Improved Method for the Retrieval of Extinction Coefficient Profile by Regularization Techniques
In this work, we revise the retrieval of extinction coefficient profiles from Raman Lidar. This is an ill-posed problem, and we show that methods like Levenberg–Marquardt or Tikhonov–Phillips can be applied. We test these methods for a synthetic Lidar profile (known solution) with different noise re...
| Published in: | Remote Sensing |
|---|---|
| Main Authors: | , , , |
| Format: | Article in Journal/Newspaper |
| Language: | unknown |
| Published: |
MDPI
2025
|
| Subjects: | |
| Online Access: | https://epic.awi.de/id/eprint/60070/ https://doi.org/10.3390/rs17050841 https://hdl.handle.net/10013/epic.6d39ec9c-ad74-44cb-a671-9f573f7029cf |
| Summary: | In this work, we revise the retrieval of extinction coefficient profiles from Raman Lidar. This is an ill-posed problem, and we show that methods like Levenberg–Marquardt or Tikhonov–Phillips can be applied. We test these methods for a synthetic Lidar profile (known solution) with different noise realizations. Further, we apply these methods to three different cases of data from the Arctic: under daylight (Arctic Haze), under daylight with a high and vertically extended aerosol layer, and at nighttime with high extinction. We show that our methods work and allow a trustful derivation of extinction up to clearly higher altitudes (at about half a signal-to-noise ratio) compared with the traditional, non- regularized Ansmann solution. However, these new methods are not trivial and require a choice of parameters, which depend on the noise of the data. As the Lidar signal quality quickly decreases with range, a separation of the profile into several sub-intervals seems beneficial. |
|---|