Countering a climate of scepticism

Since the Garnaut review released its findings, climate change deniers have mounted a rearguard action.

Their cry is that Ross Garnaut is no scientist and that he uncritically uses discredited science from the Intergovernmental Panel on Climate Change.

In fact, the Garnaut review relied on the Australian climate science community to make its scientific case. And while Garnaut stands by his economics, the science community stands by its science, particularly research following on from the IPCC's fourth assessment report, released in 2007.

This research was a key input to the Garnaut review, not least the finding that emissions, greenhouse gases, temperature and sea level rise are tracking the upper limits of projections assessed by the IPCC.

Furthermore, projected emissions over the next few decades as well as the risks of severe impacts are higher than previously thought.

Some arguments from the denial community qualify as fatal errors. I will detail two, one relating to statistics and the other to models.

The first argument is that the Earth has not warmed since 1998 and the second is that the models used to project future climate are fatally flawed. My view is that anyone with a higher degree in science who maintains that the Earth stopped warming in 1998 should hand their degree back.

The only way to maintain such a position violates the basic principles of statistics: significance, use of the appropriate test, causality and independence. Free speech it may be, but scientific speak it is not.

Three surface temperature records are widely used by climate scientists. The first two show that 2005 was the warmest year on record, while the third shows that it was not quite as hot as 1998.

According to all three the past decade was the warmest on record, while 12 of the 13 warmest years on record have occurred since 1995. If we take the 31 10-year periods since 1977 - that is, 1977-87, 1978-88, and so on - the 1998-2007 trend that best fits the data is not even the lowest trend in that sequence.

Ten-year trends range from virtually no change to +0.35C a decade, averaging about 0.2C a decade over the whole sequence.

The choice of 1998 as the origin of a recent trend is cherry-picking, especially as the ElNino event that year transferred a record amount of heat from the ocean to the atmosphere.

According to NASA data, 2007, during a La Nina (normally colder than average), was similar in temperature to 1998.

When we know that two events are similar, if one associated with temporary cooling occurs a decade after another associated with temporary warming, there must be an underlying upward trend. This is science.

Second, arguments that climate models are fatally flawed means jettisoning the science that underpins the models, which in turn means having to jettison our understanding of environmental and earth science.

The development of climate models, earth system science and global observing systems has been inextricably linked since the 1950s.

This understanding underpins our standard of living and, ultimately, human interactions with the environment. It is the same science that underpins the future of the global economy.

Scientific models are used to review what we are measuring and what it tells us. For example, taking Earth's temperature is very difficult, so meteorologists measure air temperature 1.5m above the ground, using thermometers in shade.

Standards are maintained but these are not perfect. So, to create a homogenous record and extend our record to non-standard measurements taken in the past, we must make adjustments. The World Meteorological Organisation has technical standards for this task.

The second step is to link that record of air temperature with other environmental observations. Scientists develop and apply models of plant growth, of physiological response among animals, and of the physical performance of materials.

These models are validated by measurements in the lab and in the field.

All environmental observations need to be adjusted in some way, and we use scientific models to do so. Such understanding underpins weather forecasting, agricultural production systems, the management of natural hazards and so on.

These relationships are also critical to our understanding of future climate risks.

Claims that that this task is not carried out with the utmost care and with the appropriate scientific scepticism cannot be supported.

In this context, we need to understand what a climate model is: a conceptual construct of how a system works, in this case Earth. Many people, when they think of models, straight away turn to mathematical models. However, because everybody views the world as a set of concepts, we use models all the time.

A scientific model differs from an everyday conceptual model by being a systematic framework of observation, hypothesis, experiment and review of hypotheses.

The last two steps are critical in qualifying a model as scientific. Climate models are continually subjected to testing and revision in the light of new evidence. Most significantly, when statistical principles are violated to discount a scientific model, a red flag should go up: what is being violated is common sense.

Those who wish to discount this scientific effort are driven by their own ideologically dominated models, which are unreviewed, unaccountable and unverifiable. They should be disregarded.