FISHERY BULLETIN: VOL. 72. NO. 2 



four oceanographic parameters: water depth, 

 secchi disc transparency, Forel-Ule color, 

 and salinity (Table 4). The models contain selected 

 2-factor interactions formed as products between 

 parameters and treated as additional independent 

 variables. Interaction selection was based on 

 whether or not an interaction significantly in- 

 creased the precision of the estimate 0). The 

 models were constructed from data collected on 

 main days (i.e., 7 August, 25 August, and 28 Sep- 

 tember 1972) and are presented separately and in 

 combination and with and without the inclusion of 

 color as an independent variable. 



Model Testing and Interpretation 



The models were tested by playing them with 

 oceanographic data collected during commercial 

 fishing operations and main day sea-truth station 

 data stratified to include only those stations where 

 menhaden were not detected photographically 

 (Figure 4). Ideally, model products for fishing data 

 should have grouped close to 1, and products for 

 the "without fish" sea-truth stations should have 

 grouped close to 0; obviously, this type of grouping 

 is not demonstrated in Figure 4, indicating a gen- 

 eral lack of accuracy and precision in the models. 

 Product populations, however, are significantly 



1-0 1,0 



MODEL I'm: DICTIONS 



4 



Figure 4. — Histogram plots of "with fish" (shaded) and "with- 

 out fish" (unshaded) model products. 



Table 4. — Empirical regression models which predict menhaden distribution (D) in the ERTS-1 study area. 



B = water depth (m) S = salinity (ppt) 



T = Secchi disc transparency (m) C = Forel-Ule color (units) 



BT, BS, ST, CT, and CS = interactions formed as the products of the respective parameters. 



D1 7 Aug. 82 D = 1.9959-0 06645+0 7453 7-0.68208- 



0.0233S7 - 0.01448 7 + 02308S 



D2 25 Aug 42 D = 5.1537 - 0.1740S - 0.91957 - 0.0371C - 



0.43508 + 0.0502S7 - 0.12438 7 + 0.01958S 



D3 28 Sept. 73 D = 2.3473 - 0934C - 0.81178 - 0.0358S 7 - 



0.0007CS + 0.0528C7 + 0.05168 7 + 0.02358S 



D4 7 and 25 124 D = 2.4691 - 0855S + 0.3948 7 - 0.64778 - 



Aug 0.0054S 7 - 0.04418 7 + 0.02238S 



D5 7 Aug and 155 D = 1.8559-0 05775 + 0,56047-0 69548- 



28 Sept. 0.0191S7 - 0.00798 7 + 0,023285 



D6 25 Aug and 115 D =2.9396-0.10245 + 0.15227-0 74868- 



28 Sept. 0.00265 7 - 05478 7 + 026885 



D7 25 Aug and 115 D = 3.6035 - 0.09875 -0 12497 - 0.0416C - 



28 Sept 0.67178 + 00875 7 - 04418 7 + 0.023485 



D8 7 and 25 Aug. 197 D = 2,3759-0 07975+0.3928 7-0.70518- 



and 28 Sept 00865 7 - 03268 7 + 024285 



0.2492 



0.3793 



0.2443 



0.3009 



02489 



0.596 



0.630 



0.409 



0.584 



0480 



3118 0.488 



3090 508 



99 



99 



90 



99 



99 



99 



99 



028560,515 



99 



386 



