SURFACE VELOCITY PROGRAMME PLANNING COMMITTEE
Report of the First Meeting
SVP - 1
and
TOGA PAN-PACIFIC SURFACE CURRENT STUDY
NOAA Atlantic Oceanographic and Meteorological Laboratory,
Miami, Florida, USA
25 - 26 April 1988
WCRP-26
(WMO/TD-NO. 323)
October 1988
WOCE International Planning Office
Wormley
Appendix I: List of Participants
Appendix II: TOGA Pan-Pacific Surface Current Study
Global surface velocity and sea surface temperature (SST) fields of known accuracy can now be measured with ARGOS tracked upper ocean drifters. The overall objectives of a WOCE/TOGA Surface Velocity Programme (SVP) are to globally describe mixed layer circulation, to provide mixed layer velocity and SST observations for testing both global and basin-wide models of surface currents and the advection of heat and for relating air-sea fluxes of heat and momentum and satellite measurements of sea-level to upper ocean dynamics. The goals of SVP are to provide:
i) Global maps of mixed layer current and SST on approximately 500 km x 500 km space scale and seasonal time scale resolution for a five-year period. With these data new atlases of world ocean surface currents and temperatures will be made. Global SST measurements of unprecedented accuracy will be provided in real time for modeling air-sea interaction in atmospheric and oceanic climate models. SVP drifters can be used as platforms for enhanced global sea level pressure measurements, and pressure gauges will be developed for this purpose.
ii) Surface eddy statistics and wind-driven currents. A global distribution of horizontal kinetic energy and particle diffusivities due to eddy and wind stirring will be computed and related to eddy energy. Inter-comparisons of particle diffusivities in eddy resolving models will be made directly with SVP data.
iii) Studies of ocean surface processes. Surface drift will delineate areas of ocean surface divergence and convergence. Horizontal advection of heat, essential to climate models of the surface layers, will be measured. Using satellite derived sea level and winds, the momentum balance of the turbulent upper ocean can be studied and modeled.
The resources required for the WOCE/TOGA SVP are 3300 ARGOS tracked mixed layer drifters over a five-year period in the 1990's. The drifter will consist of a small float with a telemetry unit and thermistor on the surface and a large semi-rigid drogue of known dimension with a drag center at 15 m depth. The slip characteristics of the drifters, as a function of wind speed, will be directly measured in TOGA and pre-WOCE field studies. Real-time distribution of SST will be through GTS. Drifters are initially expected to survive for 1.5 years. Evaluation of drifter survival and deterioration will be made throughout the WOCE/TOGA period with a goal of obtaining a 2 year life time drifter. Deployment of drifters will be from ships of opportunity; each drifter will be encased in a 15-20 kg soluble cardboard container, and can be launched singlehandedly by an able-bodied person.
The deployment of drifters will be coordinated and the data will be quality controlled at a TROPICAL DRIFTER CENTER (TOGA/TDC), at the Atlantic Oceanographic and Meteorological Laboratory (AOML) in Miami, FL, and a GLOBAL DRIFTER CENTER (GDC), the location of which is to be selected by WOCE/IPO in the near future. The TOGA/TDC will also serve the real-time data needs of TOGA for tropical El Niño modeling and prediction and quality control of the GTS reports. The GDC will focus on WOCE/Core I, II, III data requirements in the mid-latitude and polar oceans and on the drifter design, deployment and international coordination. The quality controlled data will be filed at and distributed by the Marine Environmental Data Service (MEDS) in Canada, an IOC/WMO responsible data center.
This report presents the detailed scientific rationale for surface velocity measurements in WOCE and discusses the specific activities which lead up to and which will comprise a global SVP commencing in 1991. A TOGA Pan-Pacific Surface Current Study which began in Summer 1988 with a goal of maintaining a 230 drifter population in the tropical Pacific in 1988-1990 period will be shortly reviewed. The most immediate needs for WOCE are to conduct a heavy weather drifter calibration study in Winter of 1989-1990. Surface pressure sensors need to be adapted to low-cost drifters. Theoretical work on interpretation of Lagrangian data is encouraged.
The broad scientific objectives, rationale and implementation plans for WOCE and TOGA have been published for a number of years (WCRP-12, WCP-62) and these are reviewed and updated periodically by their respective Science Steering Groups, (SSGs). An element of each program is the measurement of surface velocity by ARGOS tracked drifters. In no other ocean observing system to be deployed in 1990's, are the objectives of both programs as closely linked by the natural phenomena as in the stated requirement to obtain the mixed layer velocity field over basin wide scales. For TOGA the focus is the tropics, but drifters deployed there will transit rapidly along western boundaries into mid-latitudes and the WOCE domain. Conversely drifters deployed into midlatitude southern oceans will be drawn into the tropical regions by powerful, large-scale South Equatorial Currents. The most cost effective and rational mode to proceed to a global study is for both WOCE and TOGA to closely coordinate their strategies in SVP. The least time consuming way to accomplish this for the ocean scientists in both programs is to form a joint plan of action for SVP for the 1989-1997 period during which both WOCE and TOGA will deploy ocean observing systems. To this end, the duly constituted WOCE/SVP Planning Committee, all Principal Investigators of the ongoing TOGA Pan-Pacific Surface Current Study and a group of invited experts met at the Atlantic Oceanographic and Meteorological Laboratory of NOAA in Miami, Florida on 25 and 26 April, 1988.
This is the first report of the WOCE/SVP Planning Committee and a planning progress report of the TOGA Pan Pacific Surface Current Study. Herein is found the science and the implementation plan for a SVP of 1990's in greater detail than exists in the existing WOCE or TOGA documents. Thus this document is meant to serve as a supplement to WOCE and TOGA planning documents and also as an expanded, but handy, companion for individuals or groups who will write yet more specific proposals to respective national funding agencies for participation in the global SVP.
The participants of the meeting are listed in Appendix I. We all are grateful for Dr. Don Hansen and AOML/NOAA for serving as our hosts in Miami.
Satellite tracked, surface drifters have been used in a large number of experiments to follow water particles in the surface velocity field. Early investigations include the NORPAX experiment (Figure 1), (McNally et al., 1983), Richardson's (1980, 1983) work in the Gulf Stream and its rings and over a major portion of the North Atlantic Subtropical gyre (Figure 2), and the Australian studies of the Tasman Sea (Cresswell, 1977; etc.). One of the more ambitious use of drifters occurred at the transition from the NIMBUS system to the current ARGOS tracking system. During FGGE over 300 drifters were released in the Southern Oceans (Daniault and Menard, 1985; Hofmann 1985; Patterson, 1985). More recently there have been substantial drifter deployments in nearly all oceans. Mid to high latitude experiments have occurred in the North Atlantic Current (Krauss and Käse, 1984; Krauss and Böning, 1987), the North Pacific (Emery, et al., 1985), the California current (Niiler et al., 1988), and the South Atlantic (Olson and Evans, 1986; Olson et al., 1988; Figure 3). In all equatorial oceans significant arrays have either been analyzed (Molinari, 1983; Hansen and Paul, 1984; Richardson and Reverdin, 1987) or are currently in the field (Figure 14). (See Appendix II).
Use of drifters presently offers exciting and novel new views of the surface layers of the ocean. Drifters have given us the only technique to continuously follow the drift of strong ocean eddies such as those spawned by the Gulf Stream, Kuroshio, and Agulhas currents. Drifters loop around these eddies providing a measure of their rotation velocity, their advection velocity, and changes in characteristics. The movement of drifters in these energetic eddies have revealed new information about these current systems. Thirty years ago we did not know that the Sargasso Sea and Slope water region of the North Atlantic were populated by numerous energetic current rings pinched off from the main Gulf Stream. Today the number, movement, and life history of these rings is relatively well known due in large part to the use of drifters to follow them. In the Gulf Stream region, drifter trajectories have given us a new way of visualizing the flow and a quantitative measure of its mean and eddy energies. We expect that WOCE drifters will reveal scientifically interesting trajectories and velocity statistics in the world ocean and shed light on the standing population of rings and eddies in other areas.
Drifter trajectories and satellite IR images of the ocean are a fascinating combination of data because both provide a visualization of ocean currents and resulting surface temperature patterns. Both techniques revealed how intricately complex ocean motion is, and resulted in entirely new concepts of how the ocean works. One of the major results of the WOCE drifter experiment will be a new visualization of the entire world ocean. One will be able to visually (and statistically of course) compare different regions by simply looking at the patterns of drifter trajectories there. How fast did they go, how similar are the trajectories, do they loop, are they trapped to bathymetric features, etc. Maps of trajectories will reveal some very puzzling phenomena as have trajectories in the past. These could lead to new ways of thinking about the ocean, to new studies, and a much greater understanding about how the whole ocean works.
One of the most interesting results of the SEQUAL-FOCAL drifter program was the clear picture of the retroflection of the North Brazil Current (NBC) into the NECC during the fall. There were only ~ 5 drifters in this region and time but their trajectories all agree in showing that the whole N. Brazil current retroflects and that the western NECC meanders swiftly eastward. This region in the western Tropical Atlantic was found to be as eddy energetic (EKE) as the high eddy average of the Gulf Stream downstream from Hatteras. During spring the drifters in the
Figure 1. Combined trajectories or NORPAX drifters for period October 1976 and August 1977. Solid dots indicate initial positions, solid triangles indicate last position (McNally, 1981). These drifters demonstrate coherent motion on 500 km x 500 km spatial scale patterns.
Figure 2. Mean velocity derived from drifters in the North Atlantic (upper panel) and velocity variance principals axis (lower panel). Global characterizations of surface velocity similar to this description can be derived from WOCE-TOGA drifters (Richardson, 1983).
Figure 3. South Atlantic surface dynamic topography with respect to 2500 db (Levitus, 1982) (solid contours) and drifter velocities averaged over 56 day intervals (arrows). (Olson, et al., 1988). The drifters are seen to move to the left of the geostrophic current derived from surface dynamic topography.
Figure 4. Average surface current vectors in central and eastern Tropical Pacific from EPOCS/NORPAX drifter data (Hansen and Paul, 1984).
NBC continued up the coast into the Guyana and Caribbean Currents. These results are significant for two reasons: 1) all maps of ocean currents based on ship drifts show the NBC continuing northward year round within only part of it retroflecting into the NECC during fall; 2) models show that the retroflection is part of the seasonal modulation of northwest heat flux through the tropics which is part of the large scale meridional overturning in the ocean (northward heat flux reverses during fall, and heat is spread in equatorial band). The retroflection was not observed by other techniques except the CZCS. Images of the Amazon plume show that during fall it moves north then east around the retroflection while during spring it continues northwest up the coast. The two techniques combined, drifters and CZCS images, have given us a new look at both the retroflection of NBC and the fate of Amazon water in the North Atlantic. WOCE drifters trajectories combined with remote sensing win reveal other new current patterns in the world ocean.
In the north Pacific the west-wind drift appears in ship drift charts as a broad large scale flow to the east which bifurcates smoothly at about 50° latitude, part of it going north into the Alaskan Stream and part of it going south forming the California Current along the North American coast. Releases of clusters of drifters in 1987 in this drift show (Figure 10) that an accelerated flow or the Alaskan Stream begins to form at least 900 km to the west of the North American coast. Such a midlatitude, mid-ocean current formation is not currently part of any broad-based observational or theoretical framework, and was discovered by drifter deployments.
Recent eddy resolving hydrographic surveys, moored current observations and surveys of currents with expendable velocity profilers reveal time-dependent patterns of strong east-west flows north of Hawaii which have east-west scales larger than 1000 km and north-south scales of only 10 km (Talley and deSzoeke, 1986). Similar vacillations appear in the mid-latitude North Atlantic (Schmitz, 1978). The strength of these flows on the surface are about 15 cm/sec with time scales of about 50-60 days or much longer than mesoscale eddies. Understanding the family of surface eddy motions in the mid-oceans is still far from complete and can be enormously enhanced by direct measurements of flow by drifters.
The SVP provides oceanographers with an opportunity to design a global ocean observing system. The FGGE drifter program was designed as a sea level pressure observing system, in which surface velocity was of no special importance. The careful calibration of the new drifters and the extensive reseeding in the SVP will yield reliable estimates of the spatial and temporal variability of kinetic energy and absolute (single particle) diffusivity. Repeated pair-releases in selected regions will provide unique observations of the spatial covariance of surface velocity, and relative (two-particle) diffusivities. These data are essential for inverse modeling and turbulence closure theory.
The large numbers of drifters envisaged by the SVP means that our Lagrangian view of the inhomogeneous and nonstationary ocean will no longer be anecdotal. Rather, it will be possible to consider circulation in terms of transition probabilities from place to place. These may be more meaningful diagnostics for eddy-resolving ocean circulation models than detailed synoptic maps.
ARGOS tracked surface drifters have become a readily available and rather commonly used tool for studies of large scale and remote upper ocean circulation. Their use in obtaining significant new data on ocean circulation is mandated in both WOCE and TOGA plans.
2.2 The Surface Velocity Measurement Problem
Upper ocean drifters are designed to measure Lagrangian water particle trajectories at a fixed level. Considerable effort has been expended towards the problems associated with the interpretation of drifter data and Lagrangian measurements in general. It is important to outline some of the difficulties in drifter data interpretation as these can impact use in WOCE and TOGA in a significant way: 1) Slip of a drifter with respect to its drogue drag center can be due to windage on the exposed surface float portion of the drifter and, more significantly due to rectified horizontal forces caused by the interaction of the float and drogue tether with surface gravity waves and shear in the mixed layer; 2) Surface drifters are only pseudo-Lagrangian since they follow particles at a specified level of the ocean rather than the three dimensional flow of water parcels; 3) The model intercomparisons are made difficult with Lagrangian data stemming from the problems of casting the equations of motion into Lagrangian frames and simulating particle trajectories in models employing finite grids. An assessment of each of these points involving both technical and theoretical aspects is given below.
Comparing surface drifter velocity measurements with velocity measurements from upper ocean moored current meters and acoustic doppler profilers have been difficult to make, because drifters do not sample the same location very long as do the Eulerian instruments. Richardson and Wooding (1985) have compared a number of trajectories of window-shade drogues at 20 m depth with progressive vector diagrams from moored current meter at 20 m depth in the Atlantic North Equatorial Countercurrent. The drifter and current meter data agree within 6 cm/sec in a 30 cm/s mean flow when their drifter is within 30 km of the mooring. A down-wind "slip" of 4 cm/sec is computed from a comparison of progressive vector diagrams for a 6 m/sec wind speed. These comparisons suggest that the FGGE generation of drifters are capable of measuring velocity with errors of 4-10 cm/sec. However, since most of the global ocean large scale flows are less than 10 cm/sec, drifters with better water following capabilities than have been deployed to date are required for WOCE/TOGA objectives.
Since 1983 efforts have been under way to design, test and deploy drifters with low slippage characteristics. The size of the surface flotation has been significantly reduced from FGGE and NORPAX models and semi-rigid drogues, which retain their shape under strong upper ocean shears have replaced parachutes and window-shades. In addition, measurements of flow relative to the drag center of drifters in different shear and wind conditions in the open ocean have been done (Niiler et al., 1987), resulting in drifters whose slip can be reduced to 0.2% of the wind speed (see Section 4.2.1). Although further tests (and perhaps design changes) are needed as part of the pre-WOCE activities, expectations are that drifters with 1-2 year life will be built whose slip can be corrected to less than ± 2 cm/sec (with known winds to ± 4 m/sec) in ± 2 x 10**-2/sec shear conditions in the upper 20 m of the ocean.
Our second problem arises because drifters follow a horizontal level rather than the three dimensional path of the fluid parcels, and thus their sampling is biased toward convergent ocean areas. This occurs from scales of Langmuir cells to Ekman divergences on gyre scales. If this bias is recognized at the outset, difficulties which arise in data interpretation can be circumvented by a proper deployment strategy, posing the questions addressed in the analysis correctly, and by analysing errors which would be introduced by biased sampling. Drogues can be made which are wider than the narrow (2 m wide) Langmuir cells and drifters can be deployed in areas into which they are unlikely to drift. Therefore the dynamical issues can be analyzed even though the sampling problem is somewhat complicated by the tendency of drifters to spend considerable time in convergent flow (Hofmann, 1985). The question of deployment pattern and number of drifters necessary to obtain statistically meaningful results in a region has been explored by Davis (1983) and McPhaden et al. (1984) and is taken up again in Section 3.
Finally, the prospect of a large drifter data set necessitates a careful review of methods for estimating Eulerian and Lagrangian parameters. It is useful to recall the ideal case of homogeneous, incompressible flow: the Lagrangian and Eulerian fields of any scalar or vector quantity have identical single-point, single time probability distributions. Note that isotropy need not be assumed. It follows that, for example, the ensemble-averaged Eulerian heat flux may be estimated by time-averaging the Lagrangian heat flux measured by a single drifter. Such an estimate could be obtained by a drifter provided it does not leave a region of local homogeneity, during an interval containing several d egrees of freedom of the flux stochastic process (that is, several decorrelation times). The typical 500 km x 500 km cell used in SVP design calculations in Section 3 will contain a drifter with a mean velocity of 15 cm/s for up to 30 days, or about 5 integral time scales. The drifter will not diffuse out of the cell in 30 days starting from the center, provided the mixing length does not exceed 120 km (assuming an r.m.s. velocity of 15 cm/s). This cell size is on the other hand small enough to justify assumptions of local homogeneity, except near boundary current regions and the like. The assumption of statistical stationarity is also plausible, for time intervals of up to one month. The statistical reliability of these estimates would be improved by averaging over other drifters happening to be in the cell at the same time. It would be dangerous to compute the divergences of heat fluxes estimated in this manner without further analyses of the local homogeneity assumption. Likewise the assumption of incompressibility requires careful consideration when estimating surface velocity parameters. This has some specific solutions, an example of which is given in Section 2.3.3.
The analysis of Lagrangian data has a long history. Recent advances have been spurred by technological availability of data from acoustically tracked floats such as SOFAR and RAFOS and satellite drifters of the type proposed here. The pseudo-Lagrangian problem of dynamical equations is ameliorated by taking into account the constraint imposed in the equations of motion for constant level particles in three dimensional flows (Okubo, 1970; Olson and Backus, 1985). The problem of converting between the Lagrangian and Eularian frames has received considerable attention by various authors but is still not possible in a general form (see Davis 1982, 1983).
In summary, some problems of accuracy and sampling by Lagrangian drifters have a solution. Others, which have dynamical interpretation, require more sophisticated research than has been done to date. These latter problems will be addressed by SVP scientists during and after the field program.
The following is a summary of objectives of surface velocity measurements which have been garnered from WOCE and TOGA science planning documents. These are reproduced here in summary forms for conciseness. In a select number of objectives, an expanded development is presented.
Global surface circulation has been derived from traditional ship drift observations. Because ship drift is estimated to the nearest 1/2 knot, only strong currents are adequately mapped. Also ship drift is well sampled only in major shipping lanes. Ship drift efforts are largely unknown. Drifter-derived maps of the North Atlantic Subtropical Gyre (Richardson, 1983), the Circumpolar Current (Patterson, 1985), the eastern tropical Pacific (Hansen and Paul, 1984), and the South Atlantic (Figueroa and Olson, 1989) have recently been published.
A multi-year global mapping of the mixed layer currents is proposed. Such a data set will make possible new atlases of world ocean surface currents with the annual cycle well resolved. With the addition of satellite surface slope and wind stress, numerous modeling studies of surface dynamical and kinematical balances will also be possible. The surface current measurements, when taken together with the mid-depth float experiments and the hydrographic program, provide a global data base for an improved description of circulation in both the deep geostrophic and shallow non-geostrophic regions of the ocean. A circulation model for the water column, together with an assessment of mixing will allow the development of more accurate models of the influx and fate of geochemical species in the oceans. Temperature measurements on the drifters provide a calibration data set for surface SST fields from satellites, a crucial input for surface heat flux estimates, and a globally uniform data set for monitoring long term surface layer thermal variability. From measurements of SST and surface velocity, surface heat advection can be estimated.
The accuracy of an average velocity estimate is roughly equal to the standard deviation of eddy fluctuations divided by the number of independent observations averaged. The number of independent observations are for a record length L and Lagrangian integral time scale T, approximately equal to the square root of LT**-1. Observations at the surface indicate an integral time scale between 8 and 16 days (the T here is twice what is often reported as the integral scale). Thus a 5 year record will provide an equivalent of more than 15 independent observations and reduce the uncertainty of the associated mean flow to less than 7% of the eddy variability. For example in an area with a 15 cm/sec eddy motion, the mean flow can be determined to an uncertainty of about 1 cm/sec. This is a value commensurate with the accuracy with which the drifters follow water. As a rough guide, about 5 years of observations in each of 1100 resolution cells will be requested to adequately map surface velocity (WCRP-11).
The SVP array offers a larger number of floating platforms from which atmospheric pressure can potentially be measured. FGGE demonstrated that atmospheric pressure could be accurately measured from drifters which maintain a pressure port continuously above the waves. On the SVP drifters small floatation element submerges often and the technique for accurately measuring atmospheric pressure from such floats needs to be developed. The SVP has adopted the development of the technique of atmospheric pressure measurements from submerging floats as a problem to be solved in the period 1990-1992, so atmospheric pressure sensor arrays can be deployed in SVP in 1993.
2.3.2 Eddy Statistics and Wind Driven Currents
Ship drift charts, hydrographic surveys, and satellite altimeter data have delineated fifteenfold changes in surface eddy energy across major ocean gyres (Emery, 1983; Cheney et al., 1983). In the areas of strong currents, drifter data have been used to describe variability in more detail than ship drift has afforded (Richardson, 1983; Hansen and Paul, 1984), and in the southern oceans where no ship observations exist, drifters have provided the first look at the circulation patterns. The time variability of surface currents reflects both the wind driven circulation and surface eddy currents. A global surface wind stress data set will permit separation of the surface velocity into wind-coherent, and wind-incoherent (eddy) components. The kinematic properties (Eulerian Reynold's stress ellipses) of the variable currents will be compared with the eddy properties obtained with floats and satellite altimeters. Because of wind driving, the net single particle dispersions on the surface will be estimated and are expected to be different from the subsurface quantities.
2.3.3 WOCE Model Intercomparison
While several model simulations have shown skill at estimating thermocline variations and the distribution of eddy energy throughout the upper parts of the ocean interior, there is a need for improvements in ocean surface parametrization. Specifically surface current strength depends upon how parametrization of vertical mixing is done. A surface drifter data set is one of the few means of testing such parametrizations on the scale of ocean gyres. Such a velocity data set could also be assimilated into models. The drifter data set will be particularly useful for model verification. Specific examples of how model intercomparisons might be done are given in Appendix II.
(a) Momentum Balance
A fundamental problem of interest is an analysis of the surface momentum field and here SVP data is useful. First, drifter trajectories are differenced in time to produce horizontal velocities at various locations and times. The information derived then includes a set of velocities at a set of time/space points determined by the flow, given the initial position of the drifter.
Averages of these data can be viewed in two ways: 1) As a Lagrangian average; or 2) as a mixed Lagrangian/Eulerian average. Acceleration following a particle however has a simpler Lagrangian/Eulerian interpretation. Consider the momentum equation written in the traditional fashion
where 
is the Lagrangian or total derivative following a two
dimensional particle path and the second term on the left is the Coriolis acceleration. The forces
shown on the right hand side are the pressure gradient force, the force due to turbulent momentum
convergence, and the force due to the fact that drogues on fixed horizontal levels do not follow
three dimensional paths. The data provide both terms on the left hand side of the equation directly
which are associated with the horizontal motion.
The direct analysis of the equation following a drifter is related to the typical specification of the right hand side of the equation, (i.e. those due to pressure gradient and vertical processes) by a Jacobian which provides the mapping of the spatial and temporal nature of the forcing terms with respect to the trajectory of the drifter. While the complex nature of this sort of analysis is the reason that the typical oceanographic problem is performed in an Eulerian coordinate system, the problem of using Lagrangian data, at least at the level of analysing drifter data, can be done in a mixed Lagrangian/Eulerian frame as discussed in the example below.
A set of surface mixed layer drifters launched in the South Atlantic in 1983-85 (Figure 3) are
used here to demonstrate the combination of such a data set with climatological estimates of the surface
stress and pressure fields. To consider the balance of forces responsible for the observed drifter
motions, a data analysis was carried out on a merged drifter velocity, dynamic height and wind data set.
For this analysis the following steps were performed: the drifter velocities were time averaged to derive
a mean velocity and location, and then the mean pressure gradient and wind stress for each mean drifter
location were taken from the Levitus (1982) set and the Hellerman and Rosenstein (1983) seasonal wind
stress data, respectively. The dynamic height increases following the drifter trajectories, which is what
one expects based on the Ekman convergence of mass in the center of a warm water gyre. It was found that
the dynamic height increase following drifters becomes equal to the gyre topography change at time scales
larger than half a year and for drift lengths scales comparable to the gyre (3000 km). Thus, after
averaging the drifter velocities over 56-day intervals, which suppresses eddy noise, the surface drift
is plotted along with the geostrophic current and the wind stress along the path of the drifter. The
result is plotted in Figure 5. Note that the drift exceeds geostrophy and is to the left of dynamic
height contours as it should be in the southern hemisphere. The residual
vector, 
, is approximately 90° to the left of the wind stress
direction. This is in agreement with what would be expected in a slab Ekman layer, and is perhaps
fortuitous as these drifters are drogued at about 1 m depth in the wave zone. The FGGE drifters
reflect the Ekman veering the least, as might be expected, since they have a greater tendency to
slip in the direction of the wind because they are either undrogued or have small drogues in comparison
with their surface floats.
The computations above are of interest in the sense that they give time scales for surface gyre forcing and a consistent picture of the combined geostrophic/Ekman flow in the ocean. In WOCE, analyses like this can be done on space and time scales which also resolve the temporal variations in the surface flow. In WOCE the altimeter will provide the time-dependent sea-level pressure gradients, although the dynamic height climatology and limited synoptic surveys will be important in application of TOPEX/POSEIDON to WOCE goals. The availability of scatterometer and calibrated atmospheric model boundary layer winds also allow for a temporal decomposition of the surface layer Ekman response. This is where the TOPEX/POSEIDON mission, coupled with the availability of wind fields ERS-I and improved atmospheric boundary layer models provides a capability to discern the dynamics behind evolution of the surface field on time scales greater than two weeks.
(b) Particle Diffusivities
Single particle diffusivities estimates have been made from drifters drogued at 100 m depth in the North Atlantic Current (Krauss and Böning 1987), and drogued at 15 m depth in the California Current (Poulain and Niiler, 1988). In the California Current the surface particle diffusivity is a linear function of the eddy kinetic energy. This is similar to single particle diffusivity estimates in the deep western North Atlantic from SOFAR floats (Rossby, et al., 1983). Figure 6 shows the California Current data. The slope of the line is the scale-independent Lagrangian time scale, which is 4 days in the surface layers of the east Pacific. More local estimates can be made of the differences between surface and deep layer particle diffusions with the SVP and WOCE float data, as well as more general descriptions of the global surface fields of diffusion rates. For example, the data of Krauss and Böning shows that in the North Atlantic Current the Lagrangian time scale at 100 m depth is a function of eddy energy.
Single particle statistics are essential for the development of turbulence parametrization schemes in ocean climate models. Even if a scheme cannot be developed, so that the turbulent stresses must remain as unresolved noise in a larger-scale model, the noise statistics (space-time autocovariances) must be specified in order to carry out least-squares data assimilation or inverse calculations using the model. A realistic model of "noise" is important because inverses with white noise can produce physically unacceptable (spikey) assimilations (Bennett and Budgell, 1987,1988).
Figure 5. Decomposition of South Atlantic drifters into a geostrophic component based on the Levitus (1982) hydrographic data set and a residual drift. Results are scaled such that the geostrophic component has unit length. For comparison purposes the direction of the wind stress vector averaged over the drifter trajectories is also shown. Note that the residual drift is to the left of the wind stress as expected in a slab Ekman flow. Results suggest the residual flow is consistent with a 40 m Ekman depth.
Figure 6. Single particle diffusivities in the California Current east of Baja California as a function of eddy energy (squares represent meridional diffusivity and triangles represent zonal diffusivity). The solid lines are least square fits. The average slope is 4.6 days, which is the average Lagrangian time scale (Poulain and Niiler, 1988).
(c) Surface Heat Advection
On climate time scales the ocean affects the atmosphere principally through changes of SST. Thus a principal scientific objective of WCRP is to accurately measure a global SST and to correctly model its seasonal and longer time scale evolution. Horizontal heat advection by ocean surface currents is an important component of the surface energy balance which determines the SST evolution. Figure 7 shows an estimate of the tropical Pacific surface mean heat advection based on historical ship drift and R. Reynold's SST climatology on a 2° Lat x 10° Long resolution. Here heat flux convergence due to advection is about 30% of the convergence due to surface flux in the upper 50 m. From SST and surface velocity, surface heat advection by major ocean current systems can be estimated. Satellite-tracked drifting buoys data have already been used in understanding the role of advection in the heat budget of the mixed layer. For instance, Hansen and Paul (1984) found, from drifter observations, that anti-cyclonic eddies associated with instability waves in the eastern tropical Pacific contribute significantly to equatorially convergent heat transport. Figure 8 shows this eddy heat flux whose convergence can cause a 2.5°C rise in SST per month. (Horizontal eddy momentum flux convergence can also cause a stress comparable to the surface stress). Anomalous advection and the associated heat transport during El Niño events has been shown to be critical in the surface heat budget and consequently in SST prediction (Halpern et al., 1983). In the tropical Atlantic, Molinari (1983) has used surface drifters to relate the seasonal evolution of SST fields to surface energy fluxes.
In ocean models, surface heat advection is very difficult to estimate, because the strength of the surface current depends upon vertical mixing parametrizations in the upper ocean. One objective of the SVP is to test and enhance ocean climate modelling ability to both advect heat in a realistic fashion, thus enhancing the models to produce realistic maps of SST, and produce correct interaction with the atmosphere.
2.3.5 Objectives of the TOGA Pan-Pacific Surface Current Study
The TOGA Pan-Pacific Surface Current Study commenced in 1988 to deploy 230 drifters in the Tropical Pacific. It is anticipated that this project will continue as pan of the SVP in the 1990-1997 time frame. Appendix II is the full program document for a ready reference.
Figure 7. The annual mean ship drift (arrow) and SST gradient (bar) in the tropical Pacific on 4° latitude by 10° longitude resolution (upper panel) and the resulting annual mean heat convergence (lower panel). Unit contours in lower panel would be 20w/m² contours, if the ocean were of uniform velocity and temperature over upper 50 m.
Figure 8. Meridional profiles of eddy flux of momentum and heat as measured from drifters in the eastern tropical Pacific (see Figure 4 for longitudinal extent of averaging) (Hansen and Paul, 1984).
The scales of variability of surface drift are suggested by a variety of drifter deployments. The following set of observations lead to the preliminary choice of 500 km x 500 km global sampling of surface drift. 1) Drifters in the Northeast Pacific have discernable wind driven Ekman currents which have scales similar to major storm systems on monthly time scales - or 500 km. 2) On scales smaller than 500 km, the velocity fields are dominated by mesoscale eddies, as has been shown in the South Atlantic. On scales larger than 500 km, the drifters respond to persistent, dynamic topographic variations. 3) A 500 km size box is big enough to contain a drifter for a month (less than 17 cm/sec velocity) but small enough to resolve the basin scale inhomogeneities of circulation. 4) The 500 km scale in some places such as the tropics will be interpreted as 1000 km east-west and 250 km north-south because for persistent equatorial current systems, like those along the equator, east-west scales have been shown to be larger than north-south scales. 5) The estimate of the numbers required are in keeping with a uniform distribution on 500 km scale. In some ocean areas, drifters tend to congregate and the density numbers increase. The deployment scheme, thus, will not necessarily be uniform. The estimate of the number of drifters required to satisfy the science needs of SVP is derived below. This estimate will be further refined by specific proposals.
Let the drifter density be designated by Pi, for each of several areas Ai of the global ocean. The total number of buoys required for that area is
Mi = Pi Ai (D/L),
where D is the duration of WOCE and L is a design drifter lifetime in years. In general, we assume that in the beginning of WOCE L = 1.5 and D = 4-5. At the end of WOCE, L ~ 2. Thus, either 3 or 2 arrays will be deployed, depending upon when the area is seeded. The present life design is 1.5 years, based on anticipated drogue loss, and losses due to piracy and grounding. Values of Mi will be estimated for various regions and will be summarized on Table 3.1.
To accomplish the science objectives of the TOGA Pan-Pacific Surface Current Study, an array of drifters will be maintained in the area 15°N to 15°S, and from 80°W to 126°E. The desired sampling density can be accomplished with 230 resident drifters. By late 1990, a recommendation will be made to WOCE and TOGA on the merits of continuation of a tropical circulation study. We assume that this program will be successful and the sampling program will remain about the same as was in place in 1988-89. By 1990 the required renewal rates will have been determined empirically. The present estimate is that 230 drifters will be required in each array of 1 1/2 year duration. A total of Mi=690 drifters will be required.
The North Pacific Ocean from 15°N to about 35°N averages about 12,000 km in east-west extent. Due to strong zonal flows and eddy mixing related to their mesoscale components it is believed that this is a relatively inhomogeneous region in north south direction so a 400 km north to south drifter spacing will barely be sufficient. However, east-west scales are long, thus 600 km spacing is desired; so an array of Mi = 132, needs to be maintained. North of 35°N the average east-west extent is only about 6,000 km, but because of the predominance of wind forcing a 500 km spacing in both dimensions is called for. The number of resident buoys required is 68, so a total of 200 are to be maintained in the North Pacific. The recent Ocean Storms experiment with 60 km spacing in the Gulf of Alaska, should allow for more meaningful interpretations of the WOCE data set in this area ( See Figure 10). Past and future Japanese buoy deployments will serve a similar purpose in the vicinity of the Kuroshio and its extension.
The region of the Pacific between 15°S and 35°S extends about 10,000 km east-west and 2100 km north-south. There is very little experience on which to base a drifter array design, but FGGE drifters in this region do display some zonality. Therefore, a drifter spacing at 500 km north-south and 500 km east-west is chosen. This coverage would be achieved with a maintained array of 90 drifters. Three deployments are planned, so Mi = 270.
The Core I global coverage of drifters in the Atlantic is to extend from 60°N to 30°S, an area of about 50 million sq km. This consists of approximately 240 boxes 500 km square. Thus Mi = 720 drifters, consisting of 3 deployments, are needed to maintain an even population at a 500 km spacing. Because of the very different flow regimes of the Atlantic and the use of volunteer ships of opportunity, the deployments will not be uniform. We expect more frequent deployments at closer spacing in the swift western boundary currents like the North Brazil current and the Gulf Stream, and less frequent deployments at wider spacing in the more sluggish mid-ocean regions. We will be able to achieve our desired array more easily in the Atlantic through a judicious deployment strategy, made possible because of the extensive experience already gained in this ocean.
The design of a WOCE drifting buoy array in the Southern Oceans is based on a number of published analyses such as those based on the extensive 1979 drifting buoy deployments. Studies have shown the sampling to be adequate for mapping large scale flows and eddy energies from March through April when the density was highest, with about 200 drifters reporting. These studies have determined the surface circulation, and the kinetic energy distributions of the Southern Oceans, (Daniault and Menard, 1985; Patterson, 1985). To duplicate this density of drifters, a 200 drifter array in the Southern Ocean must be maintained. We here desire to increase that resolution to be comparable with the sampling in other oceans.
From studies of the large-scale, low frequency motions of the FGGE drifting buoy array (Large and van Loon, 1989), marked zonation (Hofmann, 1985) and areas with large 6 month period motions, demand a drifter spacing of only 2 degrees in latitude. However, the zonal scales are very large and, as in Tropics, a 10-degree longitude spacing should provide adequate sampling over most of this ocean. This spacing would be provided by about 344 buoys for the oceans south of 35°S. Initial placement might be more dense in areas where bottom effects are known to be important and can lead to shorter spatial scales. Two array deployments will be planned for the southern oceans so Mi = 688.
In addition to some FGGE drifters, there have been a number of subsequent Indian Ocean drifter deployments, particularly in the tropics. An extensive data analysis project is presently underway (R. Molinari, personal communication, 1988). The minimum number of buoys required to complete the basin wide survey and 500 km spacing in both dimensions is 151, for this 5000 km by 5000 km region. Special studies of the Arabian Sea and Somali Current outside WOCE objectives will undoubtedly include surface current studies in much higher space resolution than indicated here. Two deployments are planned for the Indian Ocean, so Mi = 302.
3.3 Enhancements To The Global Array (Core III, Gyre Dynamics Experiment)
To date the Core III process experiments that have expressed a need for surface drifters are the Ekman experiment and Subduction (see International WOCE Implementation Plan for rationale). The former asks for only about 20 drifters and the latter calls for about 100. However, these are total numbers, not per year numbers, so they are to be simply added to the annual total.
3.3.2 Basin Scale Measurements
The well measured basin concept calls for increased sampling in the Atlantic basin from about 30°S to 60°N. The global drifter array in the Atlantic is to have 500 km spacing, which is already somewhat smaller than in the Pacific. Present evidence indicates that to sample the mesoscale eddies would require much increased sampling beyond our capabilities. But a high resolution map of eddy statistics is within the scope of this project. An additional 100 buoys per array are to be used to enhance the sampling in the Atlantic basin, primarily to better study the western and eastern boundary current regimes of the Gulf Stream, the Brazil current Benguela and the North Brazil current. The latter is also connected with attempts to study and quantify cross-equatorial flow. These boundary layer deployments will ultimately feed the basin array. In addition, the above process experiments are planned for the Atlantic, so the effective drifter coverage of the Atlantic basin will be considerably better than elsewhere.
The above calculations are summarized in Table 3.1. The Global Surface Velocity Program requires an array of 1254 surface drifters to be maintained throughout WOCE. It is the assumption that in the beginning of WOCE a 1.5 year lifetime drifter will be available and at the end of WOCE, drifters last two years.
The drifter enhancements for the well measured basin amount to 300 buoys over WOCE period. The process studies need a one time deployment of 120 buoys.
The grand total of drifters required for the WOCE/TOGA surface velocity program is 3690 over 4-5 years. It is planned to start deployments in the Pacific, building on the existing array in the TOGA Pan-Pacific Surface Current Study. This is to be followed by Atlantic, Southern and Indian Oceans. On the average about 1000-1200 drifters will be in the ocean at one time. Present accounting by Service Argos indicates that about half this number of drifting buoys (of all kinds) are reporting today. This implementation plan will proceed along two lines. First, efforts will be made to upgrade the buoys presently being deployed, so that their dynamical characteristics conform to what is required for SVP. Secondly, there is a need to expand the worldwide capability of producing drifters by about a factor of two, in order to meet the requirement of drifters during WOCE. Neither of these requirements appear to present any major difficulties.
Table 3.1 Summary of Total Drifter Requirements
| Equatorial Pacific | ||
| North Pacific | ||
| South Pacific | ||
| Atlantic | ||
| Southern Ocean | ||
| Indian Ocean | ||
| Atlantic Enhancements CORE III | ||
| Process Studies | ||
4.1 Design parameters and recommendations for the WOCE/SVP "standard drifter"
I. Introduction
The success of the SVP depends on the successful acquisition of a high quality surface mixed layer velocity and temperature data set that is rigorously documented and which withstands careful scientific scrutiny. To this end, it is imperative that the velocity and temperature sensing capability of the surface mixed layer drifters to be used in the WOCE/TOGA SVP be traceable to a reference standard with an accuracy comparable to that of other more conventional current and temperature sensing devices presently in use. The purpose of this section of the SVP plan is to outline the steps to be taken to insure the data collected meets these necessarily high standards.
II. Design Parameters
A detailed set of specifications has already been written for the drifters to bc used in the TOGA Pan-Pacific Surface Current Study (Appendix II). The WOCE/TOGA SVP planning group finds these specifications to be of high quality and recommends the adoption of a similar set of specifications for the SVP drifters as well to insure optimum compatibility of the two data sets.
A. Velocity. Slip with respect to the theoretical drag center of the drogue of a SVP drifter should be correctable with known winds to less than 2 cm/s in the conditions likely to be encountered during WOCE, (a 2 cm/s slip is due to unknown shear of upper ocean). Previous data (Figure 9) have shown that a drag area ratio of 40-50 between the drogue drag area and the drag area of the surface elements exposed to surface wind and wave conditions is required to meet this specification. A rigorous calibration of all types of WOCE/SVP drifters in a wide variety of wind conditions and sea states will be required to insure that this specification can be met. Details of the methods to be used during the calibration experiments are outlined below in the calibration section.
Buoy positioning error must be less than 700 m so that they will not contribute more than 20% of the allowable slip, in a velocity determination from positions 3 days apart. Most, but not all, ARGOS positions satisfy this requirement. The accuracy of any alternate positioning methods must be demonstrated.
B. Temperature. The sea surface temperature (SST) measurement system must have an accuracy of ±0.1° C, with a sensor stability such that a drift of less than 0.1° C per year is encountered. This sensor should be mounted on the buoy in a way that maximizes the contact with the water (for the fastest response time) and minimizes the thermal conductivity between the sensor and the buoy hull (to avoid contamination via surface heating of the buoy hull and subsequent heat conduction down the hull to the temperature sensor). SST measurement can be done in the upper 1 m of the water column. When subsurface temperature is measured, the required accuracy is plus or minus 0.04° C, with sensor stability comparable to that of the SST sensor.
Figure 9. "Slip" of surface drifters at drogue center as function of wind speed and drogue to surface element drag area ratio. The diamond shapes are TRISTAR drogues with spherical surface floats; the ellipses, circles and half circles represent a DRAPER LAB pill, a sphere and hemisphere surface float, respectively, with a holey sock drogue. The numbers designate the sock diameter. The large rectangle with 72 is a 6' x 21' holey sock with spherical surface floats identical to TRISTAR. The FOCAL-SEQUAL drifter is described in the figure. The scatter in the diagram is principally a result of computation of individual measurements of "slip" over 20 min - 1 hour in specific wind-wave conditions. All data is from ocean wind conditions of less than 10 m/sec.
The drogue drag center for the WOCE/TOGA SVP drifters should be at 15 m depth, with less than plus or minus 3 m vertical excursion in a vertical current shear of 2 x 10**-2/sec. The rationale for this depth choice is as follows: The surface mixed layer over most of the world's oceans at night exceeds 20 m much of the time. The daily thermocline forms above 10 m. Most of the proposed drogues to be used in WOCE are less than 7 m long. In order to insure that the drogue is below the daily thermocline and in the upper mixed layer, it can therefore be centered no deeper than 15 m. As an upper limit, the line connecting the drogue to the surface buoy cannot be too short, or a) the buoy spends too much time under water and b) the stresses on the line from the heave of the buoy are too great and the survivability of the system is reduced. A nominal depth of 15 m is therefore the best compromise between the practical and scientific requirements of the experiment.
Previously deployed drifters have all lost their drogues eventually, with the time of this occurrence varying from a few days in the worst case to about 9 months for the best cases. A rigorous scientific interpretation of the data requires one to know whether the drogue is still attached or not, therefore, all drifters to be used in SVP must be equipped with a reliable sensor to report whether or not the drogue is still attached to the surface buoy. A number of different types of sensors have been proposed and it is not presently clear which type of sensor is best. The possible candidates include tilt sensors, strain guages, submergence counters, acoustic pinger/hydrophone arrangements, and combinations of the above. More engineering evaluation is needed in this area (see 4.2.4).
All the drifters to be used must report their data via service ARGOS. The SST data will be most useful if it is in a format compatible with the Global Telecommunications System (GTS). The present specification on the maximum reporting interval is every three days, based on the expected oceanic time scales and also on the billing system used by Service ARGOS.
The design goal for the SVP drifters is that they be able to survive and report their data in the ocean for a period of 1.5 years. The power supply for the surface buoy should be adequate to drive the buoy electronics for a period of 2 years.
The packaging of the drifters should be such that they can be deployed by one able-bodied seaman from a ship of opportunity. The system must be able to withstand the shock of striking the sea surface after free-falling a distance of 10 m. An alternative deployment scheme is to lower a drifter over the side with a disposable tether.
All investigators who wish to deploy drifters as part of the WOCE/SVP are required to calibrate their surface drifter instrumentation. They may do this on their own or send their preferred system to the WOCE Global Drifter Center (GDC) and the program will do the calibration for them on one of the scheduled calibration cruises (described below). All investigators are required to send a copy of the calibration sheets for their temperature sensing devices to the WOCE GDC.
4.2 Testing and Verification Experiments
Three calibration studies have been done in which absolute flows past drifters have been measured. These are reviewed in Figure 9. An important parameter for drifter water-following performance is the drag area ratio - R, which is the ratio of the sum of the product of drag coefficients and frontal areas of the tether and surface element to the drag coefficient times the frontal area of the drogue. This data suggests a value of R in excess of 30 is required for keeping slip below 2 cm/sec in 10 m/sec wind speed. If upper ocean shear effects are added, a value of R in excess of 50 is required.
The data on Figure 9 are quite "noisy" and come from wind (and wave) conditions below 10 m/sec and upper ocean shears less than 2 x 10**-2/sec. Tropical shears exceed 2 x 10**-2/sec and winds in mid-latitude winter exceed 10 m/sec. To assess the drifter performance capability in tropical shears which exceed 3 x 10**-2 /sec and heavy weather conditions of midlatitude winter, further performance tests are required.
4.2.2 Results of TOGA Tropical Calibration Experiment
In May 1988, TRISTAR and holey sock drogues were fitted with VMCM's on top and bottom and released in the tropical Pacific for six separate four hour periods in different shear, wind and wave conditions. Various float combinations were used. The objective of these tests was to obtain data on the slip of drifters used in TOGA Pan-Pacific Surface Current Study. These tests produced quantitative slip data in the parameter range of 18 <R <98, wind speeds of 4 - 11 m/sec and vertical shears from 0.2 to 1.5 X 10**-2/sec. The most stringent conditions for drifter performance were with 1.5 x 10**-2/sec shear and 10 m/sec wind, both acting at the same time. Under these conditions, the measured slip was 2.0 cm/sec in a drogue and float configuration with R = 97 (in the same test slips of 4.9 cm/sec and 2.7 cm/sec were recorded on drifters with R = 32 and R 50, respectively). In all 5 other tests in the tropical conditions, measured slip was less than 2 cm/sec with R>50. Slip is strongly correlated with wind speed (Niiler, et al., 1987). If wind velocity is known to within 4 m/sec, drifters currently in use in TOGA with R = 50 can be corrected to slips less than 2 cm/sec (which would be due to unknown shear alone).
While the global mean wind speed is 7 m/sec, significant seasonal and latitudinal variations exist. In the WOCE global array there will be a significant number of drifters in areas where wind speed for many seasons exceeds the global mean, for which winds drifter performance has not been measured. Drifter performance cannot be extrapolated from 10 m/sec wind conditions to more severe weather with any certainty because of the non-linear nature of the hydrodynamic drag on the drogue and the effect of breaking surface moves in exerting a biased horizontal force on the surface element. It is recommended that a heavy weather study of drifter performance be done in the winter of 1989-1990. The objective of this study would be two-fold: i) To obtain data on the slip of drifters in wind conditions from 10 m/sec - 25 m/sec and ii) evaluate the progressive wear and tear of drifter elements through one winter season.
In the tropical tests, drifters have been deployed from research vessels for a few hours test and recovered either by launching a small boat or by directly attaching a recovery line. In heavy
Figure 10. Tracks of the first 100 days of drifters deployed in October, 1987 in eastern North Pacific. Open circles mark the deployment location. This is the area proposed for WOCE heavy weather tests.
weather, recovery operations are not possible, thus drifters would have to be followed for many days through a storm and brought back on board when calmer conditions between storms prevail. We have found that launch and recovery operations can be safely done in up to 10/sec winds which are typical of north-east Pacific conditions between individual storms in October-November period.
One ideal site for the heavy weather test is the OWS-PAPA area at 50° N - 145° W. Drifters with various R ratios, can be outfitted with two current meters and released at the beginning of the Fall quarterly PAPA-LINE cruise a few hundred kilometers from shore. Upon return from the PAPA survey, the drifters can be located with ARGOS (and with H-F radio) and recovered when weather permits. Periodic recovery of operational drifter in the same vicinity in intervals of 3, 6 and 9 months would form a data base for wear and tear under sever conditions through a North Pacific winter. In this area, drifters with drogues intact are known to travel eastward and these reach shore waters in nine months, Figure 10 shows, a 100 day drift of a cloud of TRISTARS released in the OCEAN STORMS experiment in October 1987.
A joint Canadian and U.S. proposal has been prepared to conduct WOCE heavy weather tests in Fall-Winter 1989-1990.
During the first few years of WOCE/TOGA SVP a program for recovery and inspection of drifters at sea should be carried out. A drogue indicator can be designed to monitor the failure of the tether line, but it would be most difficult to ascertain the progressive deterioration of drogue elements. Therefore from a number of deployments sites, drifters would be released which have daily ARGOS reporting (recall, a 3 day transmission cycle in general is proposed). A research vessel (or a fishing charter) would be dispatched to effect the periodic recovery operations. The data on progressive deterioration is vital to the quality of the SVP data set.
The SVP drifter array can serve as platforms for ocean sensors in addition to WOCE/TOGA requirements. Data can be sent through ARGOS at no additional cost. Subsurface data links to the drogue have been developed for acquisition of a variety of data: light extinction for ocean transparency, sound monitoring for wind speed, mixed layer salinity and temperature, profiles, etc. The SVP intends to make available technical expertise and is open to possible joint ventures with GOFS, GLOBEC and other global (or local) geosciences programs in using ARGOS tracked Lagrangian drifters for platforms for acquisition of upper ocean data sets.
4.4 Operational Meteorological Drifting Buoys
It is apparent that large numbers (about 100) of surface drifting buoys to measure sea surface temperature (SST) and sea level atmospheric pressure will be deployed annually by various nations as part of their operational meteorological data collection programs. For example, TOGA maintains an array of 40 such buoys in the southern oceans. These buoys are not useful for SVP purposes in their present configurations however, since they are not acceptable current following devices, i.e. either the buoy/drogue combination has not been properly calibrated or they have no drogues at all. These buoys nevertheless represent a large untapped potential resource for supplementing the SVP Lagrangian surface velocity data set at a very low cost. The working group therefore recommends that the WOCE International Program Office (IPO) formally contact the WMO with specific recommendations regarding the changes that need to be made to the WMO surface drifters in order to make them acceptable for both WMO and WOCE purposes.
Specifically, there are two basic options: 1) The WOCE/SVP Global Drifter Center (GDC) will design a special version of the WOCE standard drifter that incorporates an atmospheric pressure sensor and make these buoys available to the WMO. The projected cost of these drifters will be less than the buoys presently in use by WMO. 2) The GDC will supply drogues for use with the existing WMO surface buoys, and calibrate the system so that its current following capabilities and errors are known. Both of these options require the development of an atmospheric pressure sensor that can withstand temporary submersion. The GDC will handle the engineering development of this sensor. Those nations presently purchasing new MET buoys should consider option 1, while those refurbishing older buoys for redeployment will find option 2 more attractive.
The MET buoy community will receive substantial benefits for their efforts in cooperating with this program. 1) A drogued buoy will remain in its deployment region for a longer time than an un-drogued buoy. 2) The drogued and calibrated buoys will return surface mixed layer current information which is useful as a forecasting aid. 3) The surface pressure data from other WOCE deployed drifting buoys can be made available to WMO, to enhance the resolution of their pressure observation system.
Since there is considerable lead time involved in discussing and implementing these changes, the WOCE IPO should initiate contact with WMO with all due expediency.
4.5 The WOCE/TOGA Standard Drifter
The SVP calibration group at GDC will calibrate or certify the calibration of all the drifters proposed for use in the program. Following the calibration experiments, the SVP will recommend and provide engineering drawings for an optimum drifter to the scientific community for use in the program. While all investigators (particularly international participants) will not be required to use the chosen configuration, it is highly recommended that they do so, to insure the uniformity of the results, and to allow the mass production (or at least mass purchasing of the necessary parts) of the drifters at greatly reduced cost (see implementation section). Examples of drifters which serve as candidates are diagrammed on Figure 11. The criteria to be used in choosing the WOCE/SVP "standard drifter" are as follows:
1. They must meet the design specifications outlined in section II above.
2. They should be constructed of materials which are easily available worldwide.
3. They must be cost effective.
Figure 11. Schematics of ARGOS tracked, mixed layer drifters to be tested as for WOCE/TOGA use.
Introduction
Although the national plans are necessarily incomplete at the time of this report, it appears that participation in the SVP will come from investigators in several countries. The TOGA Pan-Pacific Project to be conducted during 1988-1990 is a basin-scale model for the WOCE/TOGA SVP. In the Pan Pacific Project, Australia, France, Japan and U.S. investigators are collaborating in collection and analysis of surface velocity data from the tropical Pacific Ocean. In the global project envisaged for WOCE, close coordination should be maintained with the TOGA effort and the planning committee recommends that joint WOCE/TOGA implementation be established. To obtain economy of scale for equipment procurements, best coordination of buoy releases to maintain desired buoy distribution, and production of a homogeneous research-quality data set, some of the activities will be centralized in special centers. Because of the relatively greater importance of real time activities for the TOGA Program, and because of the relatively more global data equipments for WOCE, it appears appropriate to establish a Global Drifter Center (GDC) and a Tropical Drifter Center (TDC) with somewhat distinct objectives.
WOCE/TOGA SVP requires an array of nearly 1100 drifters to be maintained on a global basis. Figure 12 shows the expected time line for the implementation of SVP. Presently, the largest single population of drifters is in the TOGA Pan-Pacific Surface Current Study, where about 230 drifters are being built and deployed in 1988-89. To meet the combined WOCE/TOGA requirements, a nearly five fold increase of drifters is required. This increase can be achieved with sufficient planning and lead time. The GDC will have to become fully operational about eighteen months before the first global array would be in place.
5.1 Functions of the Global Drifter Center
The GDC will implement the qualification of drifting buoy designs established by SVP Planning Committee, coordination of procurements of drifting buoys or buoy components, contracting and acceptance of drifting buoy hardware. It will carry out the heavy weather tests and make available technical expertise for construction of drifters to various nations and scientific establishments who wish to build their own drifters.
5.1.2 Logistics of Drifter Releases
Most drifter releases will be made from volunteer ships. Figure 13 displays the merchant ships routes which are in current use in deploying XBT's in the Pacific and which will be available on about 1 month interval for use in SVP deployments. Other oceans have similar networks. The GDC will establish and maintain liaison with suitable vessels and arrange for release of drifters as necessary to meet SVP objectives. It is proposed to begin the program on a basin by basin deployment, releasing drifters in the mid-ocean from north-south units and from east-west lines in eastern and boundary currents. The objective would be to "fill" a basin over a six month time frame and then make adjustments to the array as significant losses or convergence occur.
5.1.3 Coordination with Other Users of Drifting Buoys
It is expected that there may be as many as a few hundred non-WOCE buoys working concurrently with WOCE. The GDC will encourage use of buoy configurations having performance
Figure 12. Time lines for the scheduled WCRP global drifter deployment (upper panel) and a schematic of major activities leading up to the global array (lower panel).
Figure 13. Volunteer ship of opportunity tracks in the Pacific from which drifters can be deployed at least once a month. The stippled area has higher than once a month frequency ship traffic.
parameters suitable for multiple uses of the data collected. A principle focus of this activity will be to seek better Lagrangian performance, or better calibration, of drifting meteorological buoys. It will undertake the development of a technique of measuring atmospheric pressure from submerging floats, if deemed necessary for carrying out the objective of global pressure array.
Data will be uniformly processed to research standards for use by WOCE Program participants and be submitted to the Marine Environmental Data Service (MEDS) for incorporation into the international archive system which will make them available to the general user.
The GDC will create and distribute such data products as are requested by the SVP Planning Committee.
5.2 Functions of the Tropical Drifter Center
5.2.1 Tropical Array Requirements
The Tropical Drifter Center will closely monitor the tropical drifter array in relation to requirements of the TOGA program. It also will reevaluate those requirements in the light of new data as they are acquired. Deployment requests will be made to the GDC.
Data flow from WOCE/TOGA tropical drifting buys will be monitored for the purpose of withholding from dissemination via GTS data from buoys identified as defective.
5.2.3 Real Time Data Processing
The Tropical Drifter Center will create and distribute surface current and temperature data products to be used by scientists involved in real-time aspects of the TOGA program.
The data from the tropical drifters will be uniformly processed to provide a research quality data set for use by program participants and scheduled submission to the MEDS for distribution to general users. Figure 14 displays the processed data from TDC for the period 6 March - 3 April, 1989.
Figure 14. Displacement of drifters in period 6 March - 3 April, 1989, in Pan-Pacific Surface Current Programme. The data are processed in the Tropical Drifter Center, AOML/NOAA.
5.3 WOCE/TOGA SVP Planning Committee
Terms of Reference
The terms of reference for the SVP Planning Committee was approved by WOCE SSG in November, 1987. The present constitution reflects a predominant WOCE participation with one TOGA representative. An expansion of TOGA membership is expected in 1989.
1. To be responsible to the SSG for the scientific advice necessary for designing and carrying out the global surface velocity and SST measurement programme using ARGOS tracked drifters as specified by WOCE CORE Projects I, II and III.
2. To advise on the capability of drifters to measure surface velocity and SST, the facilities required at sea and on shore for buoy deployment and on the management of data flow in real time through the GTS. Where possible this should be done in consultation and cooperation with TOGA.
3. To determine the necessary protocols that must be established to carry out the Surface Velocity Programme both in dedicated facilities for data and field programme.
4. To advise and consult with organizations and individuals involved in the operational aspects of drifter deployment, the technical aspects of data collection through ARGOS and calibration procedures to ensure that the programme is effectively carried out.
5. To advise the SSG whether proposed management structures for the operational aspects of the programme are consistent with the scientific and technical requirements necessary for the collection of a uniform high-quality data set.
6. To keep the Director of the WOCE IPO informed of significant developments in the planning of WOCE Surface Velocity Programme. The IPO will supply support staff to the Planning Committee.
Membership (1988); WOCE/SVP Planning Committee
The Marine Environmental Data Service in Canada was identified as the final archive for drifting buoy data from the WOCE/TOGA SVP. Both GDC and TDC will submit quality controlled, processed data sets on drifter location and SST to MEDS on a six month interval.
MEDS will distribute the data submitted by SVP in accordance with data availability policies and time scales as decided by WOCE and TOGA data management plans.
All SVP drifting buoy data will be transmitted on the GTS to insure that new temperature data are available to meteorological global forecast programs and to TOGA for tropical data. MEDS will copy the GTS data and produce inventory maps of both SVP and non-SVP drifting buoys. MEDS will also maintain an inventory file of buoys, owners, buoy and drogue types to further assist in the interpretation and use of non-WOCE drifter data in the SVP.
MEDS will limit the general distribution of SVP drifting buoy data copied from the GTS to conform to the data availability policies and time scales as described by WCRP requirements. Data will be made available to other WOCE and general users on request as per WCRP policies and time scales.
Data will be provided to the IODE World Data Centers in concordance with WOCE and TOGA data policies.
The WOCE/TOGA SVP will require a substantial increase in the number of buoy years presently being handled by Service ARGOS. Added to the SVP buoys WOCE plans call for a number of fixed stations, floats, etc., to relay their data via ARGOS. All of this activity, planned to start in early 1990's, will cause a most rapid increase in the number of ARGOS platforms being processed. At present tariff prices, this large number of platform years would add up to an equally large cost, which would be a significant portion of the WOCE scientific budget.
SVP has begun this process by suggesting that the drifters report only every third day thus saving two thirds of the ARGOS cost. This savings mechanism may not be optimal (for example, some drifters require initial full sampling, to be followed by a sparse sampling later) and it may be possible to negotiate some other alternative with Service ARGOS that accomplishes the same savings without having to forgo the daily fixes. ARGOS has stated its willingness to negotiate and has acknowledged the big changes that will occur with WOCE. It is therefore recommended that the WOCE International Project Office (IPO) uses presently available WOCE planning information to estimate the number of ARGOS platform years anticipated for the WOCE project along with their anticipated data rates. Using this information WOCE IPO should initiate negotiations with ARGOS via the Drifting Buoy Coordinating Panel, to evolve an appropriate pricing structure for the WOCE experiment. These negotiations should begin in the very near future as these ARGOS tariff numbers are needed for the allocation of funds for the science projects using ARGOS services. The SVP recommends that a representation of the SVP Planning Committee works closely with IPO in this matter. Since such a large increase of ARGOS usage is anticipated and since many costs at ARGOS are fixed, adjustments to the present tariff should reasonably be expected.
Bennett, A.F., 1987. A Lagrangian analysis of turbulent diffusion. Reviews of Geophysics, 92, 2907-2918.
Bennett, A.F. and W.P. Budgell, 1987. Ocean data assimilation and the Kalman Filter. J. Phys. Oceanogr., 17, 1583-1601.
Bennett, A.F. and W.P. Budgell, 1988. The Kalman smoother for a linear quasigeostrophic model. Submitted to Dyn. Atmos. Oceans.
Cheney, R.E., J.G. Marsh, and B.D. Beckley, 1983. Global mesoscale variability from collinear tracks of Seasat altimeter data. J. Geophys. Res., 88: 4343-4354.
Cresswell, G.R. 1977. Trapping of two drifting buoys by an ocean eddy. Deep-Sea Res., 24, 8022-8030.
Daniault, N. and Y. Menard, 1985. Eddy kinetic energy distribution in the Southern Ocean from altimetry and FGGE drifting buoys. J. Geophys. Res., 90, 11877-11889.
Davis, R.E. 1982. On relating Eulerian and Lagrangian velocity statistics: single particles in homogeneous flows. J. Fluid Mech., 114: 1-26.
Davis, R.E. 1983. Oceanic property transport, Lagrangian particle statistics and their prediction. J. Mar. Res., 41: 163-194.
Davis, R.E. 1985. Drifter observations of coastal surface currents during CODE: The method and descriptive view. J. Geophys. Res., 90: 4741-4755.
Emery W.J., 1983. On the geographic variability of the upper level mean and eddy fields in the North Atlantic and North Pacific, J. Phys. Oceanogr., 13, 269-291.
Emery, W.J., T.C. Royer and R.W. Reynolds 1985. The anomalous tracks of North Pacific drifting buoys 1981 to 1983, Deep-Sea Res. 32, 315-347.
Figueroa, H.A. and D.B. Olson, 1989. Lagrangian statistics in the South Atlantic as derived from SOS and FGGE drifters. Submitted to J. Mar Res.
Grundlingh, M.L. 1978. Drift of a satellite-tracked buoy in southern Agulhas Current and Agulhas Return Current. Deep-Sea Res., 25: 1209-1224.
Halpern, D., S.P. Hayes, A., Lectmaa, D.V Hansen and S.G.H. Philander. 1983. Oceanographic observations in the 1982 warming of the tropical Pacific. Science, 221: 1173-1175.
Hansen, D.V. and C. Paul, 1984. Genesis and effects of long waves in the equatorial Pacific. J. Geophys. Res., 89(10): 431-440.
Hellerman, S. and M. Rosenstein, 1983. Normal monthly wind stress over the world ocean with error estimates. J. Phys. Oceanogr., 13(7): 1093-1104.
Hofmann, E.E., 1985. The large-scale horizontal structure of the Antarctic Circumpolar Current from FGGE drifters. J. Geophys Res., 90, 7087-7097.
Imawaki, S. et al., 1988. A new method for estimating the turbulent heat flux at the bottom of the daily mixed layer. J. Geophys. Res., 93, 14,005-14,012.
Krauss, W. and R. H. Käse, 1984. Mean circulation and eddy kinetic energy in the eastern North Atlantic. J. Geophys. Res., 89(C3): 3407-3416.
Krauss, W. and C.W. Böning, 1987. Lagrangian properties of eddy fields in the northern North Atlantic as deduced from satellite-tracked buoys. J. Mar. Res, 45, 259-291.
Large, W.G., J.C. McWilliams and P.P. Niiler, 1986. Upper ocean thermal response to strong Autumnal forcing of the Northeast Pacific. J. Phys. Oceanogr., 16: 1524-1550.
Large, W.G. and H. van Loon, 1989. Large scale, low frequency motions of the 1979 FGGE drifting buoys and winds of the Southern Hemisphere. Accepted by J. Phys. Oceanogr.
Levitus, 1982. Climatological atlas of the World Ocean, NOAA Professional paper 13, U.S. Dept. of Commerce, 173 pp.
McNally, G.J., 1977. Trapping of two drifting buoys by an ocean eddy. Deep-Sea Res., 24, 1203-1209.
McNally, G.J., W.C. Patzert, A.D. Kirwan, Jr. and A.C. Vastano, 1983. The Near-Surface Circulation of the North Pacific Using Satellite Tracked Drifting. J. Geophys. Res., 88(C12): 7634-7640.
McPhaden, M.J., 1982. Variability in the central equatorial Indian Ocean, Part II: Oceanic heat and turbulent energy balances. J. Mar. Res., 40: 403-419.
McPhaden, M.J., G. Reverdin, J. Merle, Y. DuPenhoat and A. Kartavtseff, 1984. Objective analysis of simulated equatorial Atlantic Ocean data on seasonal time scales, Deep-Sea Res., 3(5): 551-569.
Molinari, R.L. 1981 Observations of near-surface currents and temperature in the central and western tropical Atlantic Ocean. J. Geophys. Res., 88(C7): 4443-4438.
Muller-Karger, F.E., C.R. McClain and P.L. Richardson, 1988. The disposal of the Amazon's water. Nature, 333, 56-59
Niiler, P.P., R.E. Davis and H.J. White, 1987. Water-following characteristics of a mixed layer drifter. Deep-Sea Res., 34, 1867-1881.
Niiler, P.P., P.-M. Poulain and L.R. Haury, 1988. Synoptic three dimensional circulation in an onshore-flowing filament of the California Current. In press Deep-Sea Res.
Okubo, A., 1970. Horizontal dispersion of floatable particles in the vicinity of velocity singularities such as convergences. Deep-Sea Res., 17: 445-454.
Olson, D.B., and R.H. Backus, 1985. The concentrating of organisms at fronts: A cold-water fish and a warm-core Gulf Stream ring. J. Mar. Res., 43(1): 113-137.
Olson, D.B. and R. Evans, 1986. Rings of the Agulhas Current. Deep-Sea Res., 33(1): 27-42.
Olson, D.B., G.P. Podesta, R.H. Evans and O.B. Brown, 1988. Temporal variations in the separation of Brazil and Malvinas Currents. Deep-Sea Res., 35(12), 1971-1990.
Olson, D.B., 1988. Surface circulation in the South Atlantic. ONR, South Atlantic Accelerated Research Initiative (ARI), South Atlantic Section, SAARI Meeting Report, May 24-26, 1988, Lamont-Doherty Geological Observatory, 5 (invited abstract).
Patterson, S.L., 1985. Surface circulation and kinetic energy distributions in the Southern Hemisphere oceans from FGGE drifting buoys. J. Phys. Oceanogr., 15, 865-884.
Poulain P.-M. and P.P. Niiler, 1988. Statistical analysis of the surface circulation in the California Current system using satellite-tracked drifters. In press Deep-Sea Res.
Richardson, P.L. 1980. Gulf Stream trajectories measured with free-drifting buoys. J. Phys. Oceanogr., 10, 90-104.
Richardson, P.L. 1983. Eddy kinetic energy in the North Atlantic from surface drifters. J. Geophys. Res., 88(C7): 4355-4367.
Richardson, P.L. and G. Reverdin, 1987. Seasonal cycle of velocity in the Atlantic North Equatorial Countercurrent as measured by surface drifters, current meters, and ship drifts. J. Geophys. Res., 92, 3691-3708.
Rossby, H.T., S.C. Riser and M.J. Mariano, 1983. Western North Atlantic - Lagrangian viewpoint. In Eddies in Marine Science, A. Robinson Editor, Springer-Verlag, New York.
Schmitz, W.J., Jr, 1978. Observations of the vertical structure of low frequency fluctuations in the western North Atlantic. J. Mar. Res., 36,, 295-310.
Talley, L.D. and R.A. de Szoeke, 1986. Spatial fluctuations north of the Hawaiian Ridge. J. Phys. Oceanogr. 16(5), 981-984.
WCRP-11 World Ocean Circulation Experiment - Implementation Plan - Scientific Background (Volume II).
WCRP-12 World Ocean Circulation Experiment - Implementation Plan - Scientific Background (Volume II) (WMO/TD - No. 243).
WCP-62. Report of the First Session of ISC/CCCO Toga Scientific Steering Group (Hamburg, 25-27, August 1983).
PARTICIPANTS: WOCE/SVP WORKSHOP, AOML, MIAMI
APRIL 25, 1988
PAN-PACIFIC
Table of Contents
3. The first hypothesis - Heat advection by the tropical Pacific surface circulation
4. The second hypothesis - The vertical turbulent flux
5. Tropical circulation model and data intercomparisons
6. Regional studies of circulation
7. The Lagrangian Drift Experiment
In the tropical Pacific the large-scale interactions between the ocean and the atmosphere are well documented. The sea surface temperature (SST) provides the most direct link in modeling the oceans' influence on the atmosphere. Thus, a principal scientific objective of the Tropical Ocean and Global Atmosphere Programme (TOGA) of the World Climate Research Programme (WCRP), is to accurately measure SST and to correctly model its seasonal and longer time-scale evolution. The two competing hypotheses on the causes of the seasonal and interannual variations in the tropical Pacific are that either surface heat advection or near-surface vertical mixing change the SST. Presently there is a lack of direct measurements to assess the relative role of each process, as well as measurements of SST to the accuracy (± 0.3°C) required for air-sea interaction model improvement in TOGA. Hydrodynamic models of the ocean also require circulation data for verification and model initialization. This document outlines an international programme to measure surface circulation, to assess the strength of vertical mixing and provide accurate in situ data for improved SST analyses in the tropical Pacific.
To accomplish these scientific objectives, it is proposed to deploy a pan-Pacific array of low cost, ARGOS tracked mixed layer drifters with temperature sensors for a three-year period. Drifters would be deployed from research vessels and the TOGA Volunteer Observing Ships Program (VOSP) in the area 15° N - 15° S, 80° W to 126° E. The desired spatial sampling density is 2° latitude x 10° longitude which can be accomplished by 230 resident drifters. A total of 450 drifters will be released over a 2-year period. Required renewal rates of the array will be determined empirically. The drifter water-following capability will be ± 2cm/sec in 7 m/sec winds, and accuracy of temperature measurements will be ± 0.1°C. The SST data will be reported in real time to the Global Telecommunications Systems (GTS), and the circulation data will be used in real time to enhance and verify models of El Niño prediction. Heat advection and mixing at the base of the daily thermocline will be computed. A variety of regional ocean circulation phenomena will be measured and described. By 1990, recommendations will be made to WCRP on the implementation of a joint tropical surface circulation study for TOGA/WOCE (World Ocean Circulation Experiment).
In TOGA, the ocean subsurface thermal structure and sea level projects have been implemented in a pan-Pacific basis. This proposed study will implement the measurement of near-surface currents in an equivalent pan-Pacific basis.
The most direct influence of the ocean on tropical atmospheric circulation is due to the variability of the sea surface temperature (SST). Hydrodynamic models of this variability demonstrate that in the equatorial Pacific the west-to-east SST gradient maintains the rising air over the Indonesian archipelago, and subsiding air over the eastern basin (Stone and Chervin, 1984). Typically, in August-January the SST gradients are the largest and in February-May, the gradients are the weakest. Substantial spatial and temporal variability of this Walker circulation in the atmosphere occurs during times when the regular seasonal east-to-west SST gradient vacillation is disrupted. During the years when cold water does not appear in the eastern equatorial Pacific and the entire basin is warm, significant world-wide deviations in normal climate patterns are observed. The warming event in the 1982-1983 period was accompanied by the largest and most devastating shifts of tropical rainfall and mid-latitude storminess in modern times (ITPO, 1985). The Tropical Ocean and Global Atmosphere programme (TOGA) has been organized to develop a deeper understanding of the physical basis of this coupled ocean-atmosphere variability, and to construct models which may lead to its prediction. The principal scientific objectives in the ocean portion of the TOGA programme are to accurately measure the SST, and to correctly, model its seasonal and longer time-scale evolution (WCP-89).
Sea surface temperature variations are caused by the horizontal movement of water, vertical mixing, and by heat exchange with the atmosphere. To assess the relative importance of these processes and to ascertain whether these are resolved in models in a realistic fashion, requires a comprehensive data set on surface water movements, mixing and heat exchange. TOGA Science Steering Group (WCP-130) has recommended an experiment in the Pacific to acquire a surface circulation data set, and this proposal outlines a programme to meet this objective. The proposed measurements programme is also designed to assess the intensity of vertical mixing at the base of the daily thermocline. Concurrently the TOGA Heat Exchange Project (THEP) will monitor surface heat exchange (Liu and Niiler, 1985).
Two hypotheses have been advanced as to why cold water occasionally does not surface in the eastern tropical Pacific in the September-January period and a warm SST episode results. The first is that a major disruption of the normal surface circulation occurs. Gill (1983) and Reed (1983) showed that the surface warming could not be caused directly by an increase of observed surface heat flux in the central equatorial Pacific during the 1972 warming event. The hot water, therefore, could have accumulated there by horizontal movement. Gill shows how the eastward heat advection due to Kelvin waves generated by the collapse of the easterlies in the western equatorial Pacific could accomplish an increase of the observed thermal energy in the central Pacific. While no direct observations exist of the surface current affecting advective warming during the 1972 El Niño as postulated in Gill's computation, evidence of such motion exists in the equatorial current measurements during the 1982-83 event (Halpern et al., 1983): zonal equatorial surface currents were reversed and the undercurrent briefly disappeared. Indirect evidence for large-scale horizontal water movements during warm episodes also comes from sea level observations. Rebert et al. (1985) show that monthly mean island sea-level variations in the tropics are an excellent measure of the steric level and heat storage. The steric level changes also mirror the depth of the main thermocline. Sea level decreases over broad areas of the western Pacific during warm episodes, in many areas this is due to the rising thermocline. Thus, large volumes of warm water must flow to the east (or poleward) accompanied with rising sea level and deepening thermocline (Pazan and White, 1987). That large-scale changes of circulation must accompany these sea level or thermocline changes is obvious, but only a few measurements have been made which can quantify these (see Section 3), and thus test the warm anomaly advection hypothesis in a quantitative fashion. The evidence of such change during the 1982-83 El Niño event in data from both moored current meters and drifting buoys is available only in the equatorial and South equatorial current in the eastern Pacific (Halpern et al., 1983). The SST anomalies have much larger north-south and east-west scales than existing current measurements.
The second hypothesis of eastern tropical Pacific warming has been succinctly stated by Newell (1986). Newell notes that the maximum surface temperature of the eastern Pacific during warm events is nearly identical to the western Pacific - or 30°C. Thus, he argues that local air-sea interaction, or the balance between evaporation and solar radiation must determine the approach to the warm conditions. This occurs in the east when the local, vertical supply of cold water and/or vertical mixing has been severely reduced. Unfortunately, this hypothesis is difficult to test because vertical motion or vertical mixing rate are difficult to measure over large areas, and estimation of oceanic heating rates requires measurement of a number of noisy meteorological parameters. However, it is a plausible hypothesis and one, as we will show (Section 4), which may be tested with recently developed techniques of analysis of the structure of the daily thermocline.
In summary, a large-scale advection of warm water to the east or a local decrease of vertical mixing have been proposed as mechanisms which contribute significantly to the development of warm episodes in the central and eastern Pacific. Herein we propose a pan-Pacific Lagrangian drift experiment as a contribution toward testing these warming hypotheses and, further, to provide a verification data set for numerical models of the Pacific basin surface circulation, and new data for an improved SST analysis.
3. The first hypothesis - Heat advection by the tropical Pacific surface circulation
Ship drift and satellite-tracked buoy drift provide the only comprehensive data set of the
large-scale surface circulation of the tropical Pacific. Figures 1a and 1b display the historical
ship drift mean velocity and the EPOCS/NORPAX satellite-tracked buoy velocity, respectively. Both
diagrams reveal the large zonal current systems - South Equatorial Current (SEC) between 10°S
and 3°N, the North Equatorial Countercurrent (NECC) between 3°N and 8°N, and the North
Equatorial Current (NEC) north of 8°N. They also show a northward drift north of the equator, and
a more subtle southward drift south of the equator. The component of surface flow parallel to the
temperature gradient advects heat, and it is this component which we seek to determine during central
and eastern Pacific warming episodes. The historical mean temperature gradient is depicted together with
the historical mean ship drift on Figure 2. It is apparent from this figure that surface heat advection
is due to both zonal and meridional components. It is known that a measure of the zonal component of
subsurface flow in the tropical Pacific can be obtained by the geostrophic method applied to subsurface
temperature profiles, or from sea level gradients. The surface geostrophic flow relative to 1000 db is
equatorward everywhere in this area (Wyrtki, 1975, and Figure 12). The actual flow, which is poleward
and driven by the wind, is opposite the direction of the geostrophic flow. Thus, for surface heat
advection determination, accurate direct flow measurements are required because indirect methods
(e.g., geostrophic methods) would yield a flow (and heat advection) of the wrong meridional sense
over large areas of the equatorial Pacific. Figure 3 is a map of the mean horizontal heat advection
of 
(SST). It is comparable in
magnitude and of same sign as the surface heat flux divergence in a 50 m deep mixed layer (Weare
et al., 1980) There are not enough data to construct this field during the warming episodes. Neither do
we know how accurate such a ship-drift derived map is. A program of direct measurements would allow
us to construct the fields of SST and
and hence compute the horizontal advection.
Figure 1a. Annual mean historical ship drift velocity in the tropical Pacific averaged on 2° latitude x 5° longitude scale.
Figure 1b. Average surface current vectors in Central and Eastern Tropical Pacific from EPOCS drifter data (Hansen and Paul, 1984).
Figure 2. The annual mean ship drift (arrow) and SST gradient (bar) in the tropical Pacific on 4° latitude by 10° longitude resolution.
Figure 3. The annual mean horizontal heat advection in the tropical Pacific computed from data in Figure 2.
Figure 4. Meridional profiles of eddy flux of momentum and heat by surface currents (Hansen and Paul, 1984).
Horizontal flow patterns are not steady and the time variable components can also transport heat. Hansen and Paul (1984) have dramatically demonstrated how eastern tropical Pacific "eddies" of 25-30 day periods (or "21-day" waves) transport heat toward the equator. Thus, depending upon over what time scale (or space scale) a mean field of horizontal advection is defined, there also can be a corresponding horizontal eddy heat convergence. The eddy transport is due to motions of spatial and temporal scales that are smaller than those of the mean motion and mean temperature. Figure 4 is the horizontal eddy momentum and heat fluxes measured by drifters in eastern Pacific in the 1978-1982 period. Convergence of the eddy heat flux, in the absence of mean advection and vertical diffusion, would cause a rise of SST of 2.5°C per month near the equator. Figure 4 demonstrates that an understanding of both low frequency (seasonal) and "eddy" (high frequency) heat convergence processes are required. The large-scale temperature gradients and surface flows are seasonally dependent, as are the mesoscale or "21-day" waves which carry the eddy heat fluxes. During the 1982-83 warming these waves were much reduced and so would be the eddy heat flux. Within the existing TOGA thermal measurement program (USTOGA 3, 1984), the eddy processes are not resolved by the large-scale geostrophic flow, thus direct measurements of flow which resolve the spatially-varying statistics of the 10-30 day period fluctuations are also essential to testing the horizontal advection hypothesis.
4. The second hypothesis - The vertical turbulent flux
The detailed discussion of the vertical mixing hypothesis is best done in the context of the surface heat balance equation. In the usual hydrodynamic notation, the heat content evolution of the very near surface layer of depth h
. (1)
The net surface heat flux is Qo, the sum of the turbulent flux and radiative fluxes; at depth z = -h,
these are Q-h. The vertically-integrated temperature is 
Tdz
and 
is the horizontal velocity averaged over depth h. We choose h to
be shallower or equal to the depth of the daily thermocline, so that 
,
<T> is the change of heat content of the daily thermocline. The amplitude of the daily cycle of
SST is between 0.2°C and 1.0°C, and the daily heat wave seldom penetrates deeper than 20 m
(although daily variations have been measured at some locations at 30 and 40 m depths). In the above
expression, we have neglected the effect of vertical advection near the surface, where w=0 and the heat
flux convergence due to 
, where and T'
are the deviations of the velocity and temperature from their respective vertical means (this latter
can be shown to be a valid neglect of terms compared to the retained terms from moored array
observations in the tropical Pacific [Imawaki et al., 1987]). Following the vertically averaged
motion, 
is the Lagrangian rate of heating.
The hypothesis expressed by Newell (1986) states that the warm episodes are approached during the condition where, on a monthly time scale, the net flux Qo ~ 0 , d<T>dt ~ 0 and the net turbulent flux and radiative flux at the base of the daily thermocline Q-h, is much reduced from its usual value. That such conditions cannot occur on a daily basis is obvious, and a detailed analysis of the daily cycle can be shown to offer a technique evaluating the monthly average of Q-h, the turbulent flux at the base of the daily thermocline.
Consider now a measurement which can resolve the daily cycle of heating and cooling of the upper 15-20 m of the water column following its mean motion. Following Imawaki et al. (1987), integrate equation (1) from local time 6:00 to 18:00, and again from 18:00 to 6:00, the morning of the next day. The heat budget equations are, with D designating the day time and N the night time averages,
respectively,
Now subtract the nightly average from the daily average, thus
Imawaki et al. (1987) show that since the latent, sensible and long-wave radiation components of surface heat flux do not change appreciably through any one 24-hour period, and since most of the turbulent flux occurs at night (Moum and Caldwell, 1985), equation (3) can be rearranged to give,
where the overbar denotes a monthly ensemble and Qshort is the monthly mean, short-wave solar flux.
Equation (4) was used by Imawaki et al. (1987) to estimate the turbulent flux at the base of the daily
thermocline in a number of locations where accurate high vertical resolution temperature time series
have been collected in the equatorial Pacific in the 1983-85 period. Figure 5 displays a 15 month time
series of
from 120°W, 0°N. This estimate agrees favorably
with the turbulent flux estimated by microstructure measurements: it is between 70w/m2 and 160w/m2.
Thus, if an accurate measurement can be made of the daily cycle of d<T>/dt the temperature change
following the water, an estimate of the space and time variability of the turbulent flux can be made,
and the second hypothesis of warm anomaly can be tested is obtained to an estimated accuracy of
± 10 w/m 2 within other TOGA-supported programs). This estimate is essential to testing the hypothesis
of the manner in which SST patterns are produced and as shown in Figure 5, significant long period
changes do occur.
5. Tropical circulation model and data intercomparisons
One of the principal TOGA objectives is to model the ocean circulation with sufficient realism that the SST fields can be predicted accurately over the tropical oceans. Presently, the best success is achieved in modeling the time variability of the depth of the main thermocline or sea level (Katz and Witte, ed, 1987). This success can be reported because there have been extensive observations of the thermal structure and sea-level in the tropical Atlantic and Pacific, by which intercomparisons can be made. The best agreement is in the area of ±5° from the equator. It appears that only low vertical resolution adiabatic changes of ocean-structure, forced with a realistic wind field, are required to achieve this agreement.
The prediction of SST, in contrast, requires realistic modeling of both surface circulation and vertical mixing. Horizontal mixing processes can be explicitly resolved to a much better degree of realism than vertical mixing by eddy-resolving models (Katz and Witte, ed, 1987). Firstly, vertical mixing directly affects SST prediction because heat is directly transferred from warm surface waters to cold water at deeper levels. Secondly, for a fixed wind field the strength of surface currents, and hence the strength of SST advection, depends crucially on how vertical mixing is parameterized in a model. If vertical mixing is strong, momentum imparted by wind is distributed over a deep column and weak surface currents result. Weak vertical mixing results in strong surface advection. In present thermodynamically active models, vertical mixing and mean surface circulation shear are also inexorably related through a local gradient Richardson number parameterization. Local winds are expected to produce a large portion of the local shear and local heating is expected to produce a large fraction of the upper ocean vertical
stratification. Thus, verifying the surface circulation in a model against data would be both an indirect test of the adequacy of mixing parameterization as well as the realism of the local air-sea fluxes supplied at the surface. Verification of these conditions at one or two points, as can be done from current meter data on the equator (which continues to be available in TOGA) is not sufficient because not enough degrees of freedom exist in the few profiles of currents or stratification to form crucial tests: discrepancies could be due to either mixing or air-sea interaction inadequacy. A pan-Pacific field of surface currents under a large variety of air-sea interaction conditions is required to span a sufficiently wide parameter range of heating, wind-stress and Coriolis force to form critical tests.
With the Equatorial Pacific Ocean Climate Study Program (EPOCS/NOAA), investigators are regularly producing operational model-generated ocean surface current fields from the Pacific. One of the important uses of a well-calibrated and well-resolved surface current data set is the evaluation of the quality of model-generated fields. It is expected that Climate analyses Center (CAC/NOAA) will continue running the Geophysical Fluid Dynamics Laboratory (GFDL/NOAA) model of the tropical Pacific on an experimental operational basis for the next three years. Tests of the use of the thermal and density data acquired in TOGA in initialization and verification of the model are underway. Scientists at the University of Miami are presently carrying out a comparison of modeled surface current maps with EPOCS surface drifter motion, and will propose to extend this work with the more extensive pan-Pacific data. Scientists at Atlantic Oceanographic and Meteorological Laboratory (AOML-NOAA) are developing methods for assimilation of drifter data in models.
Besides use of surface circulation in operational models, TOGA research mode models require evaluation. Research models run with more complete forcing fields than operational models, because post-GTS surface marine observations have gone into producing the fluxes of heat and momentum. For example, research mode models are integrated forward in time with a number of different vertical and horizontal resolutions and different parameterizations of vertical and horizontal mixing. The research objectives are to use the best surface forcing fields and under-stand what internal model parameters are most crucial in producing realistic surface currents and realistic surface heat advection - or what ocean process most crucially affects SST prediction.
In summary, the principal scientific objective of the model-data intercomparisons is to test existing operational and research model surface currents against observations, and learn how to use surface current and SST observations to enhance the model's ability to both advect heat in a realistic fashion, as well as to produce more realistic SST maps.
6. Regional studies of circulation
The pan-Pacific scientific objectives are to test the competing hypotheses of heat advection and mixing in the surface layers of the tropics. Model intercomparisons with observations will be made. These data sets on ocean velocity and surface temperature, however, are also important for describing and studying a number of more local oceanographic circulation phenomena. To address these in detail, the tropical Pacific basin is divided into eastern, central and western parts. Specific scientific proposals for data analysis and interpretation in these areas will be submitted to regional national agencies and comprise part of the scientific work statements from the principal investigators. A summary of the regional scientific questions and objectives are given below.
6.1.1. Heat advection in the "Cold Tongue"
A most salient feature of the sea surface temperature of the tropical Pacific is the development of a region of relatively cool surface water extending from the coast of South America to and along the equator to 150°W or farther during most Austral winters. In El Niño events, this cold tongue fails to develop in the usual way. A study of variations of the heat budget on annual and interannual time scales has been done using drifter data from EPOCS (Figure 6). This study indicates that zonal advection is more important than meridional advection in the region, but this result is limited by sparse data to an area integral result. Finer resolution made possible by more data might yield a different picture. Other results with EPOCS drifter data indicate that meridional advection is of major importance close to the equator (Fig 7).
6.1.2. Equatorial divergence
Closely related to meridional advection near the equator is the divergence of Ekman transport. Wind-driven surface current theory indicates, and drifter data from the EPOCS and NORPAX programs confirm, the existence of the divergence of surface currents near the equator under normal wind regimes. Using EPOCS drifter data, it has been possible to estimate climatological divergence and upwelling in the eastern Pacific (Figure 8). Given adequate data, such estimates could be made on an event and time series basis. Present evidence suggests that during El Niño and related warm events, the surface equatorial divergence is much reduced.
6.1.3. Shear-instability of waves
From June to October of most years, instability of the shear between the EUC and that part of the SEC lying north of the equator or the shear between NECC and SEC leads to generation of cusp-shaped moves on the equatorial SST front lying north of the equator. The principal feature of the surface currents associated with these waves ia s train of anticyclonic eddies, primarily in the Northern Hemisphere, with eddy velocities of up to 100 cm/s. Hansen and Paul (1984) found that these eddies are instrumental in effecting an equatorially convergent heat transport in the surface layer, counteracting a substantial part of the mean heat transport divergence by the equatorial upwelling and meridional advection (Figure 7). TROPIC HEAT program is proposing an intensive study of the instability waves with current meter moorings and other instrumentation in 1989-1990. If that study is funded, the pan-Pacific drifter project will provide information about the large-scale environment in which it is done, and the statistics of this eddy field over a number of years. Beyond that, drifters have been found to be a particularly sensitive detector of the occurence of these waves and, more emphatically, of their non-occurence. The occurence of the waves west of about 150°W is undocumented.
6.1.4. Eastern Pacific Warm Pool
Although most references in the TOGA context are to the western Pacific "Warm Pool," the eastern Pacific contains a secondary warm pool. This warm pool, extending seaward a few thousand kilometers from the coast of Ecuador or Colombia to Mexico, contains surface waters as warm (28-30°C) as those of the western Pacific, but more limited in extent, and especially in depth. It is therefore more prone to variations. One view of El Niño development is that it is a major expansion of the eastern Pacific warm pool, expecially to the south and west. The region of this warm pool is also the generation region for eastern Pacific hurricanes, and it is a major pelagic fishing area. The processes that affect the maintenance and variations of this warm pool are an important part of TOGA in the Pacific sector.
Figure 6. Lagrangian net heating rate estimated from EPOCS drifting buoy and XBT data in the region 0-10°S, 90-130°W.
Figure 7. Meridional profiles of average zonal and meridional surface currents in eastern tropical Pacific (Hansen and Paul, 1984).
Figure 8. Annual cycle of zonal (+), meridional (x), and total (-) divergence in the region ±1.5°, 80°W to 110°W. (Hansen and Paul, 1987).
Figure 9. Zonal displacements of drifting buoys in latitude band of the NECC (4° 10°N). Note the annual cycle of NECC east of 140°W. (Hansen et al., 1987).
6.1.5. The North Equatorial Countercurrent
A major source of warm water for the eastern Pacific warm pool and for the eastern Pacific generally, is the NECC. This is a major ocean current that extends one-third of the distance around the earth, and in the east undergoes a complete seasonal modulation (Figure 7). Increased focus-on the NECC is being planned within EPOCS. Drifter deployments included within the pan-Pacific projects will become an integral part of that program. The special ability of drifters to provide information about spatial patterns in surface currents will be exploited to determine the extent, speed and scales of variability in the NECC. Model results of Philander and Siegel (1985) indicate greatly enhanced flow of the NECC in the eastern Pacific during 1982, but measurements to confirm this finding are few. In the western and central Pacific, variations of the NECC can be inferred from island sea level records. In the eastern Pacific, islands are few; satellite altimetry offers some hope, but drifters are an established method of making the observations needed for understanding and testing the ability of models to simulate this current.
6.1.6. Eddies generated by coastal winds
Recent observations indicate that strong winter winds flowing through topographic lows in the central American Cordillera and across the coastline produce an unusual kind of coastal upwelling, especially in the vicinity of the Gulf of Tehuantepec and Gulf of Papagallo, and even in the Gulf of Panama. The upwelling disturbances evolve into eddy structures that can endure for many months and appear to propagate to the west. Freshly-upwelled water is readily detectable in satellite IR imagery, but within relatively short time the surface temperature signature of the eddies is lost beneath the eastern Pacific Warm Pool. Such eddies have been tracked with surface drifters over several months and 3000 km from the region of generation (Figure 10). These eddies are possibly a major source of variability in the North Equatorial Current to the west. Drifters appear to be a most appropriate tool for determining their occurrence and distribution.
6.1.7. Currents off the coast of Peru and Ecuador
Most surface current charts of the ocean indicate a Humboldt Current or Peru Current flowing northwestward generally parallel to the coast of Peru. This current is indicated to become the South Equatorial Current westward of about 100°W. Curiously no support of the existence of this current can be found in the tracks of EPOCS drifters. Results to date from EPOCS suggest that the currents off Peru and Ecuador are weak and variable, but have a slow westward or even southwestward drift. A distinct South Equatorial Current is encountered well offshore. The pattern of surface currents in this regions is of considerable importance in evaluating the possible influence of the Peruvian coastal upwelling on the eastern Pacific cold tongue.
6.2.1. SST anomalies
During ENSO events, the normal seasonal development of the SST patterns across the entire Pacific are disrupted. In the central Pacific, these anomalies are the most effective in coupling to the atmosphere, because these SST changes can effectively "trigger" the onset of atmospheric low-level convergence of moisture and upper-level heating by enhanced precipitation. The east-west movement of the SST anomaly pattern is described in Figure 11, which shows the 28.5°C isotherm location in the equatorial band. If this movement is produced by currents, a net flow to the east - or at least significant weakening of the normal westward flow - has to occur. Such anomalous patterns have not yet been observed over the broad central Pacific. The pan-Pacific network offers the opportunity of obtaining a comprehensive description of these and other patterns important in ocean-atmosphere interactions.
Figure 10. Mesoscale eddy near 10°N and possibly also near 5°N generated by coastal wind events (Hansen, private communication, 1987).
Figure 11. Longitudinal movement of the 28.5°C SST surface between 4°N and 4°S. El Niño events occur when the 28.5°C surface crosses 160°W (Diaz, 1987, private communication).
6.2.2. Equatorial mixing
East of 140°W, a principal source of surface heat transport and variability is produced by "21-day" waves. In the central Pacific, a lower frequency energy band of 40 to 60 day oscillations is evident from current meter measurements (Knox and Halpern, 1982). The horizontal eddy energy in the central tropical Pacific is unknown. The importance in mixing momentum and heat by the tropical mesoscale has already been underscorerd in the development of the eastern cold tongue, and such an evaluation must also be made in the central Pacific. Circulation measurements across the Pacific can provide a direct assessment of the importance of eddy mixing in the entire equatorial zone and an effective test can be formulated as to whether numerical models parameterize or resolve this mixing correctly.
6.2.3. Equatorial divergence
Easterly winds are equally strong across the eastern and central Pacific. The eastern Pacific exhibits what is believed to be wind-driven equatorial divergence, which results in the upwelled cold tongue. The manifestations of the cold tongue in the central Pacific are not as obvious, yet the surface divergence is presumed to exist due to the strong easterly winds. There are virtually no surface circulation measurements from which to compute or describe a surface divergence pattern.
6.2.4. The wind-driven currents
The ship meteorological observations in the central Pacific are the most dense because the merchant ship traffic between North America and Australia transits this region. In this area of the ocean, the most realistic test of the tropical wind-driven ocean currents can be made because the winds are the best resolved, as are the geostrophic currents by VSOP-XBT measurements. The geostrophic currents north of the equator imply a southward surface flow toward the equator, while historical ship drifts and EPOCS buoy motion imply a northward component (Figure 12). Thus, the wind-driven meridional component of the surface flow in the central Pacific must be larger and of the opposite sign than the geostrophic flow. Drifter movements can describe the strength of the surface circulation, and subtraction of the geostrophic flow estimates produced by TOGA thermal analysis will produce a picture of the wind-driven currents.
6.3.1. The large-scale circulation
The circulation of the western Pacific is comprised of the confluence of a number of extensive current systems. The SEC runs northwestward along the north coast of New Guinea, may cross the equator and near 5°N, 130°E, veers offshore or retroflects into the NECC. The SEC disappears during the period November to April and is replaced by the southeastward flowing New Guinea Coastal Current. This eastward flow appears to be connected in ship drift maps to the SEC which extends from the Solomon Islands east to near 155°E. The westward flowing NEC divides at the western boundary near 12°N; part flows northward into the Kuroshio and part flows southward as the Mindanao Current. Part of the Mindanao Current flows into the Celebes Sea, turns counterclockwise, and flows out through the Molucca Sea merging with the other part which veers offshore into the NEC near 5°N, 130°E. The confluence zone of the Mindanao Current and SEC at 3°N forms the origin of water flowing eastward in the NECC which crosses the whole Pacific Ocean The currents of the western equatorial Pacific are very swift, have enormous transports (e.g., 40 x 10**6 m3/sec for NECC) and large seasonal and interannual variations.
Figure 12. The geopotential anomaly contours (with arrows) relative to 1000 db for November-December (Wyrtki, 1975), and typical track of an EPOCS drifting buoy. [Hansen, private communication, 1987]. The buoy is deployed on the equator at 85°W and travels northward and westward to 11°N, 154°W in 320 days (last 160 days without a drogue). (Dots mark 20 day displacements.) The implied 1000 db reference velocity is to the south and buoy drift is to the north.
Figure 13. Schematic diagram of surface currents in Western Tropical Pacific based on ship drift measurements (Schott, 1939). Top panel shows June-Aug and bottom panel shows Jan-March periods.
Figures 13 displays this circulation schematically in the opposing seasons of the monsoon.
Most information concerning the surface circulation comes from historical ship drifts. Although these are adequate to schematically sketch the major currents, observations are too scanty to reveal much about the continuity of the currents, or their variations (mesoscale, seasonal, interannual). Thus, these currents are poorly measured and our knowledge of them is seriously deficient. The interaction of western boundary currents and equatorial circulation is unknown. What is the continuity between western boundary currents and the NECC and the Indonesian throughflow? The Mindanao Current and SEC cause a confluence in the west. What is the pattern of their retroflection into the NECC and its subsequent connection to the NEC?
Recently a few drifting buoys have been tracked in the equatorial western Pacific (Cresswell, Lindstrom, Hansen, personal communication). These have revealed interesting features of the flow such as the reversal of the SEC north of New Guinea (Figure 14), and a cyclonic gyre consisting of the Mindanao Current, NECC and NEC (Figure 15). The results to date suggest that because of the complex spatial patterns, drifters - due to their combined space-time sampling characteristics - would be an effective tool in learning about general circulation features and swift western Pacific systems.
Specific goals are to measure, describe and understand the western boundary and equatorial circulation in the western Pacific.
6.3.2. Cross-equatorial flow
The confluence in the western Pacific yields considerable cross-equatorial transport. The seasonal reversal of the SEC along northern New Guinea and along the Solomon Islands causes the cross-equatorial flow also to vary seasonally. The questions to be addressed include: What are the patterns of this flow at different seasons? Where does flow in the New Guinea Current and SECC go? Is there a time mean flow crossing the equator in the West?
6.3.3. Western Pacific warm pool
A large pool of warm water exists in the western Pacific. Eastward advection of this pool in an anomalous fashion is hypothesized to be an important part of the ENSO mechanism (see Figure 11). The north and south equatorial currents advect warm water into the pool; the Kuroshio NECC, SECC, and Indonesian Sea's throughflow advect water out of it. How is this complex circulation arranged so the warm pool is maintained in normal years, and how does it spread advected eastward during El Niño years? What is the short and long-term variability of its near-surface temperature? Present remote techniques of determining the warm pool structure (± 0.7°C) by remote techniques are not accurate enough for the TOGA requirements (-+0.3°).
6.3.5. West wind bursts and eastward jets
Westerly wind bursts with period of several weeks occur over the western Pacific. These drive an eastward jet along the equator as observed by several of EPOCS surface drifters in late 1986 and early 1987. Most drifters going west diverge from the equator; one drifter going east stayed near the equator for 2000 km (Figure 15). How far east do these jets carry the water originating in the west? What is the equatorial convergence in the surface layer in the eastward jet?
Figure 14. Trajectory of a satellite-tracked buoy in the western Tropical Pacific Ocean showing the seasonal reversal of the SEC north of New Guinea (G. Cresswell and E. Lindstrom, personal communication). This seasonal switch in current direction agrees to a large extent with historical shipdrift data for the same region.
Figure 15. Trajectories of drifting buoys deployed in western equatorial Pacific during Fall, 1986. Closed circles denote initial location, plus signs at first day or each month (Hansen, private communication, 1987).
7. The Lagrangian Drift Experiment
The scientific goal of the pan-Pacific Surface Current Study is to provide a critical test of the competing hypotheses of advection and mixing as key factors in tropical Pacific warming cycles.
The scientific objectives are to:
1. Produce accurate (± 0.3°C) maps of the monthly mean SST for atmospheric circulation modeling and climate monitoring.
2. Measure the large-scale surface transport of thermal energy across the tropical Pacific basin.
3. Describe and evaluate the role of the surface mesoscale field in transporting heat and momentum.
4. Determine the changes of heat budget of the daily cycle and its day-to-night differences. From this, using monthly mean short-wave flux data, estimate turbulent fluxes at the base of the daily thermocline.
5. Determine what are the time-dependent, wind-driven surface currents in the tropics.
6. Provide a data set for evaluating and testing numerical models of the tropical Pacific surface circulation both in real time and research mode.
Each regional area has more specific objectives to which the pan-Pacific data set will contribute. These are developed in individual proposals.
The technical objectives are to to:
1. Obtain a 2-3 year data set of Lagrangian drift at 15 m depth of the entire Tropical Pacific between 15°S and 15°N on a 2° latitude and 10° longitude resolution.
2. Obtain a selected number of 20 m daily temperature cycles.
3. Learn how to maintain a population of drifters in the desired resolution from ships of opportunity and research vessels. Develop a TOGA/WOCE plan for tropical surface current measurements for the 1990-1995 period.
4. Establish and maintain a testing and calibration program for accuracy of thermal sensors and current-following of drogues. Evaluate survivability of drogue configurations.
The Lagrangian drifters is the only practical instrument with which the surface circulation can be measured over areas as vast as the tropical Pacific. Moored current meters can provide some sparse time series; present TOGA plans are to maintain these at 165°E, 140°W and 110°W on the equator. Satellite or island sea level and thermal field can yield only the geostrophic flow which, because of the strong wind-driven component, can be opposite to the meridional surface current component derived from the historical buoy and ship drift (see Figure 12). In the past few years, the cost of Lagrangian drifters has been reduced by a factor of two or three from FGGE type drifters (to less than 3000 U.S. dollars in lots of 100 or more). These have a small surface element to reduce wind and wave loads and a large, self-deployable subsurface drogue. Model and field calibration studies indicate their slippage through the vertically-averaged flow past the drag center of the drogue can be reduced to become as little as 1-2 cm/sec in the tropical wind-wave conditions (Niiler et al., 1987). These are packaged in soluble cardboard containers which release the drifter upon coming in contact with sea water, making deployment from ships of opportunity or air deployment practicable.
In the proposed experiment, the drogue center will be at 15 m depth. Two shapes of drogues will be used: a 3.5 m diagonal TRISTAR and a long holey sock. Figure 16 shows one proposed initial specifications of each configuration. These have identical drogue to surface float and tether drag area ratios and, according to the simplest hydrodynamic scaling, should have identical water-following characteristics. The water-following characteristics of each system, however, will be assessed in drifter calibration experiments (see Section 7.4). The reason for choosing the two shapes to be manufactured by at least two different commercial and/or scientific groups is to test various techniques of construction on the survivability of drifters at sea. At the conclusion of this study, recommendations can be made as to which system (or which combination) should be used for the TOGA/WOCE objectives in the 1990 decade. Today there simply is not enough statistical information on post-FGGE, low-cost drifters, to assess their survival efficacy beyond a 9-month period. A selected number of drifters will carry a small thermistor chain, with 4 thermistors as the tether line. The thermistor chain is required for determining the daily cycle of heating and cooling, and estimation of the turbulent flux near the surface. The data sampling in the tropics will be six to seven samples per 24-hour period, every third day.
In summary, the design objectives of the drifter configurations are:
1. Slip with respect to the theoretical drag center of the drogue should be less than 2 cm/sec in tropical wind, wave and shear conditions.
2. SST-system measurement accuracy of ± 0.1°C. The SST data message should be GTS compatible. When subsurface temperature is measured, required accuracy is ± 0.04°C.
3. Drogue theoretical drag center at 15 m with less then ±3 m vertical excursion in vertical shears of 2 cm/sec/m.
4. Drogue on/off indicator should measure if tether line has severed.
5. Report of the entire daily data set from each third day, is required.
The choice of a 2° x 10° scale for the resolution of surface currents is not entirely rational over the entire tropical Pacific as we do not know what the coherence scales of Lagrangian surface currents are. The computation of the actual coherence scales of surface currents is one of the objectives of the study so a more rational plan can be made for the 1990 decade. The guidelines are, of course, the coherence scale derived from subsurface temperature measurements, which on El Niño time scales are larger than 2° x 10°, but on seasonal time scales are smaller (USTOGA-3). In the eastern Pacific "21-day" eddy field, the coherence scale depends both upon direction and time because these eddies propagate westward.
Figure 16. Schematics of ARGOS tracked, mixed layer drifters to be used in this experiment.
The choice of the accuracy of current following capability (2 cm/sec) is based on the requirement of obtaining 10% accurate advection in a 20 cm/sec current system. The choice of accuracy of 0.1°C in SST is based on the requirement of obtaining temperature gradients on 2° x 10° resolution to 10% accuracy in the central Pacific. The accuracy of ± 0.04°C for subsurface temperature measurements is so the daily cycle of 0.4°C can be resolved to 10%, and daily heat storage change to 60 w/m2 (or 10% of the signal).
The success of the design objectives will be assessed with field dat over the lifetime of this study. A number of drifter calibration experiments are planned to ascertain the absolute performance of all proposed systems in the field. A number of recoveries at sea from TOGA research vessels are planned to inspect progressive drogue deterioration patterns which do not register on a drogue tether on/off indicator.
7.3 The drifter calibration experiment
From the few field calibration studies which have been made in the mid-latitude conditions, we have found that to reduce the the slip through the water, the most important parameters in drifter design are:
1) A large drag area of the drogue compared to the drag area of the tether and surface floatation elements. To keep the slip below 2 cm/sec in 10 m/sec winds, a drag area ratio of 50 is a minimum (Niiler et al., 1987).
2) The drogue must not have excessive "sailing" character. Examples of drogues which can "sail" in laboratory conditions are "window shade" drogues, "circular cylinders (wind socks)," and "holey socks" with ratio of height to width in excess of 10 (Nath et al., 1979).
3) An internal tension should be kept in the subsurface tether and drogue to reduce "kiting" in the presence of strong shears expected near the equator and keep the drogue drag center at the design specification (here ± 3 m). At the same time, a weak negative buoyancy should be passed on to the surface element to prevent aliasing of vertical to horizontal forces by gravity waves. This may require that a subsurface float and a weight at the bottom of the drogue be used.
It is emphasized that these design concepts, though based on laboratory model and numerical model studies, require thorough field evaluation. The behavior of a drogue in sheared, wind-wave conditions of the tropical Pacific cannot be reproduced in the laboratory or a numerical model very well, because of the inherent non-linear nature of hydrodynamic drag and the inability of models to simulate the complex, three-dimensional nature of the interaction of waves with the surface floats.
As part of the proposed experiment, field calibration tests of all drogue configuration used (or proposed) will be done so their performance in following water at the theoretical drag center of the drogue in the tropical conditions are quantitatively established.
A typical calibration study would consist of a deploymemt of drifters with attached neutrally-buoyant current sensors and pressure sensors on the top and bottom of the drogue. The attending vessel would record wind, wave and 2-3 m vertical resolution shear of currents between the current meters with the doppler acoustic log (or other means). The difference of flow measurement at the current sensors would also serve as a calibration for the doppler acoustic log. A number of studies in typical wind-wave and shear conditions within the tropical Pacific would be carried out. Intercomparisons among drogues released from the same site are not planned as drogues will separate quickly in the natural ocean turbulence, and their difference of performance cannot be assessed quantitatively because upon separation, these will be in different water masses of unequal vertical shear. Valid data can in principal be obtained from this kind of intercomparison, but multiple releases of different drifters will have to be made to gather sufficient statistics. Such experiments are difficult to manage in tropical open ocean conditions. Calibration studies are a requirement established by TOGA ad hoc panel on ocean drifters (WCP-103), but the accuracy requirements for this study are more stringent than recommended there. This is due to the requirements of measuring advection and vertical turbulent fluxes. A more detailed description of calibration studies is given in individual proposals.
The area of the Pacific which is of direct interest to the proposed surface circulation study spans from 80°W to 126°E, and 15° S to 15° N and it contains about 230, 2° x 10° boxes. The objective of the experiment is to build up to a sampling array of drifters on that scale through a two-year period. The design life of the Lagrangian drifters discussed in section 7.2 is two years; however, very few drifters have existed that long. Presently, 9 months to 1 year lifetimes are common (Niiler et al., 1987). Loss rates due to biofouling, capture by ship traffic and run-up on beaches are minor in the eastern Pacific, but are expected to be greater in the western Pacific because of the density of islands and local small ship traffic. Thus, the minimum number of drifters required for this study is 230, and conservative maximum is 450. The larger number is a conservative estimate based on experience gained already on typical displacements and survival rates with drifters in the tropical Pacific, and is nearly twice the sampling density we will require. Our strategy, therefore, will have to be an evolving one in which we begin with two 50-60 buoy deployments two months apart from VOSP and research vessels, and within two years work up to the desired density. Our estimate is that 350-360 drifters will be required to achieve this objective.
In understanding the deployment strategy and sequence, constant reference will be made to Figure 17 and the mean ship drift chart Figure 1. The deployment sequence will be as follows:
1) Four (4) drifters will be deployed in April 1988 along 110°W, 15°-3°S. These will drift westward and are expected to begin to fill the ship track gap between 110°W and 150°W. They are marked with squares on Figure 17.
2) Fifty three (53) drifters will be deployed in June-July (1988) period. These are marked with circles on Figure 17. They are expected to show a predominant east-west drift for the next two months, depending upon latitude.
3) Sixty (60) drifters will be deployed two months later, in Aug-Sept 1988. These generally are released at the deployment #2 locations, and are marked with triangles in Figure 17.
4) Continued releases will be made at the eastern boundary, the eastern equator and the western boundary, throughout the next eight months (total of 37 buoys). These releases are expected to be in areas where buoys are known to exit the most rapidly. Table I describes these releases.
Figure 17. Ship of opportunity tracks available for deployment of drifting buoys with approximate frequency of occupancy (top panel). The deployment of locations for drifters April 1988 (squares). June-July 1988 (circles), and Aug-Sept 1988 (triangles). Additional deployments are planned for the western boundary, equatorial and eastern boundary areas. In the second year, deployments will be made from ship routes on top panel as empty regions occur.
Table I - Deployment vessels for drifting buoys. The symbol T is time measured in months.
Ship Name | |||||
| SIO/ORSTOM Volunteer Ships: | |||||
| Pacific Islander | 2 (150°E) | 3 (170°E) | |||
| Micro Indep. | 6 (150°E) | 3 (130° - 150°E) | |||
| Cap Anamur | 3 (170°W) | 1 (165°E), 2 (170°W) | |||
| Lillooet | 1 (170°W), 1 (160°W) | 4 (165°W) | |||
| Polynesia | 8 (150°W) | 7 (150°W) | |||
| Moana Pacific | 5 (140°W) | 7 (140°W) | |||
| Mt. Cabrite* | 9 (110°W) | ||||
| Japan Tuna | 5 (125°W) | Seal Island | 7 (100°W) | ||
| AOML Volunter Ships: | |||||
| (Ore Ship) | 5 (95°W) | 4 (95°W) | |||
| (Coastal Ecuador) | 2 (85°W) | 2 (85°W) | 2 (85°W) | 2 (85°W) | |
| (Coastal Peru) | 3 (85°W) | 2 (80°W) | 3 (80°W) | 2 (80°W) | |
| Research Vessels | |||||
| MOPS R/V (NOAA) | 3 (95°W) | ||||
| Epocs R/V | 7 (Equator) | 7 (Equator) | |||
| R/V Coriolis | 9 (165°E) | ||||
| PRC R/V: | |||||
| R/V Xiangy Anghong #14 | 6 (175°E) (Nov) | 6 (West. Pac) | |||
| 6 (140°E) (Oct) | |||||
| Totals: |
Mt. Cabrite will pre-deploy 4 buoys at T = -2 (April - May, 1988) so they can float into potential data void northeast of Tahiti.
In the first year about 150 buoys will be deployed.
As stated in Section 7.1, one technical objective of this study is to learn how to deploy and maintain an array of drifters on a 2° x 10° resolution in the complicated surface circulation of the tropical Pacific. The proposed plan is based on experience gained with drifters over the past 9 years and presently it is as rational a start for a pan-Pacific array as can be made in 1988. Numerical simulations of drifters motion are not useful as currents can, and will, depart significantly from the climatological ship drift. The release strategy of 7 north-south legs separated by 20° longitude is based on the fact that drifters move about 15-20 km/day, principally in the east-west direction and in two months they will have been displaced 8°-12° in longitude. These fill the gap between the deployment lines, and two months later, another array is released to achieve the 10° east-west resolution. The desired latitude resolution scale of 2° will not be achieved until the middle of the second year.
In the second year, more judicious releases will be made, depending upon where the gaps in buoy density occur. The first year will be done by deployment of drifters from U.S. contribution, but in the second year, significant French (50-60 drifters) and Australian (10-20 drifters) will occur. Thus, by the end of two years 350-360 drifters will have been released, filling the pan-Pacific basin to the required sampling density (with expected losses of 60% from first year deployment by the end of the second year).
The proposed schedule of deployment cannot be delayed significantly if appropriate recommendations are to be made to the TOGA/WOCE for global deployment strategy by 1990.
Several agencies in various countries will participate in this project's investigations. Numerous regional as well as basin-wide objectives have been identified. It is certain that drifting buoys nominally provided from one agency or country will, over their operating lifetime, depart the regional area of interest and become of value to others. To assure uniformity in data processing procedure, and facilitate efficient access to data and their reporting, a drifting buoys data center for this study will be established at AOML/NOAA in Miami.
At the center, raw ARGOS data will be sampled on an approximately daily basis. Buoy location maps based on weekly drifts will be distributed to project participants on a regular basis and more detailed data will be made available on a dial-up system.
The magnetic tape data received from SERVICE ARGOS will be quality-controlled and interpolated to uniform intervals at AOML. These processed data will be provided to project P.I.'s and submitted to the National Oceanographic Data Center (NODC) at the end of the study. All P.I.'s will have direct SERVICE ARGOS access to the data through the experiment number and password. The P.I.'s will collaborate in publication of descriptive data reports on a yearly basis.
SST data will be disseminated via the GTS for use in operational analysis and other TOGA purposes. In monthly intervals, the velocity data will be made available to all TOGA real time ocean model projects for initialization and verification purposes. The monthly data will be reported in the Climate Analysis Bulletin and released to any country participating in TOGA through the International TOGA Program Office (ITPO). Island nations of the Pacific have expressed special interest in these data sets. Many of these data products are currently being done at AOML/NOAA for EPOCS.
The pan-Pacific Surface Current Study will be managed by a council of P.I.'s, with ex-officio representation from ITPO (to aid in rapid distribution of the data products to the international community). The P.I.'s will elect a chair-person who will report on a regular basis to the TOGA-Science Steering Group on the overall scientific results of the study. Individual P.I.'s will report to their respective funding agencies on the specific results. P.I.'s will meet at least once each year to discuss the project results and write joint papers and data reports.
Australia's participation in the experiment, through the contribution of drifters and the payment for ARGOS tracking, will begin when Dr. Cresswell and other Australian investigators successfully acquire funds from the Australian government (by fall 1988). In preparation for and in support of this activity, the experiment's principal investigators have agreed to approach the key Australian scientific policy makers informally. The approach will include communication of an outline of the experiment, the perceived benefits to Australia and the global community, and a suggested Australian role. More formal requests will be made from the ITPO for Australian participation in 1990-1995 period for the extended TOGA/WOCE studies.
In view of both the progressive lowering in price of drifters and the proposal to only obtain tracking by ARGOS on every third day, the chances of Australia's participation from late 1988 onwards are quite reasonable. Australia's intellectual contribution to the study would come through the 5+ year commitment that Dr. Lindstrom has made to WEPOCS, and through the great value that the drifter SST measurements will have for the initialization of climate models (Meyers and Nicholls). Dr. Lindstrom has conducted two cruises into the WEPOCS area and obtained 6 months of current meter records from 150°E on the equator. He has a third cruise planned for April/May 1988 in which he and Drs. Godfrey and McDougall will further the description of the circulation and examine mixing processes.
Dr. Lindstrom plans to cooperate with his U.S. colleagues to maintain a current meter mooring at 150°E for 5 years. This would naturally lead to intercomparisons with currents measured by the drifter data set. He has already compared the track of one of Dr. Cresswell's drifters with the 6-month record from the current meter at the drogue depth. The two measurement method suggested a slab-like behavior of the upper water, with currents at the current meter being reflected in the movements of the drifter as far as 15° longitude farther to the west. Drs. Meyers and Nicholls would like to access the SST data in real time via the GTS for input into climate models. The data would be compared with XBT results and derived geostrophic currents.
In summary, Australia has scientists with a strong interest in the drifter study and they will be able to scrutinize, use and interpret the data from it. The chances of participation with Australian drifters are reasonable, with a proposed start date of late 1988.
8.2 France (Bouées dérivantes TOGA [BODEGA])
In a cooperation between LODYC (Reverdin) and ORSTOM Noumea (Henin, Picaut, Delcroix), it is planned to seed the longitude 165° between 2°N and 12°S with drifting buoys. The buoys would be launched four times a year during the TROPAC or SURTROPAC cruises, starting in mid 1989.
Specific scientific interests are in the seasonal cycle of the South Equatorial Countercurrent and in the occurence of eastward jets. Complementarity between the near surface currents obtained from the buoy drifts and the other estimates of currents from current profiling or geostrophy will be investigated. The daily temperature cycle near the surface and its seasonal cycle also are of interest. Between fifty and seventy buoys could be deployed with funding requested from various French funding agencies grouped into a Climate Programme Committee (the Programme National pour l'Etude du Climat). Modeling activities (Delecluse, LODYC) are also planned, aimed at understanding the vertical penetration of energy in the upper ocean. Specific goals are to parameterize more correctly the vertical and horizontal mixing in the upper ocean of an ocean general circulation model.
Informal contacts have been made with Japan through the Japanese TOGA-SSG representative, Prof. A. Sumi, and representatives of the Research Project on the Variations of the Atmosphere-Ocean Coupled System in the Tropical Pacific, which will become the long-term TOGA project in Japan. One of the principal components of the above proposal is "observations by drifting and moored buoys." This document will serve as a basis for further, specific exchange with Japanese P.I.'s and participation is expected from Japan during the second year of the study.
The principal investigators from the U.S. are Dr. Hansen (Atlantic Oceanographic and Meterological Laboratory/NOAA), Prof. Niiler (Scripps Institution of Oceanography), Dr. Richardson (Woods Hole Oceanographic Institution), and Dr. Hwang (Rosenstiel School of Marine and Atmospheric Sciences). The principal investigators have interests in the eastern, central and western Pacific, and in data-numerical model comparisons, respectively. EPOCS TOGA has been requested to supply 112 drifters and NSF to supply 204 drifters for this study over a two-year period.
The proposed study was a focus of discussion at the U.S. TOGA Workshop in "Dynamics of Equatorial Oceans," during August 11-15, 1987, in Honolulu, Hawaii (Katz and Witte, ed., 1987). Fifty-five tropical scientists met and concluded that:
"In conjunction with sea-level and upper ocean thermal structure, direct measurements of near-surface currents are a valuable contribution in understanding large-scale circulation throughout the tropics."
In TOGA, the thermal structure and sea level projects have been implemented on a pan-Pacific basis. This proposed study will implement the measurement of near-surface currents on an equivalent pan-Pacific basis.
Halpern, D., S.P. Hayes, A. Leetmaa, D. Hansen, S.G.H. Philander (1983). Oceanographic observations of the 1982 warming of the tropical eastern Pacific. Science. 221, 1173-1175.
Hansen, D.V. and C.A. Paul (1984). Genesis and effects of long waves in the equatorial Pacific. J. Geophys. Res., 87, 10431-10440.
Gill, A.E. (1983). An estimation of sea-level and surface-current anomalies during the 1972 El Niño and consequent thermal effects. J. Phys. Oceanogr., 1313, 586-606.
Imawaki, S. (and 9 authors) et al. (1987). A new method for estimating the turbulent heat flux at the bo
Gill, A.E. (1983). An estimation of sea-level and surface-current anomalies during the 1972 El Niño and consequent thermal effects. J. Phys. Oceanogr., 1313, 586-606.
Imawaki, S. (and 9 authors) et al. (1987). A new method for estimating the turbulent heat flux at the bottom of the daily mixed layer (submitted).
ITPO (1985). The Tropical Ocean and Global Atmosphere Programme, WMO, Geneva, Switzerland.
Katz, E.J. and J.M. Witte, ed. (1987). Further progress in equatorial oceanography. Nova Univ. Press, Ft. Lauderdale, Fla.
Knox, R.A. and D. Halpern (1982). Long range Kelvin wave propagation of transport variations into the Pacific Ocean equatorial currents. J. Mar. Res., 40, 323-339.
Liu, W.T. and P.P. Niiler (1985). TOGA Heat Exchange Project - a Summary Report. The Jet Propulsion Laboratory, Pasadena, CA. No. 85-49.
Moum, J.N. and D.R. Caldwell (1985). Local influences on shear-flow turbulence in the equatorial ocean. Science, 230, 315-316.
Nath, J.H., C. Liu, E.G. Kerut and D.M. Brooks (1979). Hydrodynamic coefficients of flexible drogues. Proc. of the Specialty Conf. Civil Eng. in Oceans IV, II, 595-610. Am. Sec. of C.E. New York.
Newell, R.E. (1986). El Niño. An approach towards equilibrium temperature in the tropical eastern Pacific, J. Phys. Oceanogr. 6, 1338-1342.
Niiler, P.P.. R.E. Davis and H.J. White (1987). Water-following characteristics of a mixed layer drifter. (In press. Deep-Sea Res.).
Pazan, S.E. and W.B. White (1987). Short-term climatic variability in the volume budget of the western tropical north Pacific Ocean during 1979-82. J. Phys. Oceanograr., 17, 4, 400-445.
Philander, S.G.H. and A.D. Siegel (1985). Simulation of El Niño 1982-83. Coupled Ocean-Stmosphere Models. J.C.J. Moum ed., Elsevier, Amsterdam. 517-531
Reed, R.K. (1983). Heat fluxes over the eastern tropical Pacific and aspects of the 1972 El Niño. J. Geophys. Res., 14, 3627-3638.
Rebert, J.P., J.R. Donguy, G. Eldin and K. Wyrtki (1985). Relations between sea level, thermocline depth, heat content and dynamic heighth in the tropical Pacific Ocean. J. Geophys. Res., 90, 11, 719-11725.
Stone, P. and R. Chervin (1984). The influence of ocean surface temperature gradient and continentality on Walker circulation. Part II. Prescribing global change. Mon. Wea. Rev. 112, 8, 1524-1534.
U.S.TOGA-3 (1984). Proceedings of the first international TOGA workshop on thermal sampling. ed. M.J. McPhaden and R.A. Taft, UCAR, Boulder, Colo.
Weare, B.C., P.T. Strub and M.D. Samuel (1980). Marine Climate Atlas of the Tropical Pacific Ocean, Contrib. to Atm. Sci. No. 20 (Univ. of Calif. at Davis).
Wyrtki, K. (1975). Fluctuations of the dynamic topography in the Pacific Ocean. J. Phys. Oceanogr., 5, 450-459.
WCP-89. Report of the second session of the JCS/CCCO TOGA Scientific Steering Group, WCRP, Geneva.
WCP-103. Report of the TOGA drifters planning meeting. WCRP, Geneva.
WCP-130. Report of the fifth session of the JCS/CCO TOGA Scientific Steering Group, WCRP, Geneva.
WCRP-1: VALIDATION OF SATELLITE PRECIPITATION MEASUREMENTS FOR THE GLOBAL PRECIPITATION CLIMATOLOGY PROJECT (Report of an International Workshop, Washington, D.C., 17-21 November 1986) (WMO/TD-No. 203)
WCRP-2: WOCE CORE PROJECT 1 PLANNING MEETING ON THE GLOBAL DESCRIPTION (Washington, D.C., USA, 10-14 November 1986) (WMO/TD-No. 205)
WCRP-3: INTERNATIONAL SATELLITE CLOUD CLIMATOLOGY PROJECT (ISCCP) WORKING GROUP ON DATA MANAGEMENT (Report of the Sixth Session, Fort Collins, USA, 16-18 June 1987) (WMO/TD-No. 210)
WCRP-4: JSC/CCCO TOGA NUMERICAL EXPERIMENTATION GROUP (Report of the First Session, Unesco, Paris, France, 25-26 June 1987) (WMO/TD No. 204)
WCRP-5: CONCEPT OF THE GLOBAL ENERGY AND WATER CYCLE EXPERIMENT (Report of the JSC Study Group on GEWEX, Montreal, Canada, 8-12 June 1987 and Pasadena, USA, 5-9 January 1988) (WMO/TD-No. 215)
WCRP-6: INTERNATIONAL WORKING GROUP ON DATA MANAGEMENT FOR THE GLOBAL PRECIPITATION CLIMATOLOGY PROJECT, (Report of the Second Session, Madison, USA, 9-11 September 1988) (WMO/TD-No. 221) (out of print)
WCRP-7: CAS GROUP OF RAPPORTEURS ON CLIMATE, (Leningrad, USSR,28 October-1 November 1985) (WMO/TD-No. 226)
WCRP-8: JSC WORKING GROUP ON LAND SURFACE PROCESSES AND CLIMATE, (Report of the Third Session, Manhattan, USA, 29 June-3 July 1987) (WMO/TD-No. 232)
WCRP-9: AEROSOLS, CLOUDS AND OTHER CLIMATICALLY IMPORTANT PARAMETERS: LIDAR APPLICATIONS AND NETWORKS, (Report of a Meeting of Experts, Geneva, Switzerland, 10-12 December 1985) (WMO/TD-No. 233)
WCRP-10: RADIATION AND CLIMATE (Report of the First Session, JSC Working Group on Radiative Fluxes, Greenbelt, USA, 14-17 December 1987) (WMO/TD-No. 235)
WCRP-11: WORLD OCEAN CIRCULATION EXPERIMENT - IMPLEMENTATION PLAN DETAILED REQUIREMENTS (Volume I) (WMO/TD-No. 242)
WCRP-12: WORLD OCEAN CIRCULATION EXPERIMENT - IMPLEMENTATION PLAN - SCIENTIFIC BACKGROUND (Volume II) (WMO/TD-No. 243)
WCRP-13: RADIATION AND CLIMATE (Report of the Seventh Session of the International Satellite Cloud Climatology Project (ISCCP) Working Group on Data Management, Banff, Canada, 6-8 July 1988) (WMO/TD-No. 252)
WCRP-14: AN EXPERIMENTAL CLOUD LIDAR PILOT STUDY (ECLIPS) (Report of the WCRP/CSIRO Workshop on Cloud Base Measurement, CSIRO, Mordialloc, Victoria, Australia, 29 February-3 March 1988) (WMO/TD-No. 251)
WCRP-15: MODELLING THE SENSITIVITY AND VARIATIONS OF THE OCEAN-ATMOSPHERE SYSTEM (Report of a Workshop at the European Centre for Medium Range Weather Forecasts, 11-13 May 1988) (WMO/TD-No. 254)
WCRP-16: GLOBAL DATA ASSIMILATION PROGRAMME FOR AIR-SEA FLUXES (Report of the JSC/CCCO Working Group on Air-Sea Fluxes, October 1988) (WMO/TD-No. 257)
WCRP-17: JSC/CCCO TOGA SCIENTIFIC STEERING GROUP (Report of the Seventh Session, Cairns, Queensland, Australia, 11-15 July 1988) (WMO/TD-No. 259)
WCRP-18: SEA ICE AND CLIMATE (Report of the Third Session of the Working Group on Sea Ice and Climate, Oslo, 31 May-3 June 1988) (WMO/TD-No. 272)
WCRP-19: THE GLOBAL PRECIPITATION CLIMATOLOGY PROJECT (Report of the Third Session of the International Working Group on Data Management, Darmstadt, FRG, 13-15 July 1988) (WMO/TD-No. 274)
WCRP-20: RADIATION AND CLIMATE (Report of the Second Session of the WCRP Working Group on Radiative Fluxes, Geneva, Switzerland, 19-21 October 1988) (WMO/TD No. 291)
WCRP-21: INTERNATIONAL WOCE SCIENTIFIC CONFERENCE (Report of theInternational WOCE Scientific Conference, Unesco, Paris, 28 November to 2 December 1988) (WMO/TD No. 295)
WCRP-22: THE GLOBAL WATER RUNOFF DATA PROJECT (Workshop on the Global Runoff Data Set and Grid estimation, Koblenz, FRG, 10-15 November 1988) (WMO/TD No. 302)
WCRP-23: WOCE SURFACE FLUX DETERMINATIONS - A STRATEGY FOR IN SITU MEASUREMENTS (Report of the Working Group on In Situ Measurements for Fluxes, La Jolla, California, USA, 27 February-3 March 1989) (WMO/TD No. 304)
WCRP-24: JSC/CCCO TOGA NUMERICAL EXPERIMENTATION GROUP (Report of the Second Session, Royal Society, London, UK, 15-16 December 1988) (WMO/TD-No. 307)
WCRP-25: GLOBAL ENERGY AND WATER CYCLE EXPERIMENT (GEWEX) (Report of the First Session of the JSC Scientific Steering Group for GEWEX, Pasadena, USA, 7-10 February 1989) (WMO/TD-No. 321)
WCRP-26: WOCE GLOBAL SURFACE VELOCITY PROGRAMME (SVP) (Workshop Report of WOCE/SVP Planning Committee and TOGA Pan-Pacific Surface Current Study, Miami, Florida, USA, 25-26 April 1988)