Precipitation Performance

I noticed the new NAM did a better job than the current NAM in forecasting precipitation over North Central Texas last Wednesday-Thursday. Since the forecast wind fields were nearly identical I wonder whether the difference was due to the nonhydrostatic nature of the new NAM, or the new vertical coordinate.

Bernard --

What was it about the precipitation performance did you think was better in the NAMX vs NAM? Was it structure? Location? Both? From what I've heard among forecasters, some like to see structure/banding/convective mode (if relevant) even if it's displaced from the verifying location. A graphic would also be helpful; they can be either attached or embedded through a link in the message.

Bill

The 12z April 5 run showed zero precip in NAM-Eta ("NAM") and light precip, convective, with a wave and front moving across north/central TX in NAM-WRF ("NAMX"). The primary difference in that run was in moisture, with an area of 10 mm more precipitable water in NAMX than NAM over northern Mexico initially. As this moisture advected northeast across TX, differences of 3 mm to 9 mm occurred ahead of the wave. (Also, due to slight location/timing differences in the wave, a narrow strip of much bigger differences shows up in a difference plot, but that's reflecting air mass change, not water content in the warm sector.)

The BMJ convective parameterization used in both NAM and NAMX (same parameterization is used) is very sensitive to moisture above the LCL (and below too, since that affects the LCL height and thus how much water is above it).

There are major differences in the way moisture is analyzed in the NAM-Eta 3d-var and the NAM-WRF GSI, with different moisture variable and more or better use of satellite data in the latter, as well as different pattern of covariances that affect the radius and shape of influence of the observation. Y'all out in the field looking at these runs should pay attention to where you think one model or the other has a better handle on initial moisture or where neither are good, and let us know!

Stephen