Supplemental_Information:
Procedures used to create or automate data
Overview
The software package developed by Michael Hutchinson at
Australian National University known as ANUDEM version 4.4,
was used to make the DEM. Four types of input data were used for the
production of the DEM: hypsography (land-surface elevation) contours;
hypsography points; hydrography; and points in large depressions.
Following ANUDEM processing, a "fill" procedure (Jenson and Domingue,
1988) was used to remove remaining depressions except for several
large, true depressions (usually playa lakes). Each of the processing
steps is described in detail below.
Hypsography Preprocessing
The U.S. Geological Survey (USGS) 1:100,000-scale Digital Line Graph (DLG) files provided the hypsography data. The DLG files were converted into
ARC/INFO coverage format along with elevation attributes associated
with the contours. Only the elevation and depression contours were
needed. Other features from the DLGs were retained but given an
elevation of 0. ANUDEM accepts contour information only in a
user-specified range of elevations, so the contours with elevations of
zero were excluded from the ANUDEM-gridding process. All the neat
lines from the DLGs were removed. Most DLG files except one
quadrangle (Hugoton, Kansas) included point elevations. The DLG point
elevations were converted into ARC/INFO point coverages, along with
elevation attributes that were associated with the points.
The ANUDEM software package is able to retain depressions
specified by the user. Thus large depressions such as playa lakes
could be retained in the DEM and not removed by the
drainage-enforcement procedure. ANUDEM requires coordinates of a
point in a depression to retain the depression. Small
depressions--less than 3 cells (180 meters) wide--are likely to be
smoothed over by the gridding algorithm, so only depressions larger
than 3 cells were specified for retention. The ARC/INFO GRID
function, SHRINK (Environmental Systems Research Institute, Inc.(ESRI), 1994)
with a distance of 3 cells, (120 meters) was used on the areas enclosed by depression contours to identify depressions large enough to retain.
Points in these large depressions were input to ANUDEM. The elevations associated with these points were set to the elevation of the surrounding depression contours minus half of the contour interval.
After the data were compiled into ARC/INFO format these data were
converted into ASCII files using the UNGENERATE command (ESRI, 1994).
Elevations were included in the hypsography and depression data as
line or point identification numbers. These files were input to
ANUDEM.
Hydrography Preprocessing
The 1:100,000-scale hydrography data were acquired as ARC/INFO
datasets. These data had been separated into 8-digit cataloging
units, then later appended into one dataset. These data were
an early release of the River-Reach File (RF-3) distributed by the
U.S. Environmental Protection Agency (Horn, 1986). Cataloging units
that included any part of Oklahoma were processed.
Four changes were made in the RF-3 dataset before use in
ANUDEM. First, many small water bodies and streams that were not
connected to the main stream network were eliminated. Larger
unconnected streams were retained.
Second, center lines were generated for all large lakes, wide
streams, and other water bodies. The polygons forming the water bodies
were removed. Because the stream center lines were used in the
creation of the DEM rather than water-body polygons, the DEM is not
flat in the areas covered by water. Some contours of the land surface
before reservoir construction were included in the DLGs and were used
with ANUDEM. Because these input data were used, the user is
cautioned that DEM elevations in areas covered by water are not
necessarily reliable.
Third, ANUDEM requires that all hydrographic lines point
downstream, so all lines pointing upstream were flipped. The TRACE
command in the ARCPLOT module of ARC/INFO and the FLIP command in the ARCEDIT module of ARC/INFO (ESRI, 1994) were used to automate this
process for lines forming streams connected with the main stream
network. However, the unconnected lines were flipped interactively if
pointing upstream.
Finally, the RF-3 data were incorrect in several places. Large
parts of several rivers and lakes were missing and two streams were
incorrectly connected at the headwaters. Also, the stream segments
in the area of the Anthon 7.5-minute quadrangle had been shifted
about 1 kilometer to the west. Corrections were made using
data extracted from the USGS 1:100,000-scale hydrography DLG.
Additional information regarding the processing of the RF-3 dataset
is contained in the documentation file for "rf3stick".
ANUDEM and Generating the DEM
The ANUDEM algorithm produces a hydrologically conditioned DEM by
interpolating elevations by using hypsography and hydrography data. ANUDEM
uses a method of drainage enforcement to remove erroneous depressions
from the DEM.
ANUDEM can use both point and contour-line data for hypsography.
Depressions also can be specified for retention using point data.
ANUDEM will remove depressions except those specified for
retention, subject to certain user-specified tolerances. ANUDEM also
uses the hydrography data in the drainage enforcement algorithm. All
hydrography lines point downstream and center lines are included in place of water-body polygons.
The drainage enforcement algorithm "significantly increase[s] the
accuracy, especially in terms of their drainage properties, of digital
elevation models" (Hutchinson, 1989). This algorithm removes
depressions only when drainage conditions contradict input elevation
data by less than a user-specified tolerance.
The interpolation method is implemented by fitting a thin-plate
spline to the data, conditioned by a surface-specific roughness
penalty. Four user-specified tolerances are used to control how the
data are interpolated. The first tolerance specifies the root mean
square residual, and is referred to as the RMS tolerance. The ANUDEM
documentation recommends this tolerance be set at 1.0 for contour
data. A second tolerance is the roughness penalty trade-off
tolerance. The ANUDEM documentation recommends this tolerance be set
to 0 if contour data are being used for input. The third tolerance is
a measure of the elevation accuracy of the data, and is referred to as
the elevation tolerance. "Data points which block drainage by no more
than this tolerance are removed" (Hutchinson, 1989). The ANUDEM
documentation recommends this tolerance be set to half the contour
interval. The fourth tolerance is a measure of total relief. This
tolerance, referred to as the relief tolerance, limits the height of
saddles above depressions that may be exits for these depressions
(Hutchinson, 1989). The ANUDEM documentation recommends this
tolerance be set to the total relief in the input dataset. The tolerances
for this project were set at 1.0 for the RMS tolerance, 0.0 for
the roughness penalty trade-off tolerance, half the contour interval
for the elevation tolerance, and 50 meters for the relief tolerance.
The entire state could not be processed at one time because of
computer storage limitations and contour interval differences. The
ANUDEM processing was done on 11 separate processing blocks. Data from quadrangles adjacent to the state boundary were
used. Different tolerances were used for each processing block according to
the contour interval. The processing blocks overlapped, in most
cases, by 12 kilometers on each side. All input hypsography and
hydrography data were appended and trimmed to cover the areas for each
processing block.
ANUDEM first creates a DEM with a very coarse cell size.
Elevations are interpolated at that cell size and depressions are
removed in an iterative process. When RMS residuals reach the
specified tolerance, the interpolation process begins again, by using a
cell size one-half the previous cell size. The process repeats until
the final user-specified cell size is reached. During each
resolution, ANUDEM applies the drainage-enforcement algorithm to
remove depressions after every five iterations. ANUDEM interpolates
iteratively at each resolution until the user-specified number of
iterations is reached or the user-specified tolerances are overcome
(Hutchinson, 1989). Forty iterations were specified for this project,
but the interpolations always were completed before reaching this
maximum. The DEM produced by ANUDEM contained floating-point
elevations in meters.
Six kilometers were trimmed from the edges of each overlapping processing block after ANUDEM processing to avoid problems introduced by
interpolations near the edges of the input datasets. The elevations
in the remaining overlapping areas were averaged together using a
distance-weighted method. By using the ARC/INFO GRID function MOSAIC,
the processing blocks were combined to create two DEMs: one covering
the Oklahoma Panhandle and the Buffalo quadrangle and the other DEM
covering the state exclusive of the panhandle.
Filling the DEMs
The resulting DEMs contained numerous depressions that had not
been removed by the drainage-enforcement algorithm by using the specified
tolerances. Most depressions in DEMs are errors resulting from the
representation of the surface in raster form (Jenson and Domingue,
1988 and Hutchinson, 1989). The presence of many small depressions
would complicate the process of watershed delineation, so the DEMs
were processed by using the ARC/INFO command FILL in the GRID module, an
implementation of the approach outlined by Jenson and Domingue (1988).
The FILL command fills depressions to pour points at the minimum
elevations along the boundaries of the drainage basins upstream from the
depressions. The identification and removal of depressions is an
iterative process. When a depression is filled the boundaries of the
filled area may create new depressions that will be filled in the next
iteration. The FILL command uses five steps. First, the direction
of flow is determined for each cell. Second, depressions or sinks are
found. Sinks are cells surrounded by cells with higher elevation
values. Third, the drainage basins of these sinks are computed.
Fourth, the depths of the sinks are found. Fifth, the sinks are
filled to the value of the lowest drainage basin boundary cell. This
five-part process is repeated until all sinks are filled (ESRI, 1994).
Cells with values of NODATA were entered at the depression centers to retain large depressions such as playa lakes. The points used to
specify depressions for retention by ANUDEM were converted to NODATA
values in the DEM. The FILL procedure does not fill areas draining
into cells containing values of NODATA. The original elevations were replaced into the cells that had been set to
NODATA after the FILL procedure.
Because of a processing error the DEM, excluding the Oklahoma
Panhandle, the NODATA cells that represented large depressions were not
given a NODATA value prior to the fill procedure. Examination of the
areas covered by filled sinks revealed two large depressions that had
been filled inadvertently. Data values for these areas were extracted
from the unfilled DEM and used to replace the values in the filled DEM
using the MERGE function of the ARC/INFO GRID module (ESRI, 1994).
The FLOWDIRECTION function of the GRID module of ARC/INFO is able to resolve single-cell depressions without filling the depressions. Therefore
single-cell depressions and the large depressions specified for
retention are the only depressions remaining in the DEM after filling.
The retained large depressions are presented in the noncontributing
areas of the "noncontr" dataset.
A problem with the input 1:100,000 hydrography data was discovered
at this processing stage. The RF-3 hydrography data in the area
covered by the Anthon 7.5-minute quadrangle were shifted from the
true positions to about 1 kilometer to the west. As a result,
many streams in the area were drawn along topographic ridges
instead of valleys. These errors in the input data resulted in
substantial errors in the DEM. The erroneous data were removed from
"/rf3stick" dataset and the correct data were inserted from USGS
1:100,000 DLG hydrography data. All the hypsography and hydrography
data for several kilometers around this area were extracted from the
datasets and input to ANUDEM. The resulting DEM was trimmed by
several cells to eliminate any edge problems. An area of NODATA
smaller than this corrected DEM was made in the large unfilled DEM.
These two DEMs were then combined by using the MOSAIC function of the
ARC/INFOs GRID module. An area larger than the area that had been
corrected with ANUDEM was processed as before with FILL. The
resulting DEM was then trimmed to a polygon that avoided all filled
depressions and combined with the filled DEM of the state using the
MERGE function of the ARC/INFOs GRID module (ESRI, 1994). MERGE combined the DEMs, giving precedence to the corrected values in the area covered by the polygon, that resulted in a revised filled DEM for the
state excluding the panhandle.
The DEM for the panhandle and Buffalo quadrangle and the DEM for
the state excluding the panhandle were then trimmed to have a
12-kilometer overlap around the southern and eastern boundaries of the
Buffalo quadrangle (fig. 1). The Oklahoma Panhandle and Buffalo quadrangle DEMs were combined by using the MOSAIC function to create a seamless statewide DEM.
The direction of overland surface-water flow for each cell (flow
direction) was computed from the statewide floating-point DEM.
The floating-point DEM requires about twice the disk
space as an integer version, for this reason, the distribution of the DEM was rounded to the nearest meter. This was accomplished by adding 0.5 meter to
every elevation and then truncating the result to integer meters.
Related spatial and tabular datasets and programs
okflowdr -- flow direction of okdem
okflowac -- reclassed flow accumulation of okdem
rf3stick -- downstream-directed stick hydrography
watershd -- Oklahoma watersheds and 8-digit hydrologic units for
the remainder of the Arkansas, Red, and White River
Basins
noncontr -- noncontributing areas in Oklahoma
References cited
Environmental Systems Research Institute, Inc. (ESRI), 1994,
Cell-based modeling with GRID 7.0.2--Hydrologic and distance
modeling tools, ARC/INFO On-line manuals: Redlands, CA.
Horn, R.C., 1986, Reach File Manual: U.S. Environmental Protection
Agency, 40p.
Hutchinson, M.F., 1989, A new procedure for gridding elevation
and stream data with automatic removal of spurious pits: Journal
of Hydrology, v.106, p. 211-232.
Jenson, S.K. and Domingue, J.O., 1988, Software tools to extract
topographic structure from digital elevation data for geographic
information system analysis: Photogrammetric Engineering and
Remote Sensing, v. 54, no. 11, p. 1,593-1,600.
National Geodetic Survey (NGS), May 1995, Data Sheets, South
Central U.S., 1 CD-ROM.
U.S. Geological Survey, 1990, Digital Elevation Models, National
Mapping Program Technical Instructions Data Users Guide 5, 51 p.
U.S. Geological Survey, 1989, Digital Line Graphs from
1:100,000-scale Maps, National Mapping Program Technical
Instructions Data Users Guide 2, 88 p.
Notes
Any use of trade, product, or firm names is for descriptive
purposes only and does not imply endorsement by the U.S. Government.
Although this Federal Geographic Data Committee-compliant metadata
file is intended to document the "okdem.bil" file, this metadata file
was developed by using the DOCUMENT program with ARC/INFO. Therefore,
this file may include some ARC/INFO-specific terminology. Users are
cautioned not to be confused by this terminology. This metadata file
should contain enough information to eliminate any confusion caused by
the use of ARC/INFO-specific terminology.