Tests of Rain Profiling Algorithms for a Spaceborne Radar Using Raincell Models and Real Data Precipitation Fields

Jacques Testud Centre d'études des Environnements Terrestre et Planétaires, Issy-les-Moulineaux, France

Search for other papers by Jacques Testud in
Current site
Google Scholar
PubMed
Close
,
Paul Amayenc Centre d'études des Environnements Terrestre et Planétaires, Issy-les-Moulineaux, France

Search for other papers by Paul Amayenc in
Current site
Google Scholar
PubMed
Close
,
Xiankang Dou Centre d'études des Environnements Terrestre et Planétaires, Issy-les-Moulineaux, France

Search for other papers by Xiankang Dou in
Current site
Google Scholar
PubMed
Close
, and
Taoufix Tani Centre d'études des Environnements Terrestre et Planétaires, Issy-les-Moulineaux, France

Search for other papers by Taoufix Tani in
Current site
Google Scholar
PubMed
Close
Restricted access

Abstract

This paper compares the performances of several algorithms for profiling the rain rate from a spaceborne radar. Three techniques are considered: single frequency (or SF) at 13.8, or 24 GHz; dual frequency (or DF) at 13.8 and 24 GHz; and dual beam (or DB) at 24 GHz. The beam footprint resolution L is taken in the range 1.5–3 km. Other characteristics of the radar are based on realistic values. In the evaluation of algorithm performances, emphasis is put on the nonuniform beam-filling effect because of its direct impact on the required cross-range resolution. Sensitivity of the algorithms to the speckle noise, the radar calibration error, and the variability in drop size distribution is also investigated.

The study is achieved by using a three-dimensional simulation tool applied, first to model raincells, and then to precipitation field from real data. Raincells with exponential–-or Gaussian–-shapes, and various rain-rate peaks and D/L ratios, where D is the raincell “diameter,” are explored. Real data consist of vertical cross sections within a tropical squall line, including deep convection and stratiform rain, observed in Florida by means of 3D scans performed by a ground-based S-band radar.

The results point out the merit-demerit balance of the various algorithms. Use of surface echo to constrain the total path attenuation is an efficient way to improve the SF algorithms. The DF algorithm has definite advantages, compared with the SF counterparts, to correct scaling errors. The DB algorithm has some specific advantages but requires higher measurement accuracy and cross-range resolution, and has poorer spatial resolution than those of other approaches. Possible impact of the results on the definition of future spaceborne radars, and their data exploitation, is also discussed.

Abstract

This paper compares the performances of several algorithms for profiling the rain rate from a spaceborne radar. Three techniques are considered: single frequency (or SF) at 13.8, or 24 GHz; dual frequency (or DF) at 13.8 and 24 GHz; and dual beam (or DB) at 24 GHz. The beam footprint resolution L is taken in the range 1.5–3 km. Other characteristics of the radar are based on realistic values. In the evaluation of algorithm performances, emphasis is put on the nonuniform beam-filling effect because of its direct impact on the required cross-range resolution. Sensitivity of the algorithms to the speckle noise, the radar calibration error, and the variability in drop size distribution is also investigated.

The study is achieved by using a three-dimensional simulation tool applied, first to model raincells, and then to precipitation field from real data. Raincells with exponential–-or Gaussian–-shapes, and various rain-rate peaks and D/L ratios, where D is the raincell “diameter,” are explored. Real data consist of vertical cross sections within a tropical squall line, including deep convection and stratiform rain, observed in Florida by means of 3D scans performed by a ground-based S-band radar.

The results point out the merit-demerit balance of the various algorithms. Use of surface echo to constrain the total path attenuation is an efficient way to improve the SF algorithms. The DF algorithm has definite advantages, compared with the SF counterparts, to correct scaling errors. The DB algorithm has some specific advantages but requires higher measurement accuracy and cross-range resolution, and has poorer spatial resolution than those of other approaches. Possible impact of the results on the definition of future spaceborne radars, and their data exploitation, is also discussed.

Save