KEMMERER ET AL : ERTS-1 MENHADEN EXPERIMENT 



oceanographic data into meaningful information 

 has proceeded slowly because of interpretation dif- 

 ficulties. Thus, reported fisheries oceano- 

 graphiuc-parameter relationship analyses de- 

 pend primarily upon sea-truth measurements. 

 An essential exception was the photographically 

 acquired menhaden distribution and abundance 

 information. 



Oceanographic Parameter-Fish 

 Distribution Relationships 



Analysis 



The distribution and abundance of menhaden 

 in the study area, principally in the Mississippi 

 Sound, can be placed into a simplified systems 

 context (Figure 2). Factors directly affecting the 

 system, i.e., the distribution and abundance of 

 menhaden, include fish input, fish output (includ- 

 ing harvest, death, and emigration), the environ- 

 ment, and the innate behavior of the menhaden 

 not directly or immediately influenced by environ- 

 mental conditions. Examples of this latter 

 factor include fish age and degree of sexual 

 maturity. This systems concept can be modified 

 slightly and expressed as an algebraic argument 

 as: 



A,.y = f(E,B,P) (1) 



where: 



A = number of menhaden schools, 



X and y = refer to school location coordinates, 

 E = environmental conditions, 

 B = innate fish behavior, and 



P = instantaneous menhaden school 

 population. 



The problem with the argument is that the de- 

 pendent variable A^.vis a function of more than 

 just the environment, E, and as such cannot be 

 solved with available information. To simplify the 

 expression, two assumptions were made. First, it 

 was assumed that 5 was constant (i.e., the innate 

 behavior of the menhaden did not vary signifi- 

 cantly) and thus could be ignored in the expres- 

 sion, an assumption which led to the development 

 of a new expression where A^^y became a function 

 of E and P alone. This assumption appeared 

 reasonable because only adult menhaden were 

 considered in the experiment while they were 

 in the Mississippi Sound, a relatively short period 

 of time. The second assumption made was that 

 A^,y could be expressed in relative terms such 

 that: 



'x,y 



= f(E) 



(2) 



This assumption permitted the normalizing of 

 Ax,y relative to P and has its roots in many fisheries 

 catch/effort related expressions. 



In the subsequent analyses, the number of 

 photographically detected menhaden schools 

 at any given point was used as an estimator of 

 Ax,y, and the total number of detected menhaden 

 schools was used as an estimator of P. If there was 

 a constant sensor-caused bias in the photography 

 data, the quotient Ax,y/P should not be affected 

 seriously, as the bias cancels. However, if the 

 bias was not constant but instead was a variable 

 function of the environment, then the bias 

 would affect the quotient. Whether or not the 

 effect would be significant would depend on the 

 magnitude and variability of the bias. 



Because of a concern about the possibility of bias 

 affecting the relationships, a second approach also 

 was used which should have reduced sensor bias. 

 A new dependent variable, D, was defined which 

 reflected only the distribution of menhaden and 

 was related to the environment as: 



D = f(E) 



(3) 



Figure 2. — Simplified systems view of the Mississippi Sound 

 menhaden population described only in terms of distribution and 

 abundance. 



Inherent in this expression is the assumption 

 that P does not affect the distribution of menha- 

 den within the extremes of P characteristic of the 



379 



