L3 product: Snow height, ice surface height and surface height

[BaptisteVandecrux]

In the L3 version of the dataset, there should be:

• SnowHeightm: Snow height above the ice surface. Never negative. Calculated as BoomHeightOverBareIcem - BoomHeightm except for accumulation sites (EGP and KAN_U) where it will need to be manually reset to 0 at the end of each melt season.
• IceSurfaceHeightm: Height of the bare ice surface. Calculated as DepthPressureTransducerAtInstallationm- DepthPressureTransducerm. Needs to be adjusted every time a new ablation hose is installed.
• SurfaceHeightm: Height of the snow/ice surface. Calculated as IceSurfaceHeight + SnowHeightm

Illustation:

Originally posted by @BaptisteVandecrux in #2 (comment)

[BaptisteVandecrux]

Toolbox for processing/adjusting these heights:
https://github.com/GEUS-PROMICE/PROMICE-AWS-toolbox

[mankoff]

These variables are now defined in the current version of the variables DB: https://github.com/GEUS-PROMICE/PROMICE-AWS-processing/blame/045e2ab972163ab2d2ebe7c223029ac9d311cb32/data/variables.csv#L52-L55

The code to populate them has not yet been added to this code, or called from the AWS-toolbox code linked above.

[PennyHow]

Update: we have actually removed these from the variables look up table. I'm not sure when this happened.

But we would still like snow height, surface height and ice surface height calculations adding to our processing in the future.

[BaptisteVandecrux]

Now that I have moved forward with more advanced cleaning of z_pt_cor, z_boom and z_stake, I think we are getting closer to being able to provide these variables.

I think it is higher priority than "enhancement" because snow depth is a WMO variable and could be distributed to many services. Good quality snow depth would also help to derive SWE from surface height.

One of the requirements are a "station type" attribute (#217) and a smarter processing pipeline, because deriving additional variable will slow down the overall processing.

[PennyHow]

If you have some scripts alredy for deriving snow height, ice surface height and surface height then please share them and we can look at how they can be more efficient, and also how we can fit them into, and improve, the pypromice processing pipeline

[BaptisteVandecrux]

Made a first implementation here: https://github.com/GEUS-Glaciology-and-Climate/pypromice/tree/surface-height-thermistor-depth-processing

Here are the results: https://github.com/GEUS-Glaciology-and-Climate/PROMICE-AWS-diagnostic/tree/main/figures/surface_heights

I call it level 4 because it uses quite heavy sequence of filters and tests to stitch the surface height as seen by the SR50s and the pressure transducer.

I can also only be run on merged l3 data: to create a continuous surface height, transmission should be aware of what was the surface height in the raw files. And if we start to bridge between the transmission and raw file processing, then we might as well run that specific processing after the merge. There are some other variables that could be processed after the merging.

It allows afterwards to derive the thermistors' depth.

[BaptisteVandecrux]

completed in #294