Feature Request

Is your feature request related to a problem? Please describe.

Our users have requested that we write a "valid time" header to our ninjogeotiff and NetCDF files and include this timestamp in the filename. This "valid time" should be as representative as possible of the measurement time for the image being produced. We have agreed to this request (at least for GEO products).

Currently, we use the reference time for satellite imagery production, or the start time when this is not available. For Meteosat, this time is not optimally representative for the measurement time in Central Europe. In the past, we have artificially added the repeat cycle interval for MSG ZDS (15 minutes) or MSG RSS (5 minutes) and called it "end of scan" in an in-house addition, but this time is not the best possible estimate for the measurement time either (and our implementation was buggy).

Describe the solution you'd like

I would like that, when writing geotiff, ninjogeotiff, or NetCDF files, it should be possible to write a "valid time" header that is representative of the measurement time for the image. This could be the measurement time for the centre or an average measurement time over the entire image (those two are probably close to each other in most cases).

Naturally, this will need to consider resampling.

Describe any changes to existing user workflow

This would be an additional feature. Therefore, it would not change existing user workflow.

Additional context

An alternative would be an ad-hoc in-house solution where we hardcode a fixed number of minutes to add depending on the production. However, a solution integrated in satpy would be easier to maintain, and possibly of use to others as well.