Energetic subthermocline currents observed east of Mindanao

Energetic subthermocline currents observed east of Mindanao

ARTICLE IN PRESS Deep-Sea Research II 52 (2005) 605–613 www.elsevier.com/locate/dsr2 Energetic subthermocline currents observed east of Mindanao E. ...

609KB Sizes 2 Downloads 65 Views

ARTICLE IN PRESS

Deep-Sea Research II 52 (2005) 605–613 www.elsevier.com/locate/dsr2

Energetic subthermocline currents observed east of Mindanao E. Firinga,, Y. Kashinob, P. Hackerc a

Department of Oceanography, University of Hawaii at Manoa, 1000 Pope Rd., Honolulu, HI 96822, USA b Japan Agency for Marine-Earth Science and Technology, Yokosuka, Japan c International Pacific Research Center, University of Hawaii at Manoa, 1680 East-West Road, Honolulu, HI 96822, USA Received 12 April 2004; accepted 7 December 2004

Abstract Two cruises of the JAMSTEC ship Kaiyo, during October 1999 and September 2000, included lowered acoustic Doppler current profiler (ADCP) measurements to 2000 m depth on zonal sections extending east from the Mindanao coast. On the second cruise, the lowered ADCP profiles were augmented by profiles to 1000 m from a 38-kHz shipboard ADCP. All zonal LADCP sections (71N, 81, 101) showed southward flow along the coast extending to at least 2000 m depth. Although the Mindanao current in the upper 500 m forms a continuous narrow stream along the coast, the deeper southward flow appears to be part of a set of subthermocline eddies within 300 km of the coast; northward flow was found 100–200 km offshore during both cruises. Currents mapped by the shipboard ADCP on the second cruise indicate that cyclonic eddies were centered near 7.51N, 1281E and 10.21N, 1271E. Maximum speeds of 0.6 m s1 were observed at 800 m depth on the 101N section, and speeds of 0.2–0.3 m s1 were found below 1500 m on all sections. r 2005 Elsevier Ltd. All rights reserved.

1. Introduction The Mindanao Current (MC) is the western boundary current carrying waters from the North Equatorial Current (NEC) southward along the Philippines coast. Part of the MC turns east to feed the North Equatorial Countercurrent (NECC), and another branch enters the Celebes Sea. A fraction of this second branch is the primary

Corresponding author. Tel.: +1 808 956 7894;

fax: +1 808 956 4104. E-mail address: efi[email protected] (E. Firing).

source of the Indonesian Throughflow. See Fine et al. (1994) for a review. Above 300 m, the structure of the MC has been described based on shipboard acoustic Doppler current profiler (ADCP) surveys. Zonal sections from a cruise in July 1988 (Lukas et al., 1991) show both the maximum speed and the vertical shear in the MC increasing downstream from 121N to 71N; the southward flow component at 300 m depth near the coast exceeded 0.6 m s1 on 121N, but dropped to about 0.1 m s1 on 71N. In the same sections, the MC core speed increased downstream, from about 0.8 m s1 on 121N to 1.3 m s1 on 71N. An ensemble of seven sections at

0967-0645/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.dsr2.2004.12.007

ARTICLE IN PRESS 606

E. Firing et al. / Deep-Sea Research II 52 (2005) 605–613

81N (Wijffels et al., 1995) shows a mean flow exceeding 0.9 m s1 at the shelf break, but a standard deviation of only 0.1–0.15 m s1 in most of the high-velocity region. The width of the mean current is 150 km at 300 m depth. Information about currents below 300 m has come primarily from geostrophic calculations. Based on three annual sections along 7.51N, Hu et al. (1991), suggested that multiple subsurface northward current cores, some extending as deep as 1000 m, are typically found offshore of the MC. They named this northward subthermocline flow the Mindanao Undercurrent (MUC). Evidence for the MUC was also found in a synoptic hydrographic survey by Lukas et al. (1991). Individual synoptic sections yield velocity estimates that may be contaminated by ageostrophic fluctuations in dynamic height, such as internal tides, and that include transient as well as mean geostrophic currents (Lukas et al., 1991; Wijffels et al., 1995). To get a better estimate of the mean geostrophic velocity field, Wijffels et al. (1995) averaged eight sections on 81N, and Qu et al. (1998) included two additional cruises near 81N in their climatological analysis of a larger region. Relative to 2000 m (their Fig. 6b), Qu et al. (1998) derived southward flow near the coast extending to 1500 m depth, with magnitude exceeding 0.05 m s1 above 800 m. About 100 km offshore they found a northward core exceeding 0.05 m s1 from 700–1300 m, and identified it as the MUC. Qu et al. (1998) concluded that the MUC is most likely a permanent feature of the intermediate circulation. In addition to geostrophy, water property distributions have been used to infer intermediate-depth flow. Reid and Mantyla (1978) pointed to the broad high-oxygen tongue near 1000 m along the western boundary of the North Pacific, together with the dynamic height field at 1000 dbar relative to 3500 dbar, as evidence for a mean northward western boundary current. High oxygen and low salinity near the 27.2 kg m3 potential density anomaly ðsy Þ mark the influence of Antarctic Intermediate Water (AAIW). Using a recent climatology, Qu et al. (1999) detected AAIW along the Philippine coast to about 121N. With additional evidence from maps of acceleration potential on 27.2 sy relative to 1200 dbar,

they concluded that the northward flow in the MUC returns offshore; the MUC is part of a local recirculation, rather than a meridionally extensive western boundary current. Wijffels et al. (1995) concluded that the velocity variance near the Mindanao coast in the 400–2000 m depth range is large relative to the mean flow, casting doubt on any estimate of the mean based on a small number of sections. In this note we present current measurements made during surveys in 1999 and 2000 that confirm their conclusion, and that provide snapshots of the horizontal and vertical structure of the velocity field.

2. Data and methods As part of the Tropical Ocean Climate Study (TOCS), the R.V. Kaiyo surveyed the Pacific lowlatitude western boundary region from October 20 to November 6, 1999 (cruise KY9909 leg 1) and from September 13 to October 1, 2000 (KY0006 leg 3). SADCP velocity measurements were made through out each cruise, and lowered ADCP (LADCP) measurements were made on CTD stations, most of which extended to 2000 m (Fig. 1). Different SADCP systems were used on the two cruises. On KY9909, a Narrow-band model VM75 ADCP made by RD Instruments (RDI) provided velocity estimates in 16-m intervals starting at 30 m and typically extending to about 550 m. Prior to KY0006, an RDI Ocean Surveyor model OS-38 ADCP was installed. Operating at 38 kHz, this instrument profiled from 37 m to about 1000 m, also in 16-m intervals; in some locations the range extended to 1200 m. On both cruises, a differential GPS receiver provided position fixes. Heading information came solely from the ship’s gyrocompass, which was therefore the largest source of error in the absolute depth-averaged velocity estimates (King et al., 2001); these errors are negligible when the ship was on station, and are probably 0.04 m s1 or less most of the time when underway, assuming 0.41 or smaller heading error remaining after calibration. If we take this as a worst case, then the error in

ARTICLE IN PRESS E. Firing et al. / Deep-Sea Research II 52 (2005) 605–613

overlapping depth ranges (about 500 m on KY9909 and 1000 m on KY0006), the rms difference between SADCP and LADCP velocities was about 0.05 m s1. Acoustic scattering is weak in this region, especially below 1000 m, so the rms error in the deep velocities probably exceeds 0.05 m s1 in this data set. Given that our present interest is in ocean features that persist for many days, high-frequency phenomena such as tides and near-inertial waves constitute noise. Although we have no independent measurements of this noise during the cruises, the spatial coherence of our measurements within each survey confirms our expectation that the noise is probably at the 0.05 m s1 level while the magnitude of our signal is 0.1–0.6 m s1.

12˚N KY9909 KY0006 LADCP

10˚N

8˚N

Mindanao

6˚N

4˚N 124˚E

128˚E 6

5

4

3

607

132˚E 2

1 0.50.1

Depth (km)

Fig. 1. Bottom topography in the Mindanao Current region as predicted by Smith and Sandwell (1997). Red lines show 1000and 2000-m contours. Cruise tracks of the R.V. Kaiyo in 1999 and 2000, and locations of LADCP profiles, are superimposed.

transport over a 400-m layer would be 1.6 Sv ð1 Sv ¼ 106 m3 s1 Þ per 100 km of cruise track. Heading calibration calculations based on the standard watertrack method (Joyce, 1989; Pollard and Read, 1989) were quite consistent throughout the cruise, indicating little compass drift. We arbitrarily suggest that a more likely uncertainty (or standard error) in transport is about half the estimated worst case: 0.8 Sv per 100 km of cruise track. Transport calculations discussed in Section 4 are consistent with this estimate. LADCP measurements on both cruises were made with a pair of Sontek 250-kHz ADCPs. The upward-looking and downward-looking instruments were configured to ping simultaneously about three times per second. The ping length was 16 m, and velocities were estimated in 8-m cells. Velocity profiles were calculated as described by Fischer and Visbeck (1993), using shear estimates from both instruments. Instrument bias was too large to permit using the shear estimate from the difference between the first-bin velocity estimates of the two instruments. Over their

3. Currents In the upper ocean, the MC is evident during both cruises as a strong, narrow, southward flow along the Mindanao coast (Fig. 2). It is seen more clearly in the 175–225-m average than in the 40–75-m average; at the shallower level on KY0006 the flow is weak in the 101N section, and the region of southward flow on the 81N and 71N sections is broadened by what appears to be an adjacent cyclonic circulation. Although one is tempted to call this the Mindanao Eddy, the eastward flow on 81N near 1301E is inconsistent with a circular eddy; rather, it appears that southward flow along the coast is retroflecting into the NECC. In any case, this flow is very shallow; only a trace remains at the 175–225-m level. On KY9909, the dominant shallow flow east of the MC is the anticyclonic Halmahera Eddy (HE), centered near 4.71N, 1301E, together with the retroflection of the South Equatorial Current into the NECC. Compared to the 40–75-m level, the HE at the 175–225-m level is smaller and its center is shifted to the southwest, somewhat as seen during the winter of 1995 and 1996 (Kashino et al., 1999). Below about 400 m (Fig. 3), a southward component of flow near the Mindanao coast is seen in all zonal sections from both cruises, but the diagonal SADCP sections from KY0006 suggest

ARTICLE IN PRESS E. Firing et al. / Deep-Sea Research II 52 (2005) 605–613

608

KY9909 SADCP 40-75m

KY0006 SADCP 40-75m

12°N

12°N

0.5 m/s

10°N

8°N

Mindanao

8°N

6°N

Mindanao

6°N

4°N 124°E

128°E

132°E

4°N 124°E

KY9909 SADCP 175-225m

128°E

132°E

KY0006 SADCP 175-225m

12°N

12°N

0.5 m/s

10°N

8°N

0.5 m/s

10°N

Mindanao

8°N

6°N

4°N 124°E

0.5 m/s

10°N

Mindanao

6°N

128°E

132°E

4°N 124°E

128°E

132°E

Fig. 2. Upper ocean currents measured by SADCP in October–November 1999 (left panels) and September 2000 (right panels). Currents are averaged over 40–75 m in the top panels and over 175–225 m in the bottom panels. Note the large change in currents with depth east of about 127.51E.

that this subthermocline flow is part of an energetic set of eddies within 300 km of the coast. Although the flow is complicated, as a first approximation one might identify two cyclonic eddies, one centered near 1281E, 7.51N, and the other near 127.11E, 10.21N. (We will refer to these as the 7.51N eddy and the 101N eddy, respectively.) The maximum speed, averaged from 750–850 m depth and over each of two consecutive

10-km sections of cruise track in the 101N eddy, exceeds 0.55 m s1. Where LADCP profiles are available on KY0006, most show similar flow from 500 m to the deepest measurements, at 2000 m; the eddies extend at least that deep, with speeds exceeding 0.2 m s1. Contoured LADCP profiles show the depth structure in more detail (Fig. 4). The 101N eddy has a strong velocity maximum near 800 m, below

ARTICLE IN PRESS E. Firing et al. / Deep-Sea Research II 52 (2005) 605–613 KY9909 SADCP 475-550m

KY0006 SADCP 475-550m

12°N

12°N

0.5 m/s

10°N

8°N

0.5 m/s

10°N

Mindanao

8°N

6°N

Mindanao

6°N

4°N 124°E

128°E

132°E

4°N 124°E

KY9909 LADCP 850-950m

132°E

KY0006 SADCP 850-950m

0.5 m/s

10°N

0.5 m/s

10°N

Mindanao

8°N

Mindanao

6°N

6°N

4°N 124°E

128°E

132°E

4°N 124°E

KY9909 LADCP 1500-1750m

128°E

132°E

KY0006 LADCP 1500-1750m

12°N

12°N

0.5 m/s

10°N

8°N

128°E

12°N

12°N

8°N

609

Mindanao

8°N

6°N

4°N 124°E

0.5 m/s

10°N

Mindanao

6°N

128°E

132°E

4°N 124°E

128°E

132°E

Fig. 3. Deep currents measured in October–November 1999 (left panels) and September 2000 (right panels). Currents averaged from 475–550 m (top row) were measured with the SADCP during both cruises. In the 850–950-m layer (middle row), SADCP observations were available only in 2000, so the LADCP measurements are shown for 1999. LADCP measurements from both cruises show flow in the 1500–1750-m layer (bottom row). For the profiles on KY9909 that did not reach 1750 m, the average from 1500 m to the bottom of the profile is shown.

ARTICLE IN PRESS E. Firing et al. / Deep-Sea Research II 52 (2005) 605–613

610 U: KY0006, 10˚N

V: KY0006, 10˚N

0 60 40 20 0 −20

1000

cm/s

depth (m)

500

−40 −60

1500

127˚E 128˚E U: KY0006, 8˚N

127˚E V: KY0006, 8˚N

128˚E

0 60 20 0 −20

1000

−40 −60

1500

0

127˚E 127.6˚E

127˚E 127.6˚E

U: KY0006, 7˚N

V: KY0006, 7˚N

60

500 depth (m)

cm/s

40

40 20 0 −20

1000

cm/s

depth (m)

500

−40 −60

1500

127˚E 127.6˚E

127˚E 127.6˚E

Fig. 4. KY0006 zonal sections of LADCP-measured velocity at 101N (top), 81N (middle) and 71N (bottom). Zonal velocity components are on the left, meridional on the right. Colors change at intervals of 0.1 m s1, with a black line on the zero contour. Light gray contour lines at 0.2 m s1 intervals for velocity components X0.6 m s1 in magnitude show the strong southward velocity in the Mindanao Current core. Triangles show station locations.

which it weakens and widens slightly down to 1200 m. From there to the deepest measurements at 2000 m, it varies little. Neither of the other two sections show an 800-m maximum; southward velocity components exceeding 0.3 m s1 are found below 1500 m at 81N, and at 71N there is a maximum near 1500 m with only a slight decrease in deeper speeds. The 71N section from KY9909

(Fig. 5) extends much farther offshore than any of the KY0006 sections. It shows southward flow exceeding 0.3 m s1 down to 1200 m at the station closest to the coast. Below 1300 m, the southward velocity maximum, again close to 0.3 m s1, extends to 2000 m one station to the east; this is still about 50 km west of the southward deep core on KY0006. Weaker northward flow, under 0.2 m s1, is found on KY9909 about 120 km east of the southward core. Although the LADCP sampling and accuracy are inadequate for reliable transport estimates, this offshore northward transport appears to be much smaller than the southward transport near the coast. Both cruises include diagonal sections running south-southeast from Mindanao. The cross-track component of velocity shows the MC heading toward the Celebes Sea (Fig. 6). On both sections, most of the transport is above 400 m, and the deeper flow toward the Celebes Sea is much weaker than the corresponding southward flow across the 71N sections (Figs. 4,5). Most of the flow across 71N below 400 m evidently turns eastward. This is not surprising, given that the entrance to the Celebes Sea is largely obstructed at 1000 m and blocked at 2000 m (Fig. 1).

4. Transport With the available measurements from these two cruises, intermediate-depth transports across sections are estimated best using the SADCP from KY0006. We choose the layer from 500 to 900 m, which is well below the upper-ocean circulation, including most of the MC, and in which the SADCP consistently provided measurements. Average temperature and sy ranges in this layer are approximately 5.2–7.5 1C and 26.95–27.3 kg m3, respectively. It excludes the North Pacific Intermediate Water core level (26.55 kg m3) found along the Mindanao coast by Bingham and Lukas (1994), but it includes the 27.2 kg m3 level of the AAIW core. The transport in the 500–900 m layer across 71N includes 8–9 Sv to the south near the coast and 7 Sv to the north offshore. Across the 81N section we also find about 8 Sv southbound near the coast

ARTICLE IN PRESS E. Firing et al. / Deep-Sea Research II 52 (2005) 605–613

611

U: KY9909, 7˚N 0 60 40 20 0 −20

1000

cm/s

depth (m)

500

−40 −60

1500

127˚E

128˚E

129˚E

130˚E

131˚E

V: KY9909, 7˚N 0 60 40 20 0 −20

1000

cm/s

depth (m)

500

−40 −60

1500

127˚E

128˚E

129˚E

130˚E

131˚E

Fig. 5. KY9909 zonal section of LADCP-measured velocity at 71N, contoured as in Fig. 4. Profiles from 127.33 to 129.25 stopped just beyond 1500 m.

and northbound offshore, with the latter occurring from about 128–1291E. Across the 101N section we find 5 Sv southbound near the coast and 8 Sv northbound offshore. We suspect this 3-Sv net northbound transport is real; it is nearly twice the 1.6 Sv transport error that would result from a heading bias of 0.21 over the 200-km track. To check our estimate of transport uncertainty, we can look at the net transport into each of the two triangles formed by the cruise track between 81N and 101N. We find a net transport of 2 Sv into the northern triangle and 3 Sv into the southern triangle, both of which are well within our error estimates of 3.4 and 4.2 Sv, respectively, given the track length for each triangle (520 and 425 km). This indicates that for each triangle the average

heading error was less than 0.21, and similar calculations from the remainder of the cruise track suggest a similarly high level of gyro compass accuracy for the cruise as a whole. To estimate the transport over a larger depth range, we simply note that the LADCP profiles suggest that the 500–900-m velocity structure is typical of the structure over the entire range from 500 to 2000 m on the 71N and 81N sections, and the strength of the eddy they cross is about 2 Sv per 100 m, so the transport in that eddy over the larger depth range is around 30 Sv. We have no way of knowing how deep the eddy extends beyond 2000 m, so we cannot estimate its total transport. Although the peak velocity in the 101N eddy is higher than that in the 7.51N eddy, it is also

ARTICLE IN PRESS E. Firing et al. / Deep-Sea Research II 52 (2005) 605–613

612

Across: KY9909, SE of Mind. 0 60 40 20 0 −20

1000

cm/s

depth (m)

500

−40 1500

−60

5˚N

6˚N

Across: KY0006, SE of Mind. 0 60 40 20 0 −20

1000

cm/s

depth (m)

500

−40 −60

1500

5˚N

6˚N

Fig. 6. Cross-track component of velocity on the diagonal section from Mindanao to 4.11N, 1281E. Negative numbers indicate flow to the west-southwest, towards the Celebes Sea.

shallower, so its total transport from 500 to 2000 m is probably smaller than that of the 7.51N eddy.

5. Discussion and conclusions The velocity measurements presented here graphically confirm the suggestion by Wijffels et al. (1995) that the Mindanao coast is a region of high velocity variance relative to the mean in the 400–2000-m depth range. The surveys indicate that the velocity field can be approximated as a set of eddies, each 200–300 km in diameter, with speeds from 0.2–0.6 m s1. We emphasize that this is a crude description, and we do not mean to

imply that these eddies necessarily are long-lived discrete entities, isolated from the surrounding flow. In the two surveys presented here, the dominant features are cyclonic, but the sampling is inadequate to conclude anything about the relative frequency of cyclonic versus anticyclonic eddies over the course of a year. Similarly, we do not know the temporal evolution of the subthermocline velocity field, the origin of the eddy energy, or how it is dissipated. The velocity field below about 300 m bears little resemblance to that above, except in the Mindanao Current itself. This suggests that the energy source of the eddies is not local wind forcing, or instability of local upper-ocean currents. The mean western boundary current along the Mindanao coast in the 400–2000 m depth range is unknown, and difficult to discern in the presence of such strong variability; it is unlikely to be the source of the eddy energy. We suspect that the deep eddies represent energy that has propagated along the coast and/or westward and downward. The velocity measurements have implications for the interpretation of geostrophic sections. First, it is clear that for synoptic sections, 2000 m is not a good reference level; velocities there may be 0.2 m s1 or more. Second, if the observed eddies are typical, then the standard deviation of deep velocity would be about 0.1 m s1, and the standard error of the mean of 10 sections would be 0.03 m s1. Therefore, it will be difficult to get an accurate estimate of the mean western boundary current system below the thermocline from synoptic sections in this region. As suggested by Wijffels et al. (1995), the variable flow near the western boundary might lead to a northward eddy flux of intermediate water properties from the southern hemisphere, contributing to the high-oxygen tongue along the boundary (Reid and Mantyla, 1978). The relative importance of eddy diffusion and advection by the mean flow remains an open question. The eddy activity undoubtedly causes zonal mixing of water properties as well, accounting for the breadth of the high-oxygen tongue (Qu et al., 1999). The magnitude of the eddy-diffusive flux depends on presently unknown characteristics of the velocity field, such as the degree to which eddies persist as

ARTICLE IN PRESS E. Firing et al. / Deep-Sea Research II 52 (2005) 605–613

closed circulations and whether they propagate along the coast.

Acknowledgments The Kaiyo TOCS cruises in October 1999 and September 2000 were joint projects of the Japan Agency for Marine-Earth Science and Technology (JAMSTEC), Japan, and Badan Pengkiajian Dan Penerapan Teknologi (BPPT), Indonesia. We thank Captains O. Yukawa and F. Saito and the crew of the R.V. Kaiyo, together with technicians from Marine Work Japan Co. Ltd. and Nihon Marine Enterprise Co. Ltd. Rahadian and Dr. Jayvee Udarbe assisted at sea. Darryl Symonds of RD Instruments helped ensure proper operation of the newly installed OS-38 ADCP. John Toole’s comments improved the manuscript. Hacker and Firing were supported by the State of Hawaii and the National Science Foundation grant OCE9730953. School of Ocean and Earth Science and Technology contribution 6476. International Pacific Research Center contribution 291. References Bingham, F.M., Lukas, R.B., 1994. The southward intrusion of North Pacific intermediate water along the Mindanao coast. Journal of Physical Oceanography 24, 141–154. Fine, R.A., Lukas, R.B., Bingham, F.M., Warner, M.J., Gammon, R.H., 1994. The western equatorial Pacific: a water mass crossroads. Journal of Geophysical Research 99, 25,063–25,080. Fischer, J., Visbeck, M., 1993. Deep velocity profiling with selfcontained ADCPs. Journal of Atmospheric and Oceanic Technology 10, 764–773.

613

Hu, D., Cui, M., Qu, T., Li, Y., 1991. A subsurface northward current off Mindanao identified by dynamic calculation. In: Takanao, K. (Ed.), Oceanography of Asian Marginal Seas. Oceanography, vol. 54. Elsevier, New York, pp. 359–365. Joyce, T.M., 1989. On in situ calibration of shipboard ADCPs. Journal of Atmospheric and Oceanic Technology 6, 169–172. Kashino, Y., Watanabe, H., Herunadi, B., Aoyama, M., Hartoyo, D., 1999. Current variability at the Pacific entrance of the Indonesian Throughflow. Journal of Geophysical Research 104, 11021–11036. King, B.A., Firing, E., Joyce, T.M., 2001. Shipboard observations during WOCE. In: Siedler, G., Church, J., Gould, J. (Eds.), Ocean Circulation and Climate. Academic Press, New York, pp. 99–122. Lukas, R.B., Firing, E., Hacker, P., Richardson, P.L., Collins, C.A., Fine, R.A., Gammon, R., 1991. Observations of the Mindanao Current during the Western Equatorial Pacific Ocean circulation study. Journal of Geophysical Research 96, 7089–7104. Pollard, R.T., Read, J., 1989. A method of calibrating shipmounted acoustic Doppler current profilers and the limitations of gyro compasses. Journal of Atmospheric and Oceanic Technology 6, 859–865. Qu, T., Mitsudera, H., Yamagata, T., 1998. On the western boundary currents in the Philippine Sea. Journal of Geophysical Research 103, 7537–7548. Qu, T., Mitsudera, H., Yamagata, T., 1999. A climatology of the circulation and water mass distribution near the Philippine coast. Journal of Physical Oceanography 29, 1488–1505. Reid, J.L., Mantyla, A.W., 1978. On the mid-depth circulation of the North Pacific Ocean. Journal of Physical Oceanography 6, 946–951. Smith, W.H.F., Sandwell, D.T., 1997. Global seafloor topography from satellite altimetry and ship depth soundings. Science 277, 1957–1962. Wijffels, S., Firing, E., Toole, J., 1995. The mean structure and variability of the Mindanao Current at 81N. Journal of Geophysical Research 100, 18421–18435.