f/EPA INTRODUCTION Environmental Monitoring Systems Laboratory P.O. Box 93478 Las Vegas NV 89193-3478 June 1991 OFFICE OF RESEARCH AND DEVELOPMENT TECHNO Geo-EAS: Software for Geostatistics United States Environmental Protection Agency The Environmental Monitoring Systems Laboratory-Las Vegas (EMSL-LV) can meet the needs of scientists who work with spatially distributed data. The complexity of contaminant distribution and migration at hazardous waste sites requires a mathematical method that is capable of interpreting raw data and converting them to useful information. Geostatistics began in the mining industry and has grown to include applications ranging from microbiology to air monitoring. Though the application of geostatistics is crucial to the delineation of buried contami- nants, not every field scientist can be expected to develop customized geostatistical SUPPORT PROJECT algorithms for individual sites. Geostaticians at the EMSL- LV developed a software package, Geo-EAS in 1988. The current version, Geo-EAS 1.2.1, was released in 1990. This program offers the environmental scientist an interactive tool for performing two-dimensional geostatistical analyses of spatially distrib- uted data. THE METHODOLOGY EQUIPMENT REQUIREMENTS Geostatistical methods are useful for site assessment and monitoring where data are collected on a spatial network of sampling loca- tions. Examples of environ- mental applications include lead and cadmium concentrations in soils surrounding smelters, and sulfate deposition in rainfall. Kriging is a weighted moving average method used to interpolate values from a data set onto a contouring grid. The kriging weights are computed from a variogram, which measures the correlation among sample values as a function of the distance and direction between samples. advantages over other inter- polation methods: Smoothing Kriging regresses estimates based on the proportion of total sample variance ac- counted for by random noise. The noisier the data set, the less representative the samples and the more they are smoothed. Declusterlng The kriging weight assigned to a sample is lowered to the degree that its information is duplicated by highly corre- lated samples. This helps mitigate the impact of oversampling hot spots. Anlsotropy When samples are highly correlated in one direction, kriging weights will be greater for samples in that direction. Precision Given a variogram represen- tative of the area to be esti- mated, kriging will compute the most precise estimates from the data. Estimation of the variogram from sample data is a critical part of a geostatistical study. Geo-EAS is designed to make it easy for the novice to use geostatistical methods and to learn by doing. It also provides sufficient power and flexibility for the experienced user to solve practical problems. Geo-EAS was designed to run under DOS on an IBM, PC, XT, AT, PS2, or compat-ible computer. Graphics sup-port is provided for Hercules, CGA, and EGA. At least 512 Kb of RAM is required, but 640 Kb is recommended. An arithmetic co-processor chip is strongly recommended due to the computationally inten-sive nature of the programs, but is nnt rpquirprl Programs may be run from floppy disk but a fixed disk is required to use the programs from the system menu. The system storage requirement is ap- proximately three megabytes. For hardcopy, a graphic printer is required. Support is provided for most plotters. Design features such as simple ASCII file formats and standardized menu screens, give Geo-EAS flexibility for future expansion. It is antici- pated that Geo-EAS will be- come a significant technology transfer mechanism for more advanced methods resulting from the EMSL-LV research and development programs Geo-EAS software and docu- mentation are public domain, and may be copied and dis- tributed freely. 0683EX91 ------- MAPS AND MENUS The Geo-EAS programs use an ASCII file structure for input. The files contain a header record, the number of variables, a list of variable names and units, and a nu- meric data table. All Geo-EAS programs are controlled interactively through menu screens which permit the user to select op- tions and enter control pa- rameters. The programs are structured to avoid a "black box" approach to data analy- sis. Several of the more complex programs permit the user to save and read param- eter files, making it easy to rerun a program. The programs DATAPREP and TRANS provide capabil- ity for manipulating Geo-EAS files. Files can be appended or merged, and variables can be created, transformed, or deleted. Transformation operations include natural log, square root, rank order, indicator, and arithmetic operations. POSTPLOT creates a map of a data variable in a Geo-EAS data file. Symbols represent- ing the quartiles of the data values or the values them- selves are plotted at the sample locations. STAT1 computes univariate statistics, such as mean and standard deviation, for vari- ables in a Geo-EAS data file, and creates histograms and probability plots. SCATTER and XYGRAPH both create x-y plots with optional linear regression for any two variables in a Geo- EAS file. SCATTER is useful for quick exploratory data analysis, while XYGRAPH provides additional capabili- ties such as multiple "y" vari- ables, and scaling options. PREVAR creates an interme- diate binary file of data pairs for use in VARIO, which com- putes and displays plots of variograms for specified dis- tance and directional limits. Variogram models can be interactively fitted to the ex- perimental points. The fitted model may be the sum of up to five independent compo- nents, which can be any com- bination of nugget, linear, spherical, exponential, or Gaussian models. XVALID is a cross-validation program which can test a variogram model by estimating values at sampled locations from sur- rounding data and comparing the estimates with known values. KRIGE provides kriged esti- mates for a two-dimensional grid of points. A shaded map of estimated values is dis- played and a Geo-EAS file of kriged grid results is gener- ated. CONREC generates contour maps from a gridded Geo- EAS data file, usually the output from KRIGE. Options are provided for contour inter- vals and labels and degree of contour line smoothing. REFERENCE: Isaaks, E. H. and R. M. Srivastava, An Introduction to Applied Geostatistics, Oxford University Press, New York, 1989. ^G^'O/V ^OiOGV AVAILABILITY: For further information about Geo-EAS, contact Dr. Evan Englund. Government agencies and academic or research institutions can obtain a copy of Geo-EAS with User's Guide at no charge by sending three Pre- formatted high-density diskettes (5-1/4" or 3-1/2") to: Dr. Evan J. Englund U.S. Environmental Protection Agency Environmental Monitoring Systems Laboratory-Las Vegas P.O. Box 93478 Las Vegas, NV 89193-3478 FAX: (702) 798-2248, FTS: 545-2248 Others can obtain a copy for a distribution charge of approximately $45 (includes diskettes, User's Guide, and USA shipping) from either: ACOGS or COGS P.O. Box 44247 P.O. Box 1317 Tucson, AZ 85733-4247 Denver, CO 80201 -1317 FAX: (602)327-7752 Phone: (303)751-8553 The Technology Support Center fact sheet series is developed and written by Clare L. Gerlach, Lockheed Engineering & Sciences Company, Las Vegas. FOR FURTHER INFORMATION: For information about the Technology Support Center at EMSL-LV. contact: Mr. Ken Brown, Manager Technology Support Center U.S. Environmental Protection Agency Environmental Monitoring Systems Laboratory-Las Vegas P.O. Box 93478 Las Vegas, NV 89193-3478 (702) 798-2270, FTS 545-2270 ------- |