Aeolus Rayleigh-channel winds in cloudy conditions

Aeolus was launched in August 2018 and is expected to be operational until 2022. Aeolus is the first Doppler wind lidar in space to measure wind profiles through Rayleigh scattering of an ultra-violet laser beam and the determination of the Doppler shift of the scattered light by molecules along the Line-Of-Sight (LOS). In addition, Mie scattering provides winds on aerosol and cloud particles. The atmosphere return signal is a small bandwidth peak (from Mie scattering) on top of a broadband spectrum (from Rayleigh scattering). The tails and central part of the spectrum are being processed separately to yield so-called Rayleigh channel and Mie channel winds respectively.

Signals in both channels are being accumulated onboard the satellite to segments of 2.85 km length along the satellite track, denoted measurements. Rayleigh winds are obtained by on-ground processing through accumulating typically 30 measurements to yield a single Rayleigh wind observation of sufficient quality for use in Numerical Weather Prediction (NWP). The vertical resolution of the horizontally projected LOS wind profiles is typically 500 m in the boundary layer, 1 km in the free-troposphere and 1.5-2 km in the stratosphere, but this can and has been changed in a flexible way during the mission.

In case of clouds and/or aerosols presence within the sensing atmospheric volume, signal from Mie scattering leaks into the Rayleigh channel signal. Since the Rayleigh-channel signal processing assumes a pure molecular signal this so-called Mie contamination causes biases in retrieved winds. This is solved through classifying measurements as either ‘clear’ or ‘cloudy’ before accumulation to observation level. Clear measurements (out of a total of 30) are accumulated to yield a Rayleigh-clear wind. This procedure has proven successful and Aeolus Rayleigh-clear winds are used operationally today by a number of meteorological centers around the world.

A similar procedure for cloudy measurements is less trivial and requires correction for Mie contamination. So far, implemented corrections were not successful in producing Rayleigh-cloudy winds of sufficient quality for use in NWP. A new correction scheme has been introduced and tested recently and proved successful to produce bias-free winds and a random error slightly larger as compared to Rayleigh-clear winds. The latter is explained by increased heterogeneous atmospheric conditions in which Rayleigh-cloudy winds are obtained. Interestingly, Rayleigh-cloudy and Mie-cloudy winds are obtained for identical atmospheric conditions and as such provide independent information on the atmospheric flow, which allows to characterize the error sources of the different types of wind observations, including instrumental/calibration errors, but also errors due to incorrect height assignment and representativity.

This paper describes the new scheme to correct Rayleigh winds for Mie contamination and its application to Aeolus data. The results show that resulting Rayleigh-cloudy winds are of good quality to be considered for operational use in NWP.