WMOD00309 - Laserfiche WebLink

Home Browse Search

The vertical resolution was "stretched," meaning the vertical grid increment was smaller at lower levels than above. In the fmal configuration, the surface Az was 100 m and Az increased to 1000 m at and above 18 km (all elevations in MSL). a. Modeling Tracer Releases The model allows the release of tracer material from one or more locations. The tracer is input as Q = oM/dt (gm hr-1) and is initially distributed over an entire grid bin of volume axAyAz = 0.1 km3. The advective transport scheme of Smolarkiewicz (1983, 1984) is used in the model which requires the manipulation of a mixing ratio by mass, q, rather than a concentration (mass per volume). Following Bruintjes (1992); aq _ Aq _ Q _N__ I at At pAxAyAz {l} where p represents the base state of density upon which the model imposes perturbations. Since all the releases were on the surface, Az = 100 m. The equation of conservation of tracer is P- dq =p- aq + yr- (p- K f(;:q) dt at H I (2) where ~ is the eddy mixing coefficient. Tracer release points were input by latitude and longitude. For the runs described in this report, the tracer releases were turned on after the model had been integrated for one hour. In the generator portion, the tracer material was kept as a mixing ratio. This was converted to ~gm m-3 in the analysis portion.. Because of the wide range of tracer concentrations derived by the model, the tracer contour interval was determined by the analysis portion of the model. b. Warm Rain Parameterization The Clark model has an option which simulates warm rain production. The term "warm rain" implies liquid condensate which can be supercooled. In its present form, a modified form of the Kessler (1969) parameterization is used. A detailed description of the application within the rnodel is given by Clark (1979). The Kessler parameterization assumes there are two forms of condensed water: cloud water and rain water. Drop-size distributions, fall velocities and conversion from cloud to rain water are calculated from empirical arguments. This so-called "bulk-parameterization" simplifies the calculation of liquid condensation. The alternative is to account for specific sizes of drops, which is far more expensive in terms of computer time and memory, though probably more accurate. The model also has an ice phase option which follows Koenig and Murray (1976); however, the author has not applied this to the cases described in this report. -2- I I I I I I I I I I I I I I I I I I I