The ThermocIine .--Completely isothermal 

 water is seldom found at any of the lightship 

 stations employing the bathythermograph. 

 From November through February, bottom 

 water tends to be slightly warmer than surface 

 water since mixing does not keep pace with 

 chilling. When vernal warning begins at the 

 surface, mixing continues to produce a similar 

 warming at depth, but again with a lag, so 

 that a negative gradient is established. The 

 thermocline thus produced is strongest during 

 July and August at a depth between 30 and 60 

 feet. At stations from Ambrose through Chesa- 

 peake the temperature gradient is from time 

 to time intensified by advection, possibly 

 through upwelling, of colder water near the 

 bottom. Destruction of the thermocline is 

 directly associated with autumn storms, often 

 connected with the passage of hurricanes. 

 If such storms occur early in the autumn, the 

 thermocline may reappear briefly before the 

 final overturn. 



Bottom Water Temperature .--Warming and 

 chilling at the bottom are not the steady 

 progressions seen at the surface. There are 

 sharp changes in rate and frequent reversals 

 in trend. When thermal stratification is pres- 

 ent, abrupt warming at the bottom is ap- 

 parently the result of wind mixing, as at the 

 time of the autumn overturn when bottom values 

 reach their maximum, sometimes rising as 

 much as 18 F. in 48 hours. Abrupt cooling 

 appears to be associated with the advection 

 of colder offshore water. 



Salinity . --From Ambrose Lightship north- 

 ward, the surface salinity minimum near 

 shore occurs in late April or early May, 

 reflecting the peak runoff augmented by snow 

 melt. At Georges Shoals, farther offshore, 

 the minimum probably appears in late summer . 

 From Barnegat Lightship southward the salin- 

 ity minima may occur at any time and nnore 

 innmediately reflect local precipitation, since 

 there is little storage by freezing on the ad- 

 jacent watersheds. At most stations. Savannah 

 excepted, the January and February surface 

 salinity readings show little year-to-year 

 difference. 



The two principal factors influencing the 

 salinity regime at the different lightship sta- 

 tions are fresh water runoff from the land 

 and incursions of highly saline oceanic water. 

 The interplay of these forces is most apparent 

 at Diamond Shoals, where Gulf Stream water 

 of 36.0 /oo often is replaced in a matter of 

 hours by water of 30.0 °/oo. 



The effect of runoff is most noticeable at 

 Portland, Ambrose, Chesapeake, and Savannah 

 stations, which are near nnajor sources of 

 outflow. The salinity fluctuations at these 

 stations run parallel to the gaged river flow 



with only a slight temporal lag. At other 

 stations further offshore the fluctuations are 

 less extreme and occur later than at the 

 inshore stations. 



Monthly mean salinity values between Am- 

 brose and Winter Quarter stations increase 

 toward the south. Bigelow (1935) drew dia- 

 grams of the locations of surface isohalines 

 for successive months, based on data from 

 sections extending offshore across the shelf 

 in this region. While his values increase 

 offshore along a given section, his isohalines 

 curve toward the coast in the Chincoteague 

 region, also showing a southward increase 

 in salinity. 



Ketchum and Keen (1955) discussed this 

 phenomenon in a study of the accumulation 

 of river water over the shelf in the same 

 region, concluding that considerable local 

 mixing across the shelf must take place. The 

 lightship data tend to confirm this conclusion. 



The annual salinity cycle at the lightship 

 stations reflects the annual precipitation- 

 runoff regime over the east coast. If dynamic 

 gradients from the shore out over the shelf 

 contribute strongly to the circulation patterns, 

 there will be large seasonal and year-to-year 

 variations in coastal currents. Southwesterly 

 movement along shore should be greatest 

 during April and May at the time of peak 

 runoff and before the summer southwesterly 

 winds become established. During the winter, 

 the dynamic gradient would be weakest. 



Meteorological Effects . --The influence of 

 weather on the hydrography of the area of 

 study has been demonstrated in several in- 

 stances. Bumpus (1960) showed that runoff is 

 a critical factor in inducing the cyclonic move- 

 ment in the Gulf of Maine, beginning in late 

 winter and early spring. The earlier postula- 

 tion of Bumpus and Pierce (1955) concerning 

 the penetration of Virginian coastal water 

 southward past Cape Hatteras has been sub- 

 stantiated (Bumpus, 1957; Chase, 1959; Day, 

 1959). Wells and Gray (1960) found a close 

 correlation between the abundance of Mytilus 

 edulis in the Beaufort, N.C. region in June 

 and the frequency of northeast storms at 

 Cape Hatteras during the reproductive period 

 of the preceding autumn. Chase has also 

 shown the effect of northeasterly wind regimes 

 on the temperature structure during the sum- 

 mer at stations between New York and Chesa- 

 peake. 



Only in the broadest terms is it possible 

 to see a direct relationship between air 

 temperature anomalies and trends in water 

 temperature, except at stations close to shore, 

 e.g., Woods Hole. Consideration of wind 

 systems, however, as in the study by Chase 



