Guideline 12: Discriminatory Ability 



The ability of the indicator to discriminate differences among sites along a known condition gradient 

 should be critically examined. This analysis should incorporate all error components relevant to the 

 program objectives, and separate extraneous variability to reveal the true environmental signal in the 

 indicator data. 



Since we are evaluating the use of dissolved oxygen concentration as an indicator of hypoxia, which is 

 defined as a DO concentration below a certain value, there is no spatial variability associated with this 

 indicator. Simply stated, using a criterion of 5 mg/L to define hypoxia, a DO concentration of 4 mg/L will 

 indicate hypoxic conditions regardless of where the sample is collected. Note that this does NOT mean that 

 adverse biological effects will always occur if the DO concentration falls below 5 mg/L. Nor does it mean 

 that a given level of nutrient enrichment will result in the same degree of hypoxia in all areas. Both of these 

 components are spatially variable and are affected by a number of environmental factors. However, they do 

 not affect the relationship between dissolved oxygen concentration (the Indicator) and hypoxia as defined 

 (the asssessment issue). 



Because of the large number of variables known to affect the dissolved oxygen concentration in sea water, 

 most of which are not routinely measured, the utility of variability component analyses is limited. However, 

 this indicator is really a direct measurement ofthe focus of the assessment question; therefore, discriminatory 

 ability is inherently high. 



Since the program's objective is to estimate the percent of estuarine area with hypoxic/anoxic condition on 

 a broad geographic scale rather than to compare individual sites, an alternative way to look at this indicator's 

 discriminatory ability is to plot out the CDF along with its confidence intervals. Figure 2-10 illustrates such 

 a plot for the EMAP Virginian Province data collected from 1990 to 1993. (See Strobel et al. [1995] for a 

 discussion of how the confidence intervals were developed). The tight 95% confidence intervals suggest 

 that this indicator, as applied, has a high degree of discriminatory ability - a relatively small shift in the curve 







2 4 6 8 



Bottom Dissolved Oxygen (mg/L) 



10 



Figure 2-10. Cumulative distribution function of bottom dissolved oxygen concentration 

 for the EMAP Virginian Province, 1990-1993. Error bars represent 95% 

 confidence intervals. 



2-14 



