This biological reference curve (see below for details) is based on dissolved oxygen 
concentration distributions at sites associated with bottom sediment-dwelling 
benthic communities scoring 3 or higher on the Chesapeake Bay benthic index of 
biotic integrity (benthic-IBI). If several of the reference segments were randomly 
removed, the regenerated reference curves do not change much, suggesting that 
within designated uses, the attainment curves for reference segments appear to be 
very similar. Although less firmly grounded, the reference curves for other desig¬ 
nated uses and other criteria also seem to be relatively stable. 
APPROACHES TO DEFINING REFERENCE CURVES 
At least three options exist for defining a reference curve (Figure VI-14). Fixed 
percentages could be selected based on a policy decision or other basis similar to the 
10 percent level of acceptable exceedances allowed in 305(b) EPA national guidance 
(Figure VI-14a; U.S. EPA 1997). Alternatively, laboratory or empirical field data 
from areas known to be unimpaired by the stressor can be used to derive a biologi¬ 
cally-based reference curve (Figure VI-14b). Even this second approach, however, 
requires technical or policy decisions regarding the acceptable level of biological 
effect. Finally, a reference curve could be established to reflect uncertainty based on 
the assumption of a normal distribution, and using observed or estimated error vari¬ 
ance for both time and space (Figure VI-14c). 
0 10 20 30 40 50 60 70 80 90 100 
Percentage of Area/Volume Exceeding the Criteria 
Figure VI-14. Three possible options for setting reference curves for application 
to the cumulative frequency distribution approach for defining criteria attainment: 
(a) fixed percentages based on policy decisions; (b) biological effects-based empirical 
field or laboratory data and; (c) observed or estimated uncertainty data. 
chapter vi • Recommended Implementation Procedures 
