You are here:
JAVA CLASSES FOR NONPROCEDURAL VARIOGRAM MONITORING
Citation:
Faulkner*, B P. JAVA CLASSES FOR NONPROCEDURAL VARIOGRAM MONITORING. JOURNAL OF COMPUTERS AND GEOSCIENCES. Elsevier Science Ltd, New York, NY, 28:387-397, (2002).
Description:
A set of Java classes was written for variogram modeling to support research for US EPA's Regional Vulnerability Assessment Program (ReVA). The modeling objectives of this research program are to use conceptual programming tools for numerical analysis for regional risk assessment. The classes presented use of object-oriented design elements, and their use is described for the benefit of programmers. To help facilitate their use, class diagrams and standard JavaDoc commenting were employed. Java's support for polymorphism and inheritance is used and these are described as ways to promote extension of these classes for other geostatistical applications. Among the advantages is the ease of programming, code reuse, and conceptual, rather that procedural implementation.
A graphical application for variogram modeling that uses the classes is also provided and described. It can also be used by non-programmers. This application uses a generalized least-squares fitting algorithm for robust parametric variogram model fitting through the variogram cloud. This feature makes this program unique from other freely available variogram modeling programs, though the classes are presented primarily so they may be extended for use in other Java programs. More traditional variogram plotting and fitting utilities are also provided. This application is graphical and platform-neutral. It uses classes of the recently proposed Java API for linear algebra, called the JAMA package.