FLOOD06805 - Laserfiche WebLink

Home Browse Search

model can produce quantitatively good precipitation estimates both in magnitude and spatial location. These results include mesoscale phenomena with considerable spatial variation. He indicated a 2.7 Ian grid had be(:n used and increased resolution would be desirable for convection. He thought increased spatial resolution would also increase precipitation in some locations. He would like to see the model simulate rain and then include the simulation of run-off. Lou Schreiner presented an outline of how hI: saw the Bureau of Reclamation using large models. He is primarily interested in estimatl:s of Probable Maximum Precipitation which is the maximum storm that nature can produCe. He would like the models to be able to help with estimates ofP:MP on the plains and in the mountains, in regard to the transposition of storms from one location to another, and in variations with elevation. He plans to utilize the existing models. John Snook has been running the CSU-RAMS and the NCAR-MM5 model in essentially an operational mode at a horizontal resolution of 10 Ian for Colorado starting each 12 hours. He used a special NOAA system to obtain data analyzed on a 10 Ian grid for initialization. At present the system is constrained by computer resources and real-time operational requirements. His evaluation is that the mesoscale models are capable of providing reliable answers to large convective storms. He agreed that the grid spacing must be smaller than 10 Ian. Harry Orville (South Dakota School of Mines and Technology) spoke primarily about the Black Hills of North Dakota which are dimensionally approximately 200 Ian in the north- south direction and 100 Ian in the east-west direction with elevations much lower than the Rocky Mountains. He showed a series of very detailed simulation results for the 1972 flood event in the Black Hills. This event had up to IS inches ofrain. He talked about 2D and 3D simulations. He thinks quite a lot ClUl be learned from the 2D simulations, but the 3D model results are needed to locate the sturm relative to the topography. c. Discussion The discussion following the presentation in<;Iuded all of the participants and was orientated to four topics which included: . capability to simulate events, . capability to verify events, . time and cost. Most of the time was spent on the first of th~se topics but a summary of each is given here. Each group using the models (Clark, RAMS, MMS) thought the models could be used to simulate large convective storms successfully. Everyone agreed the grid spacing had to be near 1 Ian. More discussion emerged regarding initialization, cloud microphysics, surface vegetation and soil moisture, and use of2D versus 3D models. 21