C-1 
appendix C 
Evaluation of Options for 
Spatial Interpolation 
Interpolation constitutes a critical element of CFD-based assessment methodology. 
It provides the spatial framework for data integration while allotting the appropriate 
weight to all data. The spatial framework consists of a grid made up of a network of 
cells that vary in size to cover the entire spatial domain. The size of the cells deter¬ 
mines the scale of the assessment; smaller and more numerous cells in a given area 
provide a more spatially detailed assessment. Estimates for all cells come from a 
spatial interpolation algorithm. 
To date, two spatial interpolation algorithms have been considered; inverse distance 
weighting (IDW) and kriging. In IDW, estimates of water quality levels are based on 
a weighted average derived from the closest measured data values. Weights depend 
upon the distance between the measurement point and the cell being estimated. Thus, 
measurements from the closest points are weighted most heavily and have the most 
influence. The second method is kriging—a well-known statistical form of spatial 
interpolation. The statistical details of kriging rest on ample research. This method, 
however, has not been used for water quality criteria. Both spatial algorithm methods 
can prove valuable for Chesapeake Bay water quality criteria assessment; one or 
both will likely be used in the future. Other methods (non-parametric regression 
methods such as Loess regression or cubic splines) are also available and could also 
be considered for future use. Further details on the IDW and kriging methods are 
provided below. 
SPATIAL INTERPOLATION NEEDS SPECIFIC TO CHESAPEAKE 
BAY WATER QUALITY CRITERIA ASSESSMENT 
The Chesapeake Bay water quality criteria were established using the spatial defini¬ 
tion of designated-use areas for the tidal waters of Chesapeake Bay (U.S. EPA 
2003a, 2003b). These spatial definitions, along with the characteristics of the Bay 
itself, present several challenges for spatial interpolation. For example, the Chesa¬ 
peake Bay shoreline is extremely complex with many small tidal tributaries, 
embayments, and inlets that occur at various scales throughout the water body. The 
small inlets present a challenge for spatial interpolation because they require 
appendix c 
Evaluation of Options for Spatial Interpolation 
