Australia’s premier research body, CSIRO, is spearheading a range of R&D projects aimed at countering risk.
In April 2005, CSIRO Mathematical and Information Sciences announced the formation of the Quantitative Risk Management group, encompassing a multidisciplinary team of mathematicians, statisticians and software engineers. Their task is to undertake dedicated risk research across areas as eclectic as finance, air traffic control and infrastructure.
Doug Shaw, acting head of the Quantitative Risk Management team, says risk research can contribute to avoiding what is avoidable and reducing the impact of what is unavoidable.
One of the first initiatives to come from the team is a project in which researchers are working with several of the major Australian banks on the development of new complex risk management software and systems, to help financial institutions better control operational, credit and market risks.
Some of the work is focussed on ensuring that the banks are able to meet the capital standards embodied under the global Basel II Accord, which is being overseen in Australia by the Australian Prudential Regulation Authority.
But researchers have also achieved significant results in the market risk area with the CSIRO team’s development of complex mathematical software for pricing exotic currency options under the brand Reditus. This has recently been integrated into the FENICS FX system that banks around the world use for pricing and analysing foreign exchange options.
“The Reditus system is proving successful and we’re working with a number of Australian banks on different risk areas,” says CSIRO’s financial risk research team member John Donnelly. “In operational risk we’re developing a customised software system which will help model all the risks that are involved in the day-to-day running of the business. It’s important for the banks because they need to comply with the risk capital guidelines detailed under Basel II.”
Meanwhile, Dr Shaw says risk research in the infrastructure area is also gaining momentum and CSIRO is working with a number of companies, including some in the mining sector, on a confidential basis.
“It’s grown out of work that’s been done over a few years into pipeline failures and developing models for predicting failures,” he says. “We’re looking at developing that sort of methodology for other assets such as those in energy industries.”
Part of the focus is also on security risk, so utilities can monitor their systems and determine statistically if there has been a sudden change that may indicate deliberate or inadvertent interference or contamination.
CSIRO is also applying its risk modelling techniques to air traffic management, using mathematical science and engineering approaches it has developed for the Australian military and converting these to civilian aviation applications.
David Peterson, business development manager for CSIRO Mathematical and Information Sciences, says the focus on airspace management risk has become increasingly important as air traffic volumes grow across Australia and airports become more congested.
“We’re looking at conflict situations that can occur between aircraft across an open and wide country,” Mr Peterson says. “Our research is focussing on the engineering development of airspace to allow different types of aircraft to operate within it efficiently. It’s putting science and engineering together to develop the airspace of the future.”
Keeping a Watch on Waterways
Risk management, no matter what the arena – financial, social or environmental – is about measuring the threat, and this is precisely what a series of new statistical models do to more effectively manage fragile waterway ecosystems.
CSIRO scientists have developed a series of complex models that help to identify and then monitor environmental threats so that timely intervention can either avert, or at least manage, long-term problems.
For example, pesticide users and regulators now have access to decision-support software that estimates the potential impact of any pesticide use on groundwater or surface-water quality before any chemicals are deployed.
The Pesticide Impact Rating Index (PIRI) takes into account a pesticide’s toxicity, chemical properties, frequency and rate of application, and then factors in local site conditions along with seasonal and soil variables.
“By assessing and ranking pesticides and their contamination risk to surrounding surface and groundwater supplies, the results allow users to select pesticides with the least detrimental impact to the local environment,” explains CSIRO scientist Dr Ray Correll, who developed the PIRI program.
CSIRO is also working with governments on a range of freshwater risk management projects in Queensland, Western Australia and South Australia, focussing on the monitoring of waterway conditions so that problems can be identified and addressed.
“We integrate quite a few different measures so you can look at the condition of the waterways, and see where the risks are in certain areas based on the signs that are evident,” says CSIRO environmental informatics leader, Dr Bronwyn Harch. “Changes in water chemistry and in the biology of the area will show where pressures are occurring and what is causing them, which will enable more effective risk management.”
Medical science’s knowledge about what is going on in the human system is advancing every day. With these advances comes the challenge of managing the vast amounts of data generated by the latest genetic analysis technologies.
For example, microarrays can determine the activity of tens of thousands of genes in a single experiment.
This is where the mathematical sciences have become an essential medical research partner and are helping to build a clearer picture of the relationship between specific genes and diseases, including cancer.
“As statisticians we have developed techniques to look at analysing extremely large data collections coming out of microarrays,” says CSIRO bioinformatics scientist Dr Glenn Stone. “With one single experiment, you can now figure out which of a person’s 44,000 genes are active in a particular tissue sample, and draw from this a picture of everything that’s going on.
“This information can then be used, for example, to pick up differences between cancerous tissue and non-cancerous tissue, or for analysing blood cells and variations that might indicate the presence of a disease.”
This research has enormous implications for medical research that is seeking to pinpoint the exact causes of particular diseases.
“With microarrays, scientists can also look at the presence or absence of certain mutations in genes in individuals in huge numbers,” Dr Stone says. “We have been working on statistical algorithms to enable the analysis of this data.”
Dr Stone says research using statistical analysis has recently identified that a certain class of lymphomas in humans is not one disease but a range of diseases. Previously they were all treated as if they were one disease.
“By differentiating the diseases using molecular markers they have the potential to find different treatments that are more effective for different people, depending on exactly which form of the disease they have,” he says. “This is one form of personalised medicine where treatment can be more effective because it is tailored to a specific sub-type of a disease.”