HISTORICAL AND RECENT-EVIDENCE OF SCOTIAN SHELF WATER ON SOUTHERN GEORGES BANK

Historical data from 1912-1987 indicate that low salinity (less than 3 2.0 psu) near-surface water occurs occasionally on southern Georges Ba nk during May, while none occurs during April. Composite monthly plots of historical near-surface salinity and estimated advection rates sho w that the southwestern Scotian Shelf is the immediate upstream source of this low salinity water. Optimally interpolated satellite-derived SST and hydrographic data reveal that very cold (less than 2.0 degrees C), low salinity (less than 32.0 psu) Scotian Shelf Water (SSW), init ially located south of the 200-m isobath off southern Georges Bank in early March 1992, moved north onto southern Georges Bank during April 1992. SSW was not as extensive during the same period in 1993, as evid enced by significantly higher temperatures and salinities. These diffe rences show large interannual variability in the transport and/or prop erties of SSW flowing onto Georges Bank. Lower (higher) salinities mea sured during spring 1992 (1993) on southern Georges Bank are consisten t with higher (lower) St Lawrence River discharge noted during spring 1991 (1992) and the similar to nine month lag between annual discharge maxima from rivers located upstream and minimum salinity at Cape Sabl e in February. However, comparisons between historical occurrences of low salinity (less than 32.0 psu) SSW on southern Georges Bank noted f or May 1966, 1971 and 1978, and cumulative St Lawrence River discharge from the spring prior to each occurrence, show no relationship. This suggests that the occurrence of low salinity water on southern Georges Bank is not directly related to variations in upstream river discharg e. Although an idealized model of wind forcing on a bank-trapped densi ty front shows that nearsurface flow is related to superposition of an Ekman layer and the along-bank gyral circulation, evidence from late winter-spring 1992 and 1993 on Georges Bank shows that except for on-b ank movement of the shelf-slope front during late spring 1992, simple Ekman theory does not explain on-bank movement of the SSW plume during 1992 nor the immobile nature of the shelf-slope front during late win ter 1992 and late winter-spring 1993. Surface heat flux and satellite- derived SST from Georges Bank and the Scotian Shelf together with a sl ab mixed-layer model allows estimation of the advective heat flux and transport of SSW flowing onto Georges Bank from the Scotian Shelf duri ng late winter-spring 1992. Mean advective heat flux onto Georges Bank during March-May 1992 is -14.5 mW cm(-2) with an estimated error of s imilar to 3.5 mW cm(-2), demonstrating that advection of cold SSW onto Georges Bank is required. Mean velocity and transport of SSW flowing onto Georges Bank during March-May 1992 are 13.4 cm s(-1) and 0.21 Sv, assuming a SSW depth (width) of 40 m (40 km), demonstrating that how of SSW onto Georges Bank was robust during the period. At this time, t he dynamics that control the flow of SSW across Northeast Channel are unknown. Published by Elsevier Science Ltd.