Rayleigh Absorption Layer

20 km

20 km

Figure 4. Schematic of the cirrus model domain (from Liu et al., 2003b).

not show any obvious differences from those obtained using the 200 x 100 m2 grid cell configuration. There are four different time steps used in this model. In the dynamic module, a time step of 1.0 s is imposed for the large time step, while 0.1 s is used for the small time step. In order to reduce the computational time spent on radiative transfer, the radiation time step is usually set larger than the dynamic time step (Lin, 1997). So, the radiative transfer module is evaluated at 30 s. This indicates that the radiative heating rates remain constant during this 30 s period. The cloud microphysics module was evaluated every 0.5 s.

Liu et al. (2003b) have summarized the many simulations performed, including four kinds of atmospheric background profiles: warm unstable, warm stable, cold unstable and cold stable. The warm cirrus profiles are based on the US Standard Spring Atmosphere at 45° N. The cold cirrus profiles are based on the US Standard Summer Atmosphere at 30° N. Surface temperature is 15°C for warm cirrus, and 31.4°C for cold cirrus. The background tropospheric temperature lapse rate is 6.5° Ckm-1 in the former standard profile, while there is more structure in the latter. The tropopause occurs at

10.5km (-56.5°C) and 15.5km (-76°C) respectively. The background tropospheric relative humidity is set to 40% with respect to a plane water surface. Temperature lapse rates for the warm unstable case are ice pseudoadiabatic from 8.0 to 8.5 km, and 1°Ckm_1 greater than ice-pseudoadiabatic from 8.5 to 9.0 km. Similarly, temperature lapse rates for the cold unstable case are ice pseudoadiabatic from 13 to 13.5 km, and 1°Ckm_1 greater than ice-pseudoadiabatic from 13.5 to 14 km. For both statically stable cases, the lapse rates in these layers are set to 8°Ckm~1. Relative humidity with respect to ice is 100% at the base of the ice-neutral layer, and increases linearly with height to 120% at the base of the conditionally unstable layer, above which it is constant.

Some care must be taken regarding solar geometry, which is defined via a specification of latitude, date and initial local solar time (LST). We choose March 21 (vernal equinox) for the warm cirrus case, and June 21 (summer solstice) for the cold cirrus case, running the simulations from 1300 to 1600 LST. Since the simulations are run for 3 h, the corresponding solar zenith angle varies from 47.8° to 69.9° for the warm cirrus cases at 45°N, and from 14.9° to 53.4° for

0 0

Post a comment