these are relict, ice-rafted deposits. Each of 

 Naidu's conclusions is examined below. 



Naidu's conclusion seems to be in part correct, 

 although his arguments need support. There are 

 few modern reports of gravel on sea ice in the 

 Beaufort Sea (Kindle, 1924; Carsola, 1954; 

 Lafond, 1949; R. Lewellen, oral communica- 

 tion). The plausibility of ice-rafting as a modern 

 transport mechanism requires an estimate of how 

 much gravel would have to be in the ice canopy at 

 any one time to account for the present accumula- 

 tion on the shelf. On the basis of observations 

 during WEBSEC 1971-73 of (1) the number of 

 gravel patches obsei-ved (there was a total of 3), 

 (2) estimates of the average amount in each oc- 

 currence (5 metric tons), (3) the area observed, 

 (4) the rate of pack movement, and (5) the cover 

 over the shelf, I estimate an accumulation rate of 

 5 mg/cm^/lO'^yr. 



This would amount to 85 mg of gravel per cm^ 

 in the 5-10 meters of Holocene sediment which is 

 found on the shelf (Reimnitz and others, 1972b 

 and unpublished data). The amount of gravel seen 

 on the shelf is conservatively estimated to be 

 20,000 mg per cm^ in the Holocene section {2% 

 gravel in 5 m of sediment). This calculation sug- 

 gests an accumulation rather more than 2 orders 

 of magnitude higher than can be accounted for by 

 my ice observations. If ice is rafting gravel at the 

 same rate today as throughout the Holocene, 

 there should be at least 100 times as much gravel 

 as I actually observed. I conclude that modern 

 ice-rafting of gravel is insignificant. 



An "inverse" sediment grading with coarser 

 (more gravel) sediments offshore would be ex- 

 pected if ice-rafting were an active agent. During 

 any melt season (the time when gravel is most 

 likely released from floating ice) on tiie average 

 more ice would be present on the outer shelf than 

 on the inner shelf (U.S. Navy Hydrographic Of- 

 fice, 1958). Therefore the number of potential 

 gravel release points would be higher on the outer 

 shelf, and greater abundance of ice-rafted gravel 

 should be expected. Thus the lack of a coarse-to- 

 fine sediment gradation as noted by Naidu (1973 

 and this volume) c(juld be used as an argument in 

 support of ice-ralting as an important mechanism 

 of sediment transport. In contrast, my studies 

 document ice-rafting as an important mechanism 



in shelf transport; however, as Naidu notes, the 

 question often is not how, bpt when the rafting 

 occurred. 



Direction of Sediment Transport 



The overall movement of ice and water in the 

 Beaufort Sea gyre off the northern coast of Alaska 

 is to the west (Campbell, 1965). On the inner 

 shelf there are significant short-term shifts in the 

 pattern of ice and water movement (table I). Here 

 the current and ice movement during the open 

 season are dominated bv the wind regime, and 

 both easterly and westerly currents up to 100 

 cm/sec are observed (Kinney and others, 1972; 

 Short, 1973; and Hufford, this volume). 



Studies on the inner shelf during the season of 

 ice cover (Barnes and Reimnitz, 1972 and unpub- 

 lished data) show an overall westward water 

 movement at very low velocities (less than 2 

 cm/sec). This is probably due to the fact that the 

 ice cover reduces wind stress as a water- 

 movement mechanism. 



The bulldozing effect of ice gouging appears 

 also to be dominantly in a westward direction 

 (Reimnitz and Barnes, 1972). As a result, sedi- 

 ments whether a) in suspension, b) resuspended 

 by currents or ice gouging, c) carried by rafting, 

 or d) bulldozed are, in toto, probably being trans- 

 ported westerly on the Alaskan shelf. The sedi- 

 ments reflect this westward transport in being 

 finer grained and better sorted to the west along 

 tlie slielf (fig. 7, 8, & 9). The westward gradation 

 is empiiasized by two factors. Firstly, the Alaskan 

 shelf is essentially closed to major sediment input 

 from the eastern end. Sediment load of the Mac- 

 kenzie River does not reach the central Beaufort 

 Sea; instead it appears to be transported eastward 

 from the delta (Collin 1962; and unpul)lished 

 ERTS-1 satellite piiotos). Secondlv. the coast ex- 

 lends further from tlie foothills of the Brooks 

 Range in a westward direction. The river sedi- 

 ments reflect this longer coastal plain traverse to 

 the west by debouching finer and more mature 

 sediments in a westward direction. 



The eastward coarsening appears to end just 

 east of Herschel Island on the western edge t)f the 

 Mackenzie Canyon (Wagner. 1972). East of this 

 region sedimentation appears to be dominated by 

 the Mackenzie River. 



191 



