Sea Surface Height - Along-Track

The TOPEX Along-Track residuals were developed by Victor Zlotnicki and Akiko Hayashi of Ocean ESIP, a member of the Earth Science Information Partner Federation, under contract to NASA.

DATA ACCESS

The along-track residual global data and read software are available at the PO.DAAC ftp site

Subsets and Plots of the along-track residual data can be accessed through the POET GUI.

Last cycle available: 364
Temporal extent: Sep 23, 1992 - Aug 11, 2002

DATASET OVERVIEW TOPEX/Poseidon

Sea surface heights in millimeters relative to 1993-1996 average
Along-track grid: fixed locations for all cycles, 6.2 km spacing along-track.
Includes shallow water points and an identifying shallow water flag.

The TOPEX/Poseidon data are based on the T/P MGDR-B released by NASA and CNES in 1997, as described in Benada, 1997; the NASA/JPL and CNES/AVISO versions of the MGDR have small differences, usually roundoff and very near land. For cycles after 133 our processing of TOPEX/ALT data uses the TOPEX-only GDRs, which are available earlier than MGDR CDs; for Poseidon/SSALT data the MGDRs are used. All corrections we apply first make the GDR data consistent with MGDR-B, then additional processing follows.

DATASET VERSION
v2000-12-20-AKH (NOTE: new READ PROGRAM required; header contains Wallops cal)

Date of Last Update
2001-03-01

SUMMARY STATISTICS PLOTS

all points, deep-water only

KNOWN PROBLEMS

v2000-12-20-AKH: none yet.
v2000-05-01-AKH: A deteriorating tape recorder causes records at the edge of a data gap to flag the record as if ALT-A were ON. This has been occurring for most of 2000. (this version only differs from 0118 in the format of the files, to speed up Fortran I/O).
v2000-01-18-AKH: the files take too long to read in Fortran; not a problem in C or IDL.
v1999-08-04-AKH: Cycles 1-132 have temperature-corrected Wallops cal applied (should have used not temp corrected). Cycles 133-169 have non-temp corrected Wallops cal (correct). Cycles 170 and later have no Wallops cal applied.
v1999-03-10-AKH: TMR drift correction not applied.
v1998-04-16-AKH: Bias of 11 mm applied to Poseidon cycles should be zero

PROCESSING HISTORY

2000-12-20: files starting with cycle 283 for the Topex (US)
altimeter, are edited with an additional flag: a record is deleted unless the flag Instrument_State=8 in the Topex GDR. This solves a problem caused by a now-degraded on-board tape recorder, and documented in email from P. Callahan dated 2000-12-12: bad data frames give a value for this flag that implies that ALT-A is working, which misleads the processing software into using constants for ALT-A instead of ALT-B.
2000-05-01 File format changed slightly, to speed up Fortran I/O.
2000-01-18 Wallops correction, non-temperature corrected, applied consistently on all cycles.
1999-08-04 Files reorganized by cycle, rather than by pass. TMR drift correction applied.
1999-03-10 Corrected instrument biases: Poseidon bias set to zero; TOPEX bias from MGDR set to zero; from GDR set to -15 mm to match MGDR.
1998-08-06 Error corrected in shallow water flagging for cycles 1-10. Added cycle 208.
1998-07-xx Shallow water points individually flagged by adding 10000 mm
1998-06-xx Shallow water points added, with a count for each lat-lon indicating how many are in shallow water.
1998-04-16 Files now include shallow water points. A shallow water flag is also included so those points can be deleted by the user, if desired.

DATA FORMAT

Units
Residual heights: in mm, Valid range: +/- 1,500 mm
Lat, lon: in degrees
Time: in seconds past Jan 1, 1992

Special Values
Value set to 32767 where data does not exist or has been edited out
Values have 10,000 mm added in shallow water

To use deep data only, exclude values h: abs(h) > 1500
To use shallow also, if(abs(h) .gt. 1500 AND h .lt. 32767) h=h-10000

Data Organization
Each file corresponds to one cycle of exactly 3127x254 points. Each group of 3127 consecutive points corresponds to one 'pass'. There are 254 passes in a cycle. Each file contains equator crossing time, the relative time (in sec) of of each 1 sec sample, and the sea level residual in mm. Latitude and longitude values (3127x254) are in a separate file, common to all cycles.

Read Software

CORRECTIONS APPLIED

The altimetric corrections and models of the ocean surface which are applied to the altimetric range data are listed below. They are explained in some detail in Chapter 3 of Benada (1997), the MGDR-B manual. All corrections are from the MGDR-B, except for the inverted barometer and the smoothed ionosphere. On TOPEX GDR data we compute the Gaspar-4 EMB from the GDR sigma-0 and SWH. When two or more options for a correction are included in MGDR-B, the chosen option is explicitly named below.

Troposphere
DRY: from FMO, on TOPEX GDR, PODAAC MGDRB or AVISO MGDRC
WET: from TMR, on TOPEX GDR, PODAAC MGDRB or AVISO MGDRC
WET: TMR drift, per Keihm et al (1999)

Ionosphere
Dual freq for TOPEX/ALT cycles, Doris for Poseidon/SSALT cycles,
from TOPEX GDR, PODAAC MGDRB or AVISO MGDRC.
Smoothed with a boxcar average over 17 points (~105 km) along track.

EM Bias
GASPAR 4 on TOPEX and Poseidon. POS tracker bias (Gaspar 4) also applied.

Instrument Bias
15 mm substracted from TOPEX/ALT values read from TOPEX GDR.
Zero bias for Poseidon/SSALT or for TOPEX/ALT read from MGDR B/C.
Wallops non-temperature-corrected cycle drift.

OCEAN SURFACE MODELS REMOVED

Tides
UT CSR 3.0
SOLID EARTH TIDES: from TOPEX GDR, PODAAC MGDRB or AVISO MGDR-C
POLE TIDE: from TOPEX GDR, PODAAC MGDRB or AVISO MGDR-C

Inverted Barometer
Converts the FMO dry tropo to pressure at the altimeter time. A daily, globally-averaged
pressure, computed from NCEP pressures, is substracted from the instantaneous value
at lat-lon.

Mean Sea Surface
RAPP Mean Sea Surface from TOPEX GDR, PODAAC MGDRB or AVISO MGDR-C

Time-Average
After along-track gridding (see below), the 4 year average between cycles 11 and 167
(1993-96) at each lat-lon is removed from the residuals for that lat-lon.

DATA EDITING/REMOVAL

NASA ALT Records are deleted if any of the following conditions are true:

1.	alt_bad2(2)=1	PA/SS not reliable -> Bad Vatt_K, bad waveforms
2.	ss_pts_avg < 7	Noisy 1 sec data
3.	residual > 3000 mm	Too far off mean ssh
4.	alt_bad1(2)=1	2 or more bad waveforms, Bad Vatt_K
5.	current_mode(0:3) < 15	Fine track only
6.	current_mode(4)=1	Threshold track
7.	current_mode(5)=1	Primary max/3
8.	current_mode(6)=1	High variability
9.	geo_bad(1)=1	Land
10.	geo_bad(7)=1	Ice
11.	Wet Trop Rad=Bad	Flagged Wet Tropo Value
12.	Dry Tropo=Bad	Flagged Dry Tropo Value
13.	No JGM-3 Orbit
14.	Spike test	Fit a quadr. to 9 consecutive residuals; if RMS of fit > 50 mm, remove two extreme values and repeat fit; delete center value if farther from quadratic fit than 300mm.
15.	Instr_state <> 8	removes records incorrectly recorder by a degrading Tape Recorder, usually near a data gap. This correction not done prior to v20001220.

Poseidon SSAlt Records are deleted if any of the following conditions are true:

RMS of 1/sec > 17.5cm

Noisy Data

n_points < 16

Number of good points

current_mode_2=1

Bad 2nd half frame

instr_state_doris

Iono Corr bad

dry1_err > 2

Bad Dry Tropo

wet_h_rad

Wed Rad does not exist

Bad/Default Values on any one of these:

HP_Sat H_alt

H_MSS

H_SET

EM_Bias_Corr_K1

Iono_Doris Dry_Corr

Wet_H_Rad

abs(corrected residual) > 3000 mm

Too Far Off Mean

geo_bad_1(1)

Land Point

10.

geo_bad_1(7)

Ice

11.

No JGM-3 Orbit

12.

Spike Test

Fit a quadr. to 9 consecutive residuals; if RMS of fit > 50 mm, remove two extreme values and repeat fit; delete center value if farther from quadratic fit than 300mm.

Either TOPEX or Poseidon records are deleted if any of the following conditions are true:

1.	TMR BAD Data
2.	TMR Rain Data

ALONG-TRACK GRIDDING

The T/P observations are reported about once per second (the time interval varies with the height of the satellite above the Earth' surface). Although the orbit repeats within +/- 1 km every 9.915625 days, the points along the ground track where sampling is reported do not repeat. We therefore interpolate sea level residuals above the mean sea surface, with all corrections applied, using an exact cubic spline to a fixed set of points along-track. The independent variable of the interpolation is the difference between observed longitude and the closest (in time) observed equatorial crossing longitude. The points were chosen so their along-track distance is approximately 6.2 km on the ocean surface.

The resulting dataset looks like a set of approximately 500,000 virtual tide gages, with a sample every 9.915625 days. From each such 'virtual tide gage', the time-average between January 1993 and December 1996 are removed, a 4-year mean. This is necessary in order to remove residual geoid signals, since at present, errors in the geoid models exceed ocean signals over much of the ocean, and removing a time-mean is more accurate than simply removing the given Mean Sea Surface.

MORE INFORMATION

The MGDR-B User Manual, Benada, 1997, has much more information about the T/P mission, various corrections, etc.

The starting times corresponding to each TOPEX cycle are found in our Table of Cycle Start Times.

Other research groups offer comparable along-track sea level residual datasets: the Goddard Space Flight Center and NOAA-Laboratory for Satellite Altimetry. Uniform space-time grids are available from this site, and from University of Texas-Austin, University of Colorado-Boulder as well as from NOAA-Laboratory for Satellite Altimetry.

Further information can be obtained from two special issues of the J. Geophysical Research, volume 99 (C12), of December 1994, and volume 100 (C12), of December 1995.

[top]


Last Updated: 2003-01-30