PCMDI Report No. 24

Documentation of the AMIP Models on the World Wide Web

by

Thomas J. Phillips
Program for Climate Model Diagnosis and Intercomparison (PCMDI)
Lawrence Livermore National Laboratory (LLNL)

August 1995
(An abridged version of this report also appeared in the June 1996 issue of the Bulletin of the American Meteorological Society, Vol. 77, pages 1191-1196. See also a similar article by the author entitled "Atmospheric Model Documentation Available Online" that is published by Eos Electronic Supplements at http://www.agu.org/eos_elec/96069e.html.)

Abstract

Table of Contents:

• 1. Introduction
• 2. The World Wide Web
• 3. Hypertext Model Documentation
• 4. Future Plans
• Acknowledgments
• References
• Appendix: Accessing AMIP Model Documentation on the WWW

1. Introduction

The intercomparison of atmospheric general circulation model (AGCM) experiments of a similar type has become an increasingly popular methodology for assessing the strengths and weaknesses of climate simulations (e.g., Cess et al. 1990, Randall et al. 1992). In such endeavors, attempts to attribute differences among the simulations to specific model properties require, as a minimum prerequisite, the accurate and comprehensive documentation of these features.

Regrettably however, atmospheric model documentation typically is fragmentary and scattered across numerous publications. It is also often inaccurate, in the sense that the pace of model development and the proliferation of new model versions usually outstrip their recorded descriptions. More often than not, the detailed configuration of a model for a particular experiment also is undocumented. In addition, there may be much unevenness in the descriptions of different facets of models (e.g., the description of atmospheric dynamics often eclipses that of surface processes). This incompleteness usually is replicated in published results of an intercomparison experiment, in that participating models' features often are summarized only perfunctorily.

By the early 1990's, developments within the World Climate Research Programme (WCRP) set the stage for redressing this state of affairs. A notable example was the 1991 launching of the Atmospheric Model Intercomparison Project (AMIP), an ambitious effort to evaluate the performance of current AGCMs in simulating the climate of the decade 1979-1988 under common specification of ocean temperatures and radiative forcings (Gates 1992). The widespread participation of international modeling groups and the unprecedented scope of model diagnosis within the AMIP made it imperative to set new standards in model documentation, as well as in a host of other arenas pertinent to model intercomparison. Crucial support by the U.S. Department of Energy also made it possible for the Program for Climate Model Diagnosis and Intercomparison (PCMDI) to meaningfully address these tasks in the course of its coordination of the AMIP on behalf of the WCRP.

Initially, documentation of the AMIP models crystallized in the form of PCMDI Report No. 18 (Phillips 1994), a summary of the numerics, dynamics, and physics of 30 participating models. The report's chief strength was that it centralized information on these models according to a common, and reasonably comprehensive framework. The bulk of the report qualitatively summarized the representation of these features in each AMIP model; tables also succinctly indicated the ways in which selected properties played out across the models.

Despite the extensive scope of PCMDI Report No. 18, it suffered from an inherent limitation, in that the printed page is not well suited for the frequent updating of information that is endemic to the AMIP. For example, since the 1994 publication of the report, the Center for Climate System Research has joined the AMIP (see list of modeling groups), and 8 other groups have repeated the intercomparison experiment with new model versions.

Such frequent changes make it impractical to issue printed revisions. The obvious need for a "living document" therefore demands the use of an electronic medium for rapid amendment and widespread dissemination of model documentation. In principle, this could be effected by physical transfer of archival media to AMIP participants (e.g., magnetic diskette or optical compact disk), but the Internet's World Wide Web offers a much less costly means for transferring information, and in "real time" as well.

2. The World Wide Web

For some years, scientists have exchanged information on the Internet, e.g., via electronic mail and newsgroups, the File Transfer Protocol (FTP), and the Gopher and Wide Area Interactive System (WAIS) Protocols. However, the recent emergence of the World Wide Web (WWW, or W3) based on the Hypertext Transfer Protocol (HTTP) has qualitatively transformed the relationship of scientists to the Internet (Berners-Lee et al. 1994).

Since the inception of the W3 only a few years ago, the application of this technology has grown exponentially (from less than 100 Web sites in 1993 to about 35,000 by mid-1995). In part, this growth is due to a server-client protocol that reprises the interaction exemplified earlier by anonymous FTP: information is served automatically "on demand" to each client. However, the phenomenal popularity of the Web is due mostly to its support of multimedia applications (text, images, sound) which can be accessed "seamlessly" across different computing platforms. The scientific utility of these capabilities is apparent. Indeed, the World Wide Web was created at CERN, the European Laboratory for Particle Physics, in order to foster collaborative exchange of scientific information in a variety of forms--see Segal 1995.

Exchange of multimedia is made possible by the Web's lingua franca, the Hypertext Markup Language (HTML). "Hypertext" is a means of conveying electronic information by first fragmenting it, and then spanning the fragments by "hyperlinks" that are activated by "point-and-click" operations. A body of information therefore can be explored "nonlinearly", according to the user's whims (Nelson 1981, Schneiderman and Kearsley 1989). In HTML, the hyperlinks may span different media (e.g., text and images) as well as different servers (hence the term "World Wide Web"). For HTML multimedia to be rendered intelligible, however, one needs W3 navigation software (see Appendix).

3. Hypertext Model Documentation

Hypertext formatting greatly enhances the utility of lengthy treatises such as PCMDI Report No. 18, and atmospheric model documentation lends itself readily to this representation (Phillips and Meyer 1990). That is, because the individual scientist often has a particular interest only in a subset of models and properties, it is natural to break the hypertext documentation according to model, with provision for rapid access of all feature descriptions.

For the AMIP models, this schema is implemented by lists of hyperlinks that map the selected features to each model (Phillips et al. 1995a). In addition, hyperlinks provide convenient cross referencing of related features within each model summary (see Display). Moreover, each citation of a reference from the model documentation literature is connected to the associated bibliographic information by means of a hyperlink. The latter feature is essential for those scientists requiring quantitative details on model algorithms and parameterizations-- information that is beyond the intended scope of the hypertext documentation. (A bibliography of references specific to each AMIP model is provided in addition to a comprehensive listing of more than 500 references. )

As in PCMDI Report No. 18, tables provide an overview of selected properties expressed across the AMIP models. A "History of Changes" page lists each update of the AMIP model documentation and a "What's New?" page advertises the most recent changes. In addition, a glossary explicates the many acronyms of the AGCM world, while hyperlinks connect to the Internet servers of these institutions.

Aside from the enhancements afforded by hypertext formatting, there are other advantages to accessing AMIP model documentation on the World Wide Web. For example, W3 browsers allow one to search on key words, and to save, annotate, and print HTML documents at the home site.

4. Future Plans

In only a few years, the advent of the World Wide Web has revolutionized the manner in which scientific information is exchanged. In the future, therefore, PCMDI plans to vigorously apply this new technology in order to further the validation, intercomparison, and improvement of global climate models.

Having pioneered in developing hypertext AGCM documentation, PCMDI will maintain its currency for the AMIP. Feature descriptions of each new model version will be made available on the Web, as the requisite output data are quality assured and archived. Moreover, each model version will be identified unambiguously ("Model Designation"), as recommended by the WCRP's Working Group on Numerical Experimentation, and will be situated historically ("Model Lineage") in relation to other models. This documentation therefore will serve as an ongoing record of international AGCM development in coming years.

In the longer term, PCMDI plans to document the properties of coupled ocean-atmosphere models that will increasingly be the focus of future intercomparison experiments. This will require development of new feature categories that are relevant for coupled models. The resulting summary descriptions of model properties also will be made available on the Web.

Finally, PCMDI intends to expand its other W3 applications. These include W3 dissemination of PCMDI software, observational data products for model validation, and publications describing the activities of the PCMDI staff. In addition, information on AMIP participants, datasets, and publications will be appropriately updated with the further progress of this intercomparison project.

Acknowledgments

I am indebted to Rita Anderson and Michael Brösius of the LLNL Technical Information Department for translating PCMDI Report No. 18 into HTML. I also gratefully acknowledge the assistance of representatives of the AMIP modeling groups in the correction and clarification of the documentation. This work was performed under the auspices of the U.S. Department of Energy, Environmental Sciences Division at the Lawrence Livermore National Laboratory (LLNL) under Contract W-7405-ENG-48.

References

Berners-Lee, T., R. Cailliau, A. Juotonen, H. Frystyk Nielsen, and A. Secret, 1994: The World Wide Web. Communicns. Assoc. Comput. Mach., 37, 76-82.

Cess, R.D, G.L. Potter, J.P. Blanchet, G.J. Boer, A.D. DelGenio, M. Deque, V. Dymnikov, V. Galin, W.L. Gates, S.J. Ghan, J.T. Kiehl, A.A. Lacis, H. Le Treut, Z.-X. Li, X.-Z. Liang, B.J. McAvaney, V.P. Meleshko, J.F.B. Mitchell, J.-J. Morcrette, D.A. Randall, L. Rikus, E. Roeckner, J.F. Royer, U. Schlese, D.A. Sheinin, A. Slingo, A.P. Sokolov, K.E. Taylor, W.M. Washington, R.T. Wetherald, I. Yagai and M.-H. Zhang, 1990: Intercomparison and interpretation of climate feedback processes in 19 atmospheric general circulation models. J. Geophys. Res., 95, 16601-16615.

Gates, W.L., 1993: AMIP: the Atmospheric Model Intercomparison Project. PCMDI Report No. 7 and Bull. Amer. Meteor. Soc., 73, 1962-1970.

Nelson, T., 1981: Literary Machines. Self-published, Swarthmore, PA.

Phillips, T. J., 1994: A summary documentation of the AMIP models. PCMDI Report No. 18, UCRL-ID-116384, Lawrence Livermore National Laboratory, Livermore, CA, 343 pp.

Phillips, T.J., 1995: AMIP information on the World Wide Web. Proceedings of the First AMIP Scientific Conference (15-19 May 1995), WCRP Report, World Meteorological Organization, Geneva (in press).

Phillips, T.J., and M.K. Meyer, 1990: A computerized database for general circulation model intercomparison studies. UCRL-JC-106079, Lawrence Livermore National Laboratory, Livermore, CA.

Phillips, T.J., R. Anderson, and M. Brösius, 1995a: Hypertext summary documentation of the AMIP models. UCRL-MI-116384, Lawrence Livermore National Laboratory, Livermore, CA (URL http://www-pcmdi.llnl.gov/projects/modeldoc/amip/).

Phillips, T.J. and other members of the PCMDI stafff, 1995b: PCMDI World Wide Web pages. UCRL-MI-119847, Lawrence Livermore National Laboratory, Livermore, CA (URL http://www-pcmdi.llnl.gov/).

Randall, D. A., R. D. Cess, J. P. Blanchet, G. Boer, A. D. DelGenio, M. Deque, V. Dymnikov, V. Galin, W. L. Gates, S. J. Ghan, J. T. Kiehl, A. A. Lacis, H. Le Treut, Z.-X. Li, X.-S. Liang, B. J. McAvaney, V. P. Meleshko, J. F. B. Mitchell, J. J. Morcrette, G. L. Potter, L. Rikus, E. Roeckner, J. F. Royer, U. Schlese, D. A. Sheinin, A. Slingo, A. P. Sokolov, K. E. Taylor, W. M. Washington, R. T. Wetherald, I. Yagai and M.H. Zhang, 1992: Intercomparison and interpretation of surface energy fluxes in atmospheric general circulation models., J. Geophys. Res., 97, 3711-3724

Schneiderman, B. and G. Kearsley, 1989: Hypertext Hands-on! An Introduction to a New Way of Organizing and Accessing Information. Addison-Wesley, Reading, MA.

Segal, B.M., 1995: A short history of Internet protocols at CERN. Accessible on the World Wide Web (URL http://ben.home.cern.ch/ben/TCPHIST.html).

Appendix: Accessing AMIP Model Documentation on the World Wide Web

a. Internet Addressing Conventions

To access information on the World Wide Web (W3), the essential first step is to obtain a connection to the Internet, either directly or through a host computer, online service provider, etc.). Each W3 server on the Internet has an associated address known as a Uniform Resource Locator (URL), which takes the form

Here the string `http' denotes the Hypertext Transfer Protocol. Every file on a W3 server is accessed according to a URL that is relative to that of the server's, e.g.

where the extension `html' identifies the file as one containing information expressed in the Hypertext Markup Language (HTML), the medium of exchange on the Web. The "home page" of a W3 server is a file that functions as a directory to other server files. (Access of the home page sometimes requires specification of a URL that includes its file name, but often the system is configured so that access of the W3 server alone is sufficient to bring up the home page.)

b. Navigation Tools

To operate on the Web, it is necessary to acquire a navigation tool, or "browser", that interprets documents written in HTML. Among the multitude of such tools presently available are the the Mosaic browser, whose development by the National Center for Supercomputing Applications (NCSA) sparked much of the early interest in the Web. Because it offers some additional enhancements, the Netscape Navigator (developed by the Netscape Communications Corporation) is now the browser of choice of most W3 users. Both Mosaic and Netscape browsers are available for use on X Window (Unix), Apple Macintosh, and Microsoft Windows computing platforms. The software also includes "viewers" and "players" for rendering images and audio applications. Users of computers that do not support multimedia may acquire a text-only browser such as Lynx, which currently runs in Unix, VMS, or DOS operating systems. A catalogue of other possible choices of navigation tools can be accessed from the home page of the World Wide Web (see Section d).

Mosaic and Lynx browsers are freely available for nonprofit applications, and an unsupported (beta) version of Netscape Navigator also can be acquired for free (the supported version being available at nominal cost). All these browsers may be downloaded via anonymous FTP, and Mosaic and Netscape browsers can be obtained on diskette as well.

Most browsers come with "help" documentation to orient the new user to basic procedures (e.g., how to "open" a URL, navigate the Web, save and print W3 files, etc.). The W3 home pages for NCSA and Netscape also provide general information on the World Wide Web and the Hypertext Markup Language, as well as directories and "search engines" that can be used to obtain URLs of interest. Much useful information of a similar type may also be obtained from the home page of the World Wide Web (see Section d).

c. Accessing PCMDI's W3 Pages

Once in possession of a W3 browser, a user can access the PCMDI home page at URL

The home page contains a directory of hyperlinks (indicated by highlighted entries); clicking on these will pull up other PCMDI Web pages containing information, software, or data pertinent to climate modeling. For example, clicking on entries labeled "Model Features Documentation" will bring up the hypertext documentation of AMIP model features. This also can be accesed directly by opening URL

Entries to other information on the AMIP (Phillips 1995) are provided on the PCMDI home page as well.

d. Address Information

Addresses and FTP/W3 servers where the Mosaic, Netscape, and Lynx W3 browsers may be obtained are listed below. The URLs of relevant pages on the World Wide Web's home server are also included.

National Center for Supercomputing Applications (NCSA Mosaic)

The University of Illinois at Urbana-Champaign

605 E. Springfield, Champaign IL 61820

e-mail: mosaic@ncsa.uiuc.edu

Anonymous FTP: ftp.ncsa.uiuc.edu/Mosaic (Use subdirectories /Unix, /Mac, or /Windows according to computing platform; see `readme' files for installation directions.)

W3 URL: http://www.ncsa.uiuc.edu/General/NCSAHome.html

Netscape Communications Corporation (Netscape Navigator)

501 E. Middlefield Road, Mountain View, CA 94043

e-mail: info@netscape.com

Anonymous FTP: ftp.netscape.com/pub/netscape (Use subdirectories /unix, /mac, or /windows according to computing platform; see `readme' files for installation directions.)

W3 URL: http://www.netscape.com

The World Wide Web Server

W3 URL of home page:
http://www.w3.org/hypertext/WWW/TheProject.html

W3 URL of catalogue on client (browser) software for Web navigation:
http://www.w3.org/hypertext/WWW/Clients.html

LLNL Disclaimers

UCRL-ID-121923