80
Q) 40
'0
:J
.->
.....
.->
j 0
-40
72
-80." 60 E 120 E 180 E 120 W 60 W 0
Longitude
Figure 13. Barotropic streamfunction, 'fib' for run I (wind only simulation). The contour labels
are in Sverdrups (1 Sv = 10 6 m 3 s-I). The arrows indicate the sense of circulation.
Boundary conditions (62)-(63) present, however, serious drawbacks. If the observed
surface value is time-dependent, the variations of the modelled variable are likely to be smaller
than those of the observed values, and there will probably be a phase lag. This may be
understood by considering a simple equation for a given variable If!,
alf!
alf!
ili + u ax = - r(lf!- If!obs) ,
(64)
where ,,-1 and u are an appropriate time scale and a constant advective velocity, respectively.
We study Fourier components of the form
(If!, If!obs) = Re{ (Y, Y obs ) exp[1 (COt - kx)] } ,
(65)
where co and k denote the angular frequency and the wavenumber, respectively. Introducing
(65) into (64), we have
IYI
argY
co- uk
argY obs - atan--r
(66)
(67)
Q) 40
'0
:J
.->
.....
.->
j 0
-40
72
-80." 60 E 120 E 180 E 120 W 60 W 0
Longitude
Figure 13. Barotropic streamfunction, 'fib' for run I (wind only simulation). The contour labels
are in Sverdrups (1 Sv = 10 6 m 3 s-I). The arrows indicate the sense of circulation.
Boundary conditions (62)-(63) present, however, serious drawbacks. If the observed
surface value is time-dependent, the variations of the modelled variable are likely to be smaller
than those of the observed values, and there will probably be a phase lag. This may be
understood by considering a simple equation for a given variable If!,
alf!
alf!
ili + u ax = - r(lf!- If!obs) ,
(64)
where ,,-1 and u are an appropriate time scale and a constant advective velocity, respectively.
We study Fourier components of the form
(If!, If!obs) = Re{ (Y, Y obs ) exp[1 (COt - kx)] } ,
(65)
where co and k denote the angular frequency and the wavenumber, respectively. Introducing
(65) into (64), we have
IYI
argY
co- uk
argY obs - atan--r
(66)
(67)