Instrumental noise and its impact on mean wind measurements

Fibre-optic based Doppler wind lidars (DL) are able to retrieve vertical profiles of kinematic quantities across the lower atmosphere with high spatio-temporal resolution. Especially short-term forecasting would benefit from assimilating their data which renders these compact systems promising candidates for operational use in future observing networks of meteorological and environmental services. Therefore, DWD includes the assessment of DLs in the effort to evaluate ground-based remote sensing systems for their operational readiness, called “Pilotstation”. Besides tests focusing on aspects such as technical reliability, uncertainty characterization, scanning strategies, and the verification of the retrieved mean wind speed and direction with the help of independent reference data from a 482 MHz radar wind profiler (RWP) and 6-hourly radiosonde (RS) ascents, DWD developed a standardized retrieval assuring a high-quality Level-2 product.

However, a prerequisite for operational applications is the robust detection of atmospheric return signals in the presence of instrumental noise. While the most common approach filters data via a fixed signal-to-noise ratio (SNR) threshold, we find the non-linear consensus method (CNS), already operational in the data processing chain of radar wind profilers, to be more efficient in the weak signal regime where it increases data availability without reducing data quality.

Here, we present results from a long-term assessment at the Lindenberg Meteorological Observatory using the RWP and RS as references and from a side-by-side comparison of eight Halo Photonics “Streamline” DLs during the FESSTVaL 2021 field experiment. We focus on the characterization of the instrumental noise and show its impact on the derived winds and the data availability.