Mm5

Hydrostatic/Non-

hydrostatic Vertical

Coordinate Equations Time differencing scheme Space differencing scheme Time step Pressure gradient force Thermal variable Turbulence parameterization

Hydrostatic

Advection form Forward-backward

4th order or Semi-Lagrangian advection Timesplitting Local reference

Ice equivalent temperature dei Based on dei and total water substances

Non-hydrostatic

Flux form

Runge-Kutta 3rd order

2nd to 6th order

Time splitting Universal reference

Potential temperature Based on temperature

Non-hydrostatic

Advection form Leap-frog

2nd order

Time splitting Universal reference

Temperature Based on temperature a a a p p p the prognostic equations for momentum; ice-equivalent potential temperature 9ei; turbulent kinetic energy (TKE); surface pressure; all phases of water, including ice, liquid, snow, rain, graupel, and vapor (Lin et al., 1983; Chern, 1994; Haines et al., 1997; Chen and Sun, 2002); and multilayers of soil temperature and wetness, etc. (Sun and Wu, 1992; Bosi-lovich and Sun, 1995; Chern and Sun, 1998; Sun and Sun, 2004; Sun and Chern, 2005). Because 0ei and total water substance, which are conserved without precipitation, are used as the prognostic variables, the model is able to include a comprehensive turbulence scheme, as discussed by Sun and Chang (1986a,b), Sun (1986, 1988, 1989), Sun (1993a,b), and Chern (1994). The PRCM also includes radiation parametri-zations (Liou et al., 1988, Chou and Suarez, 1994; Chou et al., 2001) and cumulus para-metrizations (Kuo, 1965, 1974; Anthes, 1977; Molinari 1982). The forward-backward scheme is applied in the Arakawa C grid to permit better computational accuracy and efficiency (Sun, 1980, 1984a). The fourth-order advection scheme (Sun, 1993c) is used to calculate the advection terms. A local reference is applied to calculate the pressure gradient force, which significantly reduces the error of the pressure gradient terms in the ap coordinate over steep topography (Sun, 1995a). Recently, the transport of dusts and an atmospheric chemistry module have been added by Yang (2004a,b). We are incorporating the mass-conserved, positive-definite semi-Lagrangian scheme (Sun et al., 1996; Sun and Yeh, 1997; Sun and Sun, 2004; Sun, 2007) and the sea-ice-mixed layer ocean model (Sun and Chern, 1998) into the PRCM, as shown in Fig. 1.

2.2. Weather simulations

The PRCM has been applied by Sun and Hsu (1988) to simulate cold air outbreaks over the East China Sea. They showed that the cloud, which has quite different properties than the cold air beneath the cloud base, was formed due to the warm and humid air being lifted by the cold air. This is in good agreement with observations. Hsu and Sun (1991) used the PRCM to simulate one of these cold air outbreak events, reproducing the three-dimensional mesoscale cellular convection, which has a horizontal wavelength of 20-30 km. The model has also been used to study air mass modification over Lake Michigan (Sun and Yildirim, 1989), baroclinic instability and frontogenesis (Sun, 1990a,b), as well as cyclogenesis and the life cycle of cyclones (Yildirim, 1994).

Sun and Wu (1992), applying the PRCM, were the first to successfully simulate the formation and diurnal oscillation of a dryline. Their results show that under a favorable combination of a strong soil moisture gradient, a terrain slope, and a vertical wind shear in Oklahoma and Texas in the late spring and early summer, the dryline can form within 12 hours in the absence of an initial atmospheric moisture gradient. Their results also show that the dryline moves eastward during the daytime, due to a strong mixing of the air near the dryline with the warm and dry westerly wind aloft (which descends from the Rocky Mountains) on the west side, quickly diluting the low level moist air coming from the southeast. At night, the cool, moist air on the east side continues moving into the deep, (still) hot, dry air on the west side, but is no longer vertically mixed and dissipated because the convection ceases to develop due to longwave radiative cooling at the surface. Hence, the dryline moves westward at night, as shown in Fig. 2. Their simulations (Fig. 3) also reproduce a deeper intrusion of the moisture field far above the inversion of the potential temperature field, as observed [Fig. 9 of Schaefer (1974)]. A low-level jet, low-level convergence, and strong upward motion form along the dryline, which are consistent with the inland-sea breeze theory proposed by Sun and Ogura (1979; schematic diagram in Fig. 12 of their paper). These simulations provide an explanation for the frequent occurrence of a dryline in the Great Plains in the late spring and early summer, which becomes a favorable zone for storm development.

Chern (1994) has successfully simulated the two surface low pressures observed with severe winter storms (Fig. 4) in the US, which shut down the highways in the High Plains for a week but, at that time, were poorly predicted by the NWS model. With the PRCM, Haines et al. (1997) successfully simulated the ice and super-cooled liquid water observed by aircraft, the lee vortex observed by radar, and surface precipitation of the Denver basin's Valentine's Day storm (VDS) on 13 February 1990.

Figure 2. Time-space variation of humidity qv at z ~ 10 m, contour interval of 1 g kg 1 labels scaled by 10, dryline moves eastward during daytime but retreats at night (Sun and Wu, 1992).
0 0

Post a comment