Response of a Dynamical Model to Partially Produced Imbalance in Meteorological Fields

Abstract

 Starting from a set of meteorological data, the response of a dynamical model to the artificially produced imbalance in the mass field is pursued by making use of the pseudo-forecast initialization procedure. A three-level primitive equation model in the a-coordinate system is used as the computational model.
 The imbalance is produced over a subdomain located near the center of the whole region as follows:
 Case A : The surface pressure is replaced within the subdomain by its area average over the same subdomain.
 Case B: The same as case A, but for the geopotential height at the middle level of the troposphere.
 Case C: A constant value (i.e., 5°C) is added to the temperature field at the lowest level.
 Case D: The same constant value is added to the temperature fields at all the information levels.
 As the wind field, the geostrophic wind is used, which is estimated from the geopotential height with the variable Coriolis parameter.
 During the iterative calculation mutual adjustment toward the quasi-geostrophic equilibrium is expected to take place between the mass and velocity fields.
 For the imbalance extending throughout the total air column (cases A and D), the recovery of the pattern from the unbalanced state to the original one was rapid and sufficient. The result may be explained by the essential role of the external inertiogravitational wave in the adjustment procedure. When it is deep, the imbalance is rapidly dissolved by the external mode because of the fast phase speed of the high frequency wave. The adjustment is then mostly performed under the initiative of the velocity field. Thus the restoration to the original pattern is sufficient.
 On the other hand, for the imbalance located within the air column (cases Band C), ffie convergence to the original was slow and worse than in the previous cases. In the latter cases, the internal mode of the gravity-inertia wave played the main role. Thus the adjustment procedure was dull, and initiative was given neither to the mass field nor to the wind field to yield a different state of balance from the original.
 Although the implications of the present experiment are limited, the result may be applicable to the processing of asynoptic data in weather analysis and may account for the influence of erroneous meteorological information upon numerical prediction.