n Alas - Alaska Division of Geological & Geophysical Surveys - State ...

have an easy reference to that day's data sat. Once a
bad altimetry elevation is verified, all other altimewy
elevatlons on that day can be eaaily checked and possibly
corrected. 'Ihk procedure of data storage worked
well when most calculations were performed by hand
but has proved to be even more useful on modern large
computers, although much specfalized programming
has been needed.
The earllest rnegnetic tapes of dIgi&l terrain
were received in a 'planar" format (elevations calculated
on a 0.01-in. pld [about 68 m] for a few available
1:250,000-scale maps). These tapes were used for
preliminary tests of the suitabWty of such models for
terrain correctjons, but only a limited number of tapes
were received. A much more comprehensive data set
is now available in the DEM format, which in Aleska
provides elevations on a 3- by &second geographic grid
(3 by 9 seconds north of lat 70' N.). When the tapes
are read, these elevations are combined to form files
of mean elevations for cornpartmenta of nominal 114minute
(15 by 30 second), 1-mhute (60 by 120 second),
and 3-minute (3 by 6 rnlnute) rrize, whfch are organized
into map files suitable for use by the Plouff terraincorrection
program. For computer storage, the map
files are organized into storage segments that fnclude
all the maps within each square degree, except that 3-
minute compartments are stored in I- by 24egree
"quadrangle" squares. The DEM tap- we also organized
into files of alevation data for individual square
degrees, so that the computer can calculate the segment
name, map names, and other label data fmrn the
identifytng information in each tape file. Each map
label includes the tape name and date of md'hg, in
case tape or program errom ere discovered at a later
date. Mean-compartment elevations are calculated
from elevations that have been so weighted that the
importance of elevations on compartment boundaries
and corners is proportionately decreassd. The geographic
arrangement of the DEM format thus permits
a more precise detetminatfon of mean elevation than
could be readily obtained from the planar-format
tapes. Naming and map-arrangement conventions
permit the area covered by eaoh segment to be apparent
after printing only the first two lines of any seg
ment. All the segments containing compartment
elevatlm are permanently stored on a mountable disc
from which the files needed for a terrain-correction
job may be readily retrleved,
Data-procassing procdum stin include various
steps and formats that are gradually being streamlined
and improved. The input data are still in card-image
format, and each day's data are separated into individual
data sets preceded by Iead cards providing base,
drift and altirnetry+xWrol information for that day's
data. This card-Image format still conforms to some
earlier systems, whfch limited the station identification,
data-se t dealgna tion, and source-code fields to
four characters each. Preliminary procelng provides
a comparison of the simple Bowper anomalie~ resulting
From use of altimetry and map, bench-mark, or
other elevations. Study of this comparison commonJg
results in small relocations of stations where the aldrnetry
has indicated probable errors In the positions
plotted by the field observer; geographto pOSitl0nS are
also checked by computer plots. men, a final nm of
the prelimfnary reduction program produces a listiw
of station Identifieation, latitude, longitude, and map
elevations formatted for input to the terraincorreotion
program.
The terraln-correction program uses this Hsthg,
as well as files of the l/&minute, 1-minute, and 3-
minute mean-compartment elevations obtained from
the DEM tapes. Corn~ilation of the terrain-data files
is accomp~hed partiy by an interactive selection
program and partly on the Multi- computer text
editor from files stored on a mountable disc after the
initial tape reading. This part of the compilation
procedure is now belng improved, but onae the files
are available, the terrain*orrection procedure Is es
tablished by an Interactive program that requests flle
names, radii of hand corrections, outer radius, densities,
and other parameters. The innewone corrections
may be provided by hand calculations, or the
computer will calculate a complete correction by
assuming that the innermost compartment has a gravitational
atbaation which depends on the station elevation
and the mean elevations of adjacent compartments.
The outer radius is currently limited by the
number of maps that can be handled by the program,
and mast Alaskm corrections have been made to a
radius of only 90 km (outer radius of Hayford zone N),
although a standard 166.7 km (Bayford zone 0) should
be possible in the future. The temain%orrection program
produces an output file of station names and terrain
corrections For two densities, as well es a descriptive
run summary that reports calculation parameters
along with warnings about missing maps and other
plble errors.
A supplementary program merges the output file
with the original data file so that te~aln corrections
are now part of the basic data file. The data can then
be recalculated to obtain a complete Bouguer anomaly,
which uses the altimetry elevation for the slab part of
the correction and the map elevation for the relativeterrain
part of the correction. This combination
generally provides the most contourable data, although
data from days with weather conditions unfavorable
for altlrnetry may conform better if only the map elevations
are used. Maintenance of the basic data as
individual data sets with each day's data, including the
basic base, drlft, and altlmetry control as well as teP
rain correctiona, optimizes reprocesing with other
systems of elevation control. As the file of Alaskan
gravity data gradually improves with increasing station
density, the possibilities fm recognizing errors in
earller data also steadily increase. A Final factor that
facilitates the improvement of elevation contra1 Is
that the data file also includes, for as many stations as
possible, the height difference between the measurement
site and any nearby elevatlon-mference surfaces,
such as rivers. lnkes. sea surface, highway, hilltop, or
other faturf ay eventuallg have a spot elevation
on updat
RENCES CfiED
Bernes, D. F., 1977, Bouguer gravity map of Aiaskax
U.S. Geological Survey Geophysical Investiga-
tions Map GP-913, scale 1;2,500,000.
Plouff, Donald, 1968, Digital terrain correctiona based
on geograpMc coordinate8 labs.]: Geophysics, v.
31, no. 12, p. 1208.