MPO 531, Spring 2009




  • Syllabus

    first presentations:
    Zhujun Li, precipitation initiation (Small et al. 2008)
    SiWon Song, hurricane microphysics
    Marcela Ulate Medrano, 1 vs. 2 moment schemes (Morrison et al. 2008)

    The class project:
    For this we examined consequences of WRF microphysical and cumulus parameterizations, comparing them to satellite observations garnered from the Giovanni NASA website. Marcela ran WRF simulations covering the Atlantic basin for a time period coinciding with a Saharan dust interaction with Hurricane Heleen (Sept. 12-20, 2006), and a 10-day-long Atlantic summer time period (June 16-24, 2007) thought to be climatologically typical. 3 micrphysical packages (Kessler, WSM 3-class, which has ice but no supercooled water or snow evaportation and the Morrison 2-moment scheme) and 2 cumulus parameterizations (Kain-Frisch and Betts-Miller-Janjic) were evaluated.
    project description

    Zhujun discussed the sensitivity of cloud fraction at different altitudes to microphysical parameterizations, comparing them to MODIS observations:
    writeup presentation
    Marcela examined the sensitivity of precipitation to the parameterizations, comparing model runs to TRMM and TMPA-RT:
    writeup presentation
    Eddy Hildebrand looked for hurricane-dust interactions and compared AIRS moisture/temperature to WRF values:
    writeup presentation
    Siwon assessed WRF representation of the mid-summer drought, evaluating both the 10-day June mean and a representative day.
    writeup

    Several observations we made: the Kessler scheme produces a much larger high-altitude cloud cover than the other schemes, which we attributed to the small fall speeds of liquid-only drops. In contrast, the Morrison scheme produces a conservative cloud fraction, with high cloud underestimated compared to MODIS observations. The Morrison scheme did produce precipitation values most close to TRMM-derived values. The Atlantic ITCZ was captured in the WRF simulations, though with underestimated mean precip values and at times overestimated hourly values, while the eastern Pacific ITCZ tended to be oversimulated by WRF. Simulations using Betts-Miller-Janjic match TRMM oceanic precipitation observations than simulations using Kain-Fritsch, while K-F appears to do better over land. (This suggests the cumulus parameterizations respond differently to different forcings of convections). Hurricane Helene was not captured at all by WRF. This may reflect our choice of a coarse spatial resolution (one degree), or basic modeling difficulties with cyclogenesis (requiring "bogusing" by more determined researchers). The SAL vertical structure appears captured by both AIRS and WRF near the African shore (elevated 700 mb temperatures, drier near-surface air), except that because none of the WRF simulations captured hurricane Helene, the interaction of the SAL layer with Helene was also not captured. In addition, no observational product is available with which to evaluate the SAL easterly jet, which the WRF simulations do produce.

    midterm
    final