News

Dear AVISO products users,

As you probably know, we use ECMWF atmospheric model pressure grids to derive dry troposphere and inverse barometer corrections for sea surface height measurements. Last week, on 26-Jan-2010, the ECMWF model was upgraded to a higher spatial resolution: <link http: www.ecmwf.int publications cms get ecmwfnews _blank>www.ecmwf.int/publications/cms/get/ecmwfnews/251

We have since identified an impact on the O-F/GDR and IGDR products, due to interpolation of the new higher resolution surface pressure grid onto the lower resolution grid that is still being used operationally. This anomaly has impacts on Jason-1, Jason-2 and ENVISAT products.

The affected period for those missions are: 

• For ENVISAT, FGDR and IGDR products have been affected by the ECMWF model resolution change from cycle 86 (pass 404) up to cycle 86 (pass 613).
• For Jason-1, OGDR and IGDR products have been affected by the ECMWF model resolution change from cycle 297 (pass 70) up to cycle298 (pass 3)
• For Jason-2, OGDR and IGDR products have been affected by the ECMWF model resolution change from cycle 57 (pass 197) up to cycle 58 (pass 130).

The largest impact is seen in the inverse barometer correction, which has static variations of up to 2 cm amplitude at a wavelength of ~ 50 km. This effect is most noticeable in regions within 200 km of the coast, but is not negligible outside those regions. The anomalous short-wavelength variations can be detected comparing IB values before and after the model evolution. The dry tropospheric correction is affected as well (~4 times smaller amplitude), but not the wet tropospheric correction.

The solution is to retain all the model's spectral information when interpolating from the high-resolution to low-resolution grids, which is an option when creating the operational grids. This change was implemented on 02-Feb-2010.

A smaller issue remains that the interpolation of the high-resolution ECMWF surface pressure grids to a lower resolution does not allow us to fully remove spectral leakage of land features onto the ocean. The remaining errors have a smaller amplitude (~20% of the short-wavelength features seen in the first figure), is also static in time, and almost exclusively affects regions within about 100 km of the coast. We are working on solving the remaining issue as well, and will keep you updated on our progress.

The affected IGDR files will be reprocessed for all missions. All issues related to the ECMWF model resolution change will be corrected prior to GDR production for this time period.

We would like to acknowledge our gratitude to Saleh Abdalla of ECMWF for his collaboration on this issue.

We thank you in advance for your understanding while we resolve these issues.

On behalf of the Jason-2 Measurement Systems Engineers