205
of the boundary conditions by the atmosphere, we can expect predictability due only to the slowness of the variation: we would therefore expect the
limit on the prediction time to be given by persistence, i.e. by the autocorrelation time of the boundary condition. Clearly only those mechanisms
we have discussed that intrinsically involve the ocean in a deterministic
way will lead to predictability beyond the persistence time. Thus if most
decadal-to- centennial variability is due to random forcing of the ocean by
the atmosphere, we have no hope in ever predicting the future SST beyond
persistence. If however the mechanism of variability involves the ocean either as part of a coupled mode or by internal ocean variability as described
in Sections 4 and 5, then there is some predictability of the future ocean
dynamics and we have some possibilities of predicting the future SST for
times longer than the persistence time. One way of unambiguously identifying a coupled atmosphere-ocean mode, for example, is this enhanced
predictability beyond the persistence time.
We have to know those aspects of the ocean that will affect the SST
on decadal-to-centennial time scales and initialize those aspects. Unfortunately, we don't know how to answer to this basic question. We have been
able to study, by means of tracer experiments (e.g. with Clorofluorocarbons) the inverse problem, how water proceeds from the surface into the
interior. We have not been able to answer the basic prediction question of
which water reaches the surface in a given time and how this water reaches
the surface. If the ocean can be initialized, and if the future evolution
of the coupled system is at least partially determined by this dynamical
evolution of this initial state, then we have hope of some skill in climate
prediction years into the future.
Acknowledgements
This work was supported by a grant from the NOAA/Office of Global
Programs to the Stanley P. Hayes Center of the University of Washington. MW was supported by a UCAR ocean modeling post-doctoral fellowship while 50% of F.L.Y.'s time was supported by the US Department of
Energy's NIGEC through the NIGEC Regional Center at Univ. of California, Davis (DOE Cooperative Agreement No. DE-FC03-90ER61010):
financial support does not constitute endorsement by DOE of the views
expressed in this article. The invaluable assistance of the Margaret Black
Lab is gratefully acknowledged. Conversations with Bob Charlson, Todd
of the boundary conditions by the atmosphere, we can expect predictability due only to the slowness of the variation: we would therefore expect the
limit on the prediction time to be given by persistence, i.e. by the autocorrelation time of the boundary condition. Clearly only those mechanisms
we have discussed that intrinsically involve the ocean in a deterministic
way will lead to predictability beyond the persistence time. Thus if most
decadal-to- centennial variability is due to random forcing of the ocean by
the atmosphere, we have no hope in ever predicting the future SST beyond
persistence. If however the mechanism of variability involves the ocean either as part of a coupled mode or by internal ocean variability as described
in Sections 4 and 5, then there is some predictability of the future ocean
dynamics and we have some possibilities of predicting the future SST for
times longer than the persistence time. One way of unambiguously identifying a coupled atmosphere-ocean mode, for example, is this enhanced
predictability beyond the persistence time.
We have to know those aspects of the ocean that will affect the SST
on decadal-to-centennial time scales and initialize those aspects. Unfortunately, we don't know how to answer to this basic question. We have been
able to study, by means of tracer experiments (e.g. with Clorofluorocarbons) the inverse problem, how water proceeds from the surface into the
interior. We have not been able to answer the basic prediction question of
which water reaches the surface in a given time and how this water reaches
the surface. If the ocean can be initialized, and if the future evolution
of the coupled system is at least partially determined by this dynamical
evolution of this initial state, then we have hope of some skill in climate
prediction years into the future.
Acknowledgements
This work was supported by a grant from the NOAA/Office of Global
Programs to the Stanley P. Hayes Center of the University of Washington. MW was supported by a UCAR ocean modeling post-doctoral fellowship while 50% of F.L.Y.'s time was supported by the US Department of
Energy's NIGEC through the NIGEC Regional Center at Univ. of California, Davis (DOE Cooperative Agreement No. DE-FC03-90ER61010):
financial support does not constitute endorsement by DOE of the views
expressed in this article. The invaluable assistance of the Margaret Black
Lab is gratefully acknowledged. Conversations with Bob Charlson, Todd