Spatial Analysis and Decision Assistance

SADA Version 5.0 is Almost Ready!

The next version of SADA will provide many additional features in the areas of visualization, sample design, and geospatial analysis. In the areas of visualization we will be improving the overall quality of the output in an effort to move SADA slightly in the direction of GIS. Geospatial additions include methods to characterize and plan sampling strategies for data sets that will measure both the item of interest (the contaminant of concern) and a second attribute that is correlated or in some way indicates the value of the primary attribute. For example, radionuclide measurements might be the attribute of interest and gamma counts the secondary cheaper form of information. New sampling strategies that consider both of these forms of information will be provided. The following is a brief description of a few of the improvements we intend to make.

Geostatistical Simulation

Two kinds of simulation routines will be made available: Gaussian, Indicator. Each generate equiprobable maps of contamination in two and three dimensions and permit the examination of joint probabilities. With these tools users can evaluate uncertainty in several ways. By post processing simulations one can produce distribution of volume of removal or estimate the probability that a contaminated zone of a given size exists on the site. Furthermore these simulations can be exported from SADA into external models such as fate and transport that can assess the uncertainty say that a contaminant plume reaches a certain boundary in a given time. Here one may simulate geological conditions rather than contamination.

Estimation Using Secondary Information

Markov-Bayes and various co-kriging tools join the ranks of spatial interpolators available in SADA. Each permit secondary forms of information that may be more uncertain but cheaper to acquire to be used explicitly in the spatial estimation process. In addition, a number of helping tools including correlation assessments and bivariate scatterplots assist with determining the appropriateness of secondary information.

Sample Design

We will also present a suite of new sample designs that are based on the types of joint probability assessments in simulation as well as the joint conditioning of primary and secondary data. These approaches will recognize the role of primary and secondary forms of data and seek to optimize the location and mixture of these two forms of data.

Visualization

Three areas of visualization are planned. First photographic overlays will be available. Standard formats are accepted and can come from anywhere. Or it can be as easy as just grabbing a Google Earth screen shots right off their computers of satellite or aerial images and use SADA to roughly lay them over the site. Secondly, we will provide "engineering" type graphical output with formal legends, map scales, annotations etc. Finally, we will include elevation as well.

Borehole Visualization

Version 5 now allows user to peruse through borehole results in a vertical profile view.

Elevation

Version 5 now allows user create or import elevation information. Elevation is applied to produce more realistic images where topography is variable.

Area of Concern

The area of concern and uncertainty about the location and volume of contaminated media has always been a central theme in SADA. However, now have greatly simplified how this uncertainty can be visually understood and extended it in a very practical way to deterministic interpolants as well. For geostatistical methods, normal approaches to uncertainty are used. For deterministic methods such as natural neighbor or inverse distance, the user can specify concentration intervals that express the uncertainty in the final decision. The later can also be used for geostatistical methods as well. In the following graphic, an area of concern is shown using the new approach. The dark black line shows the interior of the area of concern. In other words, the interior of this black boundary is where your contaminated media is found according to your decision criteria and selected decision scale. However, we see two other colors now. The gray part of the area of concern represents that area that definitely should be addressed. The green area represents a zone of uncertainty. By this we mean that the decision criteria is close to being violated and we cannot with great confidence (specified by user) that it is contaminated or not. The rest of ths site is outside the area of concern.

Manually Created Area of Concern

If data are insufficient to create a viable interpolation or area of concern map, users can manually create an area of concern map using simply polygon drawing tools. SADA can then calculate volumes, mass, etc from this expert judgment approach.

Cost/Benefit Analysis

The uncertainty described in the Area of Concern above translates as well to the cost benefit analysis. Here you see the relation ship between volumes of media that are 1) clearly contaminated, 2) clearly uncontaminated, and 3) contamination is uncertain. Overburden may also be calculated and presented as part of the decision analysis.

Improved MARSSIM

The MARSSIM analysis in SADA is much improved. Now users can do quick and easy MARSSIM analyses, sample designs, and quick MARSSIM checks without having to first set up formal MARSSIM analyses within SADA. This makes it much easier particularly for regulators to do quick checks.

Sum of Fractions

In line with the improvements in MARSSIM mentioned above, SADA now has the ability to do sum of fractions for MARSSIM type applications but also has been generalized for any type of application as well.

Map Manipulations

We are now providing a host of straight forward map tools that you can use on any existing geospatial contour. These range from simple algebraic functions to comparing, adding, subtracting 2 maps, to using maps in 2/3d cutting and combinations

Storing Models

You can now store models dynamically in SADA. Unlike the older results gallery, you can now recall models and continue to work with them. Furthermore you can now import models created elsewhere into SADA and continue working with them as well.

Importing Models

You can now import models created elsewhere into SADA and continue working with them as well. SADA has a very low format requirement for bringing models created outside of SADA into your SADA file. So if SADA does not have the spatial modeling tools you need, you can now create them elsewhere, import them into SADA and continue with your decision analysis or secondary sampling designs.

Better GIS Integration

SADA will have the ability to import and export raster maps, shape files, and dxf files. We can now deal with digital elevation models as well. Now users should experience a much easier transition between SADA and ESRI.

Real Time Autodocumentation

The software now presents a real time log file (in html format) showing all the types of important modeling parameters and results presented in the normal autodocumentation feature. Found under the Show The Results step, users can now scroll through the process that led to the result they now see in the graphics window.

FASTER, FASTER, FASTER

In order to accomodate the potentially enormous size of secondary data sets (think geophysics for a moment) we had to make many of SADA's import/export processes much, much faster. SADA was originally designed with smaller sites in mind. V5 is a substantial step forward in dealing with large data sets.

There are many, many other improvements planned. If you have any recommendations, please don't hesitate to suggest them at sada@tiem.utk.edu.

We are nearing the end of development. We presented a preliminary version to our federal sponsors at the end of January and plan to produce a strong beta release for everyone in April along with some detailed documentation. Stay tuned.