`rainbow/backend` may be using wrong string for `startrange` vs. `start_range`

[Steve-Roderick]

As an example, consider the RAINBOW file used in the notebook: https://github.com/openradar/xradar/blob/main/examples/notebooks/Rainbow.ipynb

filename = DATASETS.fetch("2013051000000600dBZ.vol")
ds = xr.open_dataset(filename, group="sweep_0", engine="rainbow")

Inspecting the file 2013051000000600dBZ.vol carefully:

$ grep -a startrange 2013051000000600dBZ.vol
$ # No match

Notice the following search for (start_range) instead.

$ grep -a start_range 2013051000000600dBZ.vol
         <start_range>0</start_range>
         <start_range>0</start_range>

Consider the source https://github.com/openradar/xradar/blob/main/xradar/io/backends/rainbow.py#L720

        start_range = self.root._get_rbdict_value(
            var, "startrange", default=0, dtype=float
        )

I believe the start range may not always be populated correctly and some rainbow files may not decode at all.

The rainbow spec may prescribe start_range and not startrange.

[kmuehlbauer]

@Steve-Roderick Thanks for the well-written issue. I agree, it looks like this isn't handled properly. This is probably from my impression that stoprange corresponds with startrange and not with start_range. When implementing this in wradlib back in the days, there have been no field for start_range.

@Steve-Roderick Would you be up to fixing this?

Update: I've consulted a relatively recent rainbow file format guide and it's start_range (0 by default).

[Steve-Roderick]

Thank you @kmuehlbauer for confirming this issue. I am up for fixing this issue - thanks for suggesting this. Beyond correcting the xml tag name to start_range, I will attempt to verify that the start_range is correctly applied. I believe this would be the distance in km to the center of the zeroth range bin (this start_range > 0 implies a cone of silence on the sweep with the lowest elevation angle.

[kmuehlbauer]

I believe this would be the distance in km to the center of the zeroth range bin (this start_range > 0 implies a cone of silence on the sweep with the lowest elevation angle.

Currently, we assume the (rainbow) reference plane for start_range and stoprange to be the beginning of the range bin. Indeed, for start_range > 0 this would imply a cone of silence. There is no mention in my documentation what the reference plane actually is, but we should go with the most logical approach. Please note that for xradar (internal) we use bin-center as reference.