CoAuthors

Pierre Thibaut (CLS, France); Sophie Le Gac (CNES, France); Jean-Damien Desjonquères (CNES, France); Nicolas Picot (CNES, France)

In the Jason-3 ground segment, the rain flag algorithm is similar to the one used operationally for Jason-1 and Jason-2. It is based on a dual frequency method defined by Tournadre & Morland. This method uses a threshold on the attenuation of the MLE-3-derived backscatter coefficient in Ku-band with respect to the C-band and the radiometer cloud liquid water content in order to detect measurements polluted by rain. During the first months of the Jason-3 mission, it has been shown by NOAA that mainly due to the presence of biases between measured and expected sigma naught values, the rain flag appeared to be too aggressive, editing too many measurements. Moreover, other studies have raised that on Jason-2, data are often incorrectly rain flagged in particular in coastal and sea ice regions.

For mono frequency altimeter missions such as SARAL/AltiKa, a rain flag is currently computed in the ground segment using a Matching Pursuit algorithm developed by J. Tournadre. This technique detects the coherent variations of the waveform trailing edge induced by the rain cell over-flight. In the frame of the PEACHI project, initially processing SARAL/AltiKa data and now Jason-3 data as well, a new algorithm has been developed following the same approach but using continuous wavelet transform to perform a high performance rain flagging. We propose to present in this paper, the performance analysis of this rain flagging solution based on the continuous wavelet transform. Comparisons with conventional bi-frequency method will be shown demonstrating the better performances of this algorithm.