C-6 
Up to four points are used for interpolation. If fewer than four points exist, interpo¬ 
lation is still carried out given at least one measured point. Without any measured 
data, a missing value (normally a -9) is calculated for that cell. A search radius filter 
limits the horizontal distance of monitoring data from the cell being computed. Data 
points outside the user-selected radius (normally 25,000 m or 25 km) are excluded 
from calculation. This filter ensures that only data near the location being interpo¬ 
lated are used. 
Segment and region filters have also been added. Segments are aggregations of the 
interpolator cells. For instance, eight segments make up the mainstem Chesapeake 
Bay (CB1TF, CB20H,...CB8PH). The tidal tributaries have 70 additional segments, 
created by the Chesapeake Bay Program's 2003 segmentation scheme (U.S. EPA 
2004, 2005). These segments divide the Bay into geographic areas with somewhat 
homogeneous environmental conditions. This segmentation also allows the reporting 
of results on a segment basis, revealing more localized changes compared to the 
whole Bay ecosystem. 
The region file identifies the geographic boundary that limits which monitoring 
station data are included in interpolation for a given segment (see Appendix D). The 
purpose of the data region is to select a subset of the monitoring data from the input 
data file and to use that subset for computing the values for each grid cell in a 
segment. Use of data regions ensures that the interpolator does not “reach across 
land” to obtain data from an adjacent tidal tributary—a process that would give erro¬ 
neous results. By using data regions, each segment of grid cells can be computed 
from its individual monitoring data subset. Each adjacent data region overlaps so that 
a continuous gradient—not a seam—exists across segment boundaries. Data regions 
for criteria assessment vary somewhat from the data regions in the standard interpo¬ 
lator. These new regions were developed to exclude tributary measurements from 
mainstem interpolations and to include additional observed data from Virginia. 
EVALUATION OF THE INVERSE DISTANCE 
WEIGHTING SPATIAL INTERPOLATION 
ALGORITHM FOR ASSESSING CHESAPEAKE BAY 
WATER QUALITY CRITERIA 
The current Chesapeake Bay interpolator is based on an IDW algorithm—a non- 
statistical spatial interpolator that uses observed data to calculate a weighted average 
(as a predicted value) for each location on the prediction grid (Appendix D). The 
method calculates the weight associated with a given observation as the inverse of 
the square of the distance between the prediction location and the observation. The 
IDW is a spatial interpolator; in general, such methods have provided good predic¬ 
tion maps (STAC 2006). Additionally, implementation is relatively simple since 
software exists to map IDW automatically. Further, the method does not require any 
decisions during an interpolation session. Commercial Geographic Information 
Systems (GIS) software contains IDW, requiring only GIS skills for application. 
The IDW algorithm has several advantages for use in Chesapeake Bay water quality 
criteria attainment assessment (STAC 2006). First, since it is non-statistical, the 
appendix c 
Evaluation of Options for Spatial Interpolation 
