PROGRAMS AND PLANS--Update of the National Policy to Archive Ground Water Flow and Transport Models



OFFICE OF GROUND WATER TECHNICAL MEMORANDUM NO. 00.02

Subject:  PROGRAMS AND PLANS--Update of the National Policy to
          Archive Ground Water Flow and Transport Models

INTRODUCTION

Office of Ground Water Technical Memorandum 93.01 established a policy for
archiving ground-water flow and transport models.  Although the fundamental
policy remains the same, there is a need to update the policy to account
for changes in storage technology and the wider use of supporting software
to generate input data for the models.  Rather than describing only the
changes, this memorandum contains the complete revised policy.

POLICY

Ground-water flow and transport models are an integral part of our
interpretive ground-water investigations, and the results of these models
form the basis for many of the conclusions published in U.S. Geological
Survey reports. The numerical data and related information that comprise
these models need to remain available to: (1) support and validate the
results in published reports, (2) assure that working versions of all
models are available for future scientific use, and (3) assure that the
data are available to the public when requested.  The appropriate model
data and related information are to be stored in a permanent,
well-documented manner to ensure their continued availability.

All ground-water flow and transport models that are a significant part of
ground-water investigations with completion dates of October 1993 or later
are to be included in a model archive.  The Regional Ground Water
Specialists are to act on behalf of the Office of Ground Water to assure
that all required information is present in the archive.  Status of the
archives also will be examined as a routine part of District ground-water
discipline reviews.  Implementation requirements for the archive are
presented in an Attachment to this memorandum.

The archive is for internal Water Resources Division (WRD) access and use.
The release of information from the archive is subject to compliance with
any existing WRD policies that may apply to the public release of such
information.  For example, the archive can be used to provide the public
with input data for model simulations that are described in a published
report.  Other data in the archive that were not directly used as model
input may require additional documentation prior to release.  For example,
geographic information system (GIS) data that were used to create model
input files may be stored in the archive, but such data could contain
interpretive information that was not documented in the model report.

The ground-water model archive does not relieve individual investigators of
the need to fully describe and document model analyses in their reports.




                                   William M. Alley
                                   Chief, Office of Ground Water

Attachment

WRD Distribution:  A, B, S, FO, PO
Regional Ground Water Specialists, NR, CR, SR, WR

This memorandum supersedes Office of Ground Water Technical Memorandum No.
93.01


ATTACHMENT

IMPLEMENTATION OF DISTRICT GROUND WATER MODEL ARCHIVES

Each District will set up an on-line archive on a locally based computer.
The archive must be located on disks that are routinely backed up as done
for other mission critical information.  Optical disk storage cannot be
used as a substitute for on-line storage.

The archive will consist of a main directory.  The suggested name for this
is GWMARCHIVE; however, another name can be used if desired.  A report
subdirectory, located directly below the main directory, will be
established for each published report containing a ground-water flow or
transport model analysis.  Each report subdirectory should be given a name
that clearly reflects the U.S. Geological Survey report number.

A subdirectory named CONTENTS, located immediately below each report
directory, will include one or more files that contain: (1) the full
reference for the subject report; (2) descriptions of the subdirectory
structure and of the files contained in each subdirectory, (3) descriptions
of data file formats, when appropriate; (4) the sequence of model runs; and
(5) instructions for running simulations.

The archive must include the model source codes, input files, macros and
operating files such as UNIX shell codes and personal computer batch (BAT)
files, and model output files for each simulation described in the report.
These simulations will include (when applicable) the final calibrated
steady-state and transient results and any predictive results described in
the publication.  Model results of minor importance, such as interim
calibration runs, should not be archived.  The model output files are
included to allow future verification that the model reproduces the
published results when the input files are rerun.  Input files must be
stored as read by the model.  Typically the input files will use an ASCII
format. In cases where model input files are proprietary or
machine-dependent, ASCII versions of the files should also be stored in the
archive if it is possible to generate them.  The ASCII version will make it
more likely that the files can be used on virtually any computer without
the need for specialized or proprietary software.

The storage of additional ancillary data is optional, but is strongly
encouraged.  Examples of ancillary data that might be stored are GIS data
files, data stored by pre-processor codes, or other data directly related
to the model simulations.  Where possible, these data should be stored in a
widely supported format as opposed to a proprietary format.  For example
ARC/INFO offers the ability to export GIS data stored in their proprietary
coverage format into formats that can be read by other applications.  If
ancillary data are stored elsewhere in permanent USGS storage systems, then
there is no need to duplicate the storage in the model archive.
Documentation of ancillary data must include a description, the source,
format, the version of software on which it was produced, etc.

When the input data of one model depends directly on output from another
model, both models are to be included in the archive.  If the models are
documented in separate reports, a cross-reference between the reports must
be included in the CONTENTS directory of the archive entry for each report.

The appropriate model files must reside in the District archive when the
report is submitted to the Region for approval.  If additional model
simulations are required as a contingency for approval, these simulations
should replace, or be added to the archive as appropriate.