THE NATIONAL GEODETIC SURVEY GRAVITY DATA BASE Summary In July 1978 The National Geodetic Survey (NGS) assembled a gravity data base covering the continental U.S., and Canada and the bordering continental shelf areas. At NGS this data base is used to compute the local geoid and the gravimetric deflections. Since 1978 gravity data from other sources have been added bringing the data base to a total of 1,677,370 point values. However, this version of the data base contains somewhat fewer point values because the proprietary data have been deleted. Considerable effort has been expended in editing these data in order that high quality geodetic quantities may be computed. INITIAL NGS GRAVITY DATA BASE In July 1978 the NGS completed the assembly of a gravity data base for use in computing the gravimetrically derived geodetic quantities. The data base included land gravity from the continental U.S. and Canada and sea gravity from the surrounding continental shelf area. The initial NGS gravity data base was compiled by the transcription of gravity files from the following sources (Fury, 1987): 1. Defense Mapping Agency Aerospace Center (DMAAC) 2. National Geophysical and Solar Terrestrial Data Center (NGSDC) 3. Department of Energy, Mines and Resources of the Government of Canada (CANADA) 4. Brown Geophysical Research Corporation (BROWN) 5. Georgia Institute of Technology (GEORGIA) The format, coded symbols, and contents of the record from the above sources were quite different, necessitating the data consolidation into a uniform, shortened data format. There were 1,125,138 point values in this original version of the gravity data base. In some cases it was necessary to reconstruct the gravity observation from the free air anomaly. The basic aim in the transcription of data records was to assure that every record in the data base was complete with respect to: 1. Position (latitude, longitude), 2. Observed gravity, 3. Free air anomaly and its standard error, 4. Bouguer anomaly and its standard error and 5. Observation type code. When free air or Bouguer anomalies were not contained in the input data records, they were calculated according to the formula given in the definitions. The Bouguer anomaly over ocean areas was set to zero unless water depth was indicated. Standard errors were estimated for the data types when they were not provided. The observed gravity was calculated from free air anomaly when not present in the input record. Three editing functions were performed in data transcription: 1. Error detection, 2. Reformatting and 3. Tabulation of statistical summaries. Transcription of 'DMAAC' data The first data set obtained from DMAAC (Hauer, 1975) contained 342,474 observations over the continental U.S. A reedited version of this data set, also received from DMAAC, contained 375,565 records. The distribution of observations by observation type code were: ----------------------------------------------------------------- | 1 |2 |3 |4 |5 | 6 | 7 | 8 |9 |A|B|C|D| ----------------------------------------------------------------- First data set |337847|- |- |- |- |2588|1414|464|161|-|-|-|-| Reedited data set |375565|- |- |- |- |- |- |- |- |-|-|-|-| ----------------------------------------------------------------- Observation Type - DMAAC Data The reedited data set was adopted as the standard although it contained type 1 observations only. The first data set was then matched against the reedited one with the following results: ___________________________________________ |Matched | Added | Added | Merged | |Surveys | Surveys| Records | DMAAC File | ------------------------------------------- | 532 | 14 | 6756 | 382321 | ------------------------------------------- Merged DMAAC Data Set The following editing functions were performed in data transcription: 1. The type codes were checked for validity. 2. The latitude and longitude were checked for correct sign. 3. The geographical units (degrees and minutes) and elevation units (meters) were verified. 4. It was ascertained that observed gravity was present in the record. 5. Elevations were checked for realistic values (h<4500m). 6. When the free air anomaly was not coded, both the free air anomaly and Bouguer anomaly were calculated. 7. When the Bouguer anomaly was not coded, it was calculated. There were no data records deleted from the files in the editing process, however 2,075 free air and/or Bouguer anomalies were recalculated to assure data record completeness. All anomaly values were referenced to the International Gravity Standardization Network, 1971 (IGSN 1971) (see definitions). Record Content Of 'DMAAC' Gravity File Col. Description 1 Security class code - unclassified (U) 2 Security control - unlimited dissemination (blank) 3 Geographical units code (blank = degrees & minutes) 4-10 Latitude xx xx.xx (deg min) (+ north latitude) 12-19 Longitude xxx xx.xx (deg min) (+ east longitude) 21 Gravity obs. type code (see Anomaly Computation Chart) 22 Elevation units (blank = meters) 23-29 Elevation (h) xxxxxx.x 31-35 Elevation (d) xxxx.x 37-42 Observed gravity (less 976000mgal) xxxx.x mgal 44-48 Free air anomaly xxx.x mgal 50-54 Bouguer anomaly xxx.x mgal 58-61 Source number ... (survey code) 62-75 ----- Data not processed ----- 76-77 Standard error of free air anomaly xx mgal 79-80 Standard error of Bouguer anomaly xx mgal Transcription Of 'NGSDC' Data The first data set received from NGSDC included the following NGSDC files (Marine Geophysical Data Catalog, 1975): -------------------------------- | File No. | No. of Records | --------------------------------- | 00005 | 20543 | |00076 & 00077| 145970 | | 00078 | 72009 | | 00106 | 11762 | | 00130 | 81228 | --------------------------------- | Total | 331512 | --------------------------------- NGSDC Gravity Data Files A later acquisition of marine gravity data superseded files 00076, 00077, 00078 and included the new file 00098 containing 433,831 records, all observed in coastal waters. The following editing functions were performed in data transcription: 1. It was ascertained that free air anomaly is present in the record. 2. The latitude and longitude were checked for correct sign. 3. The elevation type was set to 3. 4. The observed gravity was calculated from the free air anomaly. 5. The Bouguer anomaly was calculated when the water depth was given; otherwise, it was set to zero. Since the NGSDC files included a collection of geophysical data (i.e., depth soundings, magnetic intensity, etc.), not all records contained a gravity observation. Therefore, 117,638 records were not transcribed from the files, leaving 429,726 marine gravity observations for data base entry. The estimated standard error of observation was set as follows: ------------------------------------------ | File No. | Estimated | | | Std. Err. | ------------------------------------------ | 00005, 00098, 00106, 00130 | 7.0 mgal | ------------------------------------------ | 00076, 00077, 00078 | 3.5 mgal | ------------------------------------------ Estimated Standard Error of Gravity Observation The file numbers were substituted as the survey numbers in the data base records. Record Content of 'NGSDC' Gravity Files col. Description 1-13 ------ Data not processed ------ 14-15 Last two digits of observation year 16-17 Month 18-25 ------ Data not processed ------ 26-33 Latitude in decimal degrees xxxx xxxx 34-42 Longitude in decimal degrees xxxxx xxxx 43-57 ------ Data not processed ------ 58-62 Depth in corrected meters 63-75 ------ Data not processed ------ 76-80 Free air gravity anomaly xxxx x Transcription of 'CANADIAN' Data The data set covering Canada (Winter, 1976) contained 272,514 gravity observations. The original observations were not a homogeneous set; referring to different datums and coordinate systems however, the Earth Physics Branch transformed the data into a common reference system. The data set fell into two categories by date: 1. For data released prior to May 15, 1974, the theoretical gravity was computed on the International Ellipsoid 1930 (see definitions): The observed gravity was given in the Potsdam system and was based on a value of 980.622 gals at the National Reference Pier in Ottawa. 2. For data released after May 15, 1974, the theoretical gravity was computed on the Geodetic Reference System 1967 (see definitions). The observed gravity was based on the 1974 adjustment of the National Gravity Net which is consistent with the International Gravity Standardization Net 1971 (IGSN-71) (see definitions). This absolute system differs from the Potsdam system in Canada by 14 to 16 mgals as a function of latitude. The observed gravity was obtained by back-solving from the free air anomaly. The Bouguer anomaly was then computed using standard density values (see definitions for formulae). The data released prior to 1974 were transformed into the common reference system of the second category (IGSN-71) by the Earth Physics Branch using the approximate formula: gnew = gold - 0.95 - 13.6 * sin(lat)**2 mgals The records did not contain estimated standard errors for calculated anomalies, but did provide accuracy factors for elevations as a function of year of observation. The following tables give the estimated standard errors of anomalies entered into the NGS gravity data base based upon these provided elevation accuracies. --------------------------------------------------------------| Code | 0=unknown | 1=- 3 ft | 2=-10 ft | 3=-25 ft | 4=-100 ft | --------------------------------------------------------------| gf | 2.3 | 0.3 | 0.9 | 2.3 | 9.4 | --------------------------------------------------------------| gb | 1.5 | 0.2 | 0.6 | 1.5 | 6.0 | --------------------------------------------------------------- Accuracies For Observations Before 1969 --------------------------------------------------------------| Code | 0=unk |1=-0.1ft|2=-1ft|3=-3ft|4=-10ft|5=-25ft|6=-100ft| --------------------------------------------------------------| gf | 2.3 | 0.1 | 0.1 | 0.3 | 0.9 | 2.3 | 9.4 | --------------------------------------------------------------| gb | 1.5 | 0.1 | 0.1 | 0.2 | 0.6 | 1.5 | 6.0 | --------------------------------------------------------------- Accuracies For Observations After 1968 The records contained terrain corrections (C ), but the estimated standard errors were specified by uncertainty codes which indicated the percentage errors of the corrections. The following table shows the coded uncertainties: ------------------------------------------------------ | Uncertainty Code | 1=10% | 2=20% | 3=30% | 5=50% | ------------------------------------------------------ | Percent Factor (P) | 0.1 | 0.2 | 0.3 | 0.5 | ------------------------------------------------------ Uncertainty Of The Terrain Corrections The estimated standard error of terrain correction was then calculated by: stderrtc = P * tc Observation types were not indicated, but the water depths of the observation were coded. The water depth code was transformed into the observation type code as follows: --------------------------------------------------- | Depth Type Code | blank | W | U | I | X | --------------------------------------------------- | Observation Type| 1 | 3 | 5 | C | X | --------------------------------------------------- Depth Type Code - Observation Type Code Equivalence The following editing functions were performed in data transcription: 1. Type codes were checked for validity. 2. The latitudes and longitudes were checked for correct signs. 3. It was ascertained that observed gravity was present in the record. 4. Elevations (h) and water depth (d) were converted to meters and checked for realistic values (h<=4500m, d<=5500m). 5. When the free air anomaly was not coded, both the free air anomaly and Bouguer anomaly were calculated according to the Anomaly Computation Chart (see appendix). 6. When the Bouguer anomaly was not coded, it was calculated except when the type code was X. 7. Since the coded Bouguer anomaly included the terrain correction, the latter was subtracted from the former. The survey code was assembled by concatenating the two-digit year of observation with the three-digit project number. The editing process deleted two records from transcription; it found 4,318 records which had non-coded Bouguer anomalies, but recalculated 156 values only, the difference being X-type records with unknown water depths. The distribution of records by observation type codes are: -------------------------------------------------- | Type | 1 | 3 | 5 | C | X | -------------------------------------------------- | Records| 100204| 162855| 4968 | 323 | 4162 | -------------------------------------------------- Observation Type Codes Record Content of 'CANADIAN' Gravity File Col. Description -------------------------------------------------------------- 1-6 ---- Not processed ---- 7-8 Year of observation 9-11 Project number 13-20 Latitude in decimal degrees 22-30 Longitude in decimal degrees 31 ---- Not processed ---- 32-38 Elevation of station in feet above MSL 40 Elevation accuracy code 41 ---- Not processed ---- 42-48 Depth of water or thickness of ice in feet 49 Depth type code 50 Depth accuracy code 51-59 Observed gravity in gals 60-66 Free air anomaly in mgals 67-73 Bouguer anomaly in mgals 74-78 Terrain correction in mgals 79 Terrain correction uncertainty code 80 ---- Not processed ---- Transcription of 'BROWN' data One file of gravity observations was obtained from the Brown Geophysical Research Corporation containing 37,658 type 1 records. The following editing functions were performed in the data transcription process: 1. The latitudes and longitudes were checked for correct sign. 2. It was ascertained that observed gravity (g) was present in the record. 3. Elevations were checked for realistic values (h<=4500). 4. When the free air anomaly was not coded, both the free air and Bouguer anomalies were calculated. 5. When the Bouguer anomaly was not coded, it was calculated. There were no data records deleted from the file in the editing process, but 448 free air and/or Bouguer anomalies were recalculated. The project code has been substituted for the survey number. Record Content of 'BROWN' Gravity File Col. Description 1-3 Latitude in degrees (xxx) 4-7 Latitude in minutes (xx xx) 8-10 Longitude in degrees (xxx) 11-14 Longitude in minutes (xx xx) 15 Sign of longitude (negative west) 16-22 Topographic elevation ( xxxxx x) 23-30 Observed gravity mgals (xxxxxx xx) 31-35 Free air anomaly mgals ( xxx x) 36-40 Bouguer anomaly mgals ( xxx x) 41-58 --------Not processed------------ 59-62 Project number 63-80 --------Not processed------------ Transcription of 'GEORGIA' data One data set of gravity observations was obtained from the Georgia Institute of Technology containing 6,102 punched cards. The following editing functions were performed in data transcription: 1. The latitudes and longitudes were checked for correct sign. 2. It was ascertained the observed gravity (g) is present in the record. 3. Elevations were checked for realistic values (h<=4500m). 4. When the estimated standard error of free air anomaly was not given or the observed gravity represented ocean data (code=5) the estimated error was set to 1 mgal. 5. The estimated standard error of Bouguer anomaly was set equal to the estimated standard error of free air anomaly. 6. When the free air anomaly was not coded, both the free air and Bouguer anomalies were calculated. 7. When the Bouguer anomaly was not coded, it was calculated. 8. Alphabetic characters of the project code were changed to zero. Since the observations were referenced to the Potsdam system, the following correction was added to the observed gravity and anomalies: dg = 3.5 - 13.6 * sin(lat)**2 mgals The project code was substituted for the survey number. The transcription process deleted 3 records due to wrong signs in position, and 493 records for non-coded observations, leaving 5,606 records for data bank entry. In addition, 8 free air and/or Bouguer anomalies were recalculated. Record Content of 'GEORGIA' Gravity File Col. Description 1-3 ------------Not processed------------ 4-6 Latitude degrees (xxx) 7-10 Latitude minutes (xx xx) 12-15 Longitude degrees (xxx) 16-19 Longitude minutes (xx xx) 21 Elevation type 23-29 Topographic elevation (xxxxxx x) 37-43 Observed gravity (less 976000) mgals 44-49 Free air anomaly xxxx xx mgals 50-55 Bouguer anomaly xxxx xx mgals 58-61 Project number 62-76 ------------Not processed------------ 77 Estimated standard error of free air anomaly ( x) mgal ADDITIONS TO THE NGS GRAVITY DATA BASE Since 1978 gravity data from the following sources have been added to the NGS gravity data base: 1. U.S. Geological Survey 2. Los Alamos Scientific Laboratories 3. Purdue University 4. Hawaii Institute of Geophysics 5. Tetra Technologies Inc. 6. University of Texas at Dallas 7. Kansas Geological Survey As with the data from the previous sources the data were checked and then transcribed to conform to the NGS data format. Terrain corrections were calculated at all the points in the data base using the technique of numerical integration of flat top prisims of topography. The topographic elevations were obtained from the digital terrain model data bank of NGS which contains a point value at intersections of a geographic grid distributed at 30 arc seconds interval. The estimated accuracy of these elevations is ñ50 meters at the 0ne-sigma level. The terrain correction rounded to the nearest 0.1 mgal was added to the data base record. Considerable effort has been expended in editing these data in order that high quality geodetic quantities may be computed. Examples of this data base editing are: 1. duplicate data sets were searched for and deleted, 2. projects adjusted to local datums were shifted to the absolute datum 3. the Bouguer anomalies in the entire data base were contoured as a means of detecting data outliers and 4. the validity of the terrain corrections were checked by comparing their magnitude to the variability of topography (Fury, 1988). The gravity anomalies were then recomputed on the WGS 84 reference system which includes a correction for the atmospheric mass (WGS 84, 1987). The NGS gravity data base currently contains a total of 1,677,370 point values. Less than one-percent of these data are proprietary data and have been deleted from the data set accompanying this document. THE NGS GRAVITY DATA BASE RECORD FORMAT col. Record Content (see definitions) 1-8 Latitude (degrees * 10**(-5)) 9-17 Longitude (degrees * 10**(-5)) 18-23 Height (meters * 10**(-1)) 24-31 Observed gravity (mgals * 10**(-3)) 32-37 Depth (meters * 10**(-1)) 38-41 Sigma, observed gravity (mgals * 10**(-1)) 42-46 Terrain correction (mgals * 10**(-1)) 47-49 Sigma, terrain correction (mgals * 10**(-1)) 50-54 Survey code 55 Observation type 56-57 Observing agency 58 Edit code 59-64 Free air anomaly (mgals * 10**(-1)) 65-67 Sigma, free air anomaly (mgals * 10**(-1)) 68-73 Bouguer anomaly (mgals * 10**(-1)) 74-76 Sigma, Bouguer anomaly (mgals * 10**(-1)) 77-101 Station name DEFINITIONS observed gravity: The value of the gravitational attraction at the observation site (defined by lat,lon,hth coordinates). The data base value is the observed gravity less 978000.0 mgals to reduce the amount of storage required. sigma (gravity): The standard error (or estimated error) of the observed gravity. If it is an NGS observation then the standard error of observation is given. If the observation is by another agency and no error information is available then the observation error is estimated to be 1.0 mgals. When 0 or a value >10.0 is entered, low accuracy data is indicated. depth: The depth below ground level at which the gravity observation was made in a land gravity observation or the depth below the surface at which the observation was made in a sea gravity observation depending upon the observation code. terrain correction: The correction (to be added) to the gravity anomaly to correct for the terrain height variation in the vicinity of the gravity observation. Terrain corrections were calculated at NGS using the numerical integration of topographic prisms effects using a high density grid within a radius of 5 min. and a lesser density grid within a radius of 30 min. from the gravity station. These calculations were performed only at limited areas of the continental U.S. where the terrain effects were judged to be significant (e.g. areas of high topography), the correction term was set to zero elsewhere. sigma (terrain correction): The standard error of the terrain correction was estimated at 1.0 mgal where the terrain effects was calculated, otherwise it was set to zero. survey code: The survey identification code assigned by the submitting agency. These codes may not be entirely unique as identifiers, sometimes, in subsequent transmittals, duplicates were accumulated in the data bank; however efforts were made to identify duplicate data and remove it from the data base. observation type: The gravity observation type code (see Appendix). observing agency: The Federal, state or private agency that contributed the gravity data. The following codes are presently in use: 1. Defense Mapping Agency 2. National Geophysical and Solar Data Center; Boulder, CO 3. Geological Survey of Canada 4. Brown Geophysical Research Corporation 5. Georgia Institute of Technology 6. National Geodetic Survey 7. U.S. Geological Survey 8. Los Alamos Scientific Laboratory 9. Purdue University 10. Hawaii Institute of Geophysics 11. Tetra Technologies Inc. 12. University of Texas at Dallas 13. Kansas Geological Survey edit code: The checked/not checked status of the gravity observation. The following codes are presently in use: 0. not processed 1. preliminary rejection 2. finally rejected 3. accepted after quality check 4. corrected and accepted latitude: The approximate latitude of the gravity observation in decimal degrees; usually scaled to the nearest minute from USGS topographic maps. longitude: The approximate longitude of the gravity observation in decimal degrees; usually scaled to the nearest minute from USGS topographic maps. height: The elevation at which the observation in a land gravity observation was made or the depth of the ocean in a sea gravity observation depending upon the observation type. On land the topographic elevation is a scaled height from maps; even when the gravity measurement was done at a benchmark the elevation was recorded to 0.1 meter precision only (there is no guarantee that the benchmark height was entered into the data record. free air anomaly: The gravity anomaly (observed gravity - theoretical gravity) reduced to sea level by applying the normal gradient of gravity (0.3086 mgal/meter). sigma (free air anomaly): The estimated error of the free air anomaly. Bouguer anomaly: The free air gravity anomaly corrected for the Bouguer slab of standard density of 2.67 gr/cc. sigma (Bouguer anomaly): The estimated error of the Bouguer gravity anomaly. station name: When gravity is observed on a monumented station, the name of the station; otherwise, blank. The formulae of anomaly calculation: (Free air) fanom = gobs-(gthr-(dg/dz)*h)-((dg/dz)-4*pi*G*rhoW)*dW (Bouguer) Banom = gobs-(gthr-(dg/dz)*h)-2*pi*G*(rhoC*h -(rhoC-rhoW)*dW+(rhoC-rhoI)*dI) -((dg/dz)-4*pi*G*rhoW)*dW where: dg/dz is the vertical gravity gradient (0.3086 mgal/meter) (0.09406 mgal/foot) rhoC is the density of crustal rock (2.67 gr/cc) rhoW is the density of water (1.000 gr/cc - fresh water) (1.027 gr/cc - salt water) rhoI is the density of ice (0.917 gr/cc) dW, dI are the depth of water and thickness of ice respectively The Theoretical Gravity Formula 1930 (IUGG): gthr = 978.049 * (1.0+0.0052884*sin(lat)**2 -0.0000059*sin(2*lat)**2) gals The Theoretical Gravity Formula 1967 (GRS): gthr = 978031.85 * (1.0+0.005278895*sin(lat)**2 +0.000023462*sin(lat)**4) mgals The Theoretical Gravity Formula 1980 (GRS): gthr = 978032.67715 * (1.0+0.0052790414*sin(lat)**2 +0.0000232718*sin(lat)**4 +0.0000001262*sin(lat)**6 +).0000000007*sin(lat)**8) mgals REFERENCES Fury, R.J., Format and contents of the Unified Gravity Data Bank, NGS Memorandum, July, 1978. Fury, R.J., Completion of quality control and compression of working files of the observed gravity data bank, NGS Memorandum, Nov., 1988. Hauer, E.J., Private communication, Department of Defense, Defense Mapping Agency Aerospace Center, Gravity Services Division, Nov. 10, 1975. Marine Geophysical Data Catalog - 1975, Key to Geophysical Records Documentation No. 4, U.S. Department of Commerce, NOAA-EDS/NGSDC, Boulder, Colorado. WGS 84, WGS 84 ellipsoidal gravity formula and gravity anomaly conversion equations, DMAAC Memorandum, Aug., 1987. Winter, P.J., Private communication, Department of Energy, Mines and Resources, Gravity Division, Earth Physics Branch, Ottawa, Canada, Jan. 19, 1976. APPENDIX