160
S. George Philander
of imaginative ideas to get a program moving. Usually a brilliant person would step
in to provide a spark, a Jule Charney, or Henry Stommel, or Claes Rooth. On this
occasion in early 1990, the person who stepped in was a “suit” from Washington
who came to brief us on the type of program that politicians are likely to favor. How
did nonscientists acquire the power to influence the activities of large numbers of
scientists?
Informally organized activities, with scientists themselves doing the planning
and implementation, are feasible when a science is “small.” However, the enormous
growth of physical oceanography over the past few decades, in the number of scientists, the number of papers they write, and the resources they need, transformed
the field from a “small” into a “big” science. Such growth calls for formally structured activities that respond to demands from the sponsors of research for results
that are of practical value. These can be salutary developments, provided we take
care to distinguish between accomplishment and frantic activity—science is a highly
undemocratic activity with a few people producing most of the important results
1 —
and provided science for the sake of understanding natural phenomena continues to
receive adequate support.
Growth sometimes leads to a loss of cohesion. That this was happening became evident in the special issue of the Journal of Geophysical Research (Vol. 103,
June 29, 1998) devoted entirely to articles that review “Progress of El Ni˜ no Research
and Prediction.” Each article, written jointly by several distinguished scientists, is
excellent, but is disappointingly disconnected from the other articles. The one that
reviews oceanographic measurements makes essentially no reference to ideas presented in the paper on idealized models of El Ni˜ no. The articles on realistic models,
and on prediction, are similarly detached, from each other and from the other articles
in the special issue. When I mentioned this problem to colleagues, some denied that
there is a problem, but others, David Neelin for example, agreed that different groups
have very different perspectives on the same phenomena. In the summer of 2002 I
called a meeting in Trieste, Italy, to address a seemingly straightforward question:
given that El Ni˜ no corresponds to a natural mode of the coupled ocean–atmosphere,
is that mode (a) strongly damped so that a continual oscillation is dependent on the
presence of random disturbances? Or is it (b) self-sustaining and highly predictable?
Or is it (c) sufficiently unstable to generate secondary instabilities so that, as in the
case of weather forecasting, predictability is limited by errors in initial conditions?
Each answer had ardent supporters!
After much discussion, a few people agreed that the available data sets are
inadequate to distinguish between the three possibilities. The paucity of the data is
evident in the way we have been surprised by each El Ni˜ no that has occurred thus far.
Nobody anticipated that El Ni˜ no would be very prolonged in 1992, or that it would be
exceptionally intense in 1997. As recently as 2004 some scientists sounded an alert
that El Ni˜ no was developing, but withdrew it a few months later after nothing unusual
had happened. As in the case of weather patterns which never repeat themselves, so
each El Ni˜ no is distinct. We have not yet observed a sufficient number of realizations
S. George Philander
of imaginative ideas to get a program moving. Usually a brilliant person would step
in to provide a spark, a Jule Charney, or Henry Stommel, or Claes Rooth. On this
occasion in early 1990, the person who stepped in was a “suit” from Washington
who came to brief us on the type of program that politicians are likely to favor. How
did nonscientists acquire the power to influence the activities of large numbers of
scientists?
Informally organized activities, with scientists themselves doing the planning
and implementation, are feasible when a science is “small.” However, the enormous
growth of physical oceanography over the past few decades, in the number of scientists, the number of papers they write, and the resources they need, transformed
the field from a “small” into a “big” science. Such growth calls for formally structured activities that respond to demands from the sponsors of research for results
that are of practical value. These can be salutary developments, provided we take
care to distinguish between accomplishment and frantic activity—science is a highly
undemocratic activity with a few people producing most of the important results
1 —
and provided science for the sake of understanding natural phenomena continues to
receive adequate support.
Growth sometimes leads to a loss of cohesion. That this was happening became evident in the special issue of the Journal of Geophysical Research (Vol. 103,
June 29, 1998) devoted entirely to articles that review “Progress of El Ni˜ no Research
and Prediction.” Each article, written jointly by several distinguished scientists, is
excellent, but is disappointingly disconnected from the other articles. The one that
reviews oceanographic measurements makes essentially no reference to ideas presented in the paper on idealized models of El Ni˜ no. The articles on realistic models,
and on prediction, are similarly detached, from each other and from the other articles
in the special issue. When I mentioned this problem to colleagues, some denied that
there is a problem, but others, David Neelin for example, agreed that different groups
have very different perspectives on the same phenomena. In the summer of 2002 I
called a meeting in Trieste, Italy, to address a seemingly straightforward question:
given that El Ni˜ no corresponds to a natural mode of the coupled ocean–atmosphere,
is that mode (a) strongly damped so that a continual oscillation is dependent on the
presence of random disturbances? Or is it (b) self-sustaining and highly predictable?
Or is it (c) sufficiently unstable to generate secondary instabilities so that, as in the
case of weather forecasting, predictability is limited by errors in initial conditions?
Each answer had ardent supporters!
After much discussion, a few people agreed that the available data sets are
inadequate to distinguish between the three possibilities. The paucity of the data is
evident in the way we have been surprised by each El Ni˜ no that has occurred thus far.
Nobody anticipated that El Ni˜ no would be very prolonged in 1992, or that it would be
exceptionally intense in 1997. As recently as 2004 some scientists sounded an alert
that El Ni˜ no was developing, but withdrew it a few months later after nothing unusual
had happened. As in the case of weather patterns which never repeat themselves, so
each El Ni˜ no is distinct. We have not yet observed a sufficient number of realizations