similarities between a suspected cause and the decline in landings, 

 the cause may be further evaluated to determine whether it was 

 the actual reason for the observed problem. Generally, charac- 

 terization efforts do not establish cause-effect relationships. Ab- 

 solutes or certainties are replaced with a philosophy that accepts 

 a preponderance of evidence. 



Temporal Trend 



1950 1960 



1990 



Possible Causal Factors 



Overfishing 



Habitat Loss 



Pollution 



c 

 <-.2 

 to 2 

 J c 



1950 



\_ 



1990 1950 



1990 1960 



Years 



Years 



Years 



1990 



Determine If Further Confirming Information Available 



Figure 3.4 To suggest problem areas, temporal trends in 

 important resources may be correlated with 

 possible causal factors. Shaded zone indicates 

 comparison of possible causal factors to 

 temporal trend in fish landings. 



Spatial patterns — relative locations — can also be used to identify 

 problems and suggest possible causes. A gradient (that is, a 

 continuous increase or decrease) in the distribution of copper 

 concentrations in bottom sediments found near the head of an 

 estuary is illustrative (Figure 3.5). Close inspection of the gradient 

 might show that concentrations are highest immediately 

 downstream from a city, but that they decline both up- and 

 downstream as one moves farther away. Such a pattern suggests 

 that the copper contamination originates in the city, and that 

 subsequent analyses should attempt to identify the exact source. 



Another analytical technique to determine the potential causes of 

 environmental problems is an evaluation of relationships among 

 environmental variables. The reduction of aquatic vegetation may, 

 for instance, be cited as a priority problem in an estuary. The 

 association between declining vegetation and a second estuarine 

 variable, such as the presence of nitrogen in the water column, 

 could be examined (Figure 3.6). If there is a positive correlation, 

 nitrogen could be the culprit. 



31 



