Fragmented Climate

Fragmented Climate: In the previous two blogs (linked below) I have tried to give some insight into some of the issues that climate scientists are thinking about as we move into the future. I started from the discussions initiated at the World Modeling Summit for Climate Prediction at the European Center for Medium-range Weather Forecasts. (See also article in Nature). I provided some analysis with the discussion focused on computational resources. In that analysis I tried to show some of the tensions within the climate community and how the community operates, good or bad, efficiently or inefficiently.

Science communities tend to be fragmented for many reasons. Some argue that this arises because of the very nature of scientific discovery and the personalities of scientists. Scientists are trained to question; the very process of science requires challenging results; discovery is unpredictable and today’s discovery will be altered, perhaps rejected, by new observations and new studies. In addition to this cultural predisposition towards fragmentation, the way science is funded in the U.S. supports, if not encourages, fragmentation. There are multiple agencies that “own” the problem, or different aspects of the “problem.” Program managers within the agencies act with various degrees of autonomy. Virtually all decisions are, ultimately, based on some flavor of competition. When this is combined with the predisposition of the culture of science to reduce, to make smaller, to fragment rather than to consolidate; if there is to be a balanced program it arises out of a messy, sometimes contentious, time-consuming process.

Balance: One of the points of previous blogs was that there is a need for a balanced approach to the expenditures on climate research. This balance considers, for instance, a small number of big, unreliable computers versus a larger number of more reliable, medium size computers versus data systems versus sustained observations versus new observations versus applied research versus basic research versus education versus all of the other elements that are part of the undertaking of science.

Over the past two decades climate science became “big science.” This big science followed from increasing societal concern about the global environment. The big science was and is most defined by the cost to take observations; there is a concentration of money in space observations to support climate research, weather research and forecasting, and more general research of the Earth system. The big science is also defined by the need for big computers, consistent with the computational resources needed for several fields of research – either applied or basic research. (The biggest of all computer problems has traditionally been Stockpile Stewardship, and computational requirements for climate science are of a similar scale.) There are, consistent, with these resources a hoard of scientists and technicians and software specialists. Climate research is big science, and compared with many worthy research efforts, it is well funded (More than 15,000 scientists come to the December American Geophysical Union meeting, only one of many relevant meetings.). While well funded, there is also a mountain of work that needs to be done, that can be done, and for which the funds are not adequate. It is also true that if the funds were there the human and technical resources do not pre-exist; that is, there is not a large store of people and machines and observations on the margin that could immediately exploit additional funds.

This suggests that for climate science to move to a place to best serve society it would benefit from more formalized management, with a plan for balanced, integrated expenditures. This would suggest the need to manage the current resources more effectively. There is a need to distinguish the mission that supports the essential observations and modeling that might stand as a basis for climate services in contrast to those equally essential fundamental research topics to answer the questions that arise from the thousands of studies and papers that are published every year. There is a need for workforce development.

This need for more formalized, integrative management stands in contrast to the fragmented state of field – fragmentation that is both natural and perhaps in the best interest of basic research. The question arises whether or not the fragmentation is greater than optimal, and if so, are there ways to effectively optimize the expenditures by the reduction of fragmentation. I pose that the answer is yes, and I also know that 20 years of trying to move towards this optimization is characterized by slow, measurable, difficult progress. I know that most efforts to “bureaucratically centralize,” to reorganize, are not obviously successful; I assert that they are more negative than positive.

Hence we are faced with the need for organization in a fragmented environment that is fragmented for both good and bad reasons. The challenge is how do we sustain the healthy fragmentation, that is the diversity that supports innovation, creativity, and the scientific process, and how do we reduce the fragmentation that exists, for instance, because of bureaucratic ownership and competition for limited resources? What are the forms of management that exist between bureaucratic centralization and anarchy? The next blog in this series will explore some of the self organization efforts of the field.

r


Importance of Justification

Buying Big Computers

High End Climate Science: Development of Modeling and Related Computing Capabilities