TOPEX A & B calibrations for WAVSAT

D J T Carter

20 December 2001

1. Introduction

The TOPEX B transmitter was brought into service in February 1999, because of deterioration in the A transmitter. Checks of wave height and wind speeds against buoy data after a few month showed no significant differences between Geophysical Data Records from the two transmitters, so the calibrations applied to the A values were applied to the B values. These (derived from 1988 buoy and altimeter measurements) are, for significant wave height (Hs), from Cotton (1998):

= 1.0523 Hs(TPA)  - 0.094

[A correction for drift of -0.0004 m/day from 26 September 1996 to January 1999 were also applied to A wave heights - see Challenor & Cotton (1999).]

And for wind speed (U), from Cotton (2000):

= 0.925 U(TPA)  + 0.617

Note that wind speeds are estimated within the WAVSAT processor from the CDROM sigma-0 values -0.63 dB using the same 5th order polynomial fit to tabulated values given by Witter & Chelton (1991) as for ERS-1.

An alternative, 4th order polynomial fit to Witter & Chelton (1991) has been produced by AVISO (1996) and is used to calculate the wind speed given on the CDROM, but the differences between the two algorithms are negligible for wind speeds less than 20 m/s - see Figure 1.

Figure 1 Comparison of two algorithms for altimeter wind speed.


Cotton (2001) has now compared TOPEX B values against many more buoy values, and identified a significant difference in the calibration which should be applied to the B wind speeds.  He used buoy data from the US National Data Buoy Centre (NDBC), Environment Canada (EC) and the UK Meteorological Office (UKMO) - and found differences between them, especially in wave height calibrations.  WAVSAT has always calibrated Hs and U values by comparisons with US NDBC measurements, so for consistency will use the comparison of TOPEX B data and the NDBC data in Cotton's dataset.

The results are given below, together with a comparison between the A and B calibrations.


2. TOPEX B Calibrations

Details of the data extraction and quality control are given by Cotton (2001). He obtained 992 pairs of altimeter and US NDBC buoy data from 24 buoys in the north western North Atlantic, north eastern and central North Pacific. Buoy wind speeds are 5 minute means adjusted to 10 m above the sea surface.

Applying orthogonal distance regression (or principal component regression) to these 992 pairs gave the following results for wave height:

=1.0376 Hs(TPB)
 - 0.0674
95% conf.
1.0226 - 1.0525
-0.0996 to -0.0352  

The 95% limits include the values from Equation 1; but only just and there is no reason to expect the two to be the same, so Equation 3 would seem to be the best available calibration for B.

Results for wind speed are:

= 0.8801 U(TPB) + 0.6385
95% conf. 0.8571 - 0.9032 0.4553 - 0.8217  

However, this wind speed regression was carried out using the speeds given on the TOPEX CDROM which are calculated from the AVISO algorithm. Calculating wind speed from the CDROM sigma-0 values (adjusted by -0.63 dB) using the 5th order polynomial fit gives the results shown in Figure 2, with regression equation:

= 0.8838 U(TPB) + 0.6239

These coefficients are well within the confidence limits of those in Equation 4, but for consistency equation 5 has been used for WAVSAT calibration.

Figure 2 US NODC buoy wind speeds against TOPEX B values.


3. Comparison between A & B

Comparing the wave height calibrations for A and B (equations 1 and 3): the correction that has to be applied to values from the B transmitter which have been calibrated using the calibration derived for A is given by Equation 6. The A calibration gives a very small over-estimation of Hs for Hs> 1 m..

= 0.9860 Hs(A)  + 0.0253

The correction for wind speeds (equation 7 from equations 2 and 5) is slightly larger, with the A calibration over-estimating by about 4%.

= 0.9555 U(A)  + 0.0344


These equations are illustrated by Figure 3 which shows a plot of Hs:U from median values along a TOPEX track off Morocco, roughly from 31°N 10.6°W to 30°N 10°W, with data calibrated using equations 1 & 2 and using equations 3 & 5. The higher values of U are seen to be reduced, - also indicated by the histograms shown in Figure 4.  There are barely perceptible changes in Hs.


Figure 3  Changes in Hs and U from calibration 'A' (square) to calibration 'B' (+)





Figure 4  Wind speeds from Figure 3.



4. Conclusion

A comparison of a large number of measurements of wave height and wind speed from buoys and from TOPEX B transmitter by Cotton (2001) has produced new calibrations for these altimeter values.  Calibrations derived from the US NDBC buoy data and given above in Equations 3 and 5 will be used when analysing WAVSAT TOPEX data recorded since February 1999.

The discrepancy, found by Cotton (2001),between the US NDBC buoy measurements of wave heights and those from EC and UKMO needs further investigation.




AVISO 1996
   AVISO User handbook: Merged TOPEX/POSEIDON Products
   AVI-NT-02-101-CN, Edition 3.0, July 1996. CNES, Toulouse.

Challenor, P. G. and Cotton, P. D. 1999
   The joint calibration of altimeter and in situ wave heights.
   Preprints CLIMAR'99, WMO Workshop on Advances in Marine Climatology, 8-15 Sept. 1999, Vancouver Canada, pp 121-129. Environment Canada.

Cotton, P. D. 1998
   A feasibility study for a global satellite buoy intercalibration experiment
   SOC Research & Consultancy Report No. 26, 73pp. (Unpublished manuscript).

Cotton, P. D. 2000
   Altimeter wind speed calibration - reassessment with corrected NDBC buoy wind speeds.
   Satellite Observing Systems (Unpublished manuscript).

Cotton P. D. 2001
   TOPEX B side altimeter wind speed and significant wave height calibration (1999-2001)
   Satellite Observing Systems (Unpublished manuscript).
[At topb.html ,or the Word version, topb99-01calrep2.doc]

Witter, D. L. and Chelton, D. B. 1991
   A Geosat altimeter wind speed algorithm and a method for altimeter wind speed algorithm development.
   J. Geophys. Res, 96, 8853-8860.