262
WILLIAM B. LARGE
Figure 11 shows the northward heat transport (global and by basin)
implied by the air-sea heat fluxes of Fig. 10. In such a plot, the zonal
average heat flux gives the slope of the curves (positive for heating,
negative for cooling), while the value at a particular latitude represents
the integrated heat flux from all ocean areas farther north, or equivalently in a balanced system, minus the integrated heat flux from all
ocean areas farther south. The missing ocean heat loss via ice-ocean
fluxes are believed to be only few W/m 2 , but these would steepen the
negative slopes at high latitudes and give larger poleward maxima in
both hemispheres. Figure 11 suggests that the heating around 50 ◦ S is
approximately equal to all the heat loss farther poleward, so that very
little heat is transported across 40 ◦ S. Similarly, there is little transport across 10 ◦ S, because of nearly equal, but opposite transports in
the Pacific and Atlantic. Note that the sum of Indian plus Pacific transport gives the correct Indo-Pacific implied transport, but partitioning
between the two basins assumes no heat transport via the Indonesian
Throughflow, which is not correct.
Figure 12 shows the mean (1984 –2000) global distribution of air-sea
freshwater flux (1984 –2000), plus the climatological continental runoff
from Fig. 1. The obvious freshwater source regions (unshaded in Fig. 12)
are the intertropical convergence zones (ITCZs), the mid-latitude storm
tracks of both hemispheres, and the mouths of large rivers. There is net
water loss (evaporation) from the Arabian Sea and the subtropical gyres
of each ocean basin. These features are common to all freshwater flux
climatologies, although the magnitudes of the precipitation and evaporation differ substantially, with no way of determining reality. The zonal
averages of the freshwater flux and its components are shown in Large
and Yeager (2004). The runoff is most significant where Siberian rivers
discharge between 60 and 70 ◦ N .
Integration of fluxes in Fig. 12 gives the implied northward transport
of freshwater shown in Fig. 13. Again partitioning between the Indian
and Pacific incorrectly assumes no transport through the Indonesian
Throughflow. The total transport can be compared to several other
estimates and direct observations given be (Wijffels, 2001). In general
there is better agreement with the direct observations in the northern
Hemisphere than in the southern.
Climatologies (1984 –2000) of the zonal and meridional wind stress
components are shown in Fig. 14. The pattern and magnitudes of both
components are remarkably similar to the SOC climatology, which has
generally larger stress magnitudes than the COADS (Beranger, 1999).
Since the SOC and COADS are based on essentially the same ship observations this difference is likely due to use of a different drag coefficient
WILLIAM B. LARGE
Figure 11 shows the northward heat transport (global and by basin)
implied by the air-sea heat fluxes of Fig. 10. In such a plot, the zonal
average heat flux gives the slope of the curves (positive for heating,
negative for cooling), while the value at a particular latitude represents
the integrated heat flux from all ocean areas farther north, or equivalently in a balanced system, minus the integrated heat flux from all
ocean areas farther south. The missing ocean heat loss via ice-ocean
fluxes are believed to be only few W/m 2 , but these would steepen the
negative slopes at high latitudes and give larger poleward maxima in
both hemispheres. Figure 11 suggests that the heating around 50 ◦ S is
approximately equal to all the heat loss farther poleward, so that very
little heat is transported across 40 ◦ S. Similarly, there is little transport across 10 ◦ S, because of nearly equal, but opposite transports in
the Pacific and Atlantic. Note that the sum of Indian plus Pacific transport gives the correct Indo-Pacific implied transport, but partitioning
between the two basins assumes no heat transport via the Indonesian
Throughflow, which is not correct.
Figure 12 shows the mean (1984 –2000) global distribution of air-sea
freshwater flux (1984 –2000), plus the climatological continental runoff
from Fig. 1. The obvious freshwater source regions (unshaded in Fig. 12)
are the intertropical convergence zones (ITCZs), the mid-latitude storm
tracks of both hemispheres, and the mouths of large rivers. There is net
water loss (evaporation) from the Arabian Sea and the subtropical gyres
of each ocean basin. These features are common to all freshwater flux
climatologies, although the magnitudes of the precipitation and evaporation differ substantially, with no way of determining reality. The zonal
averages of the freshwater flux and its components are shown in Large
and Yeager (2004). The runoff is most significant where Siberian rivers
discharge between 60 and 70 ◦ N .
Integration of fluxes in Fig. 12 gives the implied northward transport
of freshwater shown in Fig. 13. Again partitioning between the Indian
and Pacific incorrectly assumes no transport through the Indonesian
Throughflow. The total transport can be compared to several other
estimates and direct observations given be (Wijffels, 2001). In general
there is better agreement with the direct observations in the northern
Hemisphere than in the southern.
Climatologies (1984 –2000) of the zonal and meridional wind stress
components are shown in Fig. 14. The pattern and magnitudes of both
components are remarkably similar to the SOC climatology, which has
generally larger stress magnitudes than the COADS (Beranger, 1999).
Since the SOC and COADS are based on essentially the same ship observations this difference is likely due to use of a different drag coefficient