Reducing the horizontal diffusion base-time form 60min to 15min. This was never documented anywhere (only for vertical diffusion) and was recently set to 60min. Since the horizontal diffusion has been tested at a time where SNAP was run with ~50km resolution meteorology, 15min timesteps were common, so the original horizontal diffusion was more similar to a setting with 15min base-time.
The bias with the 15min baseline for the etex results (above) are just slightly worse than the bias with the 60min (below), and acceptable.
This PR partly reverts yesterdays horizontal diffusion change to a more conservative approach by not changing too much in SNAP without extensive testing.
Reducing the horizontal diffusion base-time form 60min to 15min. This was never documented anywhere (only for vertical diffusion) and was recently set to 60min. Since the horizontal diffusion has been tested at a time where SNAP was run with ~50km resolution meteorology, 15min timesteps were common, so the original horizontal diffusion was more similar to a setting with 15min base-time.
The bias with the 15min baseline for the etex results (above) are just slightly worse than the bias with the 60min (below), and acceptable.
This PR partly reverts yesterdays horizontal diffusion change to a more conservative approach by not changing too much in SNAP without extensive testing.