Boundary conditions for regional models.

[Ed Lizard]

1) Secondary question, if the NAM is only trusted enough to run to 84 hours, and people make comments about not trusting it past 48 or 60, why does the DGEX run more than 84 hours?

2) More importantly, if the prior GFS is used to set the boundaries, especially since upper latitude weather sometimes retrogrades, and low level features at lower latitudes in the warm months sometimes come from the East, what happens if the solution of the NAM, maybe because of finer resolution, slightly different input data, model physics, starts to diverge significantly near the edge of the domain from the GFS feeding boundary conditions.

Wouldn't this create some kind of discontinuity if the differences were great enough?

Thanks in advance.

[dtk]

1) Secondary question, if the NAM is only trusted enough to run to 84 hours, and people make comments about not trusting it past 48 or 60, why does the DGEX run more than 84 hours?

2) More importantly, if the prior GFS is used to set the boundaries, especially since upper latitude weather sometimes retrogrades, and low level features at lower latitudes in the warm months sometimes come from the East, what happens if the solution of the NAM, maybe because of finer resolution, slightly different input data, model physics, starts to diverge significantly near the edge of the domain from the GFS feeding boundary conditions.

Wouldn't this create some kind of discontinuity if the differences were great enough?

Thanks in advance.

1. The DGEX was originally implemented to create something of a "downscaled/higher resolution (higher than the GFS)" forecast to cover the void from 84 hours to 7 or 8 days. This was all part of the push within the weather service to move to gridded products/database (NDFD). This was one means to fill the grids at realistic resolution in the medium range (particularly out west where terrain is such a vital component to sensible/surface weather).

2. This can happen on the west boundary as well (say where the GFS has a much more rigorous jet than the regional model). For this very reason, there is actually a buffer zone created (the distance/number of grid points varies by model) where blending between the GFS at the regional model boundaries and the regional model forecast itself occurs (in other words, there is wiggle room created intentionally as to avoid discontinuities, and relax the regional model back toward the boundary conditions). I'm not sure I explained this well...but I think you can get the gist....

[Ed Lizard]

1. The DGEX was originally implemented to create something of a "downscaled/higher resolution (higher than the GFS)" forecast to cover the void from 84 hours to 7 or 8 days. This was all part of the push within the weather service to move to gridded products/database (NDFD). This was one means to fill the grids at realistic resolution in the medium range (particularly out west where terrain is such a vital component to sensible/surface weather).

2. This can happen on the west boundary as well (say where the GFS has a much more rigorous jet than the regional model). For this very reason, there is actually a buffer zone created (the distance/number of grid points varies by model) where blending between the GFS at the regional model boundaries and the regional model forecast itself occurs (in other words, there is wiggle room created intentionally as to avoid discontinuities, and relax the regional model back toward the boundary conditions). I'm not sure I explained this well...but I think you can get the gist....

Thanks.