Most of the precipitation in these "precip bombs" falls from the grid-scale scheme. Illustrating this is a plot of the one-hour precipitation in hundredths of an inch for the event over Kansas. The blue amounts are the total one-hour precipitation and the green amounts are the contribution from the convective parameterization. This is raw model output plotted from the BUFR data using the grid points nearest the indicated stations.
Over a three-hour period, the convective parameterization dropped 0.93" of rain on SLN, but the grid-scale scheme dumped 3.31", for a total of 4.24". Over the entire ten-hour event, the total forecast for SLN was 5.37" of which 1.68" was from the convective parameterization.
The events were characterized by the convective parameterization generating light precipitation at first, then the grid-scale scheme later kicked in, quickly intensifying up to a deluge, then dropping off at a station as the storm moved on, while the convective parameterization continued in small amounts for a few more hours. You can see this in the hourly sequence of maps like the above.
Click the Display Graphic button to the left to display the loop for the SD
event the first night. Note that there were no BUFR stations where the model
generated the most precipitation.
Click the Display Graphic button to the left to display the loop for the KS
event the second night