for the months of June through September for a three-year period for all Chesapeake 
Bay Program segments with deep-water designated use habitats would be extracted 
and evaluated individually using the cumulative frequency distribution approach. 
The cumulative frequency distribution attainment curve would be calculated (and 
plotted, if desired) and compared to the appropriate reference curve for the desig¬ 
nated use and season using the statistical test described earlier. If the two curves are 
significantly different, then the segment/designated use is considered out of attain¬ 
ment, and failing by the amount defined by the area between the two curves. 
Seven-Day Mean Attainment Procedure. The 7-day time frame is much shorter 
than the temporal scale of the current baywide water quality monitoring programs, 
and statistical forecasting models are necessary to assess criteria of this duration. 
The proposed approach, referred to as the spectral analysis approach in this chapter 
and discussed in more detail below, uses long-term, low-frequency data from the 
monitoring program and shorter-term, high-frequency data from in situ semi-contin¬ 
uous monitors to synthesize a data set that incorporates both long- and short-term 
patterns of variability. The synthetic data set is created at user-specified time inter¬ 
vals, e.g., weekly, daily and hourly. The minimum interval will depend on the 
interval length of the continuous data. The synthetic data set is then analyzed at the 
appropriate temporal scale, which in this case is seven days. At present there are 
insufficient high-frequency data and insufficient validation of the approach to 
recommend its implementation. For now, attainment of 7-day mean criteria should 
not be assessed unless data are available for a specific location/segment at a temporal 
scale consistent with the 7-day duration. 
One-day Mean Attainment Procedure. The 1-day attainment procedure is the 
same as the 7-day mean procedure described above. For now, attainment of the 
1-day mean criteria should not be assessed unless data are available for a specific 
location/segment at a temporal scale consistent with the 1-day duration. 
Instantaneous Minimum Attainment Procedure. Again, the instantaneous 
minimum time frame is much shorter than is currently sampled. The spectral 
analysis approach presented above is one way to estimate attainment of these 
dissolved oxygen criteria. Another approach, referred to as the logistic regression 
approach in this chapter and described in more detail below, applies by restating the 
criterion in slightly different temporal terms. An instantaneous minimum implies 
that the criterion is not met if dissolved oxygen concentrations are below the crite¬ 
rion value at any time. The logistic regression approach estimates the relative 
frequency or percent of time that a region falls below a specified concentration based 
on the empirical relationship between seasonal or monthly mean values and the 
percent of dissolved oxygen concentrations above or below the specified level as 
observed in the historical data record (of the Chesapeake Bay water quality moni¬ 
toring program). This method has been applied experimentally with reasonable 
results (Jordan et al. 1992) and can approximate criteria exceedance/attainment 
frequency. However, at this time the method has not been adequately validated to 
recommend implementation for formally assessing criteria attainment. Attainment of 
chapter vi • Recommended Implementation Procedures 
