Abstract's details

Rain flags for SWIM on-board CFOSAT: methods and assessment

Bruno PICARD (Fluctus SAS, France)


Mathilde Siméon (CLS, France); Jean-Christophe Poisson (CLS, France); Jean-Alexis Daguzé (CLS, France); Victor Gressani (IFREMER, France); Annabelle Ollivier (CLS, France); Cédric Tourain (CNES, France)

Event: 2019 Ocean Surface Topography Science Team Meeting

Session: CFOSAT

Presentation type: Type Oral

Contribution: not provided


CFOSAT is a joint mission of the Chinese (CNSA) and French (CNES) space agencies with the goal to monitor the ocean surface winds and waves and to provide information on related ocean and atmospheric science and applications.

CFOSAT caries two radar instruments: SWIM (Surface Waves Investigation and Monitoring), a wave scatterometer supplied by CNES; and SCAT (wind SCAT terometer), a wind-field scatterometer supplied by CNSA. SWIM’s 6 rotating beams enable it to measure wave properties (direction, wavelength, amplitude and partitions), while SCAT measures wind intensity and direction.
SWIM is a new CNES Ku-band radar instrument, based on the technology of a spaceborne radar altimeter. It is the first ever space radar concept that is mainly dedicated to the measurement of ocean waves directional spectra and surface wind velocities through multi-azimuth and multi-incidence observations. Orbiting on a 519 km sun-synchronous orbit, its multiple Ku-band (13.575 GHz) beams illuminating from nadir to 10º incidence and scanning the whole azimuth angles (0-360º) provide with a 180 km wide swath and a quasi-global coverage of the planet between the latitudes of ±82,5º.

When it is well known on classical Ku-band altimetry missions, the impact of the atmosphere on the wave spectra is still to be assessed.

For the current presentation, we will focus on the worked performed in the frame of CNES SALP - CaSyS (Calval Systematic for SWIM) project concerning the impact of rain event on the data quality. As for classical altimetry, rain flagging is an important step in the calibration/validation process: this latter aims at characterizing the performance of the products, providing a point-by-point assessment of the quality of each individual observation.

For nadir measurements, a wavelet transform approach is applied to identify rain cells, inspired by the method developed for AltiKa [Tournadre et al. 2015]. . As the waveform shape is directly impacted when the radar footprint crosses a rain cell, along-track variations of waveform parameters are detected using a continuous wavelet transform and a threshold is used to discriminate detected events that are related to rainy conditions.

For off-nadir measurements, another statistical method (v_geo flag) is applied and computed for each beam. Thresholds are applied on the low frequency variability of the sigma0 profiles, whose shape is also directly impacted by the local attenuations of the radar footprint.

In order to set up a qualitative and quantitative reference for the study of the impact of rain on observations, a “combined” rain rate product is built from observations provided by instruments dedicated to the atmosphere monitoring. The goal is to create a reference allowing to distinguish as well as possible between rain and no rain event and to potentially fix an upper limit to the rain rate above which the impact on SWIM measurements is critical. Three different data sources have been used: SSMI-S (Special Sensor Microwave Imager-Sounder) on-board the F17 satellites under the Defense Meteorological Satellite Program (DMSP), Windsat radiometer on-board Coriolis and GMI on-board the Global Precipitation Mission. SWIM data have been systematically collocated with the rain data from these three missions and the closest observation in time is kept, together with the time lag between the altimeter and the rain observation. Taking into account the correlation time of rain events, only the collocation with a time lag lower than 5 minutes are kept.

The wavelet-based flag and the v-geo flag are both validated against this database. Then, the number of observations impacted by rain events are characterized, in terms of geographical distribution of their occurrences.

Oral presentation show times:

Room Start Date End Date
The Monroe Hub Thu, Oct 24 2019,11:40 Thu, Oct 24 2019,12:00
Fluctus SAS