Chapter 5 Climate Research

WCRP Activities > Climate Variability and Predictability (CLIVAR)

The Climate Variability and Predictability (CLIVAR) program is the main focus in WCRP for studies of climate variability, extending effective predictions of climate variations, and refining estimates of anthropogenic climate change. The program exploits the 'memory' of the oceans to enhance our understanding of the coupled behaviour of the rapidly-changing atmosphere and the slowly-varying land surface, ocean and ice masses. The program builds on the successfully completed Tropical Ocean-Global Atmosphere (TOGA) and WOCE programs.

Improving our understanding of past climate provides a foundation for our understanding of the current and future climates. Research at the Australian Institute of Marine Science (AIMS) extends the climate record back several centuries using data from long-lived massive corals. The development of a machine for measuring luminescence in coral is providing quantitative information on the past occurrence of freshwater inputs into the Great Barrier Reef, and hence on past regional rainfall and river flow. The Australian Coral Records Research Group (AUSCORE) continues to share material and facilities to provide indices of past climate variability from coral skeletons. Outcomes of the collaboration include new insights into the Little Ice Age climate of the southwest Pacific, evidence of a change in land-derived sediment load following European settlement in the region, and evidence of variability in the El Niņo - Southern Oscillation teleconnections.

Research at the University of Newcastle has calibrated monthly-resolution ice core data with instrumental records of sea-level pressure, geopotential height in the lower troposphere and wind fields for the southern extratropics. From this work, reconstructions of large-scale indices of southern hemisphere climate have been produced. Studies of geochemistry and growth patterns of coral micro-atolls in the Southern Cook Islands are also being developed to reconstruct records of multi-decadal sea level and sea-surface temperature for the past millennium. The group has also been developing a reconstruction of the sea wave climatology for the western Tasman Sea using data on longshore sediment transport and coastal stability.

There have been several studies on regional climate variability across the Australian region. Research at AIMS has focused on the variability of the climate of tropical Australia and has included the development of indices for assessing impacts on tropical ecosystems. The nature and causes of tropical Indian Ocean seasurface temperature variations and their role on Indonesian and Australian rainfall variability have been investigated in BMRC.

The first five-year program of the Indian Ocean Climate Initiative (IOCI), involving collaboration between the Western Australian Government, BMRC, CAR and CSIRO Land and Water (CLW), was completed in December 2002, and a new program commenced in July 2003. A comprehensive summary of all the results of the first program was published in September 2002. The program included studies of climate variability, climate prediction, and climate change. For example, CAR conducted modelling simulations of the impact of greenhouse-induced climate change on the southwest of Western Australia (Figure 5.8).

Figure 5.8. Average May-October rainfall 1976-2001 as a percentage of 1925-75 rainfall.

Statistical downscaling provides a means to estimate local climate variables from the output of large-scale climate models (Figure 5.9). CLW and BMRC have compared different techniques for downscaling climate model results for the southwest of Western Australia and for the Murray Darling Basin. This work has been complemented by regional climate modelling simulations at CAR, using a limited-area model nested in the CAR global climate model.

Figure 5.9. A schematic diagram describing the statistical downscaling approach. GCMs provide useful predictions for large-scale atmospheric patterns (lower part). Details contained within a grid box (upper part) are influenced by local features beyond the resolution of current global climate models.

The Bureau of Rural Sciences (BRS) has developed software systems for mapping and analysing climate variables on regional scales across Australia to support policy development in Government. In CMR there has been work on the development of systems to predict seasonal variations in agricultural variables, such as plant growth, based on correlations with sea-surface temperature. The work has included evaluation of the economic value of the predictions to agricultural enterprises.

The operational seasonal prediction system used in the National Climate Centre (NCC) has been extended to include temperature outlooks, in addition to rainfall. Research in BMRC is also exploring the feasibility of statistical predictions at monthly and sub-monthly time scales. In collaboration with the Indonesian Meteorological Service (BMG), BMRC has developed a statistical rainfall prediction scheme for Indonesia for periods up to six months ahead, based on data from 63 stations across the country.

Much of the Australian research related to CLIVAR is associated with the development and application of climate models. The joint Bureau of Meteorology - CSIRO High Performance Computing and Communications Centre (HPCCC) provides a common platform for supercomputing. In particular, it has facilitated collaboration between BMRC and CMR on the development of a coupled atmosphere-ocean model for seasonal prediction research and operations. Since October 2002, the coupled model has been run each day, initialised with both ocean and atmospheric data, in the Bureau of Meteorology, to construct an evolving ensemble of eight-month predictions. The ocean component of the model continues to be refined in CMR, with an emphasis on improving the model physics. The model was used in 2002 to provide accurate and timely guidance of the maturing and decay phase of the El Niņo event (Figure 5.10).

Figure 5.10. A series of forecasts (broken lines) of central equatorial Pacific Ocean sea surface temperatures (SSTs) compared to the actual evolution of SSTs in that region from August 2002 through to Autumn 2003. The forecasts, from the new Bureau-CSIRO coupled ocean atmosphere climate model, show that the demise of the El Niņo was foreshadowed as early as October 2002 and forecast accuracy improved as the lead time decreased.

Seasonal predictions from climate models are dependent upon the representation of land-surface processes. A study has been carried out in BMRC to determine the sensitivity of seasonal predictions during the 1997-98 El Niņo event to the representation of soil moisture. This study relates to the degree of predictability of climate variations, and there has been a range of activities in BMRC aimed at identifying the potentially predictable component of interannual variations from the background weather noise. The research gives an indication of the extent to which models respond realistically to external forcings, and of the nature of slowly varying internal dynamical processes in models, compared with observations.

Both the atmospheric model and the coupled model have been used in BMRC to investigate the simulation and prediction of tropical intra-seasonal variations known as the Madden-Julian Oscillation (MJO). These studies are complemented by the development of a statistical technique to predict the MJO and its impacts. Indices involving rainfall statistics and the monsoon onset can be predicted with useful accuracy about 15 days ahead.

International fora for identifying and resolving uncertainties in climate models have been provided a series of model intercomparisons, under the auspices of WCRP. Scientists in BMRC and CAR have been variously involved in the design and coordination of several projects, such as the Atmospheric Model Intercomparison Project (AMIP), the Coupled Model Intercomparison Project (CMIP) and the Climate of the Twentieth Century (C20C). Several sub-projects of AMIP-2 have been carried out in features such as diurnal variations of tropical convection and the impact of land surface processes on climate predictability.

The CSIRO Mark 3 climate model has been developed by CAR, and it has been used in a range of studies, including assessments of potential changes to Australian climate under enhanced global greenhouse gas concentration. The model includes sophisticated representations of processes such as cloud and sea ice. Model simulations have been compared with observations of changes in sea-surface temperature in the eastern Pacific Ocean over the last thirty years to assess potential changes in the strength of El Niņo events under climate change conditions. The model has also been used to examine the impact of climate change on other climate features in our region, including the mid-latitude highpressure belt and the sea-surface temperature of the eastern Indian Ocean.

At Melbourne University, there is a continuing program of large-scale numerical modelling focused on Australian regional problems. The work includes studies of the variability of southern hemisphere climate, tropical cyclones, and Antarctic weather systems, as well as studies to assist in the interpretation of isotopes in ice cores and corals.

A PC-based software package, Diagnose, has been developed in BMRC to allow rapid real-time analysis of Australian climate anomalies. The database of rainfall and temperature data is updated each month, so that the current anomalies can be compared with gridded data over the last fifty years. The system was used to monitor rainfall and temperature during the drought of 2002-03. The relationship between the severity of the drought and temperature trends in recent decades has been considered in the context of regional climate change.

CAR and BMRC collaborated with the National Centre for Epidemiology and Population Health (NCEPH) to conduct a study on the potential impacts of climate change on human health in Australia (Figure 5.11). Research at Macquarie University has examined links between human health and climate variability, especially links between Ross River fever and the Quasi-Biennial Oscillation. There have also been studies on the impact of climate change on allergenic plants in Australia, and on the potential impact of climate change on medication side effects.

Figure 5.11. Pathways by which climate change affects human health.

The fourth Asia Pacific Network (APN) workshop on trends in climate extremes was held in December 2002 by BMRC in collaboration with CAR and the NCC. The participants continued their analysis of trends in extreme rainfall and temperature events across the region, and also commenced a joint analysis of meta-data associated with their daily climate data.