Supercooled liquid water cloud classification using lidar backscatter peak properties
The use of depolarization lidar to measure atmospheric volume depolarization ratio (VDR) is a common technique to classify cloud phase (liquid or ice). Previous work using a machine learning framework, applied to peak properties derived from co-polarised attenuated backscatter data, has been demonst...
| Main Authors: | , , |
|---|---|
| Format: | Text |
| Language: | English |
| Published: |
2023
|
| Subjects: | |
| Online Access: | https://doi.org/10.5194/egusphere-2023-1085 https://egusphere.copernicus.org/preprints/2023/egusphere-2023-1085/ |
| Summary: | The use of depolarization lidar to measure atmospheric volume depolarization ratio (VDR) is a common technique to classify cloud phase (liquid or ice). Previous work using a machine learning framework, applied to peak properties derived from co-polarised attenuated backscatter data, has been demonstrated to effectively detect supercooled liquid water containing clouds (SLCC). However, the training data from Davis Station, Antarctica, includes no warm liquid water clouds (WLCC), potentially limiting the model’s accuracy in regions where WLCC are present. In this work, we apply the Davis model to a 9-month Micro Pulse Lidar dataset collected in Christchurch, New Zealand, a location which includes WLCC. We then evaluate the results relative to a reference VDR cloud phase mask. We found that the Davis model performed relatively poorly at detecting SLCC with an accuracy of 0.62, often misclassifying WLCC as SLCC. We then trained a new model, using data from Christchurch, to perform SLCC detection on the same set of co-polarized attenuated backscatter peak properties. Our new model performed well, with accuracy scores as high as 0.89, highlighting the effectiveness of the machine learning technique when appropriate training data relevant to the location is used. |
|---|