Search Results

You are looking at 1 - 2 of 2 items for

  • Author or Editor: Susumu Tabata x
  • Refine by Access: All Content x
Clear All Modify Search
Richard E. Thomson and Susumu Tabata


Thirty-year time series of hydrographic observations from Ocean Station PAPA and Line ‘P’ are used to estimate secular trends in monthly mean steric sea level heights relative to depths of 100 and 1000 decibars in the northeast Pacific Ocean. Linear trends at station ‘P’ (50°N, 145°W) indicate that steric heights relative to the 1000 db (approx. 1000 m) level are rising at a rate of 1.1 mm yr−1, comparable with the Order 1 mm yr−1 global trends suggested by analysis of selected long-term coastal tide gauge records. Approximately 67% of the increase in steric levels is due to thermosteric change at depths below 100 m, the smaller 33% contribution from the halosteric component apeasrs to be confined to the upper 100 m. Steric height trends at fine ‘P’ locations are also of order 1 mm yr−1 but, in contrast to station ‘P’ trends, arise mainly through the halosteric component.

Confidence levels for the linear trends an calculated in two ways. (i) using the Student-t test assuming that cub monthly observation is a statistically independent sample; and (ii) using the Student-t test in conjunction with the effective number of degrees of freedom derived from integral time scales. For station ‘P’, trends based on (i) are reliable to the 99% confidence level while for line ‘P’ only stations on the eastern portion of the fine have significant trends relative to the 1000 db level. Confidence levels obtained from (i) fail to take into consideration the long-term fluctuations in steric level records. To obtain more reliable estimates of the confidence intervals, we use integral time scales to determine the effective number of degrees of freedom for each monthly time series. Subsequent recalculation of trend-line confidence intervals indicates that the total steric height trends at Station ‘P’ remain significant at the 90% confidence level. The halosteric trend relative to 100 db is significant at 90% while the thermosteric trend relative to 1000 db is marginally significant at 70 to 80%. With the exception of stations 5 and 6, trends for line ‘P’ stations are no longer significant above the 70% level. The lower statistical reliability in the line ‘P’ trends is due, in part, to the sparse sampling rate relative to station ‘P’. We conclude that steric sea levels in the northeast Pacific are rising, at approximately 1 mm yr−1 and that this increase may be associated with a combined regional warming of the deeper waters and dilution of the surface waters. Although the observed trends appear to be linked to climate-induced eustatic changes in global sea level, the records are not of adequate length or spatial coverage to rule out effects of shifting regional circulation patterns.

Full access
Warren B. White and Susumu Tabata


For the period 1959–81, quasi-zonal hydrographic sections have been made between the southern coast of Vancouver Island (49°N, 126°W) and Ocean Station P (50°N, 145°W) approximately every two months. Along this section (called Line P) hydrographic-STD stations were repeated at nearly the same locations, approximately 2° of longitude apart and at closer intervals near the coast. As such, 13 stations in all were made nearly every two months over the 23-year period. Thus, the Line P dataset has particular value in verifying the existence of interannual baroclinic long waves near 50°N. In previous studies, White, Kang and Magaard, and White and Saur, had found evidence of annual and interannual baroclinic long-wave activity in the eastern midlatitude North Pacific over the range 20°–40°N, but earlier attempts could not discover these westward-traveling waves along Line P. In this study, both spectral analysis and complex empirical orthogonal function analysis are used to do just that. Concentrating upon the depth of the σt = 26.8 density surface, which lay just below the main halocline but within the main pycnocline, the time-longitude matrix of interannual anomalies about the long-term mean annual cycle display westward propagation of much longer period (2–5 years), with speeds similar to baroclinic Rossby waves, over the entire length of the section. The zonal wavenumber/frequency spectrum of this dataset finds the maximum spectral energy density overlying the linear Rossby wave dispersion curve. The first two complex EOFs both show westward propagation, one with larger period/wavelength than the other, both together explaining 46% of the total variance of the 23-year record. Recombining the time-longitude matrix from only these first two complex EOFs shows that the interannual baroclinic long waves tend to be associated with ENSO events that have now been found to occur in both the ocean and the atmosphere throughout the Pacific basin every 2–5 years.

Full access