Search Results

You are looking at 1 - 9 of 9 items for :

  • Author or Editor: T. P. Barnett x
  • Journal of Physical Oceanography x
  • Refine by Access: Content accessible to me x
Clear All Modify Search
T. P. Barnett

Abstract

Long-range Naval aircraft using AXBT's obtained meridional temperature sections from the central Pacific Ocean along 158 and 170°W between 30 and 50°N at approximately monthly intervals between November 1974 and April 1977 (29 months). Analyses of these sections show that the seasonal cycle in the central ocean is confined largely to the upper 100 m. The phase of the seasonal cycle increases with depth so that at 100 m it lags the surface by three months. Exceptions to the above statements occur in two narrow bands centered on latitudes 42 and 36°N where the seasonal cycle apparently penetrates in phase to the limit of observations (300 m) except in the interval 100–150 m which lags the surface by 1–2 months. Approximately 90–95% of the variance in the seasonal change of heat storage in the study region can be accounted for by air/sea heat exchange and some type of vertical mixing. Horizontal and vertical advection were of limited and little use, respectively, in reproducing the seasonal cycle variance.

The spatially coherent features of the non-seasonal, or anomaly, field were confined largely to the upper 100 m, with small spatial scales dominating the variance field below 100 m. This implies that the often studied sea surface temperature anomalies are associated with a thermal structure largely confined to the mixed layer. It was not possible to explain quantitatively a significant portion of the variance of the heat storage anomalies in terms of currently available estimates of air/sea heat exchange and advective processes. This result is partially due to noise introduced into the heat budget calculations by sampling variability. However, the major reason for the result appears to be poor estimates of the heat budget source terms. Estimates of these source terms must be substantially improved if quantitative understanding of interannual ocean variability is to be achieved.

Full access
T. P. Barnett

Abstract

Advanced statistical techniques have been used to conduct a study of the relationships between ocean and atmosphere variables in the tropical Pacific Ocean. The results of the study show that the ocean variables can hindcast features of the trade wind field (TWF) variability several months into the future. The results are compatible with the notion that the Hadley and Walker cells are associated with east-west and north-south sea surface temperature (SST) gradients as suggested by Bjerknes. However, the level of skill in even specifying the strength of these cells is small, suggesting mechanisms other than those associated with SST are responsible for much of the observed variability in those systems.

TWF predictors can hindcast themselves and ocean variables at both short lead times and lead times near one year. The main components of the wind field responsible for this long-lead-time skill are associated with the southeast trades and a near-equatorial band in the western Pacific. The results also suggest that large El Niño events are predictable a year in advance. This conclusion was supported by an independent test which successfully forecast SST anomalies off Peru one year in advance for the period 1976–79.

The results of the study have been used to test specific ideas and scenarios regarding the physical mechanisms responsible for large-scale air-sea interactions in the tropical Pacific. The results add new ideas and’ additional depth to previous work plus help formulate a more cohesive description of large-scale events. During regimes of cold equator SST the northeast and southeast trade wind field merge and extend into the western Pacific. At these times there is no clear minimum in zonal component of the wind system over the western and central ocean between 15°N and 15°S. During warm equatorial SST situations the two TWF's are separated as evidenced by a strong minimum or even reversal of the zonal component in the region noted above. The associated changes in zonal stress and wind stress curl between warm and cold events is very large. During the transition from cold to warm equator the equatorial SST anomalies move westward from South America to near the dateline. The wind anomalies in the equatorial region move simultaneously from the western Pacific to the central ocean. Both types of anomaly meet in the central ocean during the summer/fall season.

Full access
T. P. Barnett

Abstract

Numerous hypotheses have been proposed to explain interannual changes in equatorial water temperatures. It is shown that many of these hypotheses can be tested by expressing them in terms of a statistical-dynamical model based on the heat balance equation. The ability of the resulting model to account for variance in a 20-year record of observed water temperature provides a hitherto unavailable, quantitative measure of the hypothesis consistency.

Field data were used in conjunction with the models to show that the following general conclusions are consistent with available observations: 1) The advective terms (both horizontal and vertical) in the heat balance equation account for 30–50% of the variance in records of interannual changes in near-equatorial SST. 2) The advective changes are closely related to significant changes in the trade wind field, particularly those occurring near the equator and just west of the dateline, as well as major changes in sea level across the entire Pacific Basin.

Specific hypotheses about interannual changes in water temperature were tested. The following conclusions were found to be consistent with the available data: 1) Eastward advection of heat by the North Equatorial Countercurrent is far more important (29% of the variance) to the heat balance of the eastern tropical Pacific than local heating or consequences of long-term variations in the Northeast Trades. 2) At Talara, Peru, 48% of the SST variance was predictable one month in advance using basin-wide fluctuations in sea level as predictors. This suggests the importance to the heat balance off Peru of eastward advection of heat by currents or wave phenomena. Of less importance (14%) was trans-equatorial flow across the Galapagos front. Upwelling induced by local changes in the wind stress was not important, on the interannual time scale, in the estimate of SST. 3) Temperature changes in the central equatorial Pacific (Christmas Island) were consistent with the mechanisms of local upwelling at the equator (25%) and advection from the east (33%).

Full access
T. P. Barnett
and
W. C. Patzert

Abstract

Long-range P-3 aircraft have been used to occupy two 4000 km long sections from 20°N to 17°S along 150 and 158°W in the central equatorial Pacific. The temperature field along these sections was measured at approximately weekly intervals for three months (November 1977–January 1978). The principal meridional scales of variability derived from this data set suggest highly coherent fluctuations in the upper ocean thermal structure within ±10° of the equator. The region of coherent variability extends across the equator, across boundaries of major current systems and through the ITCZ. The time scale of these fluctuations is of order 2–3 months. Variability in the zonal direction is also coherent, although of smaller amplitude, with space scales of at least 2000 km and time scales of several months. These latter modes of variability are suggestive of propagating disturbances, although that hypothesis could not be proved with the current data set. The observed oceanic variability at 150°W was not closely related with changes in the local wind stress or curl of the wind stress field in the central tropical Pacific, suggesting that a large part of the observed oceanic fluctuations may not have been “locally” forced.

The relation between transports in the North Equatorial Countercurrent derived from the AXBT data and T/S relations agreed well with similar transport estimates obtained from hydrographic observations and current meter arrays. This suggests that the scales of variability we have observed in the temperature field may also apply, in part, to the zonal velocity and transport fields.

Full access
W. B. White
and
T. P. Barnett

Abstract

During the autumn and winter seasons, large amounts of heat are given up to the atmosphere at the subarctic frontal zone off the east coast of Asia. According to Fisher, the Laplacian of this heat flux (∇2 Q) is related to increases in the intensity of the relative vorticity in the westerly wind regime. This increase is related to a similar increase in strength of the wind stress curl, which thereby increases the Sverdrup transport of the subarctic and subtropical gyres. The increase in transport in turn intensifies ∇2 Q at the subarctic frontal zone via geostrophic adjustment. This coupling of atmospheric relative vorticity, Sverdrup transport, and ∇2 Q results in the intensification of the relative vorticity of both fluid media that can be checked only by an instability in either one or the other media. This mutual interaction of the ocean and atmosphere is termed a servomechanism, the natural time scales of which are determined by a mathematical development wherein the vertically integrated vorticity equations of the ocean and atmosphere are coupled by their interaction at the naviface. This coupling leads to a single wave equation for the ocean/atmosphere system, the solutions of which are Rossby waves modulated by exp[(1+it)], where α depends upon the coupling parameters. Normal values of α are found to produce an e-folding increase in the vorticity of the ocean/atmosphere system in less than two months. For anomalously high values of α, the increase in vorticity can be extreme, possibly leading to the formation of a barotropic instability in the atmospheric medium. These theoretical results are illustrated using geophysical data from 1950–60 and are used to explain the events that triggered the unusual ocean/atmospheric vorticity state that existed in the North Pacific between 1956 and 1958.

Full access
S. Hasselmann
,
K. Hasselmann
,
J. H. Allender
, and
T. P. Barnett

Abstract

Four different parameterizations of the nonlinear energy transfer S nl in a surface wave spectrum are in investigated. Two parameterizations are based on a relatively small number of parameters and are useful primarily for application in parametrical or hybrid wave models. In the first parameterization, shape-distortion parameters are introduced to relate the distribution S nl for different values of the peak-enhancement parameter γ. The second parameterization is based on an EOF expansion of a set of S nl computed for a number of different spectral distributions. The remaining two parameterizations represent operator forms that contain the same number of free parameters as used to describe he wave spectrum. Such parameterizations with a matched number of input and output parameters are required for numerical stability in high-resolution discrete spectral models. A cubic, fourth-order diffusion-operator expression derived by a local-interaction expansion is found to be useful for understanding many of the properties of S nl , but is regarded as too inaccurate in detail for application in most wave models. The best results are achieved with a discrete-interaction operator parameterization, in which a single interaction configuration, together with its mirror image (representing a two-dimensional continuum of interactions with respect to a variable reference wavenumber scale and direction) is used to simulate the net effect of the full five-dimensional interaction continuum.

Full access
T. P. Barnett
,
R. A. Knox
, and
R. A. Weller

Abstract

During January and February 1974 the NORPAX POLE experiment was carried out in the central Pacific to begin collection of data needed to design a large-scale ocean/atmosphere monitoring program. This paper describe features of the ocean temperature field observed during POLE within a region of about 400 km in diameter centered near 35°N, 155°W. The temperature field, which was approximately stationary during the month-long experiment, was dominated by a strong north-south gradient as expected. The east-west gradient was negligible. Superimposed on this mean field was energetic noise with typical rms isotherm displacements of 25 m near the bottom of the mixed layer. The characteristic horizontal scale of this noise was 50 km near the surface although the field appeared to be anisotropic. The energy, scale length and degree of anisotropy all decrease with depth. The implications of these observations to a sampling strategy are discussed as are other conclusions drawn from a statistical analysis of the temperature data.

Full access
Russ Davis
,
T. P. Barnett
, and
C. S. Cox

Abstract

Variability of near-surface Currents over a 20-Day period in a 15O km diameter region of the central North Pacific is described using vertical profiles from a current meter and the tracks of 25 drifting buoys. Energetic fluctuations of order 0.10 m s−1 having time scales of a few days and vertical scales in excess of 100 m were found, apparently coherent with the wind forcing. Buoy tracks disclose a small-scale (<15 km) short-period (less than a few days) variability with speeds of the order 0.05 m s−1 and an energetic mesoscale motion with speeds of the order 0.07 m s−1, space scales of the order 40 km and time scales exceeding 20 days. Additionally, the difference between the mean current observed over the experiment. having a speed of aboutO.03 m s−1, and the climatological norm inferred from ship-drift. with a speed of about 0.10 m s−1, suggests a larger scale variability not adequately resolved.

Full access
T. P. Barnett
,
R. W. Preisendorfer
,
L. M. Goldstein
, and
K. Hasselmann

Abstract

Methods of estimating the significance of optimal regression models selected from a model hierarchy proposed by Barnett and Hasselmann (1979) are reexamined allowing for the multiple-candidate nature of the selection criteria. It is found that the single-candidate models' significance value previously used can over- or underestimate the true multiple-candidate significance level of the selected model depending on the selection criteria used. A number of possible selection strategies to remove these problems are discussed and evaluated both theoretically and by Monte Carlo simulators.

Full access