Clark FNOC Elevation, Terrain, and Surface Characteristics

Principal Investigators:

Summary:

Primary References:

Clark FNOC Elevation, Terrain, and Surface Characteristics
 

• Integrated dataset
• Design
• Source
• Additional references
• File lists

Dataset Description

INTEGRATED DATA­SET

Data­Set Citation:

Projection:

Spatial Representation:

Temporal Representation:

Data Representation:

Layers and Attributes:

Dataset Description

DESIGN

Variables:

Origin:

Geographic Reference:

Geographic Coverage: Global

Geographic Sampling:

Time Period:

Temporal Sampling:

Dataset Description

SOURCE

Source Data Citation:

Contributor:

U.S. Navy, Fleet Numerical Oceanographic Center

Monterey, CA 93943 USA

Distributor:

Date of Production:

Lineage & Contacts:

Dataset Description

ADDITIONAL REFERENCES

Dataset Description

FILE LISTS

• Dataset Files
• Reprint Files

Clark FNOC Elevation, Terrain, and Surface Characteristics
 

• Elevation
• Primary and Secondary Surface Type
• Ocean/Land Mask
• Direction of Ridges
• Water and Urban Cover

Elevation

Description:

 Structure:

 Series:

 System Files:

 Notes:

 Primary and Secondary Surface Type

Description:

 Structure:

 Series:

 System Files:

 Notes:

 Ocean/Land Mask

Description:

 Structure:

 Series:

 System Files:

 Notes:

 Direction of Ridges

Description:

 Structure:

 Series:

 System Files:

 Notes:

Water and Urban Cover

Description:

 Structure:

 Series:

 System Files:

 Notes:

Clark FNOC Elevation, Terrain, and Surface Characteristics
 

John J. Kineman

National Geophysical Data Center

Boulder, CO

SUMMARY

The Navy Fleet Numerical Oceanography Center (NFNOC) began creating the original 10-minute terrain data set in the mid 1960's. Work extended into the early 1970's. The main sources for the data were the U.S. Department of Defense Operational Navigation Charts (ONC), scale 1:1,000,000. For certain regions ONC charts were not available; for such areas selected charts from the Jet Navigation Charts and World Aeronautical Charts were used. The charts were hand read out to forms by employees, and then read by optical character reader to tape. The values were estimates from contour lines. Isometric graphs were made for quality control, checking terrain features. Later, other errors were corrected by the National Center for Atmospheric Research in Boulder, Colorado.

Source data from NCAR were stored in binary compressed format. This was converted at NGDC to an ASCII format for products distributed on tape with latitude/longitude encoding, grouped in 5-degree squares. A series of quality checks were performed on the data in 1985 to test for internal consistency between the various parameters. Numerous errors were noted, especially in the southern hemisphere. Some of the errors could be corrected by comparison between data layers. Others were flagged with a bad-data code. Errors that could not be detected by comparing the fields were not tested for.

These data were later converted to a raster data file by mapping the lat/long coordinates into the corresponding 10-minute grid and separating the variables into individual spatial layers. The resulting files are provided here, with the addition of an "ocean mask" file produced from a combination of variables in the FNOC data-set. This land/ocean mask is provided for convenience in masking other data for display, but also for use in evaluating results of ocean masking in the Olson data-sets (Chapter A05).

Users should take special caution in using these data, as many errors have been noted. The most commonly used portion of the data-set is the elevation values, and there has been progress in improving the other values. New source tapes were obtained in the hopes that some of the errors were introduced during it's long lineage and could be removed by comparison with the originals; however this approach has not proven to be feasible due to uncertainties in the source tapes.

Several efforts are underway to obtain improved topography and to derive corrections from other data. An improved version of the 10-minute data is expected for the next CD-ROM (Disc B).

COMPILATION OF EXISTING DOCUMENTATION

BACKGROUND

The 10-minute Terrain data set distributed by NGDC was originally produced by the U.S. Navy, Fleet Numerical Oceanography Center in Monterey, CA. The data were then transferred to the National Center for Atmospheric Research (NCAR) in Boulder, CO (Dennis Josephs), where some corrections were made in the elevation values. NCAR was then the source of NGDC's version. Various quality checks were run on the data at NGDC in 1986, revealing numerous errors (about 16,000), mostly in the Southern Hemisphere and mostly in the attribute fields. A second version of the data set was then sent from NCAR. The new version had fewer errors (about 9,000), still mostly in the attribute values for the Southern Hemisphere. These errors resulted from 10 automated tests, as listed below:

1) Range (from max and min) excludes mode
2) Sea height and type mismatch (between elevation and primary or secondary type code)
3) Ridge azimuth out of limit (i.e. < 0 or > 18)
4) Number of Ridges out of limits
5) Secondary type undefined
6) Primary type undefined
7) % water out of limits
8) % urban out of limits
9) Max/min out of limit (exceeds lowest and highest known elevations on earth).
10) Sea type-code discrepancy (between primary and secondary type codes)

* summaries of these error checks are available from NGDC

In the case of elevation data, corrections were made based on other data sources, but in the case of the attribute fields, data errors that had no obvious correction were flagged (all bits on = nine's in the NGDC ASCII format, or 255 in the current one-byte integer format) to indicate bad data.

Attempts were made to retrieve original or near-original copies of the data set to see if errors may have been introduced during processing at one or more locations. After several years, an "original" tape emerged from Mr. Leo Clark at the Navy FNOC. Other copies were also obtained from various sources. NGDC will continue to evaluate the various versions, and plans to produce an up-dated version the 10-minute data set using all the latest information and reference to other data. Meanwhile, we are convinced that the data represented here, although known to have numerous errors, is the "best available" version for now.

The appendices (A and B) contain information from existing documentation of the previous (source) versions of the data set, eliminating tape and other format descriptions that are irrelevant to the current data structure and format (described elsewhere in this manual).

DESCRIPTION OF VARIABLES:

Significant Ridges:

Subjective estimate of number of ridges and their orientation in tens of degrees (00-18).

Terrain Elevation:

Elevation in meters (converted from original elevation in 100's of feet, rounded to the nearest 30 meters). Ocean are coded as zero elevation, however not all zero values indicate ocean. Refer to the Primary or Secondary terrain characteristics (or the special Ocean mask, which was created from the Primary terrain characteristics code 62). Inland water bodies are coded with the elevation of the water surface (except in minimum field where it is always zero), except when the water body is below sea level, in which case the surface elevation is used (another exception to this is the Caspian Sea, which is coded as zero).

Characteristics of the Terrain (Primary and Secondary):

0 : salt or lake bed.
1 : flat or relatively flat.
2 : desert (or, for latitudes greater than 70N, glaciers or permanent ice).
3 : marsh.
4 : lake country or atoll.
5 : major valleys or river beds.
6 : isolated mountains, ridge or peak.
7 : low mountains or hills.
8 : average mountains.
9 : extremely rugged mountains.
[31] : Ocean (used in Secondary type only)
[62] : Ocean (used in Primary type only)

Percentage of Water:

For ocean areas at sea level the value is 100, for all other areas the range is 00 to 99 (large lakes or inland seas will not be coded as 100).

Percentage of Urban Development:

Not updated (reflects highly subjective judgments from the maps used).

APPENDIX A

EXCERPTS FROM NCAR DOCUMENTATION PACKAGE:

"Data Format for Global 10­Minute Elevation Data from the U.S. Navy"

Dennis Joseph

NCAR, Data Support Section

April 1982, update Dec 1984

Global elevation data at a resolution of 10-minutes were prepared by the Navy Fleet Numerical Oceanography Center at Monterey. For each 10X10 minute area, the set includes modal elevation, minimum elevation, maximum elevation, orientation of ridges, terrain characteristics, and urban development. This is archived by the NCAR Data Support Section (DSS) in a packed binary format. Parameters available are identical to those described in documents by Meteorological International, Inc. and the Fleet Numerical Oceanography Center, but the DSS has made some [format] changes to the set. The information content of the original set has been preserved entirely. Each 64 bit group [in the NCAR/DSS data set] has the following format:

Bits Code Description

1-6 RR Estimate of the number of significant ridges
7-12 DD General direction of ridges
13-21 HMO Terrain elevation - Modal height.
22-30 HHI Terrain elevation - Maximum height.
31-39 NLO [HLO] Terrain elevation - Minimum height.
40-45 C1 Primary characteristics of terrain.
46-50 C2 Secondary characteristics of the terrain.
51-57 WWW Percentage of water surface.
58-64 URB Percentage of urban development.

General Information:

The modal terrain height has been contoured at NCAR, and major problems identified in these plots have been corrected by the Navy. Distribution summaries of all parameters indicate that there are still some invalid elevation values and unexplained code values, especially in the terrain characteristics field. As of this data no further information is available on these problems. Occasional occurrences of full range values (all bits on) are assumed to indicate missing data.

The data distribution summaries showed a strong tendency for elevation values to cluster around multiples of 500 feet. This is probably due to the contour intervals in the original maps or to some other characteristic in the method of reading map values.

The true resolution of the data is reduced to 20 minutes poleward of 70 degrees latitude, but data values are still present for each 10 minute [cell]. More information on the original data format, the sources of the data, the methods of reading the data, and the routines designed for the Navy to read the original format is available in documentation by the Fleet Numerical Oceanography Center and Meteorology International Inc.

General Information Update - DEC 1984:

Various users have noted a large number of bad data points in the minimum elevations. A few bad points in the modal and maximum elevations have also been identified. An attempt has been made to remove these bad points and replace them with estimated values. The minimum values from a previous edition of the data were found to have many fewer problems and these values were used for all minimum elevations north of 30 South. Checks for unreasonable values and gradients were run and comparisons of min, mode, and max were made. The results of these tests were manually inspected and were estimates seemed better than the original values, they were inserted in the set. Checks were run on the minimum, modal, and maximum elevation only. No checks were run on the other parameters.

There are most likely still some erroneous values in the set, but most of the totally unreasonable values have been removed. Note that the minimum elevations are coded as zero for all water surfaces regardless of the true elevation of the water surface (even when this surface is below sea level). In some areas the elevation values are constant over one degree areas indicating that the resolution is not truly 10 minutes in those areas. In general, the modal elevations seem to be more reliable than the minimum or maximum.

This corrected set will be the primary archive set, and the uncorrected earlier versions are available on request.

Roughness computations:

Stephano Tibaldi, European Center for Medium Range Weather Forecasting, has used these elevations to compute estimates of surface roughness. His method for computing roughness length (Z) over a user-defined area containing multiple 10' [cells] is given on the following page. Note that his relative maxima are determined by examining the 8 surrounding [cells]. When looking at data which are poleward of 70 degrees, use every other point to compensate for true resolution of 20'.