K-6 
the distribution of scores in each category within a segment was compared for each 
habitat to the distribution of scores for the good reference condition. Under the null 
hypothesis (Ho) of no impairment, the two populations (segment and reference) 
were considered to have the same underlying multinomial distributions of samples 
among the ordered categories. The assessment of impairment was based on a one¬ 
sided exact test of Ho against the alternative hypothesis that the segment had a 
distribution shifted towards lower B-IBI scores than for the reference condition. The 
ranking was done separately by habitat, and then combined across habitats. 
Segments with a minimum of 10 samples for which the test was significant at the 1 % 
alpha level and 90% power, were considered impaired under this method. 
3.4. BENTHIC DIAGNOSTIC TOOL 
The benthic diagnostic tool allows environmental managers to identify potential 
sources of anthropogenic stress to benthic communities within Chesapeake Bay. The 
development and application of the tool was described in detail in Dauer et al. (2002, 
2005). The benthic diagnostic tool is based on a linear discriminant function that 
classifies sites in Chesapeake Bay identified as having degraded benthic communi¬ 
ties into categories distinguished by the type of stress experienced by those 
communities. Presently, the function is capable of discriminating contaminated sites 
from sites affected by all other potential sources of stress in any of the seven benthic 
habitat types of Chesapeake Bay. Sites are classified into two groups: 1) a contami¬ 
nant group and 2) the other group representing all other potential sources of stress 
(eutrophication, low dissolved oxygen, etc.). This function is a linear combination of 
variables that includes over 60 measures of diversity, dominance, and function of 
benthic communities. The score for the function is used to calculate the probabilities 
that a sample is drawn from both groups and the sample is assigned to the group to 
which it has the highest probability of belonging. These probabilities are typically 
referred to as posterior probabilities of group membership. 
For this assessment, sites with B-IBI scores < 2.7 were defined as “degraded” for 
benthic diagnostic tool application purposes. A score of 2.7 is used in the Chesa¬ 
peake Bay benthic monitoring programs to define benthic community degradation. 
This cutoff value may differ from the threshold used by the Degraded Area method 
to determine proportion of sites with degraded benthic communities, but it should be 
very close to that threshold. Because cutoff values differ, diagnostic tool percentages 
should only be used as a general guide for identifying potential causes of degrada¬ 
tion. For each “degraded” site, benthic metric values were submitted to the function 
and posterior probabilities of group membership calculated. Posterior probabilities 
for impaired segments were then used to identify the most likely source of stress 
affecting benthic communities in these segments. Sites with posterior probabilities 
of membership in the contaminant group that were greater than 0.50 were classified 
as putatively contaminated. 
3.5. INSUFFICIENT AND EXCESSIVE ABUNDANCE OR BIOMASS 
Insufficient and excessive abundance or biomass was determined from the abun¬ 
dance and biomass metric scores for all sites not classified as putatively 
appendix k 
2006 303(d) Assessment Methods tor Chesapeake Bay Benthos 
