74
J. S. Godfrey
advances in tracking changes in trace gases—both in the atmosphere and from ice
cores. We have watched three generations of global coupled models be built, each
with higher resolution and better physics than the last; and we have seen marked
improvements in simulating the present-day climate, in each new generation. We
have heard results of various process studies—airplanes examining cloud nucleation,
studies of gas emissions by cows, response of plants to increased CO 2 , our own
TOGA-COARE—and seen changes implemented in the models as a result. We have
also seen this group of scientists respond to pretty vigorous criticisms, making changes
based on those they judged to be valid, and simply wearing those they did not. I have
not been a major contributor, which entitles me to say that I have been extremely
impressed by the rigorous approach taken throughout this collective scientific effort.
This experience has provided one example of a phenomenon which I suspect
we will see repeated many times in coming decades, as greenhouse warming rates
increase. This phenomenon leads me to believe that development and verification
of good ocean heat flux products—and work to make the ocean and atmospheric
components of the coupled models also match observed ocean heat fluxes—should
be a high-priority item for the climate science community. This phenomenon is as
follows.
From 1911 to 1975, the net river inflow to the water supply of Perth, Western
Australia fluctuated considerably on decadal time scales, but about a rather steady
mean. In 1975 that inflow fell. As time went on, this looked less and less like an
ordinary fluctuation and more like a long-term change. As of now, it looks like a step
function; the average inflow from 1975 through 2004 is very close to half that from
1911 to 1975.
By the late 1990s, the Perth Water Board asked the Bureau of Meteorology and
CSIRO to advise them: Is this due to natural climate variability? Or to greenhouse
warming? Or to some other causes such as changing land use?
After a lengthy study of the output of nine global climate models with and
without greenhouse gas increases, and “downscaling” statistical studies based on how
observed rainfall at a point depended on large-scale weather patterns, the scientists
involved gave their report. In essence they concluded that changes in land usage
probably played a minor role. They said that climate variability played a major role,
especially in the sharp onset of the drought. They noted that seven out of nine models
showed Australian rainfall decrease (along with the warming!) to have a maximum
in southwest Western Australia, as observed; though no model showed a drying rate
nearly as fast as has been observed in that region. They did not offer an opinion on
whether this meant that the drying was primarily due to natural variability, which
for some reason mimicked the modelled greenhouse signal; or that there was some
inadequacy (local or global), in all of the present generation of greenhouse models, to
properly predict the large rate of drying in this region (which is small on a global scale).
Almost inevitably, the scientists were caught in a difficult situation, since they were all
well aware of the limitations—and huge complexity—of the models they were using
to make an assessment, on which about half a billion investment dollars depended.
J. S. Godfrey
advances in tracking changes in trace gases—both in the atmosphere and from ice
cores. We have watched three generations of global coupled models be built, each
with higher resolution and better physics than the last; and we have seen marked
improvements in simulating the present-day climate, in each new generation. We
have heard results of various process studies—airplanes examining cloud nucleation,
studies of gas emissions by cows, response of plants to increased CO 2 , our own
TOGA-COARE—and seen changes implemented in the models as a result. We have
also seen this group of scientists respond to pretty vigorous criticisms, making changes
based on those they judged to be valid, and simply wearing those they did not. I have
not been a major contributor, which entitles me to say that I have been extremely
impressed by the rigorous approach taken throughout this collective scientific effort.
This experience has provided one example of a phenomenon which I suspect
we will see repeated many times in coming decades, as greenhouse warming rates
increase. This phenomenon leads me to believe that development and verification
of good ocean heat flux products—and work to make the ocean and atmospheric
components of the coupled models also match observed ocean heat fluxes—should
be a high-priority item for the climate science community. This phenomenon is as
follows.
From 1911 to 1975, the net river inflow to the water supply of Perth, Western
Australia fluctuated considerably on decadal time scales, but about a rather steady
mean. In 1975 that inflow fell. As time went on, this looked less and less like an
ordinary fluctuation and more like a long-term change. As of now, it looks like a step
function; the average inflow from 1975 through 2004 is very close to half that from
1911 to 1975.
By the late 1990s, the Perth Water Board asked the Bureau of Meteorology and
CSIRO to advise them: Is this due to natural climate variability? Or to greenhouse
warming? Or to some other causes such as changing land use?
After a lengthy study of the output of nine global climate models with and
without greenhouse gas increases, and “downscaling” statistical studies based on how
observed rainfall at a point depended on large-scale weather patterns, the scientists
involved gave their report. In essence they concluded that changes in land usage
probably played a minor role. They said that climate variability played a major role,
especially in the sharp onset of the drought. They noted that seven out of nine models
showed Australian rainfall decrease (along with the warming!) to have a maximum
in southwest Western Australia, as observed; though no model showed a drying rate
nearly as fast as has been observed in that region. They did not offer an opinion on
whether this meant that the drying was primarily due to natural variability, which
for some reason mimicked the modelled greenhouse signal; or that there was some
inadequacy (local or global), in all of the present generation of greenhouse models, to
properly predict the large rate of drying in this region (which is small on a global scale).
Almost inevitably, the scientists were caught in a difficult situation, since they were all
well aware of the limitations—and huge complexity—of the models they were using
to make an assessment, on which about half a billion investment dollars depended.