106
W.T. Liu et al.
reasons are related to the directional difference between wind and stress. The procedure to “select” the stress from multi-solutions with ambiguous direction needs
modification in two steps. The first is testing various “nudging” fields that are more
relevant to stress than wind. Our feasibility studies suggest that such changes in the
nudging field is significant in the region where strong ocean current exists and stress
should point to the direction of the vector difference between wind and current. A
more flexible median filter should also be designed to accommodate the small spatial
scale of stress as compared with winds.
The development of GMF-S could follow closely those for GMF-W, and our feasibility study using a small sample of stress vectors and “nudging” data that include
ocean current are successful. One of the reason usually given for promoting scatterometer as a wind sensor instead of stress sensor is that there are more wind than
stress measurements to develop and calibrate the GMF. Such explanation runs in
contrast to the rationale behind using U N as the geophysical product because of its
unambiguous relation to stress. To provide each U N for development or calibration
the GMF from measured wind, stress or U ∗ has to be computed first as discussed in
Section 6.2. There are as many stress available as U N .
The stress prepared from wind in such way is not ideal because it addresses
only the stability problem but does not include current information. Such deficiency may be somewhat alleviated through the ambiguity removal process by using
more appropriate filter size and nudging with the vector difference between wind
and optimal surface current information that is available. Ocean surface current
measurements are very sparse. The current velocity has been derived from Argos
satellite collections of the displacements of drifters with drogues centered at 15 mdepth (Niiler, 2001). Ocean surface currents are also provided by the Ocean Surface
Currents Analyses – Realtime program, using a combination of scatterometer and
altimeter data (Lagerloef et al., 1999). Global, high frequency current data are only
available from numerical models, such as, Estimating the Circulation and Climate
of the Ocean (Fukumori, 2002) and Regional Ocean Modeling System (Chao
et al., 2009).
6.5 Potential Oceanic Feedback
Although Pacanowski (1987) showed the importance of the feedback of ocean current on stress in a numerical experiment more than two decades ago, many ocean
scientists are still forcing their ocean models using stress that is entirely determined
from the wind field, independent of local changes caused by ocean. With the availability of QuikSCAT data, Polito et al. (2001) discussed the surface current and
temperature feedback to the stress forcing, and Pezzi et al. (2004) demonstrated
negative feedback with ocean general circulation model, at the tropical instability
waves. With the recent demonstration of spatial coherence between scatterometer
measurement and T s over extensive ocean regions, as discussed in Section 6.3, many
numerical studies of T s feedback to stress forcing were performed with numerical
models (e.g. Seo et al., 2007; Spall, 2007; Hogg et al., 2009). The measurement of
W.T. Liu et al.
reasons are related to the directional difference between wind and stress. The procedure to “select” the stress from multi-solutions with ambiguous direction needs
modification in two steps. The first is testing various “nudging” fields that are more
relevant to stress than wind. Our feasibility studies suggest that such changes in the
nudging field is significant in the region where strong ocean current exists and stress
should point to the direction of the vector difference between wind and current. A
more flexible median filter should also be designed to accommodate the small spatial
scale of stress as compared with winds.
The development of GMF-S could follow closely those for GMF-W, and our feasibility study using a small sample of stress vectors and “nudging” data that include
ocean current are successful. One of the reason usually given for promoting scatterometer as a wind sensor instead of stress sensor is that there are more wind than
stress measurements to develop and calibrate the GMF. Such explanation runs in
contrast to the rationale behind using U N as the geophysical product because of its
unambiguous relation to stress. To provide each U N for development or calibration
the GMF from measured wind, stress or U ∗ has to be computed first as discussed in
Section 6.2. There are as many stress available as U N .
The stress prepared from wind in such way is not ideal because it addresses
only the stability problem but does not include current information. Such deficiency may be somewhat alleviated through the ambiguity removal process by using
more appropriate filter size and nudging with the vector difference between wind
and optimal surface current information that is available. Ocean surface current
measurements are very sparse. The current velocity has been derived from Argos
satellite collections of the displacements of drifters with drogues centered at 15 mdepth (Niiler, 2001). Ocean surface currents are also provided by the Ocean Surface
Currents Analyses – Realtime program, using a combination of scatterometer and
altimeter data (Lagerloef et al., 1999). Global, high frequency current data are only
available from numerical models, such as, Estimating the Circulation and Climate
of the Ocean (Fukumori, 2002) and Regional Ocean Modeling System (Chao
et al., 2009).
6.5 Potential Oceanic Feedback
Although Pacanowski (1987) showed the importance of the feedback of ocean current on stress in a numerical experiment more than two decades ago, many ocean
scientists are still forcing their ocean models using stress that is entirely determined
from the wind field, independent of local changes caused by ocean. With the availability of QuikSCAT data, Polito et al. (2001) discussed the surface current and
temperature feedback to the stress forcing, and Pezzi et al. (2004) demonstrated
negative feedback with ocean general circulation model, at the tropical instability
waves. With the recent demonstration of spatial coherence between scatterometer
measurement and T s over extensive ocean regions, as discussed in Section 6.3, many
numerical studies of T s feedback to stress forcing were performed with numerical
models (e.g. Seo et al., 2007; Spall, 2007; Hogg et al., 2009). The measurement of