The coast along southern California and Baja Califor- 

 nia is characterized by winds favorable for upwelling 

 throughout the year. Peak values of surface wind stress 

 occur in April and May. Values exceeding 0.5 dyne cm -2 

 are evident from February to July. A local maximum im- 

 mediately north of Punta Eugenia is easily observed in 

 May. This feature is consistent throughout the year. A 

 region of local wind stress minima is indicated along the 

 coast south of Point Conception. This feature cor- 

 responds in location with the semipermanent cyclonic 

 eddy which dominates the ocean surface circulation in 

 the Southern California Bight (Reid et al. 1958). 



Spatial and Temporal Variability 



Spatial and seasonal variability of the monthly dis- 

 tributions will be described in terms of standard errors of 

 the means (Equation (3)), constancy of the wind stress, 

 and frequency diagrams for selected 1-degree square 

 areas and months. In general, these data indicate greater 

 variability in magnitude and direction in the region 

 north of Cape Mendocino than in the area to the south. 

 Summer distributions are characterized by well-defined 

 mean directions and magnitudes. Broad-banded fre- 

 quency histograms are typical of winter months. 



The degree of variability in direction and magnitude of 

 the surface wind stress is reflected in the spatial dis- 

 tributions of the standard errors of the means. In regions 

 outside of the primary shipping lanes (see Fig. 2), 

 monthly mean distributions are based on approximately 

 the same number of observations per 1-degree square 

 area. If one assumes that there are no systematic sam- 

 pling errors in these regions, then the magnitudes of the 

 computed standard errors of the means should be a func- 

 tion of the wind stress variability. Accordingly, standard 

 errors to the north of Cape Mendocino are larger by a fac- 

 tor of 2 to 3 (Appendix II) than those to the south of Cape 

 Mendocino. Thus, the data imply a greater degree of 

 variation in the direction and magnitude of the surface 

 wind stress off Oregon and Washington than off Califor- 

 nia. 



A contrast between winter and summer conditions is 

 also evident. South of Cape Mendocino, magnitudes of 

 the standard errors of the means remain nearly constant 

 throughout the year. Off Oregon and Washington, the 

 computed values decrease to a minimum during the 

 summer, and increase to a maximum during the winter. 

 Typical values range from 0.10 dyne cm -2 in June to 0.30 

 dyne cm" 2 in December. 



Similar features of the large-scale temporal and spatial 

 variations are evident in monthly distributions of con- 

 stancy of the wind stress as defined by the ratio of the 

 magnitude of the average stress to the average mag- 

 nitude of the stress. Figures 7 and 8 show the patterns for 

 June and December. During December, values greater 

 than 0.5 occur south of Point Conception. To the north, 

 wind stress constancy decreases to relatively small 

 values, implying a high degree of directional variability. 

 An increase in directional constancy is indicated during 



June. Values greater than 0.50 extend from Baja Califor- 

 nia to Vancouver Island. Values less than 0.50 occur only 

 in the northwest section of the grid. South of Cape Men- 

 docino, the ratios approach a value of 1.0, implying very 

 little variation in wind stress direction. 



The patterns of wind stress constancy indicate pos- j 

 sible error in estimates of the total mechnical energy 

 transfer at the air-sea interface. In the region off Califor- 

 nia, the magnitude of the mean vector wind stress 

 appears to be a good estimator of the total stress acting 

 on the sea surface. However, in the northern regions of 

 the data grid, these data show that the magnitude of the 

 mean vector wind stress underestimates the total stress 

 acting on the sea surface. This has important im- 

 plications for mixed-layer modelling, in which the input 

 of turbulent energy depends on the total stress acting on 

 the sea surface (Denman 1973), the direction of the sur- 

 face stress, and the rate at which the stress is applied to 

 the sea surface (Niiler 1975). 



The large-scale spatial and temporal variations in sur- 

 face wind stress distributions are finally described in 

 terms of selected frequency diagrams. Data for the 10 

 squares indicated in Figure 1 are displayed in Figures 9 

 to 12. Figures 9 and 10 show data for June and Decem- 

 ber taken from the five squares indicated in the inset. 

 Similar data at five different locations are presented for 

 July and January in Figures 11 and 12. 5 



Available wind stress data have been classed by direc- 

 tion and magnitude for each 1-degree square area and 

 month. Relative frequencies have been determined for 16 

 direction bands and for 11 magnitude bands. A category 

 for calm winds also includes variable winds. The direc- 

 tion bands are 22.5° wide and the magnitude bands cor- 

 respond to equivalent wind speed intervals of 2.0 m s -1 . 



The relative frequency surfaces shown in these figures 

 display contours of percentage of total reports falling 

 within a given direction and magnitude band. The 2.5% 

 and 7.5% contours are indicated by a solid line. Dashed 

 line contours indicate relative frequencies of 5.0% and 

 10.0%. The mean vector magnitude and direction is in- 

 dicated by an arrow. Note that directions are defined 

 with the oceanographic convention (i.e., the direction 

 toward which the wind blows). 



The histograms shown to the right of each frequency 

 surface display relative frequencies for magnitude at the 

 top, and direction at the bottom. Relative percent is 

 labelled along the ordinate. Midpoints of the magnitude 

 and direction class intervals are labelled on the abscissa. 



The contrast between winter and summer dis- 

 tributions in the southern section of the grid appears as a 

 change in the character of the frequency surfaces (Figs. 9, ' 

 10). In general, the number of contours indicated for 

 June is greater than the number of contours appearing in 

 December. Evidently, the wind stress is relatively con- 

 stant in magnitude and direction during the summer 



'Wind roses for all 1-degree squares within the area north of lat. 34°N 

 may be found in Climatic study of the near coastal zone; West Coast of 

 the United States, published by the Director, Naval Oceanography and 

 Meteorology, June 1976, 133 p. 



12 



