Add support for CAMELS DE dataset #253
Conversation
| return df | ||
|
|
||
|
|
||
| def load_camels_de_timeseries(data_dir: Path, basin: str) -> pd.DataFrame: |
There was a problem hiding this comment.
for multi-basin neural network training, you usually want to work with mm/day discharge and not cms. I am not fully sure how the data is provided in CAMELS-DE but I think it is cms. A few of our other classes (e.g. check CAMELSUS) have a conversion from volumetric to area-normalized units implemented. You might want to check if that is needed.
There was a problem hiding this comment.
Thanks! I’m attaching an image from the paper by R. Loritz et al.:CAMELS_DE. As you can see, both discharge variables are provided: mm/day and cms. So I believe no conversion should be necessary in this case.
Co-authored-by: Martin Gauch <15731649+gauchm@users.noreply.github.com>
There was a problem hiding this comment.
Sorry, I missed one thing: in datasetzoo/init.get_dataset, the docstring is outdated (and was already outdated before...)
)Can you change this to something like Currently implemented datasets are those listed in datasetzoo.get_dataset. The 'generic' dataset class can be used for any kind of dataset as long as it is the correct format.
Thanks!
|
I think that's it, if you need anything else, let me know. |
CAMELS DE dataset by [#]_
References
.. [#] Loritz, R., Dolich, A., Acuña Espinoza, E., Ebeling, P., Guse, B., Götte, J., Hassler, S. K., Hauffe, C.,
Heidbüchel, I., Kiesel, J., Mälicke, M., Müller-Thomy,., Stölzle, M., and Tarasova, L. (2024). CAMELS-DE:
hydro-meteorological time series and attributes for 1582 catchments in germany. Earth System Science Data,
https://doi.org/10.5194/essd-16-5625-2024, 2024.