The first injection was 5 m above the bottom in 6.8°C water on the seaward side of the front. Over 4 days the dye tagged water warmed roughly linearly to 7.4°C implying a cross-frontal Lagrangian velocity of 1.6 cm/s. The mean along-bank velocity was 4.1 cm/s
A second dye injection was in 7.0°C water in virtually the same location with respect to the tidal front but the subsequent spreading was strikingly different. Temperature change of the dye tagged water over 4 days was negligible, <<0.1°C, implying no cross-frontal flow. The hydrographic difference was the proximity of the foot of the shelfbreak front now only 5 km off bank of the dye patch, in contrast to > 20 km during injection #1. This suggests a divergence or null in the cross-bank flow near the midpoint between the tidal and shelfbreak fronts.
The third injection was in 6.85°C water seaward of the tidal front on the northeast peak at 66.9W. The dye tagged water warmed rapidly and linearly to 8.5°C in just over 2 days implying a cross-frontal velocity of 3.7 cm/s. There is an indication that once dye tagged water passed entirely through the front and mixed throughout the water column over the bank it began to flow off-bank in the pycnocline.
From the increase in the cross-bank width of the dye patch we estimate a Fickian lateral diffusivity of 50-75 m²/s on the south flank and 75-100 m²/s at the northeast peak.
The dye injection experiment has provided unambiguous evidence of an on-bank flow through the tidal front in the bottom mixed layer of Georges Bank.