Geosat JGM-3 GDRs

4. How to Construct Corrected Sea Height

As discussed in this chapter, a number of corrections must be applied to the altimeter sea height, H, to obtain the most accurate value:

The first 8 corrections (in all caps) are the primary ones. Of these, the first 7 are provided explicitly in each 1-sec record, and the eighth (IB: inverted barometer) can be computed from the 1-sec values of DRY_NCEP.

The last 3 corrections (in lower case) are secondary, more slowly varying terms and are provided in separate tables found on this CD-ROM.

4(a) Primary Corrections to H

The fundamental altimeter measurement contained in the GDRs is the sea height, H. This quantity is the difference between the satellite's orbital height above the reference ellipsoid (ORB) and the altimeter's range measurement (not explicitly provided in the GDR, but defined as the distance from the altimeter to the sea surface, determined from the round-trip travel time of the radar pulse). Thus the sea height is the distance (positive upwards) from the reference ellipsoid to the instantaneous sea surface. Both the 1-sec averaged height (H) and the 10 Hz heights (H1 - H10) are sea heights formed from "orbit-range" differences. The GDR heights have been corrected for instrumental effects, but not for environmental or geophysical influences. The sense of all the corrections above is that they should be subtracted from the sea height.

Instrumental effects and timing bias -

The instrumental effects which have already been applied include the center-of-gravity, FM crosstalk, ATT/SWH bias, and Doppler corrections (refer to JHU/APL ,1985, for a further discussion of these corrections) Additionally, a 5.0 msec timing bias (a bias between the orbit determination timing and the radar altimeter timing) has been applied while switching to the new JGM-3 orbits. The time tags on the GDR records were increased by 5 msec prior to computing the orbit height at the corrected GDR time.

Wet troposphere -

The radar pulse is slowed down by water vapor in the troposphere. The user has a choice of three corrections: WET_NCEP, WET_NVAP, or WET_T/S.

Recommendation - Use WET_NCEP

WET_NVAP is based on a monthly climatology and is suitable for checking the validity of the other two wets, or can be used as a backup if the primary wet correction is missing. However, it is not suitable, for most applications, as the primary wet correction. WET_NCEP, based on the NCEP/NCAR reanalysis meteorological model, is computed from synoptic model grids and should always be available on the GDR. Its advantage is full coverage in space/time and the synoptic nature of the grids. The model grid size of 2.5 degrees, however, makes it a relatively low-resolution wet correction. WET_T/S is computed from satellite measurements of integrated water vapor, based on the Tiros Operational Vertical Sounder data (prior to 7/9/87), and subsequently from the Special Sensor Microwave Imager data. The TOVS/SSMI data, in the form of 1 degree grids, have more spatial structure than the NCEP model data, but suffer from data gaps and aliasing from non-synoptic temporal compositing. Over the duration of the GM+ERM missions, WET_NCEP will guarantee a consistent correction with some loss of resolution and perhaps model bias. We believe WET_NCEP is superior to WET_T/S from the TOVS time period, but is probably not as good as WET_T/S based on the SSMI portion of the data.

Dry troposphere -

The radar pulse is slowed down by air molecules in the troposphere. The user has a choice of two corrections: DRY_NCEP and DRY_ECMWF.

Recommendation - Use DRY_NCEP

The two dry corrections on the GDR are comparable, with DRY_NCEP based on a more modern meteorological model. The presence of the new DRY_NCEP, and the older DRY_ECMWF (present on the previous T2 GDRs), allows the user a measure of verification on the dry correction values.

Ionosphere -

The radar pulse is slowed down by the ionosphere. Only one ionosphere correction, IONO, is provided. It is derived from the IRI95 model.

Tides -

The altimeter height measurement must be corrected for several effects unrelated to the propagation of the radar pulse. The largest of these are tides, which include the familiar ocean tide (O_TID), the solid earth tide (S_TID), and the load tide (L_TID, deformation of the solid earth caused by the ocean tide).

Sea state bias -

Sea state bias is another surface effect, related to wave height, wind speed, and satellite attitude, which should be applied to correct the height measurements. The SSB term provided in the GDR was derived by Gaspar, Ogor, and Hamdaoui (1996) based on Geosat crossover differences.

Local inverse barometer -

The response of the sea surface to changes in atmospheric pressure, the so-called "inverse barometer" correction, has a large effect on measured surface height. The simplest form for this correction is a purely local response of the sea surface to atmospheric pressure at the measurement point. Though not explicitly provided in the GDR records, the local inverse barometer correction can be easily computed from the dry correction, by first computing sea level pressure:

P (mbar) = - DRY_NCEP (mm) / (2.277 * ( 1.0 + 0.0026 * cos ( 2.0 * LAT ) )

The inverted barometer correction is then:

IB (mm) = -9.948 * ( P (mbar) - 1013.3 )

4(b) Secondary Corrections to H

In addition to the first order corrections supplied in each 1-sec record, we have included three "secondary" corrections via external files. These are located in the scnd_cxn folder on the first CD of the GM and ERM data sets. The secondary corrections are relatively small (< 3 cm peak-to-peak) and vary more slowly in time than the primary corrections. It is therefore more efficient to provide them in the form of look-up tables.

Global inverse barometer -

The glo_ib correction is provided at 6-hour intervals and augments the local inverse barometer correction, IB. The values in the file GLO_IB.TXT should be subtracted from the sea height, H.

The true response of the sea surface to changes in atmospheric pressure must account for changes in the globally averaged (over-ocean) sea level pressure (Ponte et al., 1991). The glo_ib values were derived from the NCEP sea level pressure grids by (1) averaging over all ocean points, (2) computing a mean inverse barometer value using the IB formula above, and (3) changing the sign to enable the correction to be applied by subtracting it from H (the same convention as all other corrections). The variations of the 6-hourly values, as well as a smoothed version of this correction, are shown in the figure below.

Oscillator drift and internal calibration -

"hcal" and "uso" are refinements of two instrumental corrections previously applied to the sea surface heights on the GDRs. Both are supplied in one file, HCAL_USO.TXT. Each line in this file contains a daily value of hcal and uso (George Hayne and David Hancock, NASA/ WFF, personal communication). Both values provide the difference between these instrumental corrections as already applied and the new estimate of the correction. As with the primary corrections, they should be subtracted from the sea height H.

The hcal correction is based on the internal calibration measurement of the altimeter electronics, primarily reflecting the internal temperature of the altimeter system. This correction shows large semi-annual fluctuations.

The uso correction is based on the altimeter's ultra-stable-oscillator (USO) which monitors the frequency of the altimeter's radar pulse generator. Changes in the USO frequency have a direct effect on the altimeter range measurement. The uso correction also removes a single step correction, applied on 9/12/87, which was a coarse correction applied to the GDRs to account for changes in the USO frequency.