Search Results

You are looking at 121 - 130 of 957 items for :

  • Regression analysis x
  • Journal of Physical Oceanography x
  • All content x
Clear All
Anand Gnanadesikan, Cassidy M. Speller, Grace Ringlein, John San Soucie, Jordan Thomas, and Marie-Aude Pradal

Weddell Sea over ~19 years (corresponding to half of the cycle of 57 years). Extending this to more modes results in a drop of the time scale of the oscillation to 53 years when 10 modes are used in the regression and 51 years when 12 modes are used, but does not yield additional physical insight about how different regions are coupled. As we are primarily using the POP analysis to identify key regions that are involved with the oscillation, we do not explore these results further here. b

Restricted access
Adriana Huyer

May 1977 bythe Coastal Upwelling Ecosystems Analysis (CUEA)program. The observations were concentrated in theregion just south of Cabo Nazca and includedrepeated occupations of a central hydrographic section (the "C-line") extending approximately southwestward from the coast (Fig. 1). Thirty-seven sections were concentrated in three periods: 27 April23 June 1976, 27 July- 14 August 1976 and 5 March16 May 1977 (Fig. 2). Almost all of these sectionsincluded CTD casts at each station, down to 500 mor

Full access
N. A. Bray and N. P. Fofonoff

from one time window to the next are largecompared with the GPE within each window, indicating the presence of scales larger than the station grid. An analysis of errors has been made to show the sensitivity of the estimates to data accuracy andsampling frequency.1. Introduction Nearly geostrophic flow in the ocean is maintainedby potential energy stored as vertical displacementsof density surfaces relative to level (geopotential)surfaces. If the rate of dissipation of kinetic energyis less

Full access
Xiao-Hai Yan, Young-Heon Jo, W. Timothy Liu, and Ming-Xia He

), respectively. Here M is the number of spatial points x . We defined f  ( x, t ) as η ′ T and g ( x, t ) as the SST anomaly SST  ′ in this study. By determining a regression coefficient between f  ( x, t ) and g ( x, t ), we estimated a new thermal steric height anomaly from SST  ′. The details of the application of this methodology are discussed in section 3 . c. Complex singular value decomposition (CSVD) Barnett (1983) applied complex EOF (CEOF) analysis using a covariance matrix to

Full access
Allan J. Clarke and Marcelo Dottori

1. Introduction Analysis of sea level, sea surface temperature (SST), and thermocline observations ( Enfield and Allen 1980 ; Chelton and Davis 1982 ; Kessler 1990 ) has shown that much of the interannual variability along the California coast originates along the equator; California sea level and SST tend to be higher than normal and the thermocline depth tends to be greater during El Niño, while during La Niña sea level and SST tend to be lower than normal and the thermocline shoals. Based

Full access
M. C. Gregg

separation of the profiles into mean andvarying components. From analysis of the temperaturefluctuations it was learned that 1) A distinct break-in-slope occurs between 0.06and 0.1 cpm in the temperature and displacementspectra. This is particularly convincing since it occurredin data taken with different instrument systems andprocessed by different algorithms. The spectra, in loglog coordinates, showed good fits to regression lines oneither side of the change in slope. For 0.002 x< k <x (0

Full access
Peter P. Chu, Leonid M. Ivanov, and Oleg V. Melnichenko

systems is larger than the scale of the continental shelf, the horizontal mean wind speed and direction from seven buoys are used for analysis. 3. OSD method A low-pass (40 h) filter is used to get rid of energetic shelf oscillations caused by tides, inertial currents, and sea breeze effects from the original current mooring and drifter observational data ( Chen et al. 1996 ; DiMarco et al. 2000 ). The filtered data are reconstructed (0.1 o × 0.1° grid) using the OSD method ( Chu et al. 2003a

Full access
Ned P. Smith

time series from station N was 404 days long, from 14 March 1996 to 22 April 1997. Because of gaps in the record at station S, a 266-day time period from 14 March to 5 December 1996 was selected for analysis. Mean depths at the northern and southern study sites were 3.2 and 2.0 m, respectively. Currents were recorded hourly in 2–12 layers, according to the mean water depth and the tidal and nontidal rise and fall in sea level. Layers were 29 cm thick, and the lowest layer was 69–98 cm above the

Full access
K. Hasselmann, W. Sell, D. B. Ross, and P. Müller

given in Table 1. Also shown arethe corresponding values for a composite data set consisting of a superposition of all data with the exceptionof the sets marked with an asterisk (which are discussedbelow). (The last two column groups in the table referto the energy-scale variables a and ~, which will bediscussed in Section 10.) For each data set the dependence of the shape parameters s= 'r, ~, a~, X on the stateof development of the wave field was investigated byfitting least-square regression

Full access
B. L. Blackford

for four different values of v are shown.664 JOURNAL OF PHYSICAL OCEANOGRAPHY VOLUME8used to carry out a linear regression (from t = 0to t = 60 h) between the sinusoidal wind stresspulse and the resulting current amplitude, assuming only .one wind stress pulse in a 60 h interval.The analysis was done for each value of v and theresulting linear relationships were

Full access