Due to the current gap in continued funding from the U.S. National Science Foundation (NSF), the NSF Unidata Program Center has temporarily paused most operations. See NSF Unidata Pause in Most Operations for details.

UNIDATA

A Paradigm for Inclusion of Current Geographical Data in Meteorological Education

Mitchell S. Baltuch
Robert N. Green

Unidata Program Center
Boulder, Colorado

1.0 Introduction

The Unidata Program1 was established in 1982 to respond to the need of universities and the National Science Foundation (NSF) for better access to weather data. The program is administered by the Unidata Program Center (UPC) within the University Corporation for Atmospheric Research (UCAR). The UPC assists the university community in receiving data, and supplies a suite of software packages to ingest, store, view and manipulate it (Fulker, 1989). Unidata has been providing access to current weather since 1985 and is approaching 100 universities today.

In an attempt to increase access to weather data by a diversity of academic audiences, a model for weather data distribution among various university departments was developed at the UPC (Domenico, 1991). It is referred to as the Unidata Campus Weather Display system (UCWD) and it allows for the automated distribution of data and graphical products using existing campus networks and readily available software tool kits.

This paper describes a refinement of the UCWD model to facilitate the automation and presentation of graphical weather products for use in the classroom. It permits generation of graphics based on nearly current weather data so that up-to-date examples can be used by professors to supplement lectures and lab sessions.

2.0 Overview of Need

Traditionally, professors teaching meteorology have used historical data within a classroom setting. Current data has been available to both students and faculty only outside of the classroom in a centralized map room. While this fulfills the baseline need of most meteorological courses with an associated laboratory section, general science education would benefit greatly from allowing students to learn with current weather data. The use of such data lends an immediacy to the learning experience that historical data can not match.

Another view of the need for current data in the class room is to make available data that has not been calibrated or massaged by other scientists and technicians. The belief that such data machinations should be performed, and the data presented only in a user-friendly way, is now being considered a erroneous idea by many (Serafin, 1991). The use by the professor and student of current data in its raw form to create products minimizes this problem.

3.0 Methods

Until recently, a number of roadblocks have existed to the use of current data in the classroom. The lack of integrated communications networks and the cost of computer systems have placed the financial resources necessary to making the data widely available on a near real-time basis out of reach to all but the largest atmospheric science departments. However, in the last two years the availability of low-cost UNIX2 workstations coupled with the expanding availability of campus- wide, integrated networking has changed this to the point where the integration of current data in the class room is not only technically feasible, but also financially attainable.

Along with the decrease in hardware costs has come a parallel increase in the availability of freely available software, which allows for the simplified creation of graphical images in a number of standard formats. This fact, plus the availability of a number of software tools from the UPC, has made the creation of the UCWD possible.

3.1 Goals

Two goals were identified in the implementation of the paradigm:

For the UCWD model to be of value in the classroom, the data provided must be as near real-time as possible. To meet this goal, the model makes use of automatically scheduled batch jobs with the UNIX "cron" utility.

To allow acceptance of the model by as wide an audience as possible requires that it be easy to learn and use. Software designed to be both general and flexible usually has a long learning curve due to the multiplicity of tasks that it can perform. Relatively unskilled users will find this intimidating and, therefore, not make use of the applications. Extensive use of application wrappers, implemented as UNIX shell scripts, overcomes this difficulty. The scripts mask the implementation details from the end-user and minimize the application knowledge required to access and use the powerful tools.

3.2 Model Elements

There are four main elements to the UCWD model as used within the classroom context:

The UPC assists its university users in accessing data by negotiating group discount rates with contracted weather data providers, currently Zephyr Weather Information Services, Inc. The data are transmitted via communications satellite, and universities may select which data streams are most appropriate for their needs. Data circuits currently supported by the Unidata software suite include the Domestic Data Service, the International Data Service, the Numerical Product Service and a special Unidata channel.

The Unidata channel carries a range of data prepared at the University of Wisconsin-Madison under contract with the UPC. These data include satellite and radar images, as well as selected conventional meteorological data used by the Unidata McIDAS software.

To ingest and store the incoming data, the UPC provides a software package called the Local Data Manager (LDM). This LDM handles the ingestion of site-select able products from the incoming raw data in real-time and the subsequent storage or delivery of the products.

Decoder software (available in the Purdue University Weather Processor (WXP) suite) handles the parsing and decoding of weather reports from the captured raw products to a computer usable form and stored using the netCDF library. The netCDF (for Network Common Data Form) is a portable method for storing and accessing scientific data (Rew, 1990).

The graphical product creation is handled by a number of tools, including the WXP suite and a set of graphics conversion utilities provided by PBMplus. WXP is used to create X window system raster dump files of the desired meteorological products. These files are then run through appropriate PBMplus utilities to convert to the desired graphics format.

Once the graphical products have been produced, the last step in the model is to present the product to the desired audience. This can either be done interactively via X window display packages such as "xloadimage" or via shell scripts. It is also possible to automate the process with the use of shell scripts run by the "cron" utility. This is useful for continuous display of the products in a public environment, such as a campus student center.

3.3 Discussion of Tools

To understand how the various tools utilized by the model are used, a brief discussion of the main tools is presented.

3.3.1 LDM

The LDM is a component of the Scientific Data Man agement (SDM) suite of software applications provided and maintained by the UPC. It is comprised of two parts, the first being a series of ingester programs to handle the data ingest from the various weather data circuits. The second part is the LDM server, which receives the data from the ingesters and performs site specific actions on that data. These actions include storing the data, manipulating the data in some way, mailing textual data, etc.

Ingesters may communicate with more than one server, which allows for the propagation of subsets of data to various locations. Standard TCP/IP (Transmission Control Protocol/Internet Protocol) networking protocols are used over existing campus networks.

3.3.2 WXP

WXP is a software package developed at Purdue University. It consists of several programs to both decode and visualize weather data. The decoders act on the raw ingested data to create decoded files in netCDF form. The netCDF files are, in turn, used by the analysis programs in WXP to display the weather data variables in a variety of forms such as station plots, contours and textual products.

WXP programs that act on the decoded data can generate displays either to the display terminal or to an X window raster dump file. Images may be displayed in either color or monochrome.

3.3.3 PBMplus

PBMplus, a package of graphics translation utilities written by Jef Poskanzer, converts graphic files from one graphical format to another. While not public domain, the package is freely available via anonymous ftp over the internet. The utilities may be piped together to perform several operations, including the overlaying of images, sizing, color normalization, etc.

3.3.4 Unidata McIDAS

McIDAS (Sumi, 1983) has been under development since 1974 by the Space Science and Engineering Center (SSEC) at the University of Wisconsin - Madison. The Unidata Program licenses a version of McIDAS for distribution to universities. The system operates on IBM PS/2 computers running the OS/2 operating system. Since 1987 the Unidata Program has contracted with SSEC for the generation of a specialized data stream for use by the Unidata McIDAS software system.

McIDAS has the ability to display multiple graphics in display frames. These graphics include GOES satellite images, radar plots, station plots, forecasts, meteorograms, contours and plots of surface and upper-air data, etc.

3.3.5 Xloadimage

"Xloadimage" is a freely available software package that allows graphical images in various formats to be displayed on an X window environment. It can handle images on either a color or monochrome display.

3.3.6 UNIX utilities

To pull the UCWD model together, extensive use is made of standard UNIX operating system utilities. These include shell scripts to control various elements of the model, as well as "cron" which allows automation of various steps within the model.

3.3.7 UDRES (Unidata Resource library)

UDRES is a package developed at the UPC and is part of the SDM. It allows for the concept of resource values which can be retrieved by a program and can be used to define the program environment. Use of UDRES makes customization of the UCWD model available on site, node and personal user levels. This is accomplished by the use of resource files which can be accessed by both programs and shell scripts. The resources are used to set data and executable file paths, mailing lists, file naming conventions, etc., within the program being executed.

3.4 Integration of Tools

The first element of the UCWD model, data capture, is handled by the LDM ingesters and the LDM server. The ingesters monitor the various data streams and ship the data products to the server as they are received. The server places the products in the appropriate files, which reside in a data directory that can be specified by individual user sites. The following discussion will concentrate on the treatment of the second, third and fourth elements of the model.

3.4.1 Raw Data Decode / Graphical Product Creation

The second and third elements of the model, raw data decode and graphical product creation, are handled by an integrated set of shell scripts. These scripts are executed automatically by the "cron" utility on a schedule that reflects the delivery schedules of the raw data. For example, surface observations are decoded every hour, while upper-air data are decoded every 12 hours.

The decode and product generation are handled by three scripts working together. For each data type there is a corresponding executive script. This script handles the raw data decode by setting up the command line syntax for the appropriate WXP decoder and executing it. It then calls the X window dump and the graphical product creation scripts in turn with the proper command line arguments.

The X window dump script uses the appropriate WXP analysis program to create graphics dumps in xwd raster format of selected meteorological variables. These files may be viewed directly and are also used as a starting point for the translation of the images to other graphics formats. The script is structured as a large case state ment which is keyed on a product name. The names are artificial constructs used to identify the different graphics images produced. All executive scripts call this script and pass, by command line arguments, which products they want the script to generate. Product generation can be turned off by setting the appropriate resource value in the UDRES resource file to false. This way, no modification to the scripts has to be performed when a product is not wanted. The only time the scripts have to be modified is when a new product is added.

Once the X window dumps are completed the executive script executes the graphical product creation script. This script takes the specified xwd files and runs them through one or more PBMplus utilities to create files in various other graphics formats. At the UPC we create GIF (Graphical Interchange Format), PBM (Portable Bit Map), PICT (PICTure format) and VIF (VGA Image File) files for our in-house use. GIF files can be used on a variety of platforms, including PC's, Mac's and UNIX systems. PBM files can be used on UNIX workstations for X window root background displays. PICT files are used on Mac's and VIF files are for use with the McIDAS system.

3.4.2 Graphical Product Presentation

Once the graphical products are created and available for use, the final element of the model - the presentation of the products - may be handled either as a "cron" job or interactively as needed. The use of "cron" was the original thrust of the UCWD model so that display of graphical weather products could be handled for remote displays without having any human intervention. For the use of the UCWD in the classroom, interactive presentation is more appropriate.

We are assuming, for the purposes of this discussion, that appropriate hardware to allow students in a class room setting to view the images is available. This may take the form of large-screen monitors, a video projector, an overhead projection video attachment, etc.

The simplest method for using current data in the class room is to make use of the graphical products produced by the scripts described above. This is easily accomplished by using "xloadimage". "Xloadimage" can deal with images in many different graphics formats, including xwd, GIF and PBM files. By simply giving the program the file name on the command line, it can determine what format image it has and display the image to the screen. If the display device is monochrome and the image is color, "xloadimage" produces a monochrome dithered image on the display. A nice feature of "xloadimage" is the -slideshow command line switch; by listing several image file names on the command line, the presenter may step through the series by simply executing a single keystroke.

A second method is to use a shell script to access the various WXP tools to generate current images from decoded data. While this takes a slightly longer time for the display to appear, as WXP has to calculate the variables involved, it takes advantage of the latest data available without having to wait for the automated product creation to occur. In addition, multiple scripts can be written to display multiple images in a coordinated way on the screen. This allows for implementation of a "wall of maps" concept on the computer display.

It is this method that we make use of at the UPC. We have written a variety of briefing scripts which are activated from a window manager menu on our workstations. These scripts display four images tiled in each quadrant of the screen. All displayed images created by a given script cover the same geographical area, thus giving an overall picture of the weather for that region.

As an example, we have a briefing script called COLORADO, which displays a station model plot, a manually digitized radar summary, a Denver upper-air radiosonde sounding plot, and an objectively-analyzed surface wind field plot displayed as wind vectors. With a single command, a user can display a wealth of related data and provide a viewer with a comprehensive weather display for visual analyses.

Numerous other scripts are available for customized coverages of the Washington, DC area, the continental United States, Europe and the Far East. The number of scripts possible is limited only by the imagination of the user and the available data.

In practice, a professor can create a number of scripts which illustrate the subject matter being covered in a course. These scripts can then be run during the lecture for illustration purposes. Because the images are created anew each time the scripts are run, they always reflect the most current data available.

4.0 Access to Tools

5.0 Summary

Using software tools readily available to the university community, we have shown how to incorporate current weather data into the classroom experience. The use of such methods can assist in the enhanced utilization of relevant data to make the learning experience better. It allows students to relate the subject matter being taught to the real world as it currently exists, rather than exclusively looking at purely historical case studies. The use of current data can be of much help in the teaching of analytical atmospheric science and general survey courses in meteorology.

6.0 For Further Information

For further information regarding the UCWD model, or for examples of the scripts mentioned above, contact the authors by email at:
mitch@unidata.ucar.edu
green@unidata.ucar.edu

7.0 Acknowledgments

The authors would like to acknowledge the Atmospheric Sciences Division of the National Science Foundation, and Cliff Jacobs in particular, for continued support of the Unidata Program Center. We would also like to thank the authors of the various software tools used for their support of universal software applications development by making their products freely available.

8.0 References

Domenico, B., 1991: The Unidata Campus Weather Display System. Proceedings, Seventh International Conference on Interactive Information and Processing Systems for Meteorology, Oceanography, and Hydrology, AMS, 454-458.

Fulker, D. W., 1989: Seminal Software to Analyze and Manage Geoscientific Information (Unidata Building Blocks for the Scientist-Programmer). Proceedings, Fifth International Conference on Interactive Information and Processing Systems for Meteorology, Oceanography, and Hydrology, AMS, 44-49.

Rew, R.K., G.P. Davis, 1990: The Unidata netCDF: Software for Scientific Data Access. Proceedings, Sixth International Conference on Interactive Information and Processing Systems for Meteorology, Oceanography, and Hydrology, AMS, 33-40.

Serafin, R., et al, January 1991: Study on Observational Systems: A Review of Meteorological and Oceanographic Education in Observational Techniques and the Relationship to National Facilities and Needs, UCAR/AMS report.

Sumi, V.E., R. Fox, S. limaye, and W.L. Smith, 1983: McIDAS III: A modern interactive data access and analysis system. J. Clin. Appl. Meteor., 22, 766-778.

Notes

1. The Unidata Program Center is sponsored by the National Science Foundation and managed by the University Corporation for Atmospheric Research. Mention of a commercial company or product does not constitute an endorsement by the Unidata Program Center.

2. UNIX is a trademark of AT&T



This page was Webified by Jennifer Philion.
Questions or comments can be sent to <support@unidata.ucar.edu>.
This page was updated on .