SVMD-based denoising methods for differential absorption lidar retrieval of CO2 concentration profiles

Differential Absorption Lidar (DIAL) serves as a pivotal technique for profiling atmospheric CO2 concentrations, yet its efficacy is hampered by the presence of noise. Traditional denoising methods, such as Empirical Mode Decomposition (EMD) and its variant (EEMD), have been employed to mitigate this issue. However, these methods are not underpinned by a robust mathematical framework and are prone to the phenomenon of mode mixing, which can compromise the quality of signal decomposition. In this research, we present a novel denoising method for Differential Absorption Lidar (DIAL) signals, employing Successive Variational Mode Decomposition (SVMD) integrated with Pearson correlation coefficients. The algorithm initiates by decomposing the echo signal into a multitude of intrinsic mode functions (IMFs) through the SVMD process. Subsequently, Pearson correlation coefficients are utilized to quantitatively assess the degree of similarity between each IMF and the original signal. Only those IMFs that meet a pre-defined threshold of similarity are integrated back into the reconstruction process, yielding a refined, denoised signal. The efficacy of our proposed denoising methodology is substantiated through a comparative analysis with simulated DIAL echo signals. The results highlight the algorithm's ability to effectively reduce noise in echo signals, thereby improving the precision and effective range of CO2 concentration profile retrievals.