SURFACE FLUXES
251
q(z u ) = q(z q ) −
q ∗
κ
[ln(
z q
z u
) + ψ h (ζ u ) − ψ h (ζ q )] .
(51)
4) Update the neutral 10m transfer coefficients (Fig. 5), then shift them
to the measurement height, z u , and stability, ζ u , using:
C D (z u , ζ) = C DN (1 +
√
C DN
κ
[ln(z u /10m) − ψ m (ζ u )]) −2
C H (z u , ζ) = C HN (1 +
C HN
κ
√
C DH
[ln(z u /10m) − ψ h (ζ u )]) −2
C E (z u , ζ) = C EN (1 +
C EN
κ
√
C DN
[ln(z u /10m) − ψ h (ζ u )]) −2
(52)
5) Using these transfer coefficients, θ(z u ) from (50) and q(z u ) from (51),
recompute the virtual potential temperature, θ v , and update the turbulent scales in (47).
6) Start the next iteration loop at step (2) above. Over the ocean stability is usually near neutral, and 2 iterations are all that is necessary. In
very stable conditions up to five iterations may be needed. After the last
iteration the fluxes are computed from the last set of turbulence scales
according to (15).
Even with perfect atmospheric inputs and SST, the flux estimates are
only as representative of true conditions as the bulk transfer coefficients.
Figure 3 shows that after the wind speed and direction shifts the momentum (heat and moisture) flux would be underestimated by nearly a
factor of 2 (
√
2), and overestimated early in the time series. When averaged over the 3 - 7 day synoptic period, the uncertainty should diminish
as wave effects start to average out. But, because of the seasonal cycle
of storms, annual averaging is probably required before the uncertainty
in bulk fluxes is minimized.
5.
Satellite flux estimates
Satellite retrieval products are attractive for GODAE, because of the
frequent, global sampling, and the possibility of near real time delivery; continued availability notwithstanding. Over the ocean satellites
have been most successful at producing wind, radiation and precipitation fields. However, these are necessarily very indirect measures, and
in situ measurements, are essential for providing the calibration standard, at least at a few select locations. Therefore, the accuracy of these
observations will be only briefly included below, with WGASF (2000)
a more complete reference. There are schemes for estimating monthly
mean latent heat flux (evaporation), but this is likely too infrequent for
GODAE.
251
q(z u ) = q(z q ) −
q ∗
κ
[ln(
z q
z u
) + ψ h (ζ u ) − ψ h (ζ q )] .
(51)
4) Update the neutral 10m transfer coefficients (Fig. 5), then shift them
to the measurement height, z u , and stability, ζ u , using:
C D (z u , ζ) = C DN (1 +
√
C DN
κ
[ln(z u /10m) − ψ m (ζ u )]) −2
C H (z u , ζ) = C HN (1 +
C HN
κ
√
C DH
[ln(z u /10m) − ψ h (ζ u )]) −2
C E (z u , ζ) = C EN (1 +
C EN
κ
√
C DN
[ln(z u /10m) − ψ h (ζ u )]) −2
(52)
5) Using these transfer coefficients, θ(z u ) from (50) and q(z u ) from (51),
recompute the virtual potential temperature, θ v , and update the turbulent scales in (47).
6) Start the next iteration loop at step (2) above. Over the ocean stability is usually near neutral, and 2 iterations are all that is necessary. In
very stable conditions up to five iterations may be needed. After the last
iteration the fluxes are computed from the last set of turbulence scales
according to (15).
Even with perfect atmospheric inputs and SST, the flux estimates are
only as representative of true conditions as the bulk transfer coefficients.
Figure 3 shows that after the wind speed and direction shifts the momentum (heat and moisture) flux would be underestimated by nearly a
factor of 2 (
√
2), and overestimated early in the time series. When averaged over the 3 - 7 day synoptic period, the uncertainty should diminish
as wave effects start to average out. But, because of the seasonal cycle
of storms, annual averaging is probably required before the uncertainty
in bulk fluxes is minimized.
5.
Satellite flux estimates
Satellite retrieval products are attractive for GODAE, because of the
frequent, global sampling, and the possibility of near real time delivery; continued availability notwithstanding. Over the ocean satellites
have been most successful at producing wind, radiation and precipitation fields. However, these are necessarily very indirect measures, and
in situ measurements, are essential for providing the calibration standard, at least at a few select locations. Therefore, the accuracy of these
observations will be only briefly included below, with WGASF (2000)
a more complete reference. There are schemes for estimating monthly
mean latent heat flux (evaporation), but this is likely too infrequent for
GODAE.