CLIMATE Commissioner Will Steffen has backed the accuracy of sea-level rise modelling and rejected criticism that climate scientists spend too much time in front of computer screens and not enough in the field.
Professor Steffen, executive director of the Climate Change Institute at the Australian National University and a member of the Prime Minister's Independent Climate Commission, said not all climate change modelling was reliable.
"I get annoyed when people say the models are great or the models are poor," he said. "The models are improving all the time. They do some things really well. We have a ways to go on other things. That is the nature of science."
However, modelling of sea-level rises -- which has been used by climate scientists for 20 years -- was both reliable and essential, he said.
"My rule of thumb is anything to do with average temperature changes over long periods of time and large areas, the climate models are really good," Professor Steffen said.
"We have a lot of faith in their projection into the future."
Charles Finkl, the editor of the peer-reviewed Journal of Coastal Research and an internationally recognised coastal scientist, has urged scientists to spend less time building predictive models and more time in the field observing and interpreting "hard or real data".
His comments were prompted by a heated debate over the findings of a study by NSW coastal researcher Phil Watson published in the journal showing sea levels around Australia rose in the second half of last century at a slower rate than necessary over the next 90 years to reach the CSIRO's high-end projections.
John Church, CSIRO fellow with the Hobart-based Centre for Australian Weather and Climate Research, described himself as "observationalist" but said models still played an important role in his work.
"I spend all my time analysing observations," said Dr Church, whose research into sea-level rises has been cited by both the Intergovernmental Panel on Climate Change and Professor Steffen.
"But we should not throw away models. Just observations alone do not provide us the understanding that is necessary if we are to understand what the implications are and if we are to have some sort of view of what the future holds. It is true that scientists can either focus too much on observations or too much on models. It is a matter of getting the balance between those."