Simulation of Space-borne Doppler Radar Observations of Clouds
Abstract
Space-borne Doppler radar observations of clouds are scientifically very interesting, but the correct interpretation of the measurements is challenging. Global observations of vertical velocity in cloud systems in many different climate zones will improve our understanding of large-scale convective motions and cloud microphysics. Unfortunately, observed Doppler speeds are a complex weighted average over all line-of-sight velocities within the radar's field-of-view. As the large orbital speed of the radar contributes to these line-of-sight velocities, cloud inhomogeneities can cause significant deviations in Doppler speed from the true vertical speed. EarthCARE is a proposed ESA/JAXA space mission dedicated to cloud and aerosol studies. The mission will consist of a VIS and NIR imager, a 95 GHz Doppler radar, a 355 nm lidar and a broadband radiometer all on the same platform and optimized for synergetic studies. To assess the accuracy of EarthCARE's Doppler observations, a new Doppler signal simulation technique was developed that deals consistently and effectively with cloud inhomogeneity. In particular, the technique takes account of the spatial variation in the statistical properties of reflected signals as the radar moves over a cloud scene. Traditional simulation techniques assume that during a single ground-track these properties remain constant. Using realistic cloud scenes derived from ground-based Doppler radar systems, the accuracy of the 95 GHz Doppler radar proposed for the EarthCARE space-mission was assessed. It is shown that within clouds, the required accuracy of the Doppler observations is not always met. For ground-tracks of 1 km, 10 % of the observations with Z > -20 dBZ has a deviation > 1 m/s from the truth, and for ground-tracks of 10 km 30 % has a deviation > 0.2 m/s from the truth. For observations near lateral cloud boundaries, errors can easily amount to several m/s. A correction algorithm is proposed based on the horizontal gradients in observed reflectivity. Higher horizontal sampling than is presently considered is then required (e.g. 250 m instead of 1000 m). For the cases considered sofar, the algorithm seems capable of fully correcting for the biases caused by cloud inhomogeneity.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2006
- Bibcode:
- 2006AGUFM.A13I..05S
- Keywords:
-
- 0321 Cloud/radiation interaction;
- 0394 Instruments and techniques;
- 3311 Clouds and aerosols;
- 3360 Remote sensing;
- 3394 Instruments and techniques