This paper summarizes the procedures used in creating a digital geologic
coverage, the problems encountered in producing this coverage, and the
rationale for putting this information in a CD-ROM. The advantages and
disadvantages of hand-digitizing or scan ning the raw data, an outline of
digital data manipulation, and the necessary deviations from the source
document are explored and explained.
The CD-ROM environment is cost
effective and provides a more stable storage media than magnetic tape.
However, if the system is to be accessible to the greatest audience of
potential users, development of a user-friendly framework is necessary.
The reaso ns for developing such a framework and the capabilities of the
software are discussed.
The U.S. Geological Survey (USGS) has initiated the release of a new
publication format entitled the U.S. Geological Survey Digital Data Series
(DDS). The DDS, to be produced within the U.S. Geological Survey, has been
conceived for two main reasons: (1) a significant increase in the present
capacity and capability of computers to store, process, and disseminate
information through the development of compact disc-read only memory
(CD-ROM) technology; and (2) a need to address the present and future
requirements of an evolving earth-science community that desires digital
geologic and related products. These products may be in the form of area
data layers, which include geological, geochemical, and geophysical data
for map areas, or point data as represented in sample locations. This
paper describes the procedures involved in starting with a printed
geologic map, creating a digital geologic data set, and integrating this
digital data into a CD-ROM environment. The reasoning associated with each
of these steps is explored. This experience was gained in the development
of the Geology of Nevada: A Digital Representation of the 1978 Geologic
Map of Nevada, USGS DDS-2, which is the first digital geologic coverage
released on a CD-ROM.
The goal in creating a digital geologic coverage
is to provide to the user-community a geologic base map useful in studies
requiring geologic input. These studies may involve a wide array of
topics, of which geology is just one component. With the creation of a
geologic coverage in formats accessible to multiple computer and software
types and the storage on a relatively permanent media, this coverage will
be available to many disciplines over a long period of time. Just as the
storage of data has evolved
from punched cards to magnetic tape, magnetic tape storage is evolving
towards CD-ROM for many of the same reasons, such as the large
data-storage capability, the random accessibility of the data, and the
physical tability of the media over longer perio ds of time. Unlike
magnetic tapes, which have temperature and humidity storage requirements
and degrade over time, CD-ROM storage is a relatively permanent archive of
the data.
This
paper summarizes the activities of many branches within the U.S.
Geological Survey. The authors wish especially to thank Edward McFaul of
the Office of Scientific Publications and J. Nicholas Van Driel and John
W. Jones of the Geographic Information Systems (GIS) Research Laboratory
of the U.S. Geological Survey for critical reviews of this paper.
Geographic information systems (GIS) are designed to manipulate,
analyze, and display digital geographic information within a computer
environment. Components of the system usually include scanners or
digitizers to input the data, computers and software to manipulate and
analyze the data, and plotters or printers to display the data. The level
of sophistication and complexity of each of these systems is entirely
dependent upon the ability of the user to use the system to its fullest
potential and the cant. The reasoning associated with each
of these steps is explored. This expital investment by the user into acquisition and
development of the necessary hardware and software. The GIS used to create
the digital data set discussed in this paper, and the subsequent
terminology, is ARC/INFO related (Environmental Systems Research
Institute, Inc., 1990). The data within a GIS is usually in the form
of a digital map called a coverage
(fig. 1). Each coverage consists of a
set of features, represented by related files, that can be manipulated by
a data base management system. Coverage features include the fo llowing:
arcs that represent linear features and the borders of polygons; nodes
that represent arc endpoints and the locations where line features
connect; label points that represent point features and are used to attach
attributes to polygons; polygons that represent area features and are
defined by arcs; tics that are geographic control points and are used for
registration of a coverage; and annotation that is text used to label
coverage features.
Many considerations must be examined before
creating a digital coverage. It is important to define a specific
objective and product, the methods to be used to produce the digital
input, and the features to be displayed. A primary concern in reproducing
a published geologic map is managing the accompanying thematic data layers
such as topography, roads, water bodies, and cities. Two alternatives are
available. The first is to scan all linework present to produce a data set
containing information of different themes, but possibly too large and
unwieldy for the user and focused on no specific audience. All users would
be required to invest time and labor in editing this file to extract a
useful subfile. In a second approach, experience has shown that individ
ual thematic layers of data, with each layer representing a specific data
type, are more useful to users, who can then combine respective layers of
interest.
After objectives and products have been defined, the flow of
data within a GIS starts with the method used to incorporate this data
into a digital format and the prerequisite map preparation of the raw
input data
(fig. 2).
Factors that must be evaluated when considering
different methodologies include the suitability of the raw data for
digitization, the time available, and the complexity of the map. The
amount of document preparation required at this point is dependent upon
the method used to convert the
data into digital form. Some of the benefits and drawbacks of hand
digitizing versus optical scanning are shown in
figure 3.
Hand digitizing of spatial data requires little document preparation. Ideally,
the data exists on a scale-stable medium, such as mylar, and all linework
is legible. Advantages of hand digitizing include the capability to
digitize selected features, especi ally if only a limited amount of
information on the map is of interest; the capability to handle linework
of inconsistent quality; and the relatively low capital investment in the
digitizing tablet and software.
The drawbacks of hand digitizing map
linework include the time consumed on large data sets; the fatigue factor,
which results in lines being missed or imprecisely located; the loss of
detail and inflection points, which are generalized due to the density of
nodes; and the variable quality of hand digitizing in different areas of
the same map.
If the linework is to be converted into a digital format
through the use of an optical scanner, more document preparation is
necessary. The lines should be on a scale-stable media, and the line
characteristics (intensity, width) should be consistent. All unwanted
lines should be removed at this point because removing them later in the
process could be more labor intensive. The advantages of using a scanner
to create a digital data set include the consistency of data throughout
the map; the speed of digitizing a large data set in a short time; the
scanning of very complex patterns with the detail retained; and the
locating of all lines with precision.
Drawbacks of using an optical
scanner to encode digital data include the capital investment required to
both purchase and maintain the hardware and software; the need for
consistent linework to calibrate the optical head; the inability to
retrieve selecte d arcs of the scanned area; and the preparation of the
input map for scanning.
After creating the digital data set, the next
steps involved in producing the final product can be taken in several
different orders
(fig. 2). Tasks that must be performed include
transforming from scanner units to geographic coordinates; projecting the
geographic coordinates to a desired map projection; editing arcs by
deleting extraneous lines, adding new lines, and repairing breaks;
creating a polygon topology; adding labels to polygons; tagging polygon
labels with attributes; and plotting the coverage . Many of these steps
are iterative, particularly the editing, tagging, and plotting.
Although not meant to be comprehensive, a few comments concerning these
steps may be helpful. The method of transformation from scanner units, and
the projection of geographic coordinates, is dependent upon the type of
technology available. Usually, the transformation involves the assigning
of geographic-coordinate values to control points that have been located
during the initial map preparation and digitized from the original
document. These control-point geographic-coordinate values are then used
to interpolate positional x and y data for all nodes throughout the rest
of the coverage. The projection of the geographic coordinates into a
desired map projection is done by providing the central meridian, the
lines of latitude, and other parameters required
by that particular projection.
The editing of arcs by removing
extraneous lines, adding new lines, or repairing breaks is a
time-consuming, labor-intensive task. Extraneous lines may be caused by
stray marks on the map or a poorly set threshold of the scanner that
detects unwanted map features and by linking scanned pixels that are close
together but not meant to be joined, such as where two arcs are tangential
or two arcs intersect at an acute angle to result in a cross arc. The need
to add new lines may result when new data is to be added to the coverage
or the line quality of the scanned material is poor and, therefore,
missed. Repairing breaks in arcs is necessary whenever a polygon topology
is to be created with color fill used. Breaks allow polygons to leak color
fill into each other with the color attribute assigned to the first label
encountered within the involved polygons.
Adding labels to polygons
can usually be accomplished by using a single command within a GIS.
Assigning attributes to these polygon labels, however, is done by
hand-tagging these individual labels. Attributes assigned to the Nevada
digital coverage included a unique polygon color-fill number and map-unit
mnemonic. However, a wide variety of attributes may be assigned to these
label points.
Error checking of the arcs and labels within the
coverage is accomplished after a plot is produced. At this point the
process of creating a digital coverage becomes iterative, with the
continued editing of the arcs and labels, the checking of polygon attributes,
and the replotting of the coverage.
The development
of a digital coverage to represent thematic information of map areas, and
the resulting ready availability of this digital data to many different
users for many different applications, is a significant advancement for
earth science; however, digital information must be used with caution
because digital products are more easily subjected to abuse and misuse by
the uninformed user than the same data in paper media. For the digital
geologic coverage of Nevada, several deviations between the published map
and the digital representation of that published map are necessary
(fig.4). These deviations are mentioned here to relate our experiences in
creating a digital cove rage and to caution digital-coverage makers of
potential obstacles. Obvious differences include the lack of thematic
data, such as roads, topography, and cities, on the digital coverage. As
mentioned earlier, the incorporation of these individual thematic
layers would be detrimental to a user interested only in the geology.
Another difference is that Nevada water bodies, including reservoirs,
lakes, rivers, and streams, are not included on the digital geologic
coverage; therefore, map units beneath these water-covered areas had to be
inferred from the surrounding geology. Faults on the published geologic
map can be differentiated from lithologic contact lines based on the
increased line thickness of the faults; but on the digital coverage, these
lines are
the same thickness. Map inconsistencies on the original published
geologic map posed another problem where map-unit mnemonic and
color/pattern were not in agreement; however, these inconsistencies were
resolved by using the map unit that made the best geologic sense.
Misuse of digital data, a problem not unique to digital map coverages,
arises when the data are "pushed" beyond their intended purpose. For
example, the use of the Nevada digital geologic coverage at a scale
smaller than 1:500,000 (the scale at which the original map was published
and scanned) to produce a 1:250,000 product, although tempting to users,
would result in a very poor representation of the area and possible
errors.
CD-ROM technology was introduced to the U.S. Geological Survey in
1985, when a CD-ROM reader and discs were purchased for evaluation.
Especially for the USGS, the most important characteristic of CD-ROM
technology is its cost effectiveness-from the perspectives of both the
publisher/distributor and the end user-that results for two reasons:
- 1. CD-ROM is based on CD audio technology and its attendant physical
standards, which benefit from the economy of scale of that mass consumer
market; and
- 2. CD-ROM is based on an international standard (ISO-9660) that
specifies the logical structure of the data files placed on the disc to
give CD-ROM equipment independence (for example, use with MS-DOS,
Macintosh, or various workstations and minicompu ters).
In contrast,
the only other mass-storage medium with both physical and logical (ANSI)
standards, allowing computer independence, is a nine-track, half-inch
magnetic tape. This tape-drive subsystem would cost an end user in excess
of $5,000, whereas a CD- ROM-drive subsystem can be purchased for $400 or
less. More meaningful, perhaps, is that the U.S. Geological Survey's
1:2,000,000 Digital Line Graph disc is sold for $32 when the equivalent
data on magnetic tape would cost the end user $1,400!
In addition to
the cost effectiveness of CD-ROM, the technology provides several
significant benefits: (1) the data on the disc cannot be altered or
changed by the user; (2) the discs are reasonably tolerant of use in the
office environment without succumbing to "media error" problems; (3) the
polycarbonate plastic is relatively stable (50 to 100 years) and therefore
becoming an accepted archive medium; (4) the data/files can be randomly
accessed, compared to the sequential access of tape; and (5) the storage
capacity is substantial, at 680 megabytes or more.
PRODUCTS CONTAINED ON CD-ROM
The appeal of
digital coverage is that it can be used by many people, on many different
types of computers, for many reasons, to study many different kinds of
problems. Of paramount importance in the development of this data is its
distribution in a form and format to be available to the widest spectrum
of potential users. For this reason, the digital geologic coverage
produced for Nevada was released in three different formats on the same
CD-ROM. These formats are (1) ARC/INFO 5.0.1 Export format, for AR C/INFO
users; (2) Digital Line Graph-3 Optional format, for users familiar with
DLG-3 optional data; and (3) an ASCII format describing arc and polygon
attributes, for users unable to use either of the other two. The
ARC/INFO Export command converts the coverages, INFO data files, text
files, font files, and symbol-set files into an ARC/INFO interchange file
for export to another site or type of computer running ARC/INFO. These
Export interchange files contain all
coverage information and appropriate INFO file information in a
fixed-length ASCII format. The ARC/INFO Import command creates a coverage,
INFO data files, text files, font files, and symbol-set files from an
ARC/INFO interchange file that was exported from another computer running
ARC/INFO. This command is the most efficient way to restore coverages that
were generated elsewhere into an ARC/INFO environment.
The Digital
Line Graph-3 (DLG-3) Optional format has been the most widely used method
of transferring map data because it stores topology, has a record length
of 80 characters, and is the preferred method used internally by the U.S.
Geological Survey. This ability to record coordinates, feature topology,
and descriptive attributes has made the DLG-3 Optional format a popular,
public-domain format for data transfer. For more detailed information on
the use of digital line graphs, refer to USGS Data Users Guides 1-3 (U.S.
Geological Survey, 1986, 1989, 1990).
ASCII format files are also
provided on the CD-ROM as arc attribute tables (AAT) and polygon attribute
tables (PAT). These ASCII files allow users to generate a cover if they
are unable to process the other formats provided. An example of the ASCII
arc attribute table format and data set is shown in
table 1. Brief
descriptions of the items observed are also presented. It should be noted
that the MAJOR1 item is equivalent to the right polygon identification
number and the MINOR1 item is equivalent to the left polygon
identification number. An example of the ASCII polygon attribute table
format and data set is shown in
table 2. For these files, the MAJOR1 code
is equated to the unique COLR3 numeric code, and MINOR1 code equals the
COLR number.
To make the available data sets as useful as possible,
and also to enhance the efficiency with which they can be used, previewing
software is supplied to the user on the CD-ROM. Because of the possible
diversity of applications and the unlimited combinations with other types
of data, the previewing software was designed to perform very basic, but
very important, tasks. The previewing software designed by the U.S.
Geological Survey for use in conjunction with digital geologic coverages
is generic in nature, and therefore, available to all public users.
Important considerations in designing this software included the speed at
which the digital coverage could be accessed, a filtering mechanism, by
which more detailed data is displayed as an area is enlarged, and the
incorporation of thematic overlays. The speed issue was addressed by
making available individual coverages approximately equal to a two-degree
quadrangle that could be selected by quadrangle name. The filtering
mechanism, allowing display of more detailed data as an area is enlarged,
is controlled by an argument in the software that could be changed by the
user. The addition of thematic overlays to the geologic base, such as
water bodies, roads, railroads, and State and county boundaries, is
possible because these coverages are also provided.
Among the tasks
required of the previewing software is the capability to "zoom" in and out
and to "pan" left and right, up and down, to specific spots around the
coverage. This capacity results in efficient utilization of the data and
effective use of time since redisplaying of the entire coverage is
unnecessary.
Another feature important to any user is the capability
to know the location of the cursor at any time through the use of a window
that displays latitude, longitude, and map-unit mnemonic for the polygon
in which the cursor lies.
The capability to display a map-unit legend
or key on the screen simultaneously with the coverage is also deemed
important as the user must know which map units are displayed on the
screen and be able to differentiate between the units. The large number
of available map units and the limited number of colors that can be
discerned by the normal eye make distinguishing individual units
difficult. Of help in locating all map units of a specific type is a
"flash" command that causes all occurrences of the type located beneath
the cursor to turn white and flash as the key is depressed. This important
function locates occurrences of that map-unit type.
The capability for
the user to design and save area coverages unique to a particular
application is also an important feature that has been incorporated. When
a coverage is displayed, additional overlays of interest may be shown,
selected formations label ed with a mnemonic, and the colors of formations
changed. With this option, map units may be combined visually. Once this
unique coverage has been created, it can be renamed, saved, and later
recalled.
Environmental Systems Research Institute, Inc.,1990, ARC/INFO 5.0 user
manuals: Redlands, Calif. Stewart, John H., and Carlson, John E.,
1978, Geologic map of Nevada: U.S. Geological Survey, Reston, Va., scale
1:500,000.
Turner, Robert M., and Bawiec, Walter J, 1991, Digital
geologic coverage of Nevada; A digital representation of the 1978 geologic
map of Nevada: U.S. Geological Survey Digital Data Series 2.
U.S.
Geological Survey, 1986, Digital line graphs from 1:24,000 scale maps:
U.S. Geological Survey Data Users Guide 1, 109 p.
U.S. Geological
Survey, 1989, Digital line graphs from 1:100,000-scale maps: U.S.
Geological Survey Data Users Guide 2, 88 p.
U.S. Geological Survey,
1990, Digital line graphs from 1:2,000,000-scale maps: U.S. Geological
Survey Data Users Guide 3, 71 p.
Return
to Chapter D Continue
to Chapter F
Return
to Bulletin 2016 contents
Return to USGS Home Page
