gression lines by multiplying each monthly value 

 by 0.5 and by the reciprocal of the slope given in 

 Table 2. The multiplication by 0.5 is necessary be- 

 cause the value of the drag coefficient used to 

 construct Figure 4(a) is half that used for Figure 

 4(b) whereas the slopes in Table 2 were computed 

 for two series with equal drag coefficients. Multi- 

 plication by the reciprocal of the slopes then 

 serves to adjust the data so that the spatial varia- 

 tion in the underestimate of the monthly mean 

 wind stress due to computation from monthly 

 mean atmospheric pressure data as predicted in 

 Table 2 is removed. 



The magnitudes and spatial variations dis- 

 played in Figure 4(c) are more like those shown in 

 Figure 4(a) than are those in Figure 4(b). Some 

 smoothing of gradients in Figure 4(b) and (c) re- 

 sults from having data points at 3-degree intervals 

 with each point incorporating data across a 6- 

 degree area. This causes some loss in detail and 

 displacement of contours relative to Figure 4(a| 

 where data points are at 1-degree intervals and 

 incorporate data restricted to a 1-degree square 

 area. There remains however a definite shift of 

 the maximum toward southern California in Fig- 

 ure 4(c) relative to Figure 4(a). A possible cause 

 for this shift in maximum upwelling is described 

 in the following section. 



An Effect of Coastal Topography 



During the summer an intense thermal low de- 

 velops in the interior of southern California. Due 

 to the 3-degree mesh length of the computation 



grid, the meridional component of the geostrophic 

 wind at a coastal gridpoint is computed (Equation 

 2) from a continuous constant pressure gradient 

 between an offshore gridpoint and one on the con- 

 tinent (dashed line in Fig. 5). However, the south- 

 ern California coastal mountain range causes a 

 discontinuity in this pressure gradient such that 

 the gradient actually in equilibrium with the geo- 

 strophic wind (solid line in Fig. 5) may be less 

 than that used in the computations. This leads to 

 an overestimation of the geostrophic wind that 

 in turn leads to an overestimation of the upwell- 

 ing index. The assumption made in presenting 

 these series as indicators of time variations of 

 intensity of coastal upwelling is that the actual 

 pressure gradient at the coast varies in rough 

 proportion to the total onshore-offshore pressure 

 difference which is used in the computations. 



MEAN YEARLY CYCLE OF 

 INDICATED UPWELLING 



Having been designed to give a consistant in- 

 dication of temporal variations at each particular 



LOW 

 PRESSURE 



Figure 5. — Height of a constant pressure surface. The slope of the dashed line indi- 

 cates the assumed gradient used in the calculations. The slope of the solid line indicates 

 the lesser gradient which mav exist in realitv. 



8 



