70
J. S. Godfrey
A SERIOUS MISMATCH
Such results represented a major step forward towards resolving a very frustrating
situation in oceanography: namely, the significant divide between (on the one hand)
our ability to diagnose what the present state of the ocean is, and (on the other) our
relative inability to quantify the forcing functions that cause this state to change. This
is especially true of the net heat flux; and this quantity is crucial to simulating the
SST, the key ocean parameter in coupled climate models.
We have many accurate tools for measuring the present state of the ocean. We
can routinely measure T and S to accuracies like 3 parts in 100,000. Combining
this with altimetry and Argo float data can and does provide a remarkably accurate,
global picture of the state of the upper ocean, essentially in real time. The contrast
between this “conventional” oceanography and the situation with heat fluxes is quite
embarrassing. Typical net heat fluxes into the ocean might be 100 W/m
2 at seasonal
peaks, with a long-term mean of perhaps 30 W/m
2 ; so a 10 W/m
2 accuracy is indeed
modest when expressed as a fraction of the signal one is quantifying—and even
10 W/m
2 accuracy is a research quality result. It has long been known that different
global net heat flux products—obtained using raw data collected by merchant ships
over many years—differ from one another by several tens of watts per square meter
over huge areas. Some carefully prepared products show an annual mean net heat
flux of as much as 40 W/m
2 , when averaged over the entire globe—enough to heat
the top 100 m of the ocean by about 3
◦ C/year! Evidently, these represent serious
biases.
I will argue later in this essay that if we can at least partly resolve this terrible
mismatch between our knowledge of the present ocean state, and of the forcings that
cause it to change, it will help us greatly in diagnosing what greenhouse gas increases
are doing to our climate. I will also argue that climate scientists will increasingly be
called upon to do this, in a timely manner, at higher and higher spatial resolutions;
and the financial issues involved will be very large. Before addressing this, however,
I want to explain why I think we will be able to obtain heat fluxes of much better
quality than those available till recently.
HOW CAN THE TOGA–COARE ACCURACY BE
TRANSFERRED TO GLOBAL HEAT FLUX CLIMATOLOGIES?
On the face of it, it might seem impossible to achieve global heat fluxes to accuracies
like 10 W/m
2 . If we have so far verified this accuracy only at small spots in the ocean,
after a very rigorous and expensive field campaign, how can we hope to get similar
improvements routinely in places like the Antarctic, where ships hardly ever go? Yet
recent analyses which make use of satellite data and numerical weather prediction
model output suggest that we may already have a flux estimation technique which in
combination with ship observations will allow us to approach this accuracy.
J. S. Godfrey
A SERIOUS MISMATCH
Such results represented a major step forward towards resolving a very frustrating
situation in oceanography: namely, the significant divide between (on the one hand)
our ability to diagnose what the present state of the ocean is, and (on the other) our
relative inability to quantify the forcing functions that cause this state to change. This
is especially true of the net heat flux; and this quantity is crucial to simulating the
SST, the key ocean parameter in coupled climate models.
We have many accurate tools for measuring the present state of the ocean. We
can routinely measure T and S to accuracies like 3 parts in 100,000. Combining
this with altimetry and Argo float data can and does provide a remarkably accurate,
global picture of the state of the upper ocean, essentially in real time. The contrast
between this “conventional” oceanography and the situation with heat fluxes is quite
embarrassing. Typical net heat fluxes into the ocean might be 100 W/m
2 at seasonal
peaks, with a long-term mean of perhaps 30 W/m
2 ; so a 10 W/m
2 accuracy is indeed
modest when expressed as a fraction of the signal one is quantifying—and even
10 W/m
2 accuracy is a research quality result. It has long been known that different
global net heat flux products—obtained using raw data collected by merchant ships
over many years—differ from one another by several tens of watts per square meter
over huge areas. Some carefully prepared products show an annual mean net heat
flux of as much as 40 W/m
2 , when averaged over the entire globe—enough to heat
the top 100 m of the ocean by about 3
◦ C/year! Evidently, these represent serious
biases.
I will argue later in this essay that if we can at least partly resolve this terrible
mismatch between our knowledge of the present ocean state, and of the forcings that
cause it to change, it will help us greatly in diagnosing what greenhouse gas increases
are doing to our climate. I will also argue that climate scientists will increasingly be
called upon to do this, in a timely manner, at higher and higher spatial resolutions;
and the financial issues involved will be very large. Before addressing this, however,
I want to explain why I think we will be able to obtain heat fluxes of much better
quality than those available till recently.
HOW CAN THE TOGA–COARE ACCURACY BE
TRANSFERRED TO GLOBAL HEAT FLUX CLIMATOLOGIES?
On the face of it, it might seem impossible to achieve global heat fluxes to accuracies
like 10 W/m
2 . If we have so far verified this accuracy only at small spots in the ocean,
after a very rigorous and expensive field campaign, how can we hope to get similar
improvements routinely in places like the Antarctic, where ships hardly ever go? Yet
recent analyses which make use of satellite data and numerical weather prediction
model output suggest that we may already have a flux estimation technique which in
combination with ship observations will allow us to approach this accuracy.