This is a brief introduction to the basic physical oceanography of the Pacific island region - the circulation, tsunamis, waves, temperature and salinity distributions, and the forces that create these. Since the tropical Pacific contains most of the island groups, and since the dynamics and properties of the tropical oceans differ somewhat from those at higher latitudes, this chapter concerns only tropical oceanography for compactness. The tropical Pacific is usually considered to lie between the astronomically-defined tropics - the Tropic of Cancer (23°N) and the Tropic of Capricorn (23°S). There are other useful definitions of the tropics, based on how far the effect of the equator extends to the north and south in the atmosphere - approximately 20°. Within the ocean itself, the currents within 15 to 20° of the equator are oriented much more east-west than the currents at higher latitudes.

The tropics are a region of excess heating from the sun and towering cloud convection and rainfall in narrow bands across the Pacific. The warm surface waters of the tropics sustain the coral reefs which are a central feature of the islands' ecology. Compared with areas at higher latitudes, the frequency of storms and the average strength of the winds are low. As these affect the average height of surface waves, the wave climate of the tropics is relatively mild. Tsunamis (or "tidal waves") are a feature of life on a number of the Pacific islands which stand in the path of these fast-moving waves which are created by earthquakes or volcanoes.

2.a. Surface waves

The wave climate of the Pacific Islands region is dominated by long period swell reaching the area from distant storms, by relatively low amplitude, short period waves generated by more local winds, and the occasional bursts of energy associated with intense local storms.

Waves are characterized by their wavelength (distance between crests or troughs), their period (time between successive passage of a crest past a fixed point), and their height or amplitude. Each type of wave can also be characterized by its restoring force. For surface waves, the restoring force to perturbations in sea surface height is gravity, and so the waves are sometimes referred to as surface gravity waves.

Surface waves are mostly created by wind blowing across the sea surface. (The exceptions are the tides and tsunamis which are described in the following sections.) The first waves to appear in response to wind are very small "capillary" waves, with wavelengths on the order of centimeters. These are apparent in a lake when a gust of wind blows past. If the wind persists, longer and longer waves are generated. The wave heights build, proportionally to the strength of the wind and how long it blows. Local waves forced by the wind travel in the direction of the wind. The period of a wave is the time between the passing of successive crests. For wind-generated waves, periods are on the order of seconds to many minutes for the shortest to the longest waves, respectively.

A large storm generates numerous surface waves moving in all different directions under the storm. These travel away from the storm location, so if the storm is localized, waves will radiate outwards from the storm area. The longer the waves, the faster they travel. Short waves are damped out much more rapidly by friction than are long waves. Long waves generated by storms at high latitudes, such as in the Gulf of Alaska, or far south in the Antarctic, or generated by earthquakes can travel clear across the Pacific without much attenuation.

Typically the sea state (field of waves) is a jumble of waves of many different wavelengths, moving in many directions since the wind forcing can be in many different directions. In the tropical Pacific, the wave field can be thought of as a superposition of waves forced by the local trade winds - the "tradewind sea" - and waves forced by distant storms. The tradewind sea is of small amplitude, and choppy since it is produced locally by winds which shift. The long period swell from far away storms is also of relatively low amplitude in the open ocean, and much more unidirectional than the tradewind sea.

The height of waves is now measured by various satellite sensors. A measure commonly used is "significant wave height", which is the average height of the highest one-third of the waves, where the height is measured from trough to crest. NASA routinely produces maps of significant wave height from satellite altimetry information (Fig. 1). The altimeter measures the height of the sea surface, although the significant wave height is actually constructed from the properties of the radar pulse. Maps and information are available online, both for previous years and also in near real-time. Monthly analyses for the globe show that the average wave height in the tropical Pacific is typically less than 3 meters, regardless of season, whereas wave height at high latitudes in the winter hemisphere typically reaches 3 to 6 meters due to large storms (Fig. 1).

The water particles in a surface wave move in ellipses - up and forward in the direction of the wave propagation as the wave crest passes, and down and backwards as the trough passes. In deep water, waves with the longest wavelengths (distance from crest to crest) travel faster than short waves. When the wavelength becomes of the same size as the ocean bottom depth, the waves feel the bottom. The particle trajectories become more elliptical and the amplitude grows. The traveling speed of all waves becomes the same and proportional to the square root of the water depth - thus the waves travel more slowly and all together in shallower water.

As waves reach the shallow waters of a reef and island, they shoal, increase in amplitude and eventually break. The short period, tradewind sea produces relatively small surf height because of the short wavelengths. Large surf is produced by the long period swell from distant storms because of the correspondingly longer wavelength. The north shores of the Pacific islands receive this long-period swell in the northern hemisphere winter, and the south shores in the southern hemisphere winter. Wave heights of 6 meters in the surf zone are not uncommon. Winter swell on the north shore of Oahu occasionally reaches over 15 m (Flament et al., 1997).

Because the Pacific islands are small and rise steeply from the sea floor, there is little shelf area which can affect the progress of the long waves. (Continental shelves typically refract waves.) Thus the waves impinge directly on the shore or reef and do not wrap around the islands.

Breaking waves contain a lot of energy, some of which goes into production of local currents - first into longshore currents and then into rip currents which carry water back out to sea. Most of the circulation in the surf zone, and in lagoons inside reefs, is produced by breaking waves.

2.b. Tides

The complete tide is a composite of the moon (lunar) and sun (solar) tides. Considering just the moon, the gravitational between the earth and moon creates bulges of water on opposite sides of the earth. The water bulge nearest the moon/sun is due to domination of gravitational attraction over centrifugal force; the water bulge opposite the moon/sun is due to domination of centrifugal force. The two forces cancel at the earth's center. Since the earth also rotates daily, a point on the earth passes through these bulges twice a day, resulting in semi-diurnal (twice daily) components to the tide at each location. Because the moon and the sun do not generally lie over the equator, one of the bulges at a given point on the earth is larger than the other, leading to what is known as the "diurnal inequality", which lends a diurnal (daily) component to the tide.

A modulation of the tidal range results from the relative position of the moon and the sun: when the moon is new or full, the moon and the sun act together to produce larger "spring" tides; when the moon is in its first or last quarter, smaller "neap" tides occur. The cycle of spring to neap tides and back is half the 27-day period of the moon's revolution around the earth, and is known as the fortnightly cycle. The combination of diurnal, semi-diurnal and fortnightly cycles dominates variations in sea level throughout the islands.

The geometry of the oceans - the basin shape, local coastline, bays, and even harbor geometry - has a major effect on the local behavior of the tides. On scales of oceanic basins, tides exist as very long waves propagating in patterns determined by their period and the geometry of the basin. Figure 2 shows the response of the Pacific to the tidal period of 23 h 56 min, the largest diurnal (once daily) component. The tidal amplitude (Fig. 2a) is very low in the central Pacific, but is higher in the tropical region of Australia, New Guinea and Indonesia, as well as far to the north in the Gulf of Alaska and subpolar region. Lines along which high tide occurs at the same time (called phase lines - contours of constant color in Fig. 2b), converge to several points where the tidal range is zero. There are four of these points, called "amphidromes" in the Pacific: one on the North Pacific near the dateline, one near the equator in the eastern North Pacific, ond in the central South Pacific near Tahiti, and one east of New Zealand. Phase lines rotate counter-clockwise around the amphidromes in the North Pacific and clockwise around the ones in the South Pacific. For example, at the Hawaiian Islands, the offshore diurnal tide reaches the Hawai'i island first, then sweeps across Maui, O'ahu and finally Kauai.

Local bathymetry affects the ranges and phases of the tides along the shore, as the tidal waves wrap around the islands. For example, high tide at Haleiwa on the north shore of O'ahu occurs over an hour before high tide at Honolulu Harbor. Even though the tides at one point on a coastline are not in phase with those at even a nearby point, the tides at that point can be completely predicted if they are measured for several months, because the forcing which creates the tides is so extremely regular.

Tidal currents result from tidal variations of sea level, and near the shore are often stronger than the large scale circulation. Complete mapping of tidal currents requires direct measurements. As an example, the semidiurnal and diurnal tidal currents for Hawaii (Fig. 3 from Flament et al., 1997), show that the semi-diurnal and diurnal tidal currents tend to be aligned with the shoreline. Due to high variability of tidal currents around the islands, however, this statistical picture may not correspond to the flow at a particular time: tidal currents cannot be predicted as precisely as sea level. Strong swirls often result from tidal currents flowing around points and headlands, and present hazards to divers.

2.c. Tsunamis

When the seafloor is raised suddenly during a shallow earthquake, water is raised with it, producing a mound of excess water at the sea surface. Gravity collapses the mound, producing a series of waves: a tsunami. Tsunamis are gravity waves, just like those generated by the wind, but their period is much longer, on the order of 10 to 60 minutes. While earthquakes are the most common cause of tsunamis, the waves are generated by any phenomenon which rapidly changes the shape of the sea surface over a large area: volcanic eruption, landslide, even meteorite impact. Since the largest shallow earthquakes occur in the subduction zones which ring the Pacific, and since these same subduction zones are dotted with volcanoes, the tsunami hazard throughout the tropical Pacific is high.

On the open ocean, the wavelength of a tsunami may be as much as two hundred kilometers, many times greater than the ocean depth which is on the order of several kilometers. This huge wavelength means that the entire water column, from surface to bottom, is set into motion. Tsunamis therefore always behave like waves in shallow water, which means, as already discussed above, their speed is proportional to the square root of the water depth. For typical ocean depths of 5 km, a tsunami will advance at 800 km/hr, about the speed of jet aircraft. A tsunami can therefore travel from one side of the Pacific to the other in less than a day. The speed decreases rapidly as the water shoals: in 15 m of water the speed of a tsunami (or of any wave with long enough wavelength to "feel" the ocean bottom) will be only 45 km/hr.

As the tsunami slows in shoaling water its wavelength is shortened. Just as with ordinary surf, the energy of the waves must be contained in a smaller volume of water, so the waves grow in height. The maximum height the tsunami reaches on shore is called the runup. Any runup over a meter is dangerous. Waves reaching only a meter above sea level may not seem threatening, but the waves of a tsunami are unlike normal waves. Even though the wavelength has shortened, a tsunami will typically have a wavelength in excess of ten kilometers when it comes ashore. Each wave therefore floods the land (Fig. 4) as a rapidly rising tide (hence the common English term "tidal wave") lasting for several minutes. The individual waves are typically from ten minutes to a half-hour apart, so the danger period can last for hours.

Runup can vary dramatically depending on seafloor topography. Small islands with steep slopes experience little runup; wave heights there are only slightly greater than on the open ocean. For this reason the smaller Polynesian islands with steep-sided fringing or barrier reefs are only at moderate hazard from tsunamis. Such is not the case for the Hawaiian Islands or the Marquesas, however. Both of these island chains are almost devoid of barrier reefs and have broad bays exposed to the open ocean. Hilo Bay at the island of Hawaii and Tahauku Bay at Hiva Oa are especially vulnerable. During a tsunami from the Eastern Aleutians in 1946, runup exceeded 8 m at Hilo and 10 m at Tahauku; 59 people were killed in Hilo, two in Tahauku (Shepard, et al., 1950; Talandier, 1993). Similarly, any gap in a reef puts the adjacent shoreline at risk. The tsunami from the Suva earthquake of 1953 did little damage because of Fiji's extensive offshore reefs. Two villages on Viti Levu located opposite gaps in the reef, however, were extensively damaged and five people were drowned (Singh, 1991).

Tsunamis are generated by shallow earthquakes all around the Pacific, but those from earthquakes in the tropical Pacific tend to be modest in size. While such tsunamis may be devastating locally, they decay rapidly with distance and are usually not observed more than a few hundred kilometers from their sources. That is not the case with tsunamis generated by great earthquakes in the North Pacific or along the Pacific coast of South America. About half-a-dozen times a century a tsunami from one of these locations sweeps across the entire Pacific, is reflected from distant shores, and sets the entire ocean oscillating for days. The tsunami from the magnitude 9.5 Chile earthquake of 1960 (Fig. 5) caused death and destruction throughout the Pacific: Hawaii, Samoa, and Easter Island all recorded runups exceeding 4 m; 61 people were killed in Hawaii. In Japan 200 people died. A similar tsunami in 1868 from northern Chile caused extensive damage in the Austral Islands, Hawaii, Samoa, and New Zealand. There were several deaths in the Chatham Islands (Iida, et al., 1967).

3.a. Temperature.

Ocean surface temperature globally is dominated by excess heating in the tropics compared with higher latitudes, resulting mainly from higher radiation from the sun in the tropics. This leads to a sea surface temperature difference from equator to pole of about 30°C (Fig. 6). In the tropics, including the tropical Pacific, the maximum sea surface temperature is around 28°C and can rise to at most 30°C. This is considerably cooler than the maximum temperatures regularly found over land, of about 50°C. It is currently hypothesized that the main regulation on the maximum ocean temperature is through cloud formation. Cloud formation increases dramatically when the sea surface temperature is greater than about 27.5°C (Graham and Barnett, 1987). The increased cloudiness increases the albedo (reflectivity of the earth/atmosphere to space), which reduces the solar radiation reaching the sea surface (Ramanathan and Collins, 1993), and thus keeps the surface temperature from rising much more.

The sea surface temperature is not uniformly high in the tropical Pacific. A large "warm pool" is found in the central and western Pacific (Figs. 6 and 8), and also extends into the eastern Indian Ocean. Surface water in the eastern equatorial Pacific is several degrees cooler than in the west. The vertical thermal structure of the upper ocean is responsible for these differences. In the western Pacific, the surface layer, which is fairly well mixed, is approximately 100 meters thick (Fig. 8), and warmer than about 28°C. Just below this surface layer, the temperature changes rapidly downward (for instance at 100-150 meters depth in the west in Fig. 8); this is called the "thermocline". In the central and eastern Pacific, the surface layer is shallower, and so colder water and the thermocline are found closer to the surface. Upwelling in the eastern Pacific draws this cooler water to the surface, creating the equatorial "cold tongue" at the sea surface (Fig. 6). Upwelling of cold water at the equator is apparent in sections crossing the equator (Fig. 9b from Wyrtki and Kilonsky, 1984). Upwelling in the western Pacific is somewhat weaker than in the east and draws up only warm water, and so an equivalent cold tongue along the equator is absent.

Upwelling is common along the west coast of South America, off Ecuador and Peru, and along the west coast of Central and North America. As a result of both the upwelling and the eastern boundary currents which flow towards the equator in these regions, sea surface temperatures are relatively low along these coasts. The winds which create upwelling are strongest in the area just west of Costa Rica. Here the thermocline is lifted to within 10 meters of the sea surface, and is called the Costa Rican Dome (Hofmann et al., 1981).

Below the sea surface, temperature decreases to the ocean bottom (Figs. 8, 9 and 10). The most rapid change is in the upper 500 meters, in the thermocline. Changes are more gradual below this. Temperature reaches about 1.2°C in the abyssal tropical Pacific. The initial temperature and salinity of all ocean water is set at the sea surface. The sea surface temperature distribution (Fig. 6) shows that water colder than about 18°C comes from latitudes higher than about 30°, hence outside the tropics. Waters of about 4-6°C come from latitudes of about 40-45° (northern and southern hemisphere). The coldest waters flow northward from the Antarctic region. These southern hemisphere waters, which fill the Pacific below 1000 to 1500 meters, are part of a circulation which extends through all of the oceans. The deepest waters come from the the Weddell and Ross Seas of the Antarctic and the Greenland Sea just north of the North Atlantic. The North Pacific does not produce any of this deep water, and so its deep waters have traveled a long distance from their sea surface origin. These deep waters have spent about 500 years making the journey to the deep North Pacific (and slightly less time to the deep tropical Pacific). Waters which have been far from sea surface forcing (heating/cooling and evaporation/precipitation) for a long time are fairly uniform because they mix with each other. Thus the deep Pacific contains a large amount of water in a very narrow range of temperature and salinity, centered around 1.2°C and 34.70 in practical salinity units (e.g. Worthington, 1981). (Salinity is defined in the next paragraph.) This water must upwell slowly and eventually complete the overturning cycle by reaching the sea surface, perhaps very far from the deep North Pacific.

3.b. Salinity.

Sea water density depends on temperature (warm water is less dense), and also on the amount of material dissolved in the water. The latter is mostly what is referred to as "sea salt", and is a combination of various salts. The total amount of salt in the world ocean is constant on all but the longest geological timescales. However the total amount of fresh water in the ocean is not constant - it is affected by evaporation, precipitation and runoff. Hence salinity, which is more or less the grams of salts dissolved in a kilogram of seawater, varies as a result of surface freshwater inputs and exports. (Precise salinity definitions and measurement methods are described in introductory physical oceanography textbooks such as Tomczak and Godfrey, 1994 or Pickard and Emery, 1990.)

The total range of salinity in most areas of the ocean is small enough that temperature actually contributes more to sea water density differences, but salinity differences are significant and important. For instance, if saltier water lies above fresher water, then the temperature difference between the two must be large enough to ensure stability (light water over dense water).

Surface salinity in the Pacific (Fig. 7) shows clearly the net result of the atmospheric circulation described in the Climate Chapter. Cloud formation and high precipitation occur in regions of rising, humid air, which are associated with low atmospheric pressure at the sea surface, such as in the Intertropical Convergence Zone (ITCZ) at 5-10°N and subpolar regions poleward of 40° . Surface salinity is low where precipitation is high. Evaporation and hence surface salinity are high where the air is dry - regions of atmospheric divergence (high pressure zones at the surface).

Because temperature dominates the vertical density differences in the ocean, it decreases downward almost everywhere. Thus although salinity also contributes to density, the salinity distribution can be more complex, with regions of salty water lying over fresher water and vice versa (Figs. 8b and 11). Such salinity inversions are common. In cross section from south to north, the high salinity in the surface evaporation cells extends down to the thermocline. The fresher water associated with the ITCZ extends fairly deep. Below the high salinity surface water is found a layer of low salinity "intermediate water" which extends from the rainy subpolar latitudes in the south and north towards the equator. Below this, the deep Pacific is filled with relatively more saline waters originating from the deep waters around Antarctica and from the Atlantic.

Along the equator surface salinity is lowest in the western Pacific, where normally there is much more rainfall than in the central and eastern equatorial Pacific (Figs. 7 and 8b). The freshest surface water in the western equatorial Pacific actually extends only partway down into the vertically-uniform, warm surface layer with salinity increasing strongly downward midway within this uniform temperature layer. Hence the surface stratification is dominated by salinity rather than temperature (Lukas and Lindstrom, 1991). A relatively sharp north-south front separates the fresh western equatorial surface water from the more saline central Pacific surface water. During periods such as El Nino when the trade winds slacken (section 6 below), the western fresh, warm water floods eastward towards the central Pacific along the equator (Roemmich et al., 1994).

4.a. Surface circulation.

The Pacific sea surface circulation (Fig. 13) consists of two large "subtropical gyres" centered at 30°N and 30°S, which rotate clockwise in the northern hemisphere and counterclockwise in the southern hemisphere, a "subpolar gyre" centered at about 50°N and rotating counterclockwise, a major eastward flow which circles Antarctica called the "Antarctic Circumpolar Current", and complicated but predominantly zonal (east-west) currents in the tropics between about 15°N and 15°S. At the sea surface, flow is westward from 30°S up across the equator to about 5°N. This westward flow is all called the "South Equatorial Current". Between 5°N and 10°N lies a strong eastward flow, termed the North Equatorial Countercurrent. It is associated with and driven by the winds of the ITCZ. The westward flow between 10N and 30N is called the North Equatorial Current (NEC). The northern half of the NEC is actually part of the subtropical gyre and the southern half is part of the ITCZ's elongated counterclockwise flow. Sometimes a weak ITCZ (South Pacific Convergence Zone) is also present in the southern hemisphere, creating an occasional appearance of a South Equatorial Countercurrent analogous to the North Equatorial Countercurrent.

In the western tropical Pacific, the circulation is dominated by strong currents which abut the western boundary (Fig. 14, from Fine et al., 1994). Western boundary currents are a central feature of all circulation patterns worldwide. In the tropical and South Pacific, the western boundary currents are complicated by the many islands and deep ridges. Australia forms the largest single part of the boundary. In the North Pacific, the westward-flowing North Equatorial Current reaches the western boundary at Mindanao in the Philippines. It splits into a northward flow, called the Kuroshio, and a southward flow, called the Mindanao Current. The Kuroshio flows into the East China Sea and then northward to the southern end of Japan (Kyushu) where it splits into a major flow eastward along the eastern coast of Japan, and a weaker flow, called the Tsushima Current, into the Japan East Sea. The Kuroshio is one of the strongest currents in the world, similar to the Gulf Stream and the Antarctic Circumpolar Current in strength. It affects climate in Japan through its warmth and fisheries off Japan through both its warmth and relative lack of nutrients. The Mindanao Current flows southward along Mindanao and separates to flow eastward into the North Equatorial Countercurrent at about 5°N. A portion turns westward at the southern end of Mindanao and enters the Celebes Sea.

"Eddies" (circulations of about 50 to 200 km size which are often variable over a period of weeks to months) are usually found east of Mindanao and east of Halmahera. The water entering the Celebes Sea forms the beginning of flow westward through the complex of Indonesian islands, threading through to Java and thence into the Indian Ocean.

In the South Pacific, the very broad, westward-flowing South Equatorial Current reaches the western boundary through a complex of islands. The northern portion forms a northward-flowing western boundary current along New Guinea, called the New Guinea Coastal Current (Lindstrom et al., 1987; Tsuchiya et al., 1989). This flows northward to the equator. A portion of it turns eastward along the equator and apparently forms part of the eastward-flowing subsurface Equatorial Undercurrent. A portion may continue slightly northward, joined by the westward flow just north of the equator, and then turns eastward, joining the separated Mindanao Current, into the North Equatorial Countercurrent.

The remainder of the westward-flowing South Equatorial Current flows north of Fiji into the Coral Sea and reaches the western boundary at Australia. Here it turns southward into the East Australian Current, which is the western boundary current, and then flows southward to the northern tip of New Zealand. At this point, the current meanders a great deal and some portion of it separates and flows eastward just north of New Zealand as the North Cape Current. The broad flow between New Zealand and Fiji is also eastward.

The large-scale surface flow is affected only by the larger land masses, and not much by the small islands dotting the tropical and South Pacific. Intermediate and abyssal flow however are strongly affected by the ridges in which the small islands are embedded, as described next.

4.b. Subsurface equatorial circulation

In most places of the world ocean, the currents vary only gradually from surface to bottom - they are usually strongest at the surface where they are closest to the wind forcing, and gradually blend into the circulation of the abyss. However, within 2 or 3° latitude of the equator, the subsurface currents are much more complicated (Fig. 9a from Wyrtki and Kilonsky, 1984). Between 100 and 200 meters depth lies the strong eastward-flowing Equatorial Undercurrent. The undercurrent was originally discovered by Townsend Cromwell during a research expedition in the 1950's when the drogues deployed at that depth moved strongly eastward while the surface current was westward (see Knauss, 1960). In speed, the Equatorial Undercurrent matches the strongest currents in the world (> 100 cm/sec or 1 km/day). However, the undercurrent is vertically very thin (about 100 meters thick) in contrast with the other major currents such as the Kuroshio, Gulf Stream, and Antarctic Circumpolar Current which reach to the ocean bottom.

Below the undercurrent and flanking it on either side of the equator lie the North and South Subsurface Countercurrents, flowing eastward (at 2° on either side of the equator and below 150 meters depth in Fig. 9a). These were discovered by Tsuchiya (1968). Directly beneath the Equatorial Undercurrent lies a somewhat weaker westward flow, which extends to about 1000 meters depth. Below this there is a regime of the so-called "stacked jets", extending to the ocean bottom, but with vertical extent increasing towards the bottom (Firing, 1989). Farther away from the equator, between 2° and 5° latitude, the vertical structure may show only a reversal or two. Farther away from the equator than this, the vertical structure is much simpler, with the surface circulation extending to depths of 1000 to 2500 meters, and much weaker flow dominated by bottom topography below this.

The most general characteristic of circulation in the tropical Pacific is the exaggerated east-west nature compared with flow poleward of 20° latitude in both hemispheres, where "gyres" which also include more north-south flow are the norm. This zonality is characteristic of the tropical circulation in the Atlantic and Indian Oceans as well as the Pacific.

4.c. Deep circulation

With increasing depth, the surface circulation weakens and shifts latitude. In the tropics, the surface circulation signatures disappear by about 500 to 1000 meters depth. Flow beneath this is predominantly zonal (east-west) with very slight north-south movement. Various analyses show counterclockwise circulation north of the equator and clockwise circulation south of the equator, in very elongated cells between the equator and about 10° latitude. (See Reid, 1997 for an analysis of the whole of the deep circulation.) The deepest circulation is affected by the topography of the ridges and basins. Overall, there is net northward flow in a deep western boundary current, which enters the Pacific from the Antarctic east of New Zealand and passes through a deep gap near Samoa, called the Samoan Passage. It moves on northward to the equator, crossing in the western Pacific. North of the equator, a portion branches eastward to pass south of the Hawaiian Islands, and the other portion continues northward. The northward flow appears to move westward under the Kuroshio and then northward along the western boundary to the subpolar Pacific. Return flow to the south probably occurs along the East Pacific Rise in the eastern Pacific and then westward along the equator (Johnson and Toole, 1993; Firing, 1989).

4.d. Circulation near islands and island groups

The upper ocean circulation in the tropical Pacific is driven mostly by the stress from the wind. The prevailing winds in the tropical Pacific are the trades or easterlies, which blow from east to west. Together with the westerlies of higher latitudes, these force the large subtropical gyres (section 4a). The dominant influence of these gyres on the tropics is the broad-scale westward flow mentioned above, called the North Equatorial Current (north of 5°N) and the South Equatorial Current (from the equator southward). We divide the wind forcing of the tropics into two regimes - off the equator and on the equator. The difference between these is the importance of the earth's rotation to the forcing - off the equator it is very important and on the equator we can disregard it.

5.a. Wind forcing of non-equatorial flow

The mechanism for forcing the large-scale circulation by the surface wind stress is indirect, and described well in introductory texts on physical oceanography. The large scale circulation is in "geostrophic balance". This means that the currents are driven by horizontal pressure differences which are balanced by the Coriolis force, which comes from the earth's rotation. The resulting flow is exactly at right angles to the pressure difference force- in the northern hemisphere it is to the right (so flow circulates around high pressure in a clockwise direction) and in the southern hemisphere it is to the left (counterclockwise flow around a high). Near the sea surface, the pressure difference is due to small, but large-scale and long-lasting, differences in sea surface height. Over about 100 kilometers horizontal distance, which is the width of a major current such as the Kuroshio, the sea surface height difference which creates the pressure difference which drives the current is no more than 1 meter. This is of course shorter than most surface waves in mid-ocean. The distinction between the surface height difference which drives a major current and that of just a surface wave is that the wave is just passing by - it changes the sea surface height over a very small time, whereas the surface height differences which drive currents must be in place for at least several days in order to "feel" the rotation of the earth. The largest height changes drive the fastest currents, such as the Kuroshio in the Pacific and the Gulf Stream in the Atlantic. Where these flows are most vigorous, they can extend to great depth and even to the ocean bottom.

How do the winds drive this flow? The winds push on the very top of the ocean, and move the water through frictional stress. This frictional layer is referred to as the "Ekman layer" and is a total of about 20 to 100 meters deep. (How the stress is actually exerted by the wind on the ocean involves surface waves, but we will not be explicit about how.) The resulting movement of the water is to the right of the wind in the northern hemisphere and to the left in the southern hemisphere. This very thin water layer (say, order of 1 meter thick) then pushes on a thin water layer below it through friction, and so on. Each of the thin water layers pushes the one below it slightly more to the right (northern hemisphere). The frictional stress becomes smaller and smaller with depth as the energy is put into moving the water. In fact the frictional stress dies out at about 50 meters below the sea surface. Thus the winds frictionally drive only the very top of the ocean. The overall effect of the wind on this 50-meter layer is to drive a net flow of water exactly to the right of the wind (northern hemisphere). This is called "Ekman transport". It adds on to the geostrophic surface flow, which, as said above, is driven by a pressure difference. Using surface drifters which report their positions via satellite, Ralph and Niiler (1997) have mapped the average flow at 15 meters depth in the Pacific. When they subtract the fairly well-known large-scale geostrophic flow from their average flow, the resulting flow is indeed to the right of the wind in the northern hemisphere and to the left in the southern hemisphere, which substantiates the idea of Ekman transport on a large spatial scale (Fig. 15 from E. Ralph, and based on Ralph and Niiler, 1997).

Winds are highly variable in general - weather patterns come and go in a matter of days. However, winds averaged over a season or a year or many years drive the large-scale, slowly-changing ocean circulation. Because the average winds vary in strength and also in direction over a large scale, the surface layer Ekman flow varies in strength and direction. Where the surface flow converges (flows together), there must be downwelling, and where it diverges (flows apart), there must be upwelling.

In regions of downwelling, the ocean lying beneath the surface layer, and down to about 2000 meters depth, responds with slow equatorward flow. The reason for this equatorward flow is more or less angular momentum conservation - as a vertical column of water that rotates due to the earth's rotation is squashed by downwelling, it must spin more slowly. To spin more slowly it moves towards the equator where the amount of earth's rotation which projects onto the vertical column is lower. Such slow equatorward flow is found in the subtropics (20° to 40° from the equator) where there are westerlies at higher latitude and easterly trades at lower latitude. This slow equatorward flow is fed from the western boundary, by eastward flow at latitudes of 30° to 50°. The equatorward flow returns to the western boundary at latitudes of about 15° to 30° in the northern hemisphere and from 30°S to the equator in the southern hemisphere. The western boundary current which feeds this circulation in the northern hemisphere is the Kuroshio. In the southern hemisphere it is the East Australia Current.

In regions of upwelling, the underlying ocean flow is poleward, away from the equator. This occurs at high latitudes (greater than 50°) and also in the narrow band under the Intertropical Convergence Zone at about 5° to 10°N. The result is a counterclockwise circulation. In the northern North Pacific, the western boundary current which feeds this gyre is the Oyashio. In the tropics north of the equator, the currents are nearly due east-west but they do have a slight counterclockwise gyre configuration and a western boundary current. The currents in this tropical gyre are the North Equatorial Countercurrent, which flows eastward on the southern side of this cell, and the southern part of the North Equatorial Current, which flows westward on the northern side of this cell. Its western boundary current is the Mindanao Current.

5.b. Wind-driven circulation at the equator.

Directly on the equator, the effect of rotation on the circulation vanishes, and so these concepts of geostrophic and Ekman flow do not apply. At the equator, the easterly trade winds push the surface water directly from east to west. This water piles up gently in the western Pacific (0.5 meters higher there than in the eastern Pacific). Because it is higher in the west than in the east, there is a pressure difference which causes the flow just beneath the surface layer to be eastward. This strong eastward flow is the Equatorial Undercurrent.

The alternating eastward and westward jets found below the Equatorial Undercurrent on the equator die out about 2° from the equator. Their cause has not been clearly identified. However the theory of very slow waves on the equator, which move water from side to side much more than the up and down of surface waves, shows us that equatorial waves have much more complicated (reversing) vertical structure than waves off the equator. It is expected that this complex structure for the very slow waves with very long east-west wavelengths translates to complex structure in the mean currents.

Also occurring very close to the equator is northward Ekman transport north of the equator and southward Ekman transport south of the equator, due to the easterly trade winds (blowing from east to west). This causes upwelling right at the equator.

Along the equator just below the surface, waters in the east are colder than in the west. This is partially a result of the rising of the Equatorial Undercurrent from west to east in response to upwelling. Upwelling in the eastern Pacific thus accesses much cooler water than in the western Pacific, and as a result the surface waters in the east are colder than in the west. Steady trade winds, which cause equatorial upwelling, are more prevalent in the east than in the west. There is seasonality in the winds, and equatorial upwelling is weaker in the northern winter and spring, giving rise to mini-El Nino conditions (section 6) each year in the eastern equatorial Pacific.

5.c. Response to changing winds in the tropics

When the trade winds weaken or even reverse, the flow of water westward at the equator weakens or reverses and upwelling weakens or stops. Surface waters in the eastern Pacific warm significantly since upwelling is no longer bringing the cool waters to the surface. The deep warm pool in the western Pacific thins as its water sloshes eastward along the equator in the absence of the trade winds which maintain it.

5.d. Heating/cooling and evaporation/precipitation

Ocean water density is a function of temperature and salinity, and so can be changed through heating/cooling and evaporation/ precipitation. The resulting density changes can drive circulation, but density-driven flow is much weaker than that driven by the winds. However, density changes, caused mainly by fluxes at the sea surface, are the only means of forcing circulation where the indirect effect of wind forcing vanishes, as in the ocean deeper than about 2000 meters. In the upper ocean, even though density fluxes do not greatly change the flow, they do have a major effect on ocean properties and on the overlying atmosphere, which is heated from below by the ocean. The total surface heat flux into the ocean averaged over all years of data (Fig. 16 from Hsiung, 1985) shows the greatest heating along the equator and in the western warm pool region around Indonesia. The units of heating are Watts/m^2, or energy per unit area. The uncertainty in heating is about 20 Watts/m^2 and so values lower than this are not significantly different from zero. In the subtropics where the western boundary currents bring warm water to mid-latitudes, there is strong cooling. In order to maintain a fairly steady distribution of temperature, the ocean must transport heat from the areas where it gains heat to the areas where it loses it. The large arrows in Fig. 16 show the direction of heat transport in each ocean basin across the latitudes where the arrows are placed.

Large changes in the climate occur in the tropical Pacific over the course of three to seven years, as described in the chapter on climate. This phenomenon is known as El Nino and encompasses the entire tropical Pacific ocean and atmosphere. It is a truly coupled interaction between the atmosphere and ocean. Its effects reach far beyond the tropical Pacific through connections in the far-ranging atmospheric circulation.

The events that form a typical El Nino have been described by Rasmussen and Carpenter (1982), and are illustrated in Fig. 17, from the NOAA/PMEL TAO project office (McPhaden, 1997). Philander (1990) provides a textbook summary of El Nino. A Report to the Nation (NOAA OGP, 1994) provides an excellent summary as do a number of El Nino websites (see McPhaden, 1997 in the reference list).

During an El Nino, the normal easterly trade winds slacken, as indicated by a decrease in the atmospheric pressure difference between the central and western Pacific. (The pressure difference between Tahiti and Darwin, Australia, called the Southern Oscillation Index, is often taken as a measure of El Nino, hence the commonly-used name El Nino/Southern Oscillation.) The weakened trade winds result in reduced westward flow at the equator which leads to a draining of the western warm pool towards the east. Equatorial ocean upwelling is reduced, which results in warmer sea surface temperatures in the eastern Pacific. As the western warm pool cools slightly and the central and eastern equatorial Pacific warm, this further reduces the strength of the tradewinds, in other words, providing a positive feedback. The large atmospheric convection cell over Indonesia moves eastward. This results in drought in the western Pacific, including over Indonesia and Australia, and increased rainfall in the central and eastern Pacific, for instance at Christmas Island, the Galapagos and Ecuador.

The warm water in the eastern Pacific spreads to the eastern boundary and splits to flow north and south there. Upwelling off northern Peru might weaken, or just draw on the warm, nutrient-poor equatorial water. The result is a decline in production in this important fisheries area. If the El Nino is particularly strong, its effect in the ocean can reach as far north as the California coast.

The opposite phase of the El Nino is called La Nina, characterized by especially strong tradewinds, a well-developed warm pool in the western Pacific and cold water at the equator in the central and eastern Pacific, with strong rainfall in the western Pacific, little rainfall in the eastern Pacific, and major fisheries production in the eastern boundary regions.

El Nino affects mid-latitudes through "teleconnections" in the atmosphere. Changes in the western tropical Pacific reach far to the northeast and southeast through the atmosphere and directly affect climate in the coastal regions of the United States and South America.

El Nino occurs irregularly, but generally every three to seven years. Major progress has been made in predicting an El Nino about one year in advance because the sequence of events in an El Nino is often the same. Thus detection of early signs of El Nino, such as the appearance of warm water in the eastern tropical Pacific or a change in the strength of the trade winds, often allows prediction of changes in rainfall and air temperature later in the year throughout the Pacific region. A major observing network and computer modeling effort is in place to assist in observing and forecasting El Nino occurrences (see the website in the reference list under McPhaden, 1997).

The strength of El Nino varies greatly over an even more irregular time scale of about ten to thirty years. For instance El Nino's in the 1940's were strong, followed by several decades of weak events, and then followed by very strong El Nino's again in the 1980's and 1990's. Long records of El Nino's have been extracted from the reasonably long pressure records at Tahiti and Darwin, and from growth and properties of the annual accretion in coral heads in the tropical Pacific. This so-called decadal modulation of El Nino is much less well-understood than El Nino itself. One way that the decadal pattern differs from El Nino is that its amplitude is about the same in the northern and southern Pacific as in the tropics, as opposed to the much more equatorially-trapped El Nino changes (Zhang et al., 1997). This suggests that there might be feedbacks which involve a much larger region than just the tropics, although it could still be a tropically- dominated mode. Major research on these ocean-atmosphere feedbacks, which affect much longer timescales, is being planned for the next decade.

Conkright, M. E., S. Levitus and T. P. Boyer. 1994. World Ocean atlas. Vol. 1, Nutrients, NOAA Atlas NESDIS, 162 pp. [Available from NODC/NOAA, E/OC5, 1315 East-West Highway, Silver Spring, MD 20910.]

Fine, R.A., Lukas, R., Bingham, F.M., Warner, M.J. and R.H. Gammon. 1994. The Western Equatorial Pacific: A Water Mass Crossroads. J. Geophys. Res., 99:25063-25080.

Firing, E. 1989. Mean Zonal Currents Below 1500 m Near the Equator, 159W. J. Geophys. Res., 94:2023-2028.

Flament, P., Kennan, S. , Lumpkin, R. , Sawyer, M. and E.D. Stroup, 1997. "The Ocean Atlas of Hawa ii", 24"x36" poster and copyrighted website: http:/satftp.soest.hawaii.edu/atlas. University of Hawaii.

Graham, N.E. and T.P. Barnett. 1987. Sea Surface Temperature, Surface Wind Divergence, and Convection over Tropical Oceans. Science, 238:657-659.

Hofmann, E. E., Busalacchi, A.J. and J. J. O'Brien. 1981. Wind generation of the Costa Rica Dome. Science, 214:552-554.

Hsiung, J. 1985. Estimates of Global Oceanic Meridional Heat Transport. J. Phys. Oceanogr., 15:1405-1413.

Iida, K., D.C. Cox, and G. Pararas-Carayannis, 1967, Preliminary catalog of tsunamis occurring in the Pacific Ocean, Hawaii Institute of Geophysics Pub. HIG-67-10, University of Hawaii, Honolulu, 274pp.

Johnson, G.C. and J.M. Toole. 1993. Flow of Deep and Bottom Waters in the Pacific at 10°N. Deep-Sea Research, 40:371-394.

Knauss, J.A. 1960. Measurements of the Cromwell Current. Deep-Sea Research, 6:265-286.

Knauss, J. A., 1997. Introduction to Physical Oceanography. Second Edition. Prentice-Hall, New Jersey. 309 pp.

Lander, J.F., and P.A. Lockridge, 1989, United States tsunamis (including U.S. possessions) 1690-1988, Publication 41-2, National Geophysical Data Center, Boulder, CO., 265pp.

Levitus, S. and T. P. Boyer. 1994a. World Ocean atlas. Vol. 3, Salinity, NOAA Atlas NESDIS, 111 pp. [Available from NODC/NOAA, E/OC5, 1315 East-West Highway, Silver Spring, MD 20910.]

Levitus, S. and T. P. Boyer. 1994b. World Ocean atlas. Vol. 4, Temperature, NOAA Atlas NESDIS, 129 pp. [Available from NODC/NOAA, E/OC5, 1315 East-West Highway, Silver Spring, MD 20910.]

Levitus, S., Conkright, M.E., Reid, J.L. , Najjar, R.G. and A. Mantyla. 1993. Distribution of Nitrate, Phosphate and Silicate in the World Oceans. Prog. Oceanogr., 31:245-273.

Lindstrom, E., Lukas, R., Fine, R., Firing, E., Godfrey, S., Meyers, G. and M. Tsuchiya. 1987. The Western Equatorial Pacific Ocean Circulation Study. Nature, 330:533-537.

Lukas, R. and E. Lindstrom. 1991. The Mixed Layer of the Western Equatorial Pacific Ocean. J. Geophys. Res., 96:3343-3357.

McPhaden, M. J. 1997. NOAA/PMEL TAO Project Office: El Nino theme page. Website: http://www.pmel.noaa.gov/toga-tao/el-nino.

National Oceanographic and Atmospheric Administration (NOAA) Office of Global Programs. 1994. El Nino and Climate Prediction. Reports to the Nation on our Changing Planet. 25 pp.

Neshyba, S., 1987. Oceanography: Perspectives on a Fluid Earth. John Wiley and Sons, New York. 506 p.

Philander, S.G. 1990. El Nino, La Nina, and the Southern Oscillation. Academic Press, Inc., San Diego, California. 293 pp.

Pickard, G.L. and W.J. Emery. 1990. Descriptive Physical Oceanography: An Introduction. Pergamon Press, Oxford, England. 320 pp.

Price, J.M., Vanwoert, M.L. and M. Vitousek. 1994. On the Possibility of a Ridge Current Along the Hawaiian Islands. J. of Geophys. Res., 99:14101-14111.

Ralph, E. A. and P. P. Niiler, 1997. Wind-driven currents in the tropical Pacific. J. Phys. Oceanogr., in press.

Ramanathan, V. and W. Collins. 1993. A Thermostat in the Tropics. Nature, 36, 11:410-411.

Rasmusson, E.M. and T.H. Carpenter. 1982. Variations in Tropical Sea Surface Temperature and Surface Wind Fields Associated With the Southern Oscillation/El Nino. Mon. Weather Rev., 110:354-384.

Reid, J.L. 1997. On the Total Geostrophic Circulation of the Pacific Ocean: Flow Patterns, Tracers and Transports. Submitted Prog. in Oceanogr.

Reverdin, G., Morliere, A. and E. Gerard. 1991. Compagne Oceanographique Trans-Pacifique. Paris, France. 341 pp.

Roden, G.I. 1991. Effects of the Hawaiian Ridge Upon Oceanic Flow and Thermohaline Structure. Deep-Sea Res., 38:623-654.

Roemmich, D., Morris, M.Y. and W.R. Young. 1994. Equatorial Fresh Jets. J. Phys. Oceanogr., 24:540-558.

Shepard, F.P., G.A. Macdonald, and D.C. Cox. 1950. The tsunami of April 1, 1946, Bulletin of the Scripps Institution of Oceanography of the University of California, 5:391-470.

Singh, R. 1991. Tsunamis in Fiji and their effects, in Workshop on Coastal Processes in the South Pacific Island Nations, SOPAC (U.N. South Pacific Applied Geoscience Commission), Technical Bulletin 7: 107-120.

Tabata, S. 1975. The general circulation of the Pacific Ocean and a brief account of the oceanographic structure of the North Pacific Ocean. Part I - circulation and volume transports. Atmosphere, 13:133-168.

Talandier, J. 1993. Le volcanisme et la sismicit, in Atlas de la Polynsie Franaise, ORSTOM (Inst. francais de recherche scientifique pour le developpement en cooperation, Dept. en Territoires d'Outre-Mer), Paris, maps 26-27.

Talley, L.D. and R.A. deSzoeke. 1986. Spatial fluctuations north of the Hawaiian Ridge. J. Phys. Oceanogr., 16:981-984.

Talley, L.D., Joyce, T.M. and R.A. deSzoeke. 1991. Transpacific Sections at 47N and 152W: Distribution of Properties. Deep-Sea Research, 38:S63-S82.

Tomczak, M. and J.S. Godfrey. 1994. Regional Oceanography: An Introduction. Pergamon Press, England. 422 pp.

Tsuchiya, M. 1968. Upper Waters of the Intertropical Pacific Ocean. The Johns Hopkins Press, Baltimore. 50 pp.

Tsuchiya, M., Lukas, R., Fine, R.A., Firing, E. and E. Lindstrom. 1989. Source Waters of the Pacific Equatorial Undercurrent. Prog. Oceanogr., 23:101-147.

Worthington, L.V. 1981. The Water Masses of the World Ocean: Some Results of a Fine-Scale Census. Evolution of Physical Oceanography: Scientific Surveys in Honor of Henry Stommel, ed. Warren, B.A. and C. Wunsch. The MIT Press, Cambridge, Massachusetts. 42-69.

Wyrtki, K. and B. Kilonsky. 1984. Mean Water and Current Structure during the Hawaii-to-Tahiti Shuttle Experiment. J. Phys. Oceanogr., 14:242-254.

Zhang, Y, Wallace, J. M. and D. S. Battisti. 1997. ENSO-like interdecadal variability: 1900-93. J. Climate, 10, 1004-1020.