build_renewable_profiles not actually checking for water depth

[Eric-Nitschke]

Checklist

• I am using the current main branch or the latest release. Please indicate.
• I am running on an up-to-date pypsa-earth environment. Update via conda env update -f envs/environment.yaml.

Describe the Bug

When checking build_renewable_profiles.py for how the offshore raster is implemented as part of Issue #1511 I've realized something:
build_renewable_profiles.py isn't actually checking for water depth.

The code works as follows:

• it adds an excluder raster to the cutout, where only cells are allowed where the provided value is larger than -config["max_depth"]
• as default, config["max_depth"]: 50
• the provided value stems from paths.gebco, which leads to data/gebco/GEBCO_2021_TID.nc
• data/gebco/GEBCO_2021_TID.nc does not include the GEBCO Grid data, which provides seabed-data. Instead it provides Type Identifiers (TID), which tell you, how the seabed was measured. I've attached an image of how a section of the dataset looks down below. You can get the same output using the following code. Depending on your RAM might need to make the slices even smaller:

import xarray as xr

ds = xr.open_dataset(r"data\gebco\GEBCO_2021_TID.nc") 
ds["tid"].sel(lon=slice(-20,20),lat=slice(30,80)).plot()

• since all TIDs are equal or greater than 0, the whole offshore area will automatically fulfill the depth requirement.

Unless I've misread some of the code, this simply does not do what is supposed to. To fix this, the bundle_config needs to be adjusted, to actually include the seabed data.

EDIT:
This is what the data is supposed to look like (from GEBCO):

[Eric-Nitschke]

I've tried using the actual seabed data from GEBCO. This generally works, but does not solve #1511.

[Eric-Nitschke]

@ekatef I think this could be fixed by uploading the GEBCO_2024 Grid (sub-ice topo/bathy) file from their website.
Importantly it would require the following steps:

• update the file not just in the main data bundle, but also the tutorial ones
• change the output names in configs/bundle_config.yaml
• change the input of build_renewable_profiles in the Snakefile

[ekatef]

Hey @Eric-Nitschke, many thanks for investigating it!

Completely agree with the overall algorithm you suggest to fix it. As an additional point, we may need to generate tailored datasets for tutorial data bundles as the gridded dataset is significantly larger as compared with TID one.

By any chance, have you had an opportunity to test a proper gridded dataset? Would be good to be aware if any fixes are needed to ensure that the replacement works smoothly.