|
<br />I
<br />I
<br />I
<br />I
<br />I
<br />I
<br />I
<br />I
<br />I
<br />I
<br />I
<br />I
<br />I
<br />I
<br />I
<br />I
<br />I
<br />I
<br />I
<br />
<br />active participation from academia, government
<br />laboratories, and the National Weather Service.
<br />The research and development of the
<br />microphysical modeling at 808MT under Prof.
<br />Orville's leadership have clearly influenced all of
<br />the microphysical packages, leaving a lasting
<br />legacy realized in the next generation mesoscale
<br />community model.
<br />
<br />As the horizontal resolutions of these models
<br />increase, our hope is that parameterizing the
<br />effects of subgrid-scale convection will become
<br />less important. The difficulty is that convective
<br />parameterizations were designed for use at
<br />resolutions coarser than typically 20 km; however,
<br />operational models are now being run at finer
<br />resolution (e.g., the 12-km Eta). Finer resolution
<br />runs are also made daily at NCEP over several
<br />smaller subdomains (nests) using the 8-km
<br />Nonhydrosfatic Mesoscale Model (NMM; Janjic et
<br />al., 2001). In support of Homeland Security, the
<br />model will be run at 4-km resolution in case of a
<br />national emergency. We are clearly in the middle
<br />of what some have termed the "mesoscale gray
<br />area", where the validity of the underlying
<br />assumptions in physical schemes in general, and
<br />cumulus parameterizations in particular, are
<br />questionable. Some have asserted that it will not
<br />be necessary to run a cumulus parameterization
<br />at 4-km resolution, but I think this is dangerous, if
<br />not plain wrong, given that convective updrafts
<br />and downdrafts can occur at small (<1 km)
<br />horizontal scales. A significant challenge to
<br />mesoscale forecasting (or possibly for all scales of
<br />motion?) is properly representing the complex
<br />interactions between grid-resolved microphysics
<br />and parameterized convection. For operational
<br />NWP, I try to focus on a balanced approach of
<br />improving all aspects of the cloud and radiation
<br />parameterizations, which means that the
<br />microphysics will tend to be simpler with more
<br />shortcuts than that adopted by other modeling
<br />groups.
<br />
<br />What I just said may appear to be completely
<br />opposite of what I did at GSFC, and in some ways
<br />this is true! But in other ways this is not true,
<br />because I have tried to adopt a flexible, customer-
<br />driven "no one size fits all" approach to my
<br />research. Current operational models at NCEP
<br />are resource limited, such that no one particular
<br />physics parameterization should use up more
<br />than 5-10% of the CPU resources. The new Eta
<br />microphysics (Ferrier et al., 2002) uses an
<br />average of 4-5% of the total CPU time for the full
<br />
<br />domain 12-km Eta runs. There were no such
<br />restrictions in improving the GCE model at GSFC,
<br />especially since the primary purpose was to
<br />improve TRMM spaceborne retrievals.
<br />
<br />10. PERSONAL OBSERVATION
<br />
<br />I can also recall circa 1981 how excited my
<br />former advisor, Bob Houze, was about the new
<br />SDSMT microphysics package. He gave me a
<br />copy of Chen-Hung Chang's Master's thesis, which
<br />I still have on my shelf. Little did I know then
<br />where the twisting road of cloud modeling took me.
<br />I have always appreciated the clarity of the
<br />theoretical treatment and mathematical formalism
<br />in Prof. Orville's papers, which I suspect reflects
<br />his excellence as an educator.
<br />
<br />I first had the privilege of meeting Prof. Orville
<br />during his visit to GSFC back in the early 1990s. I
<br />greatly appreciated the time he took to meet with
<br />me for several hours to discuss cloud
<br />microphysics. I was struck by how organized he
<br />was, recalling how he would look up the values of
<br />specific microphysical parameters in a laminated
<br />notebook. His kindness, thoughtfulness, and
<br />openness left a lasting impression. I appreciate
<br />the honor of participating in this Symposium.
<br />
<br />11. REFERENCES
<br />
<br />Adler, R. F., H.-Y. Yeh, N. Prasad, W.-K. Tao, and J. Simpson,
<br />1991: Microwave rainfall simulations of a tropical
<br />convective system with a three-dimensional cloud model.
<br />J. Appl. Meteor., 30, 924-953.
<br />
<br />Berry, E. X., and R. L. Reinhardt, 1974: An analysis of cloud
<br />drop growth by collection: Part II. Single initial
<br />distributions. J. Atmos. Sci., 31,1825-1831.
<br />
<br />Black, T. L., 1994: The new NMC mesoscale Eta Model:
<br />Description and forecast examples. Wea. Forecasting, 9,
<br />265-278.
<br />
<br />Black, R. A., and J. Hallett, 1986: Observations of the
<br />distribution of ice in hurricanes. J. Atmos. Sci., 43, 802-
<br />822.
<br />
<br />Bohm, H. P., 1989: A general equation for the terminal fall
<br />speed of solid hydrometeors. J. Atmos. Sci., 46, 2419-
<br />2427.
<br />
<br />Brown, J. M., T. G. Smimova, S. G. Benjamin, R. Rasmussen,
<br />G. Thompson, and K. Manning, 2000: Use of a mixed-
<br />phase microphysics scheme in the operational NCEP
<br />Rapid Update Cycle. Preprints, 9th Conf. on Aviation,
<br />Range, and Aerospace Technology, AMS, Orlando, FL.
<br />
<br />Farley, R. D., 1987: Numerical modeling of hailstorms and
<br />hailstone growth. Part II: The role of low-density riming
<br />growth in hail production. J. Climate Appl. Meteor., 26,
<br />234-254.
<br />
<br />11
<br />
|