by Judith Curry
I’ve been meaning to write a post on the recent “Open Science Conference” organized by the World Climate Research Programme (WCRP). A post at RealClimate entitled “Conference Conversations” provides a starting point for my post, with this concluding sentence:
The contrast between the conversations in this meeting and what passes for serious issues in the media and blogosphere was very clear.
First, some background on the WCRP. Snippets from their website:
The WCRP was established over 30 years ago (in 1980), under the joint sponsorship of the International Council for Science (ICSU) and the World Meteorological Organization (WMO) , and, since 1993, has also been sponsored by theIntergovernmental Oceanographic Commission (IOC) of UNESCO.
The main objectives, set for the WCRP at its inception and still valid today, are to determine the predictability of climate and to determine the effect of human activities on climate.
In 2005, after 25 years of serving science and society, the WCRP, in collaboration with the broader scientific community, developed and launched the WCRP Strategy Framework 2005-2015.
WCRP organizes meetings, workshops and conferences to coordinate and facilitate climate research. The research itself is done by individual scientists working in national and regional institutes, laboratories and universities. WCRP committees, working groups and projects, assisted by the Joint Planning Staff (JPS), are the main vehicles for setting the research agenda and mobilizing the broader research community on specific activities.
During the period 1994-2003, I was quite active in the WCRP:
- Global Energy and Water Experiment (GEWEX) Radiation Panel (1994-2003 )
- GEWEX Cloud System Studies (GCSS) Science Steering Group (1998-2003 )
- Chair, GCSS Working Group on Polar Clouds (1998-2004 )
- Chair, GEWEX Radiation Panel SEAFLUX Project (1999-2004)
- Science Steering Group, Arctic Climate System (ACSYS) Programme (1994-2000)
The alphabet soup associated with the WCRP acronymology is described by this NYTimes blog post.
Open Science Conference
Goals and vision: A better understanding of the behaviour of the climate system and its interactions with other Earth system components is critical to predict its future evolution, reduce vulnerability to high impact weather and climate events, and sustain life.
This need is perhaps greater than ever before given that humans have emerged as the dominant agent of future change. Progress will require, moreover, an increasingly holistic approach across scientific disciplines, as well as an unprecedented commitment to the development of a diverse and talented future workforce. (WCRP)
It is timed to provide strategic input into the IPCC Fifth Assessment Report. (WCRP)
The OSC will appraise the current state of climate science, thereby making a measurable contribution on the Fifth Assessment Report (AR5) of the Intergovernmental Panel on Climate Change (IPCC). It will identify key opportunities and challenges in observations, modeling, analysis and process research required to understand and predict responses of the Earth as a system. (WCRP)
More the 1900 participants from 86 countries, are attending the WCRP Open Science Conference. (WCRP)
the speakers included the biggest names in climate research and many past and present IPCC authors. (RC)
The focus of the conference was on how climate research should be done in order to be of service to society. Hence, a fair bit of focus was given to how to create useful climate information, or ‘actionable science’. This is supposed to be part of a global framework for climate services (abbreviated as GFCS to make it a bit more cryptic). But there is a lot of discussion about exactly what form of information this would be, and there were many (not necessarily exclusive) ideas around. (RC)
Bruce Hewitson’s presentation
With regards to climate services, Bruce Hewitson gave a very good presentation entitled “Meeting user needs: climate service limits, ideals, & realities.” Some excerpts:
Expressions heard/seen during this week:
- “end-to-end user-needs driven”
- “actionable science”
- “co-production of services”
- “most critical need is not more science, but translation”
- “the need to provide reliable assessments of future climate changes has never been so high”
- “new confidence in regional temperature projections despite model deficiencies”
The resultant “community confusion of information”: A proliferation of portals and data sets, developed with mixed motivations, with poorly articulated uncertainties and weakly explained assumptions and dependencies, the data implied as information, displayed through confusing materials, hard to find or access, written in opaque language, and communicated by interface organizations only semi-aware of the nuances, to a user community poorly equipped to understand the information limitations. (JC comment: gets my vote for quote of the week).
- What are the consequences of knowledge gaps?
- How best does one inform decision-making under conditions of incomplete information?
- Climate services bridge communities, language, and value systems
- Scientific products are miss-aligned with most user’s decision risk framework, in which climate is only one factor
- Uncertainty language casts doubt; likelihood messages inform
- Information on the exceedance in time, space, and frequency of user-defined thresholds is powerful
- The issues of responsibility, accountability, credibility and values is largely missing from the climate services dialogue.
Point #5 is really key. The issue (#4) of user-defined thresholds is also very important, and addressing issues related to such thresholds (with historical and paleo data and models to create scenarios whereby critical thresholds might be exceeded) can actually be more straightforward than most of what is currently being provided by climate scientists.
The changing conversation
The conversation that we have been having for the last two decades is about mitigation: stabilization of atmospheric CO2 in the context of global treaties on energy policy.
The new conversation being generated by climate science organizations seems more focused on adaptation and climate services, which is associated with the seasonal to decadal time scale.
The mitigation focus is on global climate and the century time scale, whereas the adaptation focus is regional and on timescales from the seasonal to decadal.
The mitigation focus and the century time scale has sapped the community of much resources in terms of manpower and computer time. I have talked with people in leading positions at several modeling groups, and the effort that is put into the simulations for the IPCC saps 50-70% of the total manpower time and resources of the modeling center.
What if we had devoted all of those resources to making better probabilistic predictions on timescales of 2 weeks to 3-4 months? Farmers would be able to make better choices about what crops to plant. Water resource managers could make better choices. Energy generation and demand could be made more efficient. Etc. Most of the developing world doesn’t have weather forecasts beyond two days, and often these forecasts do not anticipate extreme weather events (think Pakistan floods, Severe Cyclone Nargis). Anticipating extreme weather events by a week or two, or even a few days, could make an enormous difference in the developing world.
In terms of longer timescales (decadal to century), once the focus becomes regional rather than global, historical and paleo data becomes more useful than global climate model simulations (no matter what type of “right-scaling” methods are attempted). Searching for past regional extreme events through the historical and paleo records should be the focus, rather than working to air brush the past global variability. IMO the most interesting thing from the BEST data is the large trend during the 19th century. Can we truly infer meaningful global averages from these data? I don’t know, but it provides a heck of alot of interesting information for the U.S. and Europe.
On the weather time scale and maybe out the seasonal timescale, creating pdfs from ensemble forecasts and making decisions based upon expected utility is justified. When using climate information to support decision making on decadal to century timescales, we are in a situation of “deep uncertainty;” see the previous post on Can we make good decisions under ignorance? Attempting to use climate information in the context of expected utility can lead to bad decisions; there are much better ways to approach the decision making under conditions of deep uncertainty. The expected utility approach to decision making has led climate scientists to produce pdfs that are unjustified and misleading. Looking at other decision making frameworks that are more suitable under conditions of deep uncertainty motivates a different type of analysis and emphasizes assessment of uncertainty and areas of ignorance.
IMO, the emphasis on “translation” is misguided. Clarifying the areas of ignorance and knowledge gaps and uncertainty in predictions is the absolutely first step before “translating” anything. But actually understanding the concerns of individual decision makers and broader decision making context is key to doing anything useful. IMO, academic researchers funded by NOAA or whoever is not the best model for matching useful forecast information to user needs.
An example of success in this regard is the weather risk management industry, of which the energy sector is the largest customer. A plethora of weather risk companies provide forecasts on timescales of hours to months, focusing on energy demand, wind energy generation, etc. The most competitive of these companies are identifying niches that target specific needs with innovative forecast products. Those companies that make consistently poor forecasts don’t retain their customers.
I can easily envision a broadening of the weather risk management industry to include the shorter term climate risks, but a barrier is lack of a track record for the seasonal forecast products. NOAA NCEP has a new version of its seasonal forecast product CFSv2, which is substantially improved over its predecessor. ECMWF has its new version System 4, which will release its first seasonal forecast Nov 8. So the tools are getting better, and a serious focus on making better forecasts on timescales of weeks to 3-4 months could make an enormous difference to the economy, security, and humanitarian concerns.
And finally, Hewitson’s statement: “The issues of responsibility, accountability, credibility and values is largely missing from the climate services dialogue.” At the RC post, rasmus and gavin state:
However, communicating that understanding to people who might benefit from it remains a work in progress. Communication involves an end-to-end two-way process, as opposed to simply sending off a message hoping that the recipient will understand. There are also ethical concerns linked to the context – what are the consequences of an incorrect forecast? Are there inequities in who benefits and who doesn’t? Scientists, on their own, are not necessarily well-equipped to deal with this.
Good to see “ethical concerns” mentioned in the RC post, but the RC statement does not capture Hewitson’s concern. Hewitson raises the issue of responsibility, accountability, and credibility. It was exactly these issues that Climategate called into question in the context of behaviour reflected by the emails. Scientists, and particularly the institutions that support science, should have as its top priority dealing with responsibility, accountability, and credibility. Without dealing with these issues, the concept of climate services is doomed.
In conclusion, I think the institutions that support climate science haven’t been having the right conversation over the past two decades. In contrast to the implications of the closing sentence of the RealClimate post, the blogosphere, with its diversity of venues and perspectives, is fostering a much broader conversation that has the potential to send climate science and its applications on a more useful track.
Note to gavin and rasmus: Re your statement: The contrast between the conversations in this meeting and what passes for serious issues in themedia and blogosphere was very clear. The point is that people and decision makers care more about climate on the time scale of a season and ~10 years than they do about the century time scale, which is reflected by many of the articles in the media and the blogosphere. A different conversation.