Spinning uncertainty? The IPCC extreme weather report and the media

When it comes to weather, scientists and the media have different understandings of risk. Ameel Khan

The “reasonable person” would agree that disaster risk is best avoided. Under a changing climate, how exposed people are to risk and how socially and physically vulnerable they are affects how often disasters may happen. The more we know about risk, the better we can avoid it.

The recently released IPCC Special Report on Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation Summary for Policymakers (IPCC SREX SPM) assesses adaptation measures to reduce disasters and resilience – the ability to bounce back after an event. In the main, reporting of the SREX has concentrated on what’s known about climate extremes and their relationship to disaster, and the management of disaster risk under a changing climate.

But in the articles they have published on the SREX SPM, The Australian newspaper has concentrated instead on uncertainty.

Graham Lloyd, The Australian’s environment reporter, said:

Widely-held assumptions that climate change is responsible for an upsurge in extreme drought, flood and storm events are not supported by a landmark review of the science.

And a clear climate change signal would not be evident for decades because of natural weather variability.

Despite the uncertainties, politicians - including US President Barack Obama in his address to federal parliament yesterday - continue to link major weather events directly to climate change. Greens leader Bob Brown yesterday highlighted Mr Obama’s climate change comments and said the extreme weather impacts were “not just coming, they are happening”.

Some areas have experienced more and longer droughts; others have had fewer. suburbanbloke

But what did Obama actually say? He said:

This includes the clean energy that creates green jobs and combats climate change, which cannot be denied.

We see it in the stronger fires, the devastating floods and the Pacific islands confronting rising seas.

And as countries with large carbon footprints, the United States and Australia have a special responsibility to lead.

President Obama referenced fires, floods and rising seas. Senator Brown endorsed those comments. Lloyd focuses on extreme drought, floods and storm events as being uncertain and unattributable to human agency.

This is what the SREX actually said:

There is evidence from observations gathered since 1950 of change in some extremes. Confidence in observed changes in extremes depends on the quality and quantity of data and the availability of studies analyzing these data, which vary across regions and for different extremes.

Assigning “low confidence” in observed changes of a specific extreme on regional or global scales neither implies nor excludes the possibility of changes in this extreme [my emphasis]. Extreme events are rare which means there are few data available to make assessments regarding changes in their frequency or intensity. The more rare the event the more difficult it is to identify long-term changes.

Global-scale trends in a specific extreme may be either more reliable (eg, for temperature extremes) or less reliable (eg, for droughts) than some regional-scale trends, depending on the geographical uniformity of the trends in the specific extreme.

The absence of evidence is not the evidence of absence.

Uncertainty management in the SREX describes likelihood and confidence.

Likelihood is the probability that a statement will be true or that a predicted set of events may happen.

Confidence concerns evidence (limited, medium or robust) and agreement (low, medium or high). Two sets of confidence statements are delivered. One set explicitly reflects evidence–agreement statements. The other is on a scale of very low, low, medium, high, very high. These are subjective and variable.

In summary, the Special Report says this about observed changes:

It is very likely (90-100%) warm days and nights are increasing. Cool days and nights are decreasing globally with high confidence in Europe, North America and Australia. In other regions confidence is lower due to data limitations.

Changes in tropical cyclones have low confidence. NASA/GSFC/Jeff Schmaltz/MODIS Land Rapid Response Team

Statistically significant trends in heavy precipitation events in some regions are likely (66-100%) and there are more regions where such trends have occurred than not.

Some areas of the world have experienced more intense and longer droughts (medium confidence), but other areas have experienced fewer. There is limited-to-medium evidence on flooding because of limited data and land-use influences on flooding behaviour.

Changes in tropical cyclones are given low confidence but poleward shifts in mid-latitude storm tracks have been observed with medium confidence.

Attribution of those changes to a human influence via greenhouse-induced climate change is less certain than direct observation. Anthropogenic influences on temperature effects and coastal extremes are likely (66-100%). There is medium confidence of a human influence on extreme precipitation at the global scale. Tropical cyclone attribution is given low confidence and changes at the global scale are unattributed.

This uncertainty statement in the SREX was highlighted by The Australian:

Projected changes in climate extremes under different emissions scenarios generally do not strongly diverge in the coming two to three decades, but these signals are relatively small compared to natural climate variability over this time frame. Even the sign of projected changes in some climate extremes over this time frame is uncertain.

This is what the SREX also said about future changes:

  • Virtually certain: heat extremes increase and cool extremes decrease.
  • Very likely: heat waves increase over most land areas, extreme high water levels occur with sea level rise.
  • Likely: increases in precipitation over some land areas (medium confidence that they may also accompany a mean precipitation decrease in some areas), tropical cyclone mean wind speed will increase but numbers remain unchanged or decrease.
  • High confidence: coastal inundation, instability in mountain regions including landslides.
  • Medium confidence: extra tropical cyclones will decrease, droughts will intensify in some regions, local flooding in some regions.
  • Low confidence: tornados, hail, droughts in some regions where rainfall signal is uncertain, flooding in specific regions, the direction of climate variability.
    Disaster risks have two levels of attribution. The first level separates the physical hazard from the exposure of people and places, and their underlying social and physical vulnerability. The second concerns the anthropogenic contribution to changing extremes that influence those physical hazards.

These are difficult to disentangle. Local interactions between society and environment make conclusions at the regional and global scale even more difficult.

Sometimes limited confidence comes from a lack of data, at other times from uncertainty in drivers of change and yet other times from a lack of scientific knowledge.

We’re virtually certain that heat extremes will increase. muffytyrone

The Special Report concludes that most of the observed increase in economic damage is because people and infrastructure are living in harm’s way, rather than because of changing climate extremes. It does not, however, discount changes in those extremes.

The cycle of scientific research, publication and assessment means that events occurring over the past two years will not be included in the report. This evidence is rapidly mounting. These results are from statistical tests I’ve run over the past year or so:

  • Annual rainfall in eastern Australia has increased by 14% with respect to the El Niño-Southern Oscillation Index from 1973–2010 compared to 1900–1972.
  • Increases in days per year over 35°C and 40°C in the period 1997–2011 compared to 1973–1996 at Laverton, Victoria (non-urban high quality station close to Melbourne) are significant at the 5% level.
  • Days of high fire danger index or above (also at Laverton, which has good wind records) have increased by 1996/7–2009/10 compared to 1972/73–1995/96 by 38% from 33 to 45. This change is significant to the 1% level. Recent catastrophic fires have led to the addition of a catastrophic category to the fire danger ratings.
    The large floods in northern Victoria summer 2011 were caused by natural climate variability, but the contributing spring and summer rainfall was more than 200 mm wetter than predicted by a simple statistical model. Because such events occur about every decade or so, it would take decades before a robust signal could emerge within the statistics.

The alternative is to run a set of model experiments, a time-consuming and difficult task. The rainfall originated from warmer tropical waters off north-western Australia, which in summer 2011 was at record levels. Are we willing to bet that a warmer ocean and atmosphere is not contributing to these conditions? It’s possible of course, but theory would support the opposite conclusion.

The findings of the SREX SPM are complex. Jean Palutikov of the National Climate Change Adaptation Research Facility said that they were no surprise to anyone who is familiar with climate science. And they are not. Both Obama’s and Brown’s comments can be justified by the SREX SPM, but Lloyd’s spin cannot.

On climate change, The Australian is behaving like the media equivalent of a fog machine. Its unreliable reporting should be avoided by those with an interest in factual scientific information.