Process other modes than MESO

[gerritholl]

Currently sattools.get_abi_glm_multiscenes works only for ABI MESO (M1 and M2), because the interface to glmtools hardcodes the 1-minute intervals. To make this work with CONUS or Full Disk would require to either re-run glmtools passing longer time periods than 1 minute, or integrate them inside sattools. In either case, it also requires updating the time_threshold passed to satpy.MultiScene.from_files, which is currently hardcoded for 35 seconds.

[gerritholl]

Alternatives:

• Re-run glmtools with 5-minute period for CONUS or 10-minute for full disk. In this case, we'd want that the „integration time“ matches the period between two ABI scenes.
• Increase the time threshold when creating a MultiScene from ABI + GLM, such that multiple 1-minute GLM files match each 5- or 10-minute ABI scene. I think this would result in scenes where the GLM data had a StackedAreaDefinition? Not sure how well this would work.
• Create separate MultiScene objects for ABI and for GLM, then perform the necessary processing integrating GLM FED, then create a merged MultiScene from that.
• Something else?

[gerritholl]

Increase the time threshold when creating a MultiScene from ABI + GLM, such that multiple 1-minute GLM files match each 5- or 10-minute ABI scene. I think this would result in scenes where the GLM data had a StackedAreaDefinition? Not sure how well this would work.

This leads to vertical stacking and a StackedAreaDefinition, so instead of shape shape=(1500, 2500), it becomes shape=(9000, 2500), not great.

[gerritholl]

Or I could use a MultiScene with the regular 30-second time, then do the mapping in a custom blend function?