monthly chart were averaged. Figure 7 shows 

 a comparison of the average maximum westerly 

 components with corresponding values of the 

 Westerly index for winter (Dec., Jan., Feb.) 

 and summer (June, July, Aug.) during the 

 period 1947-57. In general high values of 

 the index are associated with high values of 

 maximum gradient and in no instance are very 

 low values of index associated with very 

 high values of maximum gradient. However, 

 the latter may often exceed the former by 

 more than 50 percent. This is true of about 

 one-third of the cases for the winter season 

 and one-thirtieth for the summer season. 

 These wide differences are confined to the 

 medium and low values. The spring and fall 

 seasons, not illustrated, are similar in 

 these features. Thus the index values do 

 indeed reflect changes in intensity of at- 

 mospheric circulation but for purposes where 

 the speed at the axis of the wind stream may 

 be important the index values would not serve 

 as a quantitative expression of intensity. 



The graphs in figure 5 (g and h, page 9) 

 for the two Alaska indices illustrate a case 

 of extreme disparity between similarly ori- 

 ented measures of the geostrophic wind field 

 in adjacent areas. Although both indices 

 exhibit greater variability in winter than 

 in summer, there is little agreement in the 

 year-to-year fluctuations of monthly anoma- 

 lies. Correlation coefficients computed 

 from corresponding monthly anomalies of the 

 two Alaska indices, for each seEison, are 0.30 

 (winter), 0.13 (spring), 0.31 (summer) and 

 0.40 (fall). Since pressure gradients were 

 often measured across low centers or troughs 

 at locations 26 and 27 and always over land 

 at locations 26a and 27a, probably neither 

 index by itself is a reliable measure of 

 wind strength over the Alaska current. 



It appears reasonable to speculate that 

 the index based on the offshore locations 

 might be the more representative of the two. 

 However, a comparative study with observed 

 winds should be made to determine their 

 relative significance. 



Another pair of indices representing 

 similarly oriented wind components in adja- 

 cent areas are the California (locations 9, 

 10) and the Alternate California (locations 

 34, 35, 36). Comparing the monthly anomalies 

 of these indices for each season, as above, 

 gives moderately good correlation coeffi- 

 cients for winter (0.64), spring (0.71) and 

 fall (0.76), but shows poor correlation for 



WINTER 



(DEC -JAN -FEB) 



1947-1957 



X 



UJ 



o 



UJ 



UJ 



X 



< 



PRESSURE DIFFERENCE -MILLIBARS 



WESTERLY INDEX 



Figure /.--Pressure differences representing "maximum 

 westerly" wind components (defined in text), 

 plotted against values of westerly index for 

 (a) winter months and (b) summer months. 



summer (0.37). 



In this case also, one of the indices 

 represents pressure gradients partially ex- 

 tending over land areas (locations 34, 35, 

 36). In summer these are strongly affected 

 by high temperatures in the valleys eind 

 deserts inland from the coast. For this 

 reason gradients at locations 9 and 10 are 

 probably more representative of summer off- 

 shore winds. 



DECADE MEANS 



Comparison of the decade means listed 

 in table 1 (page 18) provides an indication 

 of the long-term stability and trends for the 



10 



