Home

Site Map

About

News

Purchase

Products

Download

Support

Search

George F. Pinder
Professor , Dept. of Civil and Environmental Engineering
Professor , Dept. of Mathematics
University of Vermont

I am very pleased to learn of the availability of the student version of Argus ONE. An exceedingly important aspect of our approach to teaching groundwater flow and transport modeling, we have fully integrated Argus ONE into one undergraduate and two graduate courses.

The ability to build sophisticated models of complex groundwater systems via an integrated, GIS based, preprocessing, processing and post processing framework has allowed me to teach twice as much material in a one-semester course as was previously possible.

Moreover, the ability to modify the software package `in house’ in response to changes in the underlying models is particularly suited to the university environment.

We appreciate the new features you continue to provide in the Argus ONE program.

Professor George F. Pinder
Dept. of Civil and Environmental Engineering
Dept. of Mathematics
Director, Research Center for Groundwater Remediation Design
University of Vermont
Burlington, Vermont
05405
Office 802 656 8697
Fax 802 656 8446
Email: pinder@emba.uvm.edu
Web sites: http://www.emba.uvm.edu/cee/faculty/pinder/pinder.html , http://www.emba.uvm.edu/~pinder/

ARGUS Interware provides ABSOLUTELY THE BEST technical support of ANY software producer we have ever dealt with!

When we began building a salt water intrusion-groundwater flow model for Guam's principal aquifer two years ago, we encountered several esoteric problems for configuring the mesh for the aquifer's unique boundary conditions and hydrogeologic features. We also faced some daunting problems in exporting the complex mesh we ultimately produced into formats our GIS system could efficiently exploit.

Though such tasks are typically very arduous and time-consuming, the ARGUS tech support staff took personal interest in the success of our project-it was like they were (and still are!) part of our team! By their close attention to our problems and consistent 24-hour-or-less responses to our questions, they saved us MONTHS of work! Coupled with the efficiency and time-saving attributes of the software itself, we have accomplished over the last year a volume work that would surely have taken at least five times as long without their software and services-if we could have done it all!

These folks are the best in the business!

John Jenson, Assoc. Prof., hydrogeology
John Jocson, Graduate Research Asst.
Water and Energy Research Institute of the Western Pacific
University of Guam

Stephen W. Wheatcraft:

Professor of Hydrogeology
University of Nevada, Reno
1997


Argus Open Numerical Environments: A Versatile and Powerful Tool for Quantitative Analysis of Groundwater Flow and Transport

The purpose of this report is to provide information about the experiences I have had with Argus Open Numerical Environments (Argus ONE), especially with respect to teaching a graduate course in groundwater modeling. Argus ONE is a versatile and powerful tool which is useful to the beginning modeling student all the way through to the most experienced modelers.

The basic purpose of Argus ONE is to provide a model creation and visualization environment. Argus ONE does everything except for actually providing the solution to the governing equations for the problem. Argus ONE is specifically designed to be independent of the actual numerical CODE that solves the problem. For this reason, a model of a groundwater system can be created in Argus ONE, and the model information can be assembled in whatever format is required for the that has been SELECTED to solve the problem. Model information in Argus ONE is created in layers that are similar to typical paradigms encountered in GIS applications. Information from layers can be communicated to other layers, either directly, or through functional relationships. Argus ONE has a large library of functions that can operate on data in layers. Data can be operated on with a complete set of mathematical functions, including just about anything you could think of in terms of trigonometric, transcendental, statistical, etc. There are also a whole host of available functions that relate to the mesh or grid.

The first and most important thing to learn about the basic paradigm of Argus ONE is to understand the various layer types, and how they work together. Fundamentally, there are two basic groups of layers: (1) mesh or grid layers and (2) all others.
The second group includes domain, density, information, map and data layers. To put it simply, the purpose of the second group of layers is to feed information to the mesh and/or grid layers. The point is that virtually all models tie material properties, boundary conditions, stresses, etc. to nodes, elements, or grid blocks. By creating layers that contain such information, the information can be communicated to the mesh or grid layer, and Argus ONE can be taught to attach information from these other layers to the mesh or grid. Argus ONE provides a powerful and flexible scripting language (a simple programming language, essentially) that allows the analyst to assemble all the model information and write it to a file, or files, that can be directly read by the numerical model that has been chosen to solve the problem.

The scripting language includes most common elements of a programming language: the ability to define and set variables; loops over nodes, elements, variables; write out data in standard fortran format (e.g., E15.7, I5, F10.4), open and close files, etc. Between the power and flexibility of the scripting language and the power and flexibility of the functions available within layers, the analyst can use Argus ONE to create very complicated 2- and 3-dimensional models with boundary and initial conditions, material properties, stresses, and variable mesh and grid densities. In short, there is virtually no limit to the details that one could include in the model.

Possibly the best thing about Argus ONE is that if it is used properly, the analyst can develop a model of a groundwater system with complete disregard to the specific numerical code that is ultimately going to be used. The result of this flexibility is that the analyst can: (a) use more than one code to solve the problem; (b) change codes in mid-stream, if that is desired; (c) use one code to solve the flow problem and a completely different one to solve the transport problem; (d) convert data from one model format to another. Moreover, the model can (and should) be developed without regard to the specific mesh or grid that will ultimately be used. When used properly, the analyst can develop all the important model features completely independent of the mesh or grid. The result of this fact is that the mesh or grid can be refined and the model run with the new mesh or grid with no change whatsoever to the rest of the model input data. Undoubtedly there are other possibilities other than the ones presented here. The only real limitation is the imagination of the analyst.

I will provide an example that came up in the modeling class that I teach. I start out by teaching the class how to use SEFTRAN. SEFTRAN is a finite element flow and transport developed by Geotrans. For a variety of reasons, it makes a good model to start out with in a modeling class. I first start by teaching them how to use SEFTRAN to do flow modeling; then we go on to learn and use MODFLOW. After that, we go back to SEFTRAN and learn to use it for transport modeling. For a final project, I assign a problem that requires both flow and transport modeling. In a one semester course, there is really not enough time to teach (as an example) MT3D, which works directly with MODFLOW, in addition to all the other things that are covered. So I had them solve the flow problem using MODFLOW, and then do the transport portion of the problem using SEFTRAN. Without Argus ONE, this would be very tedious and difficult. But with Argus ONE, they could set up layers and an export script (the term used for the process of creating model input files for specific codes) for flow using MODFLOW. But the same layers and information can be used to send the same information to SEFTRAN for transport. All that is required is a new export script, some additional information specific for the transport problem, and a mesh layer for SEFTRAN, instead of the grid layer used for MODFLOW. So Argus ONE makes it very easy to use two models that were never meant to be used together.

Another example that I just thought of relates to linking groundwater and surface water interaction. Argus ONE doesn't care whether the analyst is dealing with groundwater, surface water, or any other environment that needs to be modeled. So in the same Argus ONE modeling environment, the analyst could set up a groundwater model and a surface water model, and LINK the two. It would be a very difficult task to do this outside of Argus ONE.

Argus ONE is also capable of visualizing model results. Since the mesh or grid topology is already known to Argus ONE, usually all that is required is reading in the node and/or element dependant variables from the model results file(s). This often requires minor changes to the models with respect to how model data are output, but for anyone with even a small amount of programming experience, such changes can be accomplished in a few minutes to no more than a few hours. Visualization capabilities include contour maps, COLOR fill contour maps, 3-D surfaces, velocity vector plots, and streamline/particle tracking capabilities. Map layers also allow the import of DXF files from GIS programs so that cultural features and other identifying features can be imported and used to help understand the visualizations.

The only major drawback to using Argus ONE is that there is a fairly steep learning curve. It requires the investment of quite a bit of time to get to the point where the user is completely comfortable with the paradigm, and the various ways that the layers work together. This problem could be mitigated by the addition of a learning guide. Currently the documentation for Argus ONE consists primarily of a user's guide, which is really just a reference manual.

However, I wish to emphasize that the steep learning curve is really quite justified. There is always a trade-off, a sort of sliding scale, when learning software. It is almost always true that software that is very easy to learn is either not very powerful, or it is essentially special (or even single) purpose. Conversely, software that is relatively hard to learn tends to be powerful and flexible. There is a major difference between Argus ONE and the other pre/post processor software packages. The main difference is that all other pre/post processor software packages are designed to be used with a specific model - usually MODFLOW. So if you want the package to do something it wasn't designed for, or you want to use information that you have developed in its modeling environment with another code, you are pretty much out of luck. Again, with Argus ONE, the only limitation is your own imagination. The best analogy is that Argus ONE is to other pre/post processor software packages as a programming language is to a spreadsheet program.

A recent addition to Argus ONE addresses the steep learning curve issue in an excellent way. Argus ONE now has the capability to work with Plug-In-Extensions, or PIEs. User/developers can now develop PIEs, which are sort of like macros. The concept is flexible enough to be used for almost any purpose within the Argus ONE modeling environment. For instance, one of the first PIEs to be developed is a unit conversion PIE. Perhaps the strongest use for PIEs that has been envisioned to date is to develop PIEs that are designed to set up the appropriate layers and export scripts to work with specific models. Currently, PIEs have been developed for MODFLOW, SUTRA, and PTC (Princeton Transport Code). Others are soon to follow. The advantage here is that you can tell Argus ONE to create a new environment that is specific to a particular model, such as MODFLOW. This eliminates a lot of work for the analyst that would otherwise have to be repeated by anyone using Argus ONE with MODFLOW, or any other specific model.

By using PIEs developed for models such as MODFLOW, the need to develop all the layers and export scripts specifically tailored to the model is eliminated. Therefore an analyst who knows very little about Argus ONE itself can immediately begin using the modeling environment with the MODFLOW PIE. As a result, Argus ONE is accessible to a much wider range of users. One final advantage to Argus ONE is that it is available on most computing platforms, including PC Windows 3.1/95/NT, MacIntosh, and several unix workstations, including Silicon Graphics, Sun, Dec and IBM. Often, it is very useful to develop the model on a PC, then take it over to Argus ONE running on (for instance) a Silicon Graphics workstation, where finer meshes can be used without worrying about the extra time that would be needed on a PC to draw the graphics and/or make the calculations. (Please note that this is no longer true. Bill is victorious and only PC Windows remains., this remark is by Argus and not by the author who at the time of writing this review had the luck to have Argus ONE on his PCs and SGIs)

Allen Shapiro:

Hydrologist
US Geological Survey (USGS)


An email by Dr. Shapiro to users of a commercial Modflow package (GMS) complaining about the lack of conceptual modeling and GIS capabilities in that package.

Date: Fri, 14 Feb 1997 08:53:17 -0500
To: Cleared to protect recipients' privacy
From: "Allen M. Shapiro" <ashapiro@usgs.gov>
Subject: Comment on "Linking GMS and ARC/Info attributes" and "Conceptual" Modeling
Cc:
Bcc:
X-Attachments:

I have been a passive follower of information on the GMS mailing list through colleagues in the USGS. The recent discussion on accessing ARC/Info attributes in ground-water modeling projects and the need for what has been referred to as "conceptual" modeling has piqued my interest because I have faced similar concerns.

My concerns, however, extends beyond the use of a single ground-water flow or transport model; not only do I use MODFLOW as part of my research activities, but many other USGS models (e.g., SUTRA, MOC, MOC3D, etc.), as well as several unpublished research codes that I developed for investigations of flow and transport in fractured rock. All of these models required access to USGS data bases (ARC/Info). In addition, I wanted the ease of synthesizing geospatial information within a graphical interface, where the geospatial information was not tied to my specific choice of discretization (referred to in your messages as "conceptual" modeling).

My brief review of the available graphical interfaces in the market for SELECTED ground-water flow and transport models left me somewhat disappointed in their ability to address these issues. For this reason, I have been using Argus Open Numerical Environments (Argus ONE) as pre- and post-processing for MODFLOW, SUTRA, as well as my own unpublished research codes. Since its inception, Argus ONE has had "conceptual" modeling capability and recently they developed a DLL (dynamically-linked library) that imports shapefiles--allowing access to ARC/Info attributes!

The ability of Argus ONE to access DLL's also has made it possible for users (such as myself) to develop external codes that run with Argus ONE--these codes could be polished graphical interfaces for any model (with tab dialog boxes, pop-up menus, etc.), interpolation and geostatistical codes, and translators of geosptial data for import into the GIS and modeling environment in Argus ONE.

Recently, I completed a graphical interface for MODFLOW96 (the latest revision of the USGS model MODFLOW is described in Harbaugh and McDonald, 1996, USGS Open-file Report 96-485, and Harbaugh and McDonald, 1996, USGS Open-file Report 96-486). A documentation for the DLL for MODFLOW96 is being completed and I anticipate its release as a USGS Open-file report in the next several weeks. Also, a DLL for SUTRA was developed recently by Cliff Voss (USGS), and I have developed several geostatistical codes that also interface with Argus ONE.

If you are not aware of the product, you can look at their web site (http://www.argusone.com) or feel free to contact me regarding my experiences.

Allen M. Shapiro

U.S. Geological Survey
12201 Sunrise Valley Drive
431 National Center
Reston, VA 20192 USA

Tel: 703-648-5884
FAX: 703-648-5274

e-mail: ashapiro@usgs.gov

Victor Podsetchine:

Senior Researcher
Regional Environmental Agency of Häme
Tampere, FINLAND


The Hydrodynamic Group of the Research Unit of the Regional Environmental Agency of Häme specializes in studies of water flow and water exchange in lakes, rivers and coastal areas of seas. We also study processes of dissolved substances and suspended sediment transport and lake eutrophication.

Finite element 2D and 3D flow models are used to calculate flow patterns under different wind loadings. Simulation results are verified with ADCP, ENDECO and AANDERA current measurements.

The first step in applying of a numerical flow model is to produce a bathymetric map. A digitizing software FINGIS (Producer: Karttakeskus Fingis Team, POB 85, 00521 Helsinki, Finland) is our main tool for converting topographic maps into a digital form. The purchase of Argus Open Numerical Environments (Argus ONE) in summer 1995 filled the gap between the digitizing program and the numerical models, and simplified further developing steps. Before that we used our own console-based tools under VAX and PC environments, which lacked user-friendly graphical interface. With Argus ONE grid generation and modification became a painless and routine procedure.

Among features which we like is a navigation window, the possibility to exchange information between different types of layers and post-processing tools. Manual editing option enables users to easily modify and correct the generated grids. Argus ONE provides a convenient environment for storing all of the information required for a numerical model in a "single package". Export templates give the user a simple and efficient way to exchange information between Argus ONE and other programs.

Moving from one computer platform to another is a laborious job, which can cause a lot of troubles and headaches. We had to move from VAX and PC Windows 3.11 to WindowsNT-Alpha at the end of the last year. There were no problems with Argus ONE in this respect. With FX!32 from Digital Argus ONE runs on Alpha just like on any Intel machine.

True appreciation comes with time only. The more we use Argus ONE the more we like it.

Victor Podsetchine

Senior Researcher
Regional Environmental Agency of Häme
Tampere, FINLAND
Tel.: 358-3-2420111
Fax: 358-3-2420266
E-mail: victor.podsetchine.vyh.fi
Internet: http://wwwtest.vyh.fi/ham/wwwengl.htm
16.06.97

Douglas R. MacAyeal:

Professor
Geophysical Sciences
University of Chicago

MeshMaker does in order of 1 hour what it used to take weeks to accomplish in the "old days"...

The quality-control and data reliability engendered through the use of MeshMaker are also very critical benefits to using MeshMaker.

MeshMaker is the "saving grace" of my research. In the past, I was "mesh bound" due to the tremendous personal time cost needed to make modifications to my models. Now I'm "free" to dream, change my mind, and correct mistakes, etc. because finite element meshes are so easy to create.

Again, thanks for your fantastic service! Now I have nothing to do but work on my finite-element modelling project (to some this would be a pain, but to me it's fun).

Wow! you people have wicked-good software...

I love your software. Your software has sped up my research (and made finite-element modelling more pleasant) by an order of magnitude....

Michel J.L. Robin:

Professor
Department of Geology
University of Ottawa

"Two thumbs" is my verdict for Argus MeshMaker Pro. If you use two-dimensional finite differences or finite element numerical models, then Argus MeshMaker Pro would no doubt allow you to concentrate on modeling instead of wasting valuable time building input files.

If you are in a hurry, you can find out more about the program from the demo. It is extremely well done and gives a fairly comprehensive tour of the program.

With Argus MeshMaker Pro we can accomplish in minutes what used to take days. For us Argus MeshMaker is reminiscent of the leap from the pencil and paper to the word processor, or the leap from the slide rule to the programmable calculator.

Argus MeshMaker is a CAD-like program that is in many ways similar to Geographic Information Systems (GIS).

Argus MeshMaker's graphical user interface stands out as being both practical and intuitive. All components of the main window are conveniently visible at all times.

Information messages and warnings are generally helpful and self-explanatory.

Perhaps the most important feature of a pre-processor is its ability to interact with other commercial or in-house software. Argus MeshMaker survived our critical (and at times skeptical) evaluation of the data exchange capabilities with flying colors.

Argus MeshMaker allows us to have full control over the way data is exported by using Export templates. We found these templates extremely useful when exporting FEMs and FDGs for use with in-house numerical software.

Argus MeshMaker Pro is a well-structured, visually-oriented numerical pre-processor that is both simple to use and versatile.