NOTES 



THE MEAN ANNUAL CYCLE OF 



COASTAL UPWELLING OFF WESTERN 



NORTH AMERICA AS OBSERVED FROM 



SURFACE MEASUREMENTS 



One of the world's major upwelling regions lies off 

 the west coast of the United States and northern 

 Mexico. This paper summarizes marine surface 

 observations to describe the normal yearly cycle of 

 intensity of upwelling for the major portion of the 

 northeastern Pacific coastal upwelling region. 



Sverdrup (1938J applied Ekman's (1905) theory 

 to account for a coastal upwelling situation 

 observed off southern California. He proposed a 

 mechanism by which water is transported off- 

 shore in the surface Ekman layer due to the stress 

 of the wind on the sea surface and is replaced 

 by water upwelled from depth. Wooster and Reid 

 (1963) presented evidence that this is, indeed, 

 the dominant mechanism acting in regions of 

 slow, diffuse eastern boundary currents wherein 

 lie the major coastal upwelling areas of the world, 

 including that of the northeastern Pacific. 



Our approach is to define the mean annual 

 cycle of offshore Ekman transport along the west 

 coast of the United States and the immediately 

 adjacent regions of Canada and Mexico and to 

 correlate this with features indicative of upwell- 

 ing which appear in the long-term mean monthly 

 distributions of sea surface temperature. 



Marine surface weather observations for this 

 study were obtained from a version of the National 

 Climatic Center's tape deck of marine surface 

 observations 'Tape Data Family-11) in use at the 

 U.S. Navy Fleet Numerical Weather Central. The 

 observations in this file come primarily from 

 merchant and naval ships and sometimes contain 

 various errors in position, measurement, or 

 processing. Consequently, the sea surface 

 temperature data were subjected to an editing 

 process which consisted of two filters. First, a 

 gross error check was performed to eliminate 

 nontemperatures. Second, the data were checked 

 by comparison with a running mean of 10 reports. 

 When a report of sea surface temperature differed 

 from the running mean by greater than 9'C, the 

 report was rejected. Wind speeds of greater than 

 100 m/s were rejected. "Variable" winds (no 



direction reported, low reported speedj were 

 treated as calms. 



The Ekman transport was calculated by the 

 following procedure. The stress vector was 

 computed from each wind observation according 

 to the classical square law: 



a L) 



\vW , 



where f is the stress of the wind on the sea 

 surface, fj,^ is the density of air (0.00122 g/cm''), 

 Cd is an empirical drag coefficient '0.0013), 

 V is the observed wind velocity vector, and 

 \~i> \ is the observed wind .speed The resultant 

 Ekman wind stress transport, M, was computed 

 according to , 



M = -T 



X k 



where r is the wind stress vector, f is the Coriolis 

 parameter, and ^ is a unit vector directed 

 vertically upward. 



Figure 1 displays composite monthly values of 

 these data for the 20-yr period, 1948-67. The plot 

 on the left displays time series i.sograms of 

 offshore component of Ekman tran.sport while the 

 central plot shows .similar i.sograms of .sea sur- 

 face temperature. The coordinates are north- 

 south di.stance on the ordinate and time by month 

 on the ab.sci.s.sa. Each plot represents about 75,000 

 individual observations made within the l'^ 

 squares shown in the coastline plot to the right. 

 The number of reports per Y' square per month 

 was in the range 22 'in January off Vancouver 

 Island) to 1,884 'in October off Los Angeles). 



The sea surface temperature plot reveals a 

 normal north-south gradation of temperature and 

 a seasonal warming-cooling cycle with minima 

 from February to April and maxima in August 

 and September. The effects of upwelling are seen 

 as distortions of this general pattern. ^ 



In the northern portion. Cape Blanco to 

 Vancouver Island, offshore Ekman transport is 

 weak and occurs from about May through 



'The fine scale detail.s of the temperature distribution are 

 masked by the 1' square spatial averages. For a detailed 

 treatment of the mean temperature cycle in the southern portion 

 of the region the reader is referred to Lynn '1967;. 



843 



