A-34 
Table 4.1. Chlorophyll criteria derived by computing and upper threshold based on predicted 
means for mid-flowl960s chlorphyll data. 
Season 
Salinity 
Zone 
Mean 
Log(chl) 
GMmean 
(chi) 
Temporal 
Variance 
Spatial 
Variance 
Std 
Dev 
lOR(chl) 
Threshold 
Criterion 
log(chl) 
Threshold 
Criterion 
(chi) 
Spring 
OH 
0 7684 
5 87 
0 0233 
0 0658 
0 2985 
1.2594 
18 17 
Summer 
OH 
1 1693 
14 77 
0 0233 
0 0658 
0 2985 
1 6603 
45 74 
Spring 
MH 
0 4137 
2 59 
0 0233 
0 0658 
0 2985 
0 9047 
8 03 
Summer 
MH 
0 8626 
7.29 
0 0233 
0 0658 
0 2985 
1 3536 
22 58 
Spring 
PH 
0 1386 
1 38 
0 0233 
0 0658 
0 2985 
0 6296 
4 26 
Summer 
PH 
0 218 
1 65 
0 0233 
0 0658 
0 2985 
0 709 
5 12 
Relative to the standard reference curves, the curve based on the observed variance 
ratio for chlorophyll is more restrictive of events where large portions of the popu¬ 
lation are out of compliance. For example, the CBP standard reference (curve 1) 
would allow 40% of area to exceed the criterion threshold up to about 6% of the 
time. The proposed chlorophyll reference curve (curve 3) would restrict occurrences 
of 40% of area out of compliance to about 2% of the time. Conversely, the proposed 
curve (curve 3) allows a higher frequency of events where a small percentage of 
space in out of compliance. For example, 10% of space is allowed out of compliance 
36% of the time under the proposed curve and 27% of the time under the standard 
curve. 
While there is mathematical and statistical logic underpinning this proposed chloro¬ 
phyll reference curve, it is important to remember that it is based on parametric 
models and simplifying assumptions. It is recommended that validation exercises be 
performed to insure that the general shape of CFD curves generated from data 
collected in near reference conditions is approximated by the proposed curve. 
4.3 ACCOMMODATING SEASONALITY IN REFERENCE CURVES 
The degree of acceptable exceedance can vary with season. For example, benthos are 
less tolerant of hypoxia in warmer water temperatures. In addition, the threshold 
criterion may never be exceeded in some seasons and frequently be exceeded in 
others. By combining seasons, the acuteness of a specific seasonal exceedence is 
diluted by data from the acceptable season(s). To some extent, seasonal differences 
can be accommodated by changing the threshold criterion among seasons. However, 
there may still be a need to develop separate reference curves by season. 
5.0 REVIEW CFD STATISTICAL PROPERTIES 
INCLUDING BIAS, PRECISION, AND INFERENCE 
The CFD as an assessment tool is a relatively new and unstudied concept. Its close 
relationship to the empirical distribution function does give some insight on the 
mathematical behavior of the CFD. In this section we review some of the properties 
appendix a 
The Cumulative Frequency Diagram Method lor Determining Water Quality Attainment 
