Wind profiling for a coherent wind Doppler lidar by an auto-adaptive background subtraction approach

Area of Science:

• Atmospheric science
• Remote sensing technology
• Signal processing

Background:

• Coherent Doppler lidar (CDL) data processing faces challenges in low signal-to-noise ratio (SNR) environments.
• Drifting power spectral density in CDL data complicates accurate wind velocity estimation.
• Existing denoising methods like periodogram maximum (PM) and adaptive iteratively reweighted penalized least squares (airPLS) have limitations.

Purpose of the Study:

• To introduce and evaluate a novel auto-adaptive background subtraction (AABS) method for CDL data denoising.
• To address the challenges posed by low SNR and drifting power spectral density in CDL measurements.
• To improve the accuracy and reliability of wind profiling using CDL.

Main Methods:

• Development of the auto-adaptive background subtraction (AABS) algorithm.
• Application of AABS to simulated and real-world CDL datasets.
• Comparative analysis against established methods: periodogram maximum (PM) and adaptive iteratively reweighted penalized least squares (airPLS).

Main Results:

• AABS demonstrates superior performance in identifying reliable peaks in CDL data compared to PM and airPLS.
• The proposed AABS method significantly extends the furthest detectable range by up to 16.7% (vs. airPLS) and 40% (vs. PM).
• AABS yields lower mean wind velocity errors and standard errors than both airPLS and PM methods.

Conclusions:

• Auto-adaptive background subtraction (AABS) is an effective denoising technique for low-SNR CDL data.
• AABS improves Doppler shift estimation quality, leading to enhanced wind profiling capabilities.
• The AABS method offers a significant advancement for CDL applications requiring precise wind measurements.