Abstract's details
Jason-3: ALES vs ADAPTIVE vs MLE4, comparing the retracking solutions
CoAuthors
Event: 2022 Ocean Surface Topography Science Team Meeting
Session: Regional and Global CAL/VAL for Assembling a Climate Data Record
Presentation type: Type Poster
Contribution: PDF file
Abstract:
The official retracking solution for Jason-3 products is the MLE4 retracking. However, recent improvement using ALES (TUM/ESA) and ADAPTIVE (CNES/CLS) retracking solutions may change this reference. While the ALES retracker was specifically developed for coastal use, the ADAPTIVE retracker aims at incorporating the true instrumental PTR rather than a model.
In this context, this study was conducted, mostly over Jason-3 1Hz data, to compare these retracking solutions in plain ocean as well as coastal areas.
For this study, we built a dataset over 117 cycles from the official Jason-3 GDR product on which we interpolated the official ALES+ product distributed in the TUM ALES website. The comparison was mainly performed on the SLA products, investigating firstly the availability of the data and secondly its quality.
Concerning the availability of the data, there is an equivalent number of measures for both ALES and ADAPTIVE retrackers. In coastal areas (3-10 kms to the coastline), there is a data availability increase of 25% for the ADAPTIVE retracker and 16% for ALES with respect to MLE4. The main difference remains in the necessity to perform the calval editing steps over the ALES data to exclude all erroneous measures, while the ADAPTIVE retracker is way more restrictive in its retracking method.
When focusing on data quality, although the noise from a spectral analysis on the SLA product shows a slight improvement for ALES data, the 20Hz noise (deduced from the range std) shows a larger noise for ALES data than both ADAPTIVE and MLE4. While the spectrum result is confirmed in plain ocean (reduction of 1mm STD on the high frequency part of the SLA), the opposite is observed in coastal areas (<15km).
This analysis, completed by other calval diagnostics, will identify the respective strengths and limitations of the three retracking solutions for altimetric studies for both coastal and open ocean use.
In this context, this study was conducted, mostly over Jason-3 1Hz data, to compare these retracking solutions in plain ocean as well as coastal areas.
For this study, we built a dataset over 117 cycles from the official Jason-3 GDR product on which we interpolated the official ALES+ product distributed in the TUM ALES website. The comparison was mainly performed on the SLA products, investigating firstly the availability of the data and secondly its quality.
Concerning the availability of the data, there is an equivalent number of measures for both ALES and ADAPTIVE retrackers. In coastal areas (3-10 kms to the coastline), there is a data availability increase of 25% for the ADAPTIVE retracker and 16% for ALES with respect to MLE4. The main difference remains in the necessity to perform the calval editing steps over the ALES data to exclude all erroneous measures, while the ADAPTIVE retracker is way more restrictive in its retracking method.
When focusing on data quality, although the noise from a spectral analysis on the SLA product shows a slight improvement for ALES data, the 20Hz noise (deduced from the range std) shows a larger noise for ALES data than both ADAPTIVE and MLE4. While the spectrum result is confirmed in plain ocean (reduction of 1mm STD on the high frequency part of the SLA), the opposite is observed in coastal areas (<15km).
This analysis, completed by other calval diagnostics, will identify the respective strengths and limitations of the three retracking solutions for altimetric studies for both coastal and open ocean use.