22 
automated to complete large number of criteria assessments efficiently and routinely. 
These capabilities are not currently available using a kriging algorithm, but the 
Chesapeake Bay Program IDW interpolator is designed with these capabilities in 
mind. Thus, the EPA recommends that large-scale interpolations (segment, baywide) 
continue to be based on the fixed-stations data be performed using the Chesapeake 
Bay Program IDW interpolator. As kriging is developed further for use, this option 
may be recommended in the future. 
For the criteria assessment of small tidal tributaries and the Bay’s shallow-water 
habitats based on data from the shallow-water monitoring program, the EPA recom¬ 
mends implementation of a kriging algorithm, where possible. The shallow-water 
monitoring program yields data to assess criteria attainment in relatively few 
systems at any one time. Thus, it is possible to provide the more focused evaluations 
of individual interpolations that kriging requires. Furthermore, the intensive data 
collection provided by the shallow-water monitoring program is particularly 
conducive to detailed statistical analysis. To utilize the data’s information fully, a 
more thorough statistical interpolation procedure, such as kriging, should be imple¬ 
mented. The shallow-water systems are highly dynamic and thus better characterized 
by more intensive data collection combined with a more rigorous statistical interpo¬ 
lation algorithm. For these reasons, the EPA recommends that kriging be 
implemented, where possible, for criteria assessment based on shallow-water moni¬ 
toring data. 
Given the recommendation above, the EPA further advises that the states develop the 
expertise to perform spatial interpolation based on statistical methods. Assessment 
of the shallow waters will largely fall to the states, with some support from the 
Chesapeake Bay Program Office. Guidelines are being developed for the use of 
kriging in shallow-water criteria assessment. The procedure is detailed, however, and 
requires expertise in geographic information systems, spatial statistics, and computer 
programming. Questions remain about how best to implement kriging as an option 
for spatial interpolation. The EPA plans to provide support through the Chesapeake 
Bay Program Office to ensure that spatial interpolations based on kriging are 
performed consistently for all shallow waters of the Bay when practical. 
In general, most of the tidal waters of the Chesapeake Bay mainstem and major trib¬ 
utaries remain impaired. This judgment was confirmed by the assessments 
performed during the 2006 303(d) listing cycle and by listing decisions made prior 
to that time. The 2006 assessments indicated that many of the assessment units were 
far out of attainment with little need to confirm the conclusions through statistical 
analysis. As restoration efforts proceed and more Bay tidal waters approach attain¬ 
ment of their designated uses, then statistical procedures may become important to 
ensure that waters are properly removed from the 303(d) list as soon as possible. 
Given that it may require several years for the Bay to respond to management 
actions, there is ample time to conduct the studies necessary to develop the required 
statistical decision-making framework based on the CFD. The EPA recommends that 
assessment of criteria attainment continue as in 2006 when the decision rule was that 
chapter ii 
Refinements to the Chesapeake Bay Water Quality Criteria Assessment Methodology 
