A-6 
RECOMMENDATIONS 
The panel identified critical research tasks that need resolution in the near future. 
The following is a list of critical aspects of that needed research. These research 
tasks appear roughly in order of priority. However, it must be recognized that it is 
difficult to formulate as set of tasks that can proceed with complete independence. 
For example, research on task 1 may show that the ability to conditionally simulate 
the water quality surface is critical to resolving the sample size bias issue. This 
discovery might eliminate IDW as a choice of interpolation under task 3. The Panel 
has made significant progress on several of these research tasks and CBP is encour¬ 
aged to implement continued study in a way that maintains the momentum 
established by our panel. 
TASK 
1. Effects of Sampling Design on CFD Results 
(a) Continue simulation work to evaluate CFD bias reduction via conditional 
simulation. 
(b) Investigate conditional simulation for interpolation methods other than 
kriging—this may lead to more simulation work. 
(c) Implement and apply interpolation with condition simulation on CBP data. 
2. Statistical inference framework for the CFD 
(a) Conduct confidence interval coverage experiments. 
(b) Investigate confidence interval methods for non-kriging interpolation 
methods. 
(c) Implement and evaluate confidence interval procedures. 
3. Choice of Interpolation Method 
(a) Implement a file system and software utilizing kriging interpolation for CBP 
data. 
(b) Compare interpolations and CFDs based on kriging and inverse distance 
weighting (IDW). 
(c) Investigate nonparametric interpolation methods such as LOESS and spline 
approaches. 
4. Three-Dimensional Interpolation 
(a) Implement 2-D kriging in layers to compare to current approach of 2-D IDW 
in layers. 
(b) Conduct studies of 3-D anisotrophy in CBP data. 
(c) Investigate software for full 3-D interpolation. 
5. High Density Temporal Data 
(a) Develop methods to use these data to improve temporal aspect of CFD 
implementation. 
(b) Investigate feasibility of 4-Dimensional interpolation. 
appendix a 
The Cumulative Frequency Diagram Method for Determining Water Quality Attainment 
