The men behind rise of CSIRO’s climate modelling program

Cosmos Magazine

Cosmos

Cosmos is a quarterly science magazine. We aim to inspire curiosity in ‘The Science of Everything’ and make the world of science accessible to everyone.

By Cosmos

Meteorologist and historian Dr Ian Smith, writing in “Historical Records of Australian Science,” pays tribute to two men who developed the first long term climate model in Australia which indicated the impacts of enhanced emissions in the atmosphere. Here, for Cosmos, Smith summarises the process of model development and explains the global impact of this research.

It is 53 years since the CSIRO Division of Atmospheric Research decided to prioritise an atmospheric monitoring programme designed to extend knowledge of the chemical composition of the background atmosphere and to conduct research into if and why it might be changing.

Back then in 1972, a joint venture between CSIRO and the Bureau of Meteorology (BoM) known as the Australian Numerical Meteorological Research Centre (ANMRC) was focussed on developing a global model for use in weather forecasts.

Old photo of man with glasses on blue background
Barrie Grant. Credit: CSIRO.

However, a member of ANMRC at that time, Barrie Hunt, was also interested in applying the model to studying climate change issues involving time scales of months to centuries, rather than just days. In an interview with a newspaper in 1976, Hunt referred to a particular experiment with the model: “We have looked at volcanic eruptions with an eye to finding out more about ice ages and their causes. The computer–created volcano threw muck into the atmosphere and we charted its progress. By model day 150 the debris had reached the poles. The temperature in the tropics dropped by 0.5–1.0°C.”

This quote is of interest for several reasons. It firstly describes one of the first climate perturbation experiments ever conducted in Australia. Secondly, it foreshadows the fact that any such climate experiments would be time-consuming, extending over many (model) months and years. Thirdly, the description, ‘… threw muck into the atmosphere’, is not precise scientific language, but more than adequate for communicating concepts to the readers of a daily newspaper.

Four years later, in 1980, with increasing concern over the impact of atmospheric carbon dioxide on the climate, Hunt expanded on his vision for climate modelling using more careful language in a paper published in the Australian Journal of Physics:

“Currently, interest in climate and the prediction of climate is at an all-time high. This is partly due to the fact that with modern computers and numerical integration methods hitherto impractical techniques have now become viable …. There is a widening appreciation that mankind now has the capacity to alter inadvertently the future climate via a number of pollutants, and that this could result in undesirable climatic changes.

One year later in 1981, Hal Gordon, a colleague of Hunt’s at ANMRC, published a paper describing one of the numerical integration methods suitable for long-term climate simulations.

Old photo of bearded man in front of computer
Hal Gordon. Credit: CSIRO.

The ANMRC was closed in 1984, but independent climate modelling activities continued at the BoM,15 the University of Melbourne16 and CSIRO, where Hunt and Gordon pursued the development of a model appropriate to meet an increasing demand for climate change information.

It is unlikely that a casual observer would appreciate the task involved in writing and updating modelling software, installing it on ever faster computers, managing the storage of output data and, most importantly, ensuring the results are physically consistent. However, Hunt and Gordon were up to the task and theCSIRO model was gradually developed to the extent that by 1992, a 4-level version (often referred to CSIRO Mk1)17 was developed that could reliably be used in applied climate change studies. In particular, it provided a realistic simulation of present-day climate in the Australian region.

Old photo of man with glasses in office
Barrie Grant. Credit: CSIRO.

The relative success of the CSIRO model, and the increasing demand for information about future climate change at both national and international levels, led to the formation of the CSIRO climatic modelling program.

Hunt developed a strategic plan and, in a memo to staff on 26 April 1993, explicitly stated the objective: ‘To maintain a world class modelling facility which solves critical climatic problems in a timely fashion and enhances understanding of the climatic system.’

Two papers (Smith, 2007 and Smith, 2024) documented the legacy of this successful (possibly underrated) scientific program and the two people who were largely responsible. The former paper focussed on the achievements between 1976 and 2007 while the latter deals focussed on events thereafter.

How to build a model

Old photo of bearded man in front of computer server
Hal Gordon. Credit: CSIRO.

Imagine, many years ago, a couple of automotive mechanics coming into possession of a basic family vehicle designed for domestic use on the road and wondering if it could be modified to perform as a racing car. Now imagine they tinkered with the basic design and modified the engine, the suspension, the brakes, the steering etc. and developed a relatively well performed racing vehicle, capable of competing at an international level.

Furthermore, over time and following several successes, the two mechanics  assembled a team that continued to upgrade the vehicle in so that it continued to compete successfully for many years with international rivals at the highest levels.

This scenario, as an analogy, certainly helps to describe what happened nearly half a century ago when Hunt and Gordon modified a

using clever coding skills and access to ever more powerful computing facilities  developed a state-of-the-art global climate model capable of addressing several important questions, not the least of which was what would happen if atmospheric CO2 concentrations continued to increase.

Impacts of the climate model

Along with other international models, the CSIRO model regularly participated in several coordinated projects and informed the deliberations of the Intergovernmental Panel on Climate Change (IPCC) 5th Assessment Report pubished in 2014.

This in turn, led to what is known as the Paris agreement whereby nations committed to limiting global warming to well below +2.0 °C – a value above which global warming impacts became quite serious.

Australia’s commitment back then was to reduce greenhouse gas emissions by 26 to 28% below 2005 levels by 2030 and, later on, to 43% below 2005 levels by 2050.

The point here is that Australia was one of just twelve countries that contributed to this compilation. It is arguable that very few scientific endeavours can lay claim to having such a large societal impact. All the contributors, including the CSIRO modelling team deserve special recognition for this achievement but there is very little evidence for this.

It is doubtful that either Barrie Hunt or Hal Gordon could have foreseen such accomplishments when they first began tinkering with a relatively simple weather forecast model back in 1976. It is worth mentioning that references to the latest version of the model in the literature since 2002 is now about 10,000, with approximately 300 citations in 2024 alone. This is remarkable given that work on this model ceased more than a decade ago

Under-recognition of these two scientists is not unique and probably reflects the fact that the work of model development by research groups everywhere is intense and ongoing, with very few individuals having the opportunity to reflect on these issues (let alone write research papers) before they eventually retire. In addition, there is no agreed method for measuring the impact of this type of work which may not be apparent for many years. Hopefully, this article provides some redress.

Please login to favourite this article.