Using radar data to indicate raining status/probability at specific location

[amotl]

Just an idea by @mhoster83:

If I understood the logic of the composite radar files correctly, they split up in 900x900 pixels. It would be great if a feature would be available, that allows you to select your "Home Pixels" and to get the color value back, indicating if it is currently raining or if it starts to rain within 15 Minutes.

Originally posted by @mhoster83 in #138 (comment)

[amotl]

Do you have any thoughts on this, @kmuehlbauer or @meteoDaniel? Does it actually make sense?

[kmuehlbauer]

I was thinking about this, and this would be a nice feature. But still you would need to transfer the complete files and extract the wanted part. And 900x900 is one format, there are other formats. So I see this in a processing library.

Not to propose anything, but wradlib can do this almost out of the box. I could help with an example, though.

[gutzbenj]

@kmuehlbauer I absolutely agree that this is not the purpose of wetterdienst as a "provider" for data.

[amotl]

I was thinking about this, and this would be a nice feature.

Thanks for confirming that.

I absolutely agree that this is not the purpose of wetterdienst as a "provider" for data.

Sure, I also agree. However, it would be cool to accompany the code base using an appropriate example, where the other radar examples also build upon wetterdienst together with wradlib.

Not to propose anything, but wradlib can do this almost out of the box. I could help with an example, though.

That would be cool. Let's add a respective example to example/radar then? I am looking forward to any contribution you might be able to make on this, @kmuehlbauer.

[kmuehlbauer]

@amotl Please have a look at RADOLAN RX with wetterdienst and wradlib. I've tried to gather the last 12 available RX composites. Maybe there is an easier method, but it works.

cc @mhoster83

[amotl]

Dear Kai,

thank you so much! If you don't mind, I will add your gist to the example/radar folder of wetterdienst as outlined above.

Maybe I can find some time to make your example more interactive by obtaining the coordinates and the radius from command line arguments.

The whole radar data acquisition thing coming in through #190 has not been integrated into the command line interface yet. That would be a first step towards this.

With kind regards,
Andreas.

[kmuehlbauer]

@amotl Yeah sure, and as you pointed out #190 has to be integrated to make this work. Nice idea with the command line arguments. This could be extended to use a region of interest, but we can always add this later.

I was also thinking about adding a tutorial to wradlib-notebooks which uses wetterdienst for retrieval of DWD (radar) data (after #190 is released).

[kmuehlbauer]

I was also thinking about adding a tutorial to wradlib-notebooks which uses wetterdienst for retrieval of DWD (radar) data

Short heads-up:

• I've merged a quite nice enhancement into wradlib main branch (ENH: make wradlib.io capable of consuming file-like objects wradlib/wradlib#469). Now all reader functions can handle file-like objects as inputs, means they can consume the in-memory buffer from the wetterdienst request directly without creating temporary files.

• There's a PR pending (ENH: adapt example to use wetterdienst wradlib/wradlib-notebooks#46) which shows the usage.

• To conform with our conda-forge based workflow I've created https://github.com/conda-forge/wetterdienst-feedstock to have wetterdienst availabe as conda package. It might be worth checking if I got all dependencies right.

[meteoDaniel]

I would recommend to build a KDTRee for each grid by transforming lat/lon coordinates into radians. The same way as the get_nearest_stationworks.