glims-rgi / rgitools Goto Github PK

Noticed some artifacts in the RGI-TOPO figures of TanDEM-X 90-m DEM. I recommend the using a series of filters based on values in the AUX products. The workflow tested for HMA is in the tandemx_mask.py script here: https://github.com/dshean/tandemx.

Most of the tools are part of OGGM anyway, and the job for RGI was done in rgi_scripts.

There would be some advantage in moving some RGI scripts to an independent package (like https://github.com/ezwelty/rgi_links) but time is missing to clean all that up before RGI release...

What's certain is that RGITOPO will move to the OGGM domain, that's just much easier to maintain there (and is an OGGM product after all...)

GPL or BSD would do it.

Hi, I noticed you're using papermill for running some notebooks. I'm the author of Ploomber, a framework that allows the creation of notebook-based pipelines. Your current code is iterating over a few parameters and executing a notebook template; ploomber can efficiently parallelize and run all of them at the same time. I'm happy to open a pull request so you can see it's worth switching; let me know!

'RGI60-03.04079'

cc @matthiasdusch

CopernicusDEM is another global elevation model of interest to include and was initiated within the Copernicus framework. This elevation model will also be used for the orthorectification of Sentinel-2, and it is envisioned that USGS will implement this as well in its pipeline, so Landsat imagery will also be based on this. CopernicusDEM is based upon the WorldDEM from TanDEM (DLR), but it's quality controlled by Airbus, which detected and corrected errors in this elevation model (where infilling was done with the help of SRTM/ASTER GDEM). Only pitfall is the availability of the dataset, as it has an open dataset with a resolution of 90 meters, while a protected version of 30 meter also exists.

for more info see:
https://spacedata.copernicus.eu/web/cscda/data-offer/core-datasets/-/asset_publisher/SMEFUkzOylMw/content/optical-hr-pan-eu-coverages-hr_image_2015-?_101_INSTANCE_SMEFUkzOylMw_redirect=%2Fweb%2Fcscda%2Fdata-offer%2Fcore-datasets

https://eop-cfi.esa.int/index.php/docs-and-mission-data/dem

If there is a valley between glaciers distribution regions, the column of hypso_df in the example of glacier hypsometry will be out of order, which result in the yaxis of the figure in chaos.

In my opinion, add a step of sorting before the plot :

hypso_df = hypso_df.reindex(sorted(hypso_df.columns, key=int), axis=1)

Based on simple heuristics and OGGM's flowline algorithm, it should be possible to develop a tool able to detect glaciers that need to be divided.

For example:

• is the max elevation of a glacier on the outline of far into the geometry?
• do the centerlines have to up glacier?
• etc.

Having a flag in the RGI tagging potentially "undivided" geometries would be quite useful...

Certain links on this docpage are not working:

• The Jupyter notebook and this link in the Examples section are pointing to outdated sources
• Also the tutorial and interactive version links in the Code availability section are outdated.

I guess the notebooks(dem_comparison.ipynb and dem_soures.ipynb) to which these links are refering to are now located in the oggm tutorials repository .

This will help to pre-detect errors in RGI 7

see #39 #38

I just browsed through the example code on readthedocs, and it seems that cftime is missing in the environment, causing documentation code to throw errors basically everywhere, e.g. here:

https://rgitools.readthedocs.io/en/latest/tools.html

I was wondering where that might come from, and I guess xarray is just installed silently without cftime (since it's optional):

https://readthedocs.org/projects/rgitools/builds/11287705/

Maybe cftime should be included in requirements.txt explicitly!?

Should we add the OGGM download function (e.g. what @fmaussion is working on here OGGM/oggm#923) to RGITOOLS, so people don't have to install OGGM?

@matthiasdusch do you have the code somewhere that produced the summary statistics on https://rgitools.readthedocs.io/en/latest/dems_subregions.html ?

@afisc is taking over for RGI7 and this would help us to get started

rgitools fails to install using pip in environments with newer versions of setuptools, due to a conflict with PEP440:

I was able to fix this by changing changing line 86 from this:

FULLVERSION = rev.lstrip('v').replace(VERSION + '-', VERSION + '+')

to this:

FULLVERSION = rev.lstrip('v').replace(VERSION + '-', VERSION + '+').replace('-', '+', 1)

This fixes the version number error, and rgitools installs as expected:

It would be quite "qucik and dirty", but based on terminus elevation and some threshold one can image an algorithm computing this...

Use geopandas to create a look-up table linking RGI6 to RGI7. This can be tested with the GAMDAM inventory for example, since it will become RGI7

Hi @matthiasdusch and @fmaussion. This is fantastic! Great work, and glad to see some of this streamlined and centralized. I have scripts/libraries to work with RGI, DEMs, and other glacier mass balance tools scattered across various repos, and it would be great to incorporate some of them here.
I'll submit separate issues on specifics. But will start with this one. I don't think you have a tool to automatically fetch the latest RGI zip - the following script should do so (can be simplified):
https://github.com/dshean/demcoreg/blob/master/demcoreg/get_rgi.sh
Might be useful to include?

This would align CI with what OGGM is doing.

See http://geopandas.org/set_operations.html#set-operations-with-overlay

Add a license text to RGI-TOPO: say that we release the product under CC-BY, but that the license of the original data providers may vary. (or unclear)