BMRC contributes directly to the planned outcome of advancing meteorological science and understanding of the mechanisms of Australian weather and climate.
• The work on changes in the southern hemisphere circulation has improved our understanding of the observed changes in rainfall and streamflow that have occurred in southwest Western Australia, and of the basic meteorology of our region.
• While there is much activity around the world on predicting weather and climate, the work on predictability is fundamental in placing theoretical limits on the scope and accuracy of such predictions.
• The MJO is a dominant influence on Australian weather and climate, and is believed to play a significant role in the triggering of El Niño events. The MJO research in BMRC helps clarify the nature of this phenomenon and provides a base for the development of techniques to improve the accuracy of climate prediction in the future.

STRATEGIC RESEARCH 

Much of the strategic research in the BMRC involves development of the scientific infrastructure to manipulate and analyse meteorological and related data. Significant components of this infrastructure are the numerical models that run on the Bureau's NEC SX-6 supercomputer to simulate and predict the state of the atmosphere and ocean. Because of the Bureau's statutory responsibility to provide forecasts and warnings, these models are used routinely to predict future weather and climate variations, and so there is a major effort in BMRC to use observed data to calculate accurate and consistent estimates of the initial state of the atmosphere and ocean. This process is known as data assimilation and it is a fundamental element of modern meteorology and oceanography.

A second major aspect of strategic research is a continuing program of field studies aimed at ensuring that the Bureau remains informed of advances in observing methods, such as radars, and at improving our understanding of atmospheric processes, such as thunderstorms and turbulence.

Major developments 2003-04

• Further consolidation of the unified BMRC Atmospheric Model (BAM) with the development of version 4 (BAM4), which includes enhanced representations of cloud microphysics and the land surface. The computer code for BAM continues to be optimised for running on the Bureau's NEC SX-6 supercomputer.
• Continuation of the daily running of the ensemble prediction system based on the global NWP system (GASP), and the extension of the supporting software to provide more diagnostic products to assist forecasters in the interpretation of the model output (Figure 15).
• Upgrade of the data assimilation software system used to set the initial conditions for the numerical weather prediction (NWP) models and unification across the global (GASP) and regional (LAPS) systems. The unification has been important in ensuring the consistent application of satellite data in the models.
• Conduct of an international workshop in November to validate and consolidate BMRC plans to develop the Ensemble Kalman Filter approach to data assimilation, and the development of a prototype system.
• Demonstration of the accuracy of total precipitable water estimates using ground-based global positioning system (GPS) observations in comparison with those based on conventional radiosonde observations through a collaborative project with Curtin University.
• Ongoing development of the BMRC global atmospheric model which is the basis for the coupled ocean-atmosphere modelling system, known as the Predictive Ocean Atmosphere Model for Australia (POAMA). The POAMA, developed jointly with CSIRO, predicts global and regional climate variations up to eight months ahead and was the first model to predict the rapid decay of the 2002-03 El Niño event.
• Continued research using the POAMA coupled model on the MJO which dominates much of the weather in the Australian tropics and sub-tropics. The model is now found to generate realistic MJOs, and its predictions are being compared with those from the statistical prediction methods developed in BMRC. The research on the MJO has included a theoretical study of the physical basis of the phenomenon (Figure 16).
• Completion of a study on the role of regional warming in the decline in streamflow that has occurred in southwest Western Australia over recent decades, under the collaborative Indian Ocean Climate Initiative (IOCI) program with CSIRO and the Western Australian Government.
• Completion of a number of collaborative studies on the impacts of climate change and variability on human health and on Australian birds.
• Organisation of the fifth international workshop on monitoring climate extremes in South-East Asia and the South Pacific by BMRC in March with support from the Asia Pacific Network (APN).
• Continuation of the analysis of data collected at the BMRC field station and the USA Atmospheric Radiation Measurement (ARM) site in Darwin with a focus on probabilistic and cluster analyses to evaluate the accuracy of model simulations of tropical clouds.
• Commencement of planning for an international field experiment, called the Tropical Warm Pool International Cloud Experiment (TWP-ICE), to be conducted around the BMRC research station and the ARM site in Darwin during the monsoon season in early 2006. The experiment is aimed at improving our capability to model the overall behaviour of tropical storms, and will involve collaboration with overseas and other Australian researchers.

Figure 15. Single frame from the Bureau of Meteorology global medium range ensemble prediction system 500 hPa height 'spaghetti' animation, showing the spread of different 5-day forecasts between members of the ensemble. In these predictions for 00 UTC on 24 May 2004, differences are apparent in the predicted speed and shape of the front passing across the south of the Australian continent. Each of the curves represents a different forecast member from the ensemble; 5600 m contour line from each member is plotted.

Contribution towards outcome

• Numerical models are the basic tool for synthesising meteorological data to improve our understanding of the atmosphere and our capability to simulate and predict atmospheric behaviour. The leading-edge modelling research in BMRC ensures that Australia is able to draw effectively on overseas trends and advances, and that the Bureau has the capability to support its services at the level expected by the Australian community.
• In advancing meteorological science in Australia, the Bureau seeks to collaborate with other research agencies (especially CSIRO) and the universities in order to ensure that meteorological capability is widespread and nurtured.
• Collaboration with overseas groups on measurement programs is vital to ensuring the planned outcome, as it allows the Bureau to maintain awareness of state-of-the-art instrumentation and to influence international trends.
• Collaboration with National Meteorological Services in our region ensures that our national scientific advances can be readily placed in a regional context, and helps promote the exchange of data and information that is vital to effective meteorological science.

Figure 16. The Madden-Julian Oscillation (MJO) is a natural oscillation of tropical winds and rainfall with a typical period of about 50 days. As the MJO traverses eastward from the Indian Ocean into the west Pacific, the chance of extreme rainfall is dramatically reduced across northern Australia in the convectively suppressed phase (Phases 1-3) and dramatically increased in the convectively active phase (Phases 4-6). Phases 7 and 8 represent the progression through to the start of the cycle again (Phase 1) before the strong suppressed convection begins. The "weak" MJO phase represents the probability of extreme rainfall when the MJO is absent. The state of the MJO appears to be predictable with a lead time of up to 3 weeks, which implies predictability of extreme rainfall across northern Australia with a similar lead time.

APPLIED RESEARCH 

A key objective of BMRC is the development of advanced systems to enhance the scope and improve the performance of the operations and services of the Bureau. These systems are generally based on the strategic research of the BMRC, and they are ultimately implemented in cooperation with operational units, such as the National Meteorological and Oceanographic Operations Centre (NMOC), Regional Offices and the National Climate Centre (NCC).

Major developments 2003-04

• Extension of the Thunderstorm Interactive Forecast System (TIFS) to generate thunderstorm alerts for the aviation industry in Sydney and Brisbane (Figure 17). It is planned to extend the system to other capital cities.
• Upgrade of the tropical cyclone version of the BMRC atmospheric model (TCLAPS) to include an improved specification of the initial tropical cyclone vortex. Comparison of the operational results of the model with results from other centres demonstrates that TCLAPS is world-class. Research is also showing that the inclusion of data from new sources, such as surface winds and heating rates derived from satellite-based instruments, further improves the accuracy of the system.
• Upgrade of the Environmental Emergency Response (EER) transport and dispersion model to improve the representation of small-scale turbulence, which is important for short-range transport problems. The model has also been extended to support a volcanic ash detection system for aircraft warnings.
• Continued development of the Meteorological Archive and Retrieval System (MARS) data archive and access system to support a wide range of operational and research data, including ocean data. The web-based software, which allows browsing and retrieval of data, has also been upgraded.
• Development of applications and systems to support the weather service requirements associated with the RNDSUP program. The Nowcaster server workstation has improved system administration and diagnostics, and more detailed information is provided on the characteristics of storm cells in the radar display system (3D-RAPIC).
• Completion of research on the background errors associated with the operational sea wave model which has shown that allowing for the variation of these errors with latitude can significantly improve the accuracy of wave predictions (Figure 18).
• Implementation of a statistical forecast guidance system in the NMOC based on a consensus of forecasts from different NWP systems. The system, which is routinely verified along with the individual forecasts that make up the consensus, is found to provide accurate estimates of daily and hourly weather elements to support public, aviation and fire weather services.
• Trial of a wind change/wind gust forecast product from the MESO-LAPS NWP model in Victoria and Tasmania (Figure 19). Figure 18. The spatial scale of the background error of the global wave model (in km) is determined by comparing four years of wave model output to satellite altimeter observations. In the current operational wave forecasting and data assimilation system, the spatial scale is set to be 300 km globally. Using the new structure shown in the figure, the influence of the observations varies with latitude and the skill of the wave forecasts can be improved by up to 10%.

Figure 17. Graphical thunderstorm warning from the Thunderstorm Interactive Forecast System (TIFS) as at 3:00 am, 18 January 2004 showing the current location of a thunderstorm in southern Queensland, the expected track of the storm over the next hour, and the regions near Brisbane under threat from storms. Shading refers to the intensity of the forecasted thunderstorm.

Figure 18. The spatial scale of the background error of the global wave model (in km) is determined by comparing four years of wave model output to satellite altimeter observations. In the current operational wave forecasting and data assimilation system, the spatial scale is set to be 300 km globally. Using the new structure shown in the figure, the influence of the observations varies with latitude and the skill of the wave forecasts can be improved by up to 10%.

Figure 19. A 15-hour forecast of surface wind gusts over Tasmania for early afternoon on 15 November from the Bureau's high resolution MESO-LAPS numerical weather prediction system. On this day, more than 60 fires occurred over southeastern Tasmania under the influence of the strong and gusty winds. A long barb indicates 5 m/sec, a short barb 2.5 m/sec, and a flag 25 m/sec.

• Tropical cyclones are a major natural hazard in Australia and the modelling research in BMRC ensures that the Bureau maintains a state-of-the-art capability to predict their behaviour.
• The dispersion of airborne material is a potential threat to communities and natural systems, and it is important that the Bureau maintains the capability to predict how material, such as volcanic ash which can disable aircraft engines or bushfire smoke which can affect human health, is transported and dispersed. The improved capability to predict short-range transport is especially relevant to protecting rural communities from outbreaks of diseases such as foot-and-mouth disease.
• Meteorological research involves the handling and exchange of very substantial amounts of data, collected from in situ and satellite-based instruments as well as the output of complex computer models. The development and application of modern data-handling infrastructure is necessary to ensure efficient management of internal data requirements for research and services, and effective communication with colleagues in Australia and overseas.
• Severe weather events, such as thunderstorms, are short-lived but constitute a great natural hazard to Australian communities. The research in BMRC on the detection and prediction of such phenomena is vital to supporting the Bureau's role in providing severe weather warnings to the Australian community.