has undergone a least-squares fit of three straight 

 lines in an attempt to define the temperature structure 

 of the surface-layer, thermocline, and lower waters in 

 a mechanized and unambiguous way. 



During a traverse from KILO to JULIET, we took BT ' s 

 each thirty minutes and performed the three-line analysis 

 of each BT trace. Figure 1 shows the traverse in "BT 

 space". The results were plotted [pig. 2] in Ref. 4 

 versus an ASWEPS-type forecast of the area, and versus a 

 simple climatological forecast from H.O. Pub. 76 1. 

 The results in Fig. 2 for layer depth show that the 

 climatological forecast is as good, or better, than the 

 ASWEPS-type forecast. I must emphasize that these 

 results were for the cooling season (the cruise was in 

 September) and cannot be extrapolated or particularly 

 interpreted in terms of other seasons without a much 

 better basic understanding of the mechanisms involved. 



One phase of MILOC 64 consisted of two ships, side- 

 by-side, twelve nautical miles apart, traversing from 

 KILO to JULIET and taking sjmoptic sea-surface tempera- 

 ture and BT's each thirty minutes [pig. 3]. The SST's 

 were plotted against each other to establish the 

 coherence of the two sets of measurements and the 

 coherence was, indeed, quite good. The two sets of 

 BT's were put through our three-line-analyses^ and the 

 layer depth was plotted LFig. 4] again to establish the 

 coherence. The results were not encouraging in the 

 least; it appears that, although a sideways extra- 

 polation of SST for twelve nautical miles is 

 appropriate, it would not be prudent to expect a 

 twelve-mile extrapolation of layer depth to be too 

 meaningful. If we want to use airborne measurements of 

 SST to infer layer depth, it appears that the correlation 

 of SST and layer depth is not certain. Again, I must 

 emphasize that these conclusions apply to this region 

 (KILO to JULIET) in this season (September). 



The MILOC 's in 1965 and 1966 were at the same 

 location and were similar to that in 1964^ but were in 

 the heating season and with the addition of much more 

 meteorology and a more sophisticated approach to the 

 problem of spatial variability. In MILOC 65 we again 

 traversed KILO to JULIET and subjected the BT's obtained 

 to the three-line analyses. The resulting layer depths 

 and first-layer gradients were plotted [Fig. 5] on a 

 typical range-prediction chart in a f requency-of- 



65 



