Search Results

You are looking at 1 - 10 of 14 items for :

  • Author or Editor: E. Harrison x
  • Journal of Physical Oceanography x
  • Refine by Access: All Content x
Clear All Modify Search
D. E. Harrison

Abstract

Oceanic gyres defined by the mean zero wind-stress curl lines have been the focus of wind-driven ocean circulation theory since its beginnings. In the face of single-signed vorticity input from the curl of the wind stress over a gyre, a mechanism to balance the wind vorticity input is required if an equilibrium is to be established. Traditional models have tended to restrict their allowable physics so that a dissipation mechanism is required for equilibrium. However, dissipation is not necessary, in principle, for equilibrium for a general fluid system. In fact, it has recently been shown that lateral eddy vorticity transport between gyres can provide an important part of the vorticity tendency required for equilibration in model oceanic systems. This note examines the possibility that in the North Atlantic subtropical gyre this process also might be important. After a brief review of the equilibration mechanisms possible in a primitive equation fluid, attention is focussed on estimates of the eddy lateral transport of relative vorticity in the North Atlantic. It appears that there could be sufficient eddy transport across the Gulf Stream to balance the wind vorticity input over the gyre. Equilibrium might thus be possible, without any vorticity dissipation mechanism or without invoking higher order dynamical processes. If this mechanism is important in the ocean there should be interesting effects on the subpolar gyres, details of which depend on the circumstances surrounding the vorticity exchange process.

Full access
D. E. Harrison

Abstract

The question of the importance of mesoscale motions in the long time averaged ocean circulation is examined from the viewpoint offered by Eulerian scale estimates of the magnitudes of the explicit eddy and largest inviscid mean flow terms in the mean heat, momentum and vorticity equations. Comparisons of these estimates reveal the quantities that must be known to obtain reliable estimates of the importance of eddy terms in the mean balances. Using historical information and long time series of data from the western North Atlantic, two distinct regimes (“near field” and “mid-ocean”) are identified for this ocean region and the appropriate term comparisons are made for each regime. From estimates of the reliability of the ocean values used in these comparisons the robustness of the comparisons is examined. The momentum and vorticity equation estimates suggest that terms based on the eddy Reynolds stress can generally be neglected compared to terms involving f 0 and β in both the near field of the Gulf Stream and the mid-ocean. In the near field, mean advective terms appear to be at least as important as the eddy terms, but the eddy terms dominate these advective ones in the mid-ocean. The heat equation comparisons suggest that the eddy term is comparable to the mean horizontal advection of heat in the mid-ocean but is of somewhat reduced importance in the near field. Some remarks on the generality of results from numerical ocean models that contain mesoscale motions to the question of eddy importance in the ocean are offered.

Full access
D. E. Harrison

Abstract

The importance of mesoscale eddies in the basin energy budgets of closed-basin numerical model oceanic systems that attempt to resolve such motions varies greatly from calculation to calculation. In existing calculations, eddy importance has been found to depend strongly on the dissipation mechanism(s) selected. These energy budget results can be understood by examination of how eddy and mean flow kinetic energy are dissipated in the long-time mean in the different regions of the model flow. Scale analysis arguments are presented, assuming that the characteristics of the flows satisfy certain mild quasi-oceanic constraints, to investigate these dissipation terms. From these scale. estimates it appears that many of the model ocean results can he understood in terms of a nondimensional parameter that measures the relative importance of horizontal and bottom friction dissipation. When horizontal friction dissipation dominates, eddies can only be of modest importance in basin energy budgets, but when bottom friction dissipation dominates, eddies generally must be important. This follows simply from the assumed flow characteristics. The implications of these results on the interpretation of present modeling results are described.

Full access
D. E. Harrison

Abstract

Several experiments using an ocean general circulation model have been carried out in order to explore the degree to which the oceanic waveguide response during the 1982–83 ENSO event was 1ocally and remotely forced. Experiments in which the chosen monthly mean surface stress field was imposed only within three degrees of the equator (3°N/S) and within seven degrees of the equator (7°N/S) reveal that the 7°N/S winds reproduce the equatorial results of the full winds case to within differences small compared to the variability of interest The 3°N/S winds case reproduce equatorial dynamic height acceptably, but introduces errors in SST and upper-ocean currents that approach the ENSO signal. A 7°N–S experiment in which the meridional stress is set to zero (NOYST) snows that meridional stress plays a nontrivial, but not dominant role, in the 1982–83 model behavior; errors generally are comparable to those of the 3°N/S case. A final experiment, in which the 1982–83 winds were imposed west of the dateline and climatological winds were imposed east of 170°W (WPAC), illustrates the extent to which the central and eastern Pacific were forced by winds in the western Pacific. While there is nontrivial remote forcing, the locally forced variability is roughly twice as great.

Implications for coupled ocean-atmosphere modeling and for design of future surface wind stress monitoring arrays for ENSO prediction are considered.

Full access
D. E. Harrison

Abstract

A method for evaluating the utility of a given parameterization field in a spatially inhomogeneous circulation is described. This method is used to examine the mean field diffusion parameterization of the heat, momentum and vorticity time deviation eddy terms from a mesoscale resolution numerical ocean circulation experiment. The diffusion model does not satisfactorily describe the eddy terms for any field throughout any of the different regions of the model flow.

Full access
D. E. Harrison
and
Robert H. Heinmiller

Abstract

We discuss the upper ocean mesoscale temperature variability field as sampled by XBT's in the joint US-USSR POLYMODE Synoptic Experiment during July 1977 to July 1978. There is a “background” persistent, relatively weak and rather large-scale pattern of variability, of alternating warm and cold areas which propagate nearly westward. For the background, very little point correlation between surface temperatures and temperatures beneath the mixed layer is found, but sub-mixed-layer temperatures above and below the “18°C water” show negative correlation. We also find a number of smaller scale, apparently discrete, features which are somewhat more intense than the background. They exhibit little consistency in direction of propagation, and pattern propagation, rather than pattern evolution, as generally observed. The time-average depth field of the 15°C isotherm exhibits spatial variations comparable in magnitude to its pointwise standard deviation values. Removing the mean by linear fits in latitude and/or longitude produces horizontal auto-correlation functions substantially different from those based on removing the time-averaged field. The statistics of the horizontal correlation are not satisfactorily determined by this data set.

Full access
Paul S. Schopf
and
D. E. Harrison

Abstract

We present results from three numerical model experiments designed to study the thermal and hydrodynamics changes associated with downwelling Kelvin wave passage and east coastal reflection along and near the equator. The model employs primitive equation dynamics in two active layers and a full thermodynamics equation, so that sea surface temperature, thermocline displacement and sea level are each independently predicted. Wind and thermal finding are used. The surface layer is a slab mixed layer using Kraus and Turner-style bulk physics. Kelvin waves are excited by introducing a westerly wind anomaly in the western part of the basin, and the temperature and current changes caused by the waves are studied as the wave fronts propagate through the circulation forced by three different mean wind fields: no mean winds, southerly men winds and easterly mean winds. The wave-induced changes depend strongly on the conditions that prevail when the waves are forced. Anomalous advection of the existing SST field is the primary SST change mechanism. The two internal Kelvin-wave modes allowed by the model sometimes induce comparable temperature changes near the east coast MA sometimes the effect of one mode substantially dominates that of the other. The shear mode wave does not always propagate to the east coast; it can be destroyed by nonlinear effects associated with the meridional circulation along the equator. Temperature changes near the east coast, similar in magnitude to those observed in the early stages of El Niño events, are caused in the mean southerly wind case, but no broad westward tongue appears latter on. The implications of these results on existing models of El Niño and for future model studies are examined.

Full access
D. E. Harrison
and
W. R. Holland

Abstract

The dynamical balances of the mean flow of a numerical model ocean general circulation experiment are examined through evaluation of regional vorticity budgets. The instantaneous flow is strongly time dependent and the effect of eddy terms in the mean budgets is of primary interest. Budgets have been computed over volumes ranging in size from less than that of a typical model eddy up to an entire wind-driven gyre, using time series of 5 and 10 year durations. The statistical reliability of terms in the budgets varies significantly with the region size; over regions the size of an eddy or smaller the reliability is often poor, but over the selected larger regions it is satisfactory. The final analysis regions are selected by requiring that each be identified clearly with some part of the mean flow and that cancellation of the locally dominant terms within each region be minimized whenever possible.

The primary mechanism for balancing the wind-stress curl vorticity input in each half basin is found to be horizontal transport of relative vorticity by the eddies across the zero wind-stress curl latitude that separates the distinct flow systems of the two half basins. However, net meridional eddy vorticity transport is generally unimportant away from the half-basin boundary latitude. Eddy horizontal transports over the analysis regions, away from the western part of the zero wind-stress curl latitude, also tend to be small. The transport flow budgets and upper layer budgets tend to be similar. The deep-layer flow is qualitatively different from these flows, a separate set of analysis regions is needed to study it, and the deep budgets are different in several respects. Away from the boundary currents and internal jets the volume integral analog of the classical geostrophic balance—vortex stretching balancing advection of planetary vorticity—holds very well. In particular, over the interior of each gyre, the net input of vorticity by the, wind balances the loss by advection of planetary vorticity to better than 10%. This result is quite different from the conclusion that would he drawn from examination of the vorticity balance at a point over much of the interior, where the divergence of the eddy relative vorticity flux is often large (but of limited statistical reliability). The eddy heat-flux divergence plays an important role in establishing the interfacial vertical velocity contribution to vortex stretching in some of the regions, and appears essential in forcing one of the deep flow currents. No simple summary of the bound current and jet region budgets can be offered, except that mean nonlinear transport often dominates eddy horizontal transport and that frictional effects can be quite small. These results are compared with classical wind-driven ocean circulation ideas and the strengths and limitations of this type of analysis for studying eddy-mean flow interaction are discussed.

Full access
D. E. Harrison
and
A. P. Craig

Abstract

A hindcast of the 1982–1983 ENSO event using a primitive equation ocean circulation model forced by monthly mean wind stresses based on the SADLER pseudostress fields shows very good agreement with observations at 0°, 159°W between June 1982 and March 1983. The hindcast experiment is analyzed to explore the processes that caused the large accelerations, decelerations, and thermal changes observed during this time. Several hindcast experiments incorporating variations of the SADLER wind field and several idealized experiments incorporating a western Pacific westerly wind event are analyzed and compared with the 1982–1983 SADLER hindcast to explore the importance of local and remote forcing, the relative importance of zonal and meridional wind stress changes, and the dynamical signatures of the processes at work. Meridional wind stress changes have little effect on either the zonal velocity or temperature fields. Local zonal wind stress variations can account for the qualitative changes in the upper-ocean zonal flow, but cannot reproduce the observed thermal changes or the timing and quantitative evolution of the zonal flow. Remote forcing is needed to account for these latter aspects of the observations. Eastward-propagating Kelvin response appears to be quite important, but westward-propagating Rossby variance forced during 1982 from east of 160°W does not appear to play any significant role. The idealized remote-forcing experiments indicate that westerly events can account for the variability not explained by local forcing; the essential aspect is how the forcing projects onto the vertical modes defined by the stratification under the forcing at the time of the wind event. Modes higher than the first and second can be strongly forced and the sum over modes produces vertical structures in the near field of the forcing similar to those observed. Simple linear Kelvin mode ideas thus are useful for understanding the response to remotely forced variability. However, nonlinear processes affect the quantitative response, both by changing the stratification under the forcing region as the forcing event proceeds (and thereby altering the modal projection of the forcing) and through zonal advection and interaction between the response and the background mean flow. The dynamical balance of terms for zonal momentum in the SADLER hindcast is quite complex and the difficulty of identifying remote forcing from the balance of terms, even during periods when remote forcing is the primary agent of change, is discussed. This detailed study of a particularly interesting period of equatorial flow and thermal variability illustrates the many processes at work on the equator in the central Pacific during periods of substantial local and remote wind stress variability. It also illustrates some of the challenges that might be encountered in interpreting the results of an oceanic local dynamics experiment under conditions like these.

Full access
D. E. Harrison
and
A. R. Robinson

Abstract

A simple linear model of the barotropic basin response to forcing imposed along the northern boundary is described. The dependence on latitude of the response may include oscillatory behavior or not, depending on whether the forcing frequency is smaller or greater than the fundamental free basin mode frequency. When oscillatory behavior is found, the forced solution may resemble oceanic mesoscale eddies. The relevance of this simple model to a description of the eddy fields of several mesoscale resolution general ocean circulation numerical experiments is examined. It is found that a single term of the analytical solution can very well describe the numerically produced eddy fields, away from the regions of strong currents. The possibility that this general mechanism might account for the existence of mesoscale eddies in the ocean is briefly discussed.

Full access