E-1 
appendix 
Potential Methods for Assessing 
Shorter Duration Dissolved 
Oxygen Criteria 
POTENTIAL METHODS 
The 2003 Chesapeake Bay water-quality criteria document described three alternatives 
for assessing attainment of the short duration dissolved oxygen criteria (U.S. EPA 
2003). Those include: 1) logistic regression; 2) a time series statistical method and 3) 
continuous dissolved oxygen data collection using meters that are deployed for an 
extended period of time. Each of these approaches has strengths and drawbacks. 
Appropriate implementation of logistic regression or time series statistical methods 
may require continuous dissolved oxygen data. To develop the full capacity to assess 
the shorter duration dissolved oxygen criteria—7-day mean, 1-day mean and instanta¬ 
neous minimum, EPA recommends a phased approach in which the methods that are 
easiest to implement are employed initially while continuing to work on development 
and implementation of the more detailed and/or expensive methods. 
LOGISTIC REGRESSION 
The instantaneous minimum criteria imply the requirement that waters within the 
respective designated use be at or above the defined concentration everywhere all the 
time. Stated in this way, the logistic regression approach clearly has application to 
the challenge of assessing attainment of instantaneous minimum criteria. In the 
context of criteria attainment, logistic equations are developed from the long term 
dissolved oxygen data record, which predict the probability that the defined criteria 
concentrations were met, based on observed monthly mean concentration. 
The logistic regression approach utilizes a well-established statistical procedure 
(U.S. EPA 2004) and has been employed in the past in Chesapeake Bay to estimate 
instantaneous minima (Jordan et al. 1992). It is relatively simple to use and only 
requires regular updating to keep the predictive models relevant to current condi¬ 
tions. The limitation of this approach is that it is based on an extrapolation of the 
fixed-station data and is likely to have higher error than the other methods. 
appendix e 
Potential Methods for Assessing Shorter Duration Dissolved Oxygen Criteria 
