19 
samples collected in space (i.e., number of sampling stations), this indicates that they 
are to provide a statistical framework that truly accounts for error in Chesapeake Bay 
water quality criteria assessments. 
Steps 7 and 8—Plotting and Comparing the Curves 
When comparing the assessment curve to the reference curve, the issue of uncer¬ 
tainty becomes most important. The preceding discussion clearly indicates that 
uncertainty in the assessment curve represents an accumulation of uncertainty gener¬ 
ated in and propagated through the preceding steps. If the reference curve is 
biologically based, it is derived under the same system of error propagation. Devel¬ 
oping the statistical algorithms to quantify this uncertainty poses a challenge. 
Even if the uncertainty can be properly quantified, the issue of who gets the benefit 
of doubt due to this uncertainty can prove difficult to resolve. 
This problem of uncertainty in the regulatory process is widespread and not limited 
to the CFD approach. Nonetheless, it must be dealt with. One option is to require that 
the assessment curve be significantly above the reference curve to establish non¬ 
attainment. This option protects the regulated party from being deemed out of 
attainment due to random effects. If assessment CFD curves are not accurately deter¬ 
mined, however, it could lead to poor protection of environmental health and 
designated uses. A second option is to require that the assessment curve be signifi¬ 
cantly below the reference curve to establish attainment. This option results in strong 
protection of the environmental resource, but could lead to the regulated party imple¬ 
menting unnecessary and expensive management actions. 
Some compromise between these extremes is needed. The simplest compromise is 
to ignore variability and compare the assessment curve to the reference curve. As 
long as unbiased estimation is implemented for both the assessment curve and the 
reference curve, this third option will result in roughly equal numbers of false posi¬ 
tive (declaring non-attainment when, in fact, compliance exists) and false negative 
(declaring attainment when, in fact, non-attainment exists) results. This last 
approach is balanced and the one currently recommended by EPA. Under this 
approach, however, no mechanism exists to motivate error reduction by improving 
the data sets on which the criteria assessments are based. 
RESULTS OF THE SCIENTIFIC EVALUATION 
Beginning in fall 2005, the Chesapeake Bay Program’s Scientific and Technical 
Advisory Committee (STAC) appointed a panel of scientists to evaluate and refine 
the CFD water quality criteria assessment methodology. Evaluations included tests 
on the effects of: 1) sample densities in time and space; 2) varying levels of attain¬ 
ment; and 3) varying degrees of spatial and temporal covariance. Appendix A 
provides a complete copy of the panel’s final report while Appendix C offers a narra¬ 
tive evaluation of the options for spatial interpolation. 
chapter ii 
Refinements to the Chesapeake Bay Water Quality Criteria Assessment Methodology 
