

2B& 







ST 



195 















_l 



U-J 



19a- 













,-' 





LU 

 1 





cr 



LU 

 CO 



165 











































O 



LU 

 t— 



CO 



_3 

 -) 



a 

 cr 



ISfl- 











.. 















175- 











L7B 





The second period analyzed was centered during the upwelling 

 period and covered 6 mo, April through August, during the years 

 1964-78. During this period, dynamic height remains the primary 

 predictor but at weaker correlation, r = 0.60. Atmospheric pressure, 

 SST, and meridional wind stress are secondary predictors and in total 

 account for 66% of the variability of monthly sea level. 



Thus, some seasonal change in the processes affecting sea level is 

 indicated, with dynamic height accounting for most of the sea level 

 variability in both the upwelling and Davidson Current periods. 

 Meridional wind stress is also important during both periods but 

 more so during the upwelling than Davidson Current period. Atmo- 

 spheric pressure and SST explain an additional portion of the sea 

 level variability during the upwelling period. The greater amount of 

 explained variance in winter than summer suggests that conditions in 

 winter are dominated by changes in the structure of the water column 

 whereas upwelling in summer causes complicated effects on sea 

 level. 



Figure 10. — Seasonal cycle of sea level and dynamic height near Monterey, 

 Calif. Sea level data are for 1963-78 and shown as dotted line and dynamic 

 height data are for 1968-77 and shown as dashed line. Ranges of monthly sea 

 levels are shown by vertical bars. 



Sea level and dynamic height are also in good agreement in a time 

 series sense. Figure 1 1 shows the time series of weekly mean sea level, 

 calculated from the hourly data, and individual dynamic height calcula- 

 tions relative to 200 and 400 db. The figure shows that both sea levels 

 and dynamic heights were higher than normal during 1969-70, 

 1972-73. and 1976. which were periods of El Nino activity in the east- 

 ern tropical Pacific. Sea levels and dynamic heights were both also 

 near or below normal during anti-El Nino periods. Because of the close 

 agreement between seasonal cycles of sea level and dynamic height, 

 and because of the high correlation of sea level at Monterey with that at 

 adjacent stations, dynamic height and sea level variations may both 

 reflect variations in the alongshore geostrophic current flow. To show 

 this, one would have to show that fluctuations of sea level were correl- 

 ated with fluctuations of slope of dynamic height normal to the coast- 

 line. Suitable data for this may be available but this was felt to be 

 beyond the scope of this report. 



The regression formula indicates that the response of sea level to 

 changes in atmospheric pressure is -1.67 cm mb whereas a purely 

 hydrostatic response would be -1 .00 em'mb. This higher than theoreti- 

 cal pressure response coefficient is poorly understood but is possibly 

 due to reinforcement of the local pressure effect by a larger scale. 

 dynamic aspect of the atmospheric pressure systems themselves. Saur 

 (1962) and Roden ( 1960) analyzed monthly tide data from stations to 

 the north and south of Monterey and found similar larger than expected 

 pressure response coefficients. 



Because of the significant seasonal changes in the oceanic and 

 atmospheric regimes near Monterey, we might expect to observe sea- 

 sonal changes in the processes affecting sea level. To define these 

 seasonal changes, the ocean and atmospheric variables were ana- 

 lyzed separately for the two major periods, the Davidson Current and 

 the upwelling periods (Table 3). 



Sea level changes during the Davidson Current period were ana- 

 lyzed using data from 5 mo, October through February', for the years 

 1963-78. The results of multiple regression analysis indicate that 

 dynamic height and meridional wind stress are major predictors of 

 sea level during this period, explaining 82% of the variance of 

 monthly sea level anomalies. During this period, dynamic height and 

 sea level are strongly correlated, r = 0.87. 



Spectral Analysis 



In the previous section, it was shown that most of the variance of 

 monthly sea level anomalies can be explained by monthly anomalies 

 of dynamic height, surface atmospheric pressure, SST, and meridio- 

 nal wind stress. However, important variations in these processes 

 occur on time scales shorter than a month. To determine how the var- 

 iance of sea level is distributed with frequency over time-periods of 

 days to weeks, auto- and cross-spectra were calculated for 6-h obser- 

 vations of sea level, atmospheric pressure, and meridional wind 

 stress. Spectra of dynamic height and SST were not computed 

 because the required data were too sparse. 



To prepare these data for spectral analysis, it was necessary to sub- 

 sample the hourly sea level series at the 6-h period of the available 

 surface atmospheric pressure and meridional wind stress data. 

 Atmospheric pressure and meridional wind stress were calculated as 

 described previously on a 6-h basis for the period 1 January 1967 

 through 31 August 1976 for a point approximately 14 km west of the 

 Monterey tide station (Fig. 1). Hourly sea level data for the same 

 time period were low-pass filtered to remove the diurnal, semi- 

 diurnal, and other short-term tidal components and were sub- 

 sampled at 6-h intervals. A complete description of the low-pass 

 filter used is given by Godin (1966). All data series were then 

 detrended by subtracting their 30-d running mean to produce band- 

 passed series. The response function for the 30-d running mean is 

 shown in Figure 12. 



Atmospheric pressure, wind stress, and sea level data (unadjusted 

 for pressure effects) were analyzed during the winter storm season (1 

 November to 8 March) and the upwelling period (1 April to 8 

 August) for the years 1967-76. The definition of these periods is 

 somewhat arbitrary but was based on visual interpretation of time 

 series of sea level and wind stress and on the requirement that the 

 number of data points used in the spectral analysis be a power of 2. 

 Since the periods are normally 3-4 mo long. 512 data points (128 d) 

 were used. A fast Fourier transform spectrum analysis with a triangu- 

 lar data window was used and the spectra were averaged for all avail- 

 able years. The frequency bandwidth is 0.04 cycles per day (cpd) and 

 the number of degrees of freedom is 90 for the winter period and 100 

 for the upwelling period. 



The spectral relationships between sea level and atmospheric pres- 

 sure are discussed first. In the low frequency region, the winter pe- 

 riod spectra (Fig. 13) are three to four times more energetic than the 

 upwelling period spectra (Fig. 14), indicating the effects of intense 

 winter storm events. The largest sea level and pressure fluctuations 



16 



