CHAPTER 3 



presence of spurious values in a relatively small sample. 

 A smoothed line was then computed by fitting a 5th-order 

 polynomial to each set of data using the least-square 

 method. This degree was chosen to resolve as many long- 

 term fluctuations as the computer and available software 

 permitted. 



The 1876-1976 mean for the northwest area is about 

 0.5° C cooler than the mean for the 1949-73 reference 

 period, but the standard deviation is essentially the same. 

 The 1896-1976 mean for the southwest area is about 0.4° C 

 lower than the 1949-73 mean. Some cyclical behavior is 

 apparent in the values for both areas, with the amplitude 

 of the oscillation being more pronounced in February than 

 in April. The temperature record for the two areas shows 

 generally good coherence. About 1970, there is a strong 

 warming trend in both sets of data, amounting to about 

 2° C. Namias and Dickson (1976) have shown that this 

 warming extended into the Atlantic over a broad zone. 



The record of mean monthly air temperatures for Feb- 

 ruary and March at New York City's Central Park for 

 1876 to 1976 is presented for comparison (fig. 3-4). The 

 mean temperature for February 1976 was 4.4° C, a value 

 exceeded only once (4.5° C in February 1954) during the 

 previous 100 years. Although March 1976 was not so near 

 to a record as February, the mean of 6.9° C was exceeded 

 only eight times in the previous 100 years. Thus it rep- 

 resents a substantial positive departure from the long-term 

 mean. Furthermore, when the averages for February and 

 March are combined, the resulting 2-month average for 

 1976 is the second warmest in 100 years, being exceeded 

 only by February-March 1945. The records for sea-surface 

 temperatures in New York Bight and air temperatures in 

 Central Park correlate well, particularly when their 

 smoothed (best 5th-order fit) curves are compared, as 

 might be expected. 



The sea-surface temperatures (fig. 3-3) during late win- 

 ter and early spring of 1976 in the western sector of the 

 Bight were among the highest recorded over a long period 

 of years. Because of this, it was thought that some cor- 

 relation could be established between the abnormally high 

 late winter and early spring sea-surface temperatures and 

 the unusually low oxygen concentrations subsequently 

 observed. However, relatively high sea-surface tempera- 

 tures also occurred during the several years before 1976, 

 and in the early 1950s and late 1940s, without the kind of 

 large-scale ecosystem disruption observed in summer 

 1976. Although fishkills were documented less extensively 

 before the 1976 event, those recorded in the area were 

 limited in extent. (See chapter 1.) 



Thus, it seems likely that other factors besides higher 

 sea-surface temperatures acted to produce the anoxia. The 

 simultaneous occurrence of anomalous wind conditions in 

 the Bight appears to have been one such factor. The min- 

 imum cyclonic activity between February and June 1976 



may have contributed directly to the early development 

 of stratification by significantly reducing vertical mixing 

 and contributing to shallow, intense stratification in the 

 water column (Elsberry and Garwood 1978). Although 

 1974 and 1975 sea-surface temperatures were similarly 

 warm, more than twice as many storm systems crossed 

 the Bight in those years as did in 1976 (fig. 3-1). 



SURFACE WIND FIELD 



The surface wind field analysis included observations 

 at four land stations (Atlantic City, Newark. John F. Ken- 

 nedy International Airport, and Westhampton Beach. 

 L.I.), two stations at lights (Ambrose and Barnegat), and 

 two environmental buoys (EB34 and EB41) (fig. 3-5). 

 These data are contained in Local Climatological Data 

 (LCD) summaries, original observation forms on file at 

 NCC, and computer products based on data received from 

 the National Data Buoy Office and routinely archived at 

 the National Climatic Center. The observations for Feb- 

 ruary through August 1976 were averaged by 10-day seg- 

 ments for each station and plotted as a time series of mean 

 resultant wind vectors for the four stations at sea and 

 Westhampton Beach (fig. 3-6). The 10-day interval was 

 chosen to account, as much as possible, for within-month 

 variations in wind regime while minimizing the number 

 of computations required. Ten days also is roughly equiv- 

 alent to the mean time interval between the passage of 

 cyclonic disturbances in the New York Bight area during 

 the spring months. The wind vectors were subjectively 

 analyzed to give mean flow patterns for each 10-day seg- 

 ment over the 7 months. From these the mean direction 

 and speed at each of six grid points (identified in fig. 3-5), 

 plus Westhampton Beach, were used to compute the mean 

 surface wind stress for each month February through June, 

 using the bulk aerodynamic formula 



t = pC„|V|V. 



:i) 



where V is the 10-day mean vector wind, "p is the mean 

 air density (1.26 x 10 ' g/cm') calculated from the av- 

 erage surface air pressure and temperature field over the 

 Bight and C„ is the drag coefficient. A value of 2.6 x 

 10 ' for Co was used; this relatively large value for the 

 drag coefficient was adopted primarily to compensate for 

 the smoothing inherent in the use of time-and space-av- 

 eraged winds (see Bakun 1973). The stress vectors for 

 each 10-day interval at each grid point were averaged to 

 produce monthly mean values for February through June. 

 These in turn were spatially averaged to obtain an estimate 

 by month of the magnitude and direction of the surface 

 wind stress over the Bight. 



In figure 3-7, surface wind patterns for February 

 through August 1976 are compared with the long-term 



55 



