Application of an Objective Analysis Scheme to Mesoscale Observational Network Design

摘要

'Objective analysis' is a programmable method for estimating grid point values by transforming (interpolating) data from observations at irregularly spaced points into data at the points of a regular grid mesh. The gridded information can be used for diagnostic studies or in numerical weather production models. The objective of this thesis was to estimate the probable effectiveness of the proposed surface observational network (PAM (Portable Automated Mesonet) II) which will be deployed as part of the winter 1986 GALE field project. This network is to contribute to the study of mesoscale features within developing east coast cyclones and the coastal front. The process of spatial weighted-averaging interpolation of surface data was investigated by using the Barnes objective analysis scheme. Tests were conducted to determine the limitations of the Barnes analysis method and evaluate how well this technique reconstructs meteorological fields. A method was developed which enables an optimum weight parameter to be determined for a given data distribution. Barnes objective analyses were generally more accurate than kinematic fields based on detailed hand analyses. The objective analysis scheme effectively filtered random errors typical on instrumental and exposure effects from the analysis. However, interpolation near the edge of a data domain and extrapolation into data void regions led to erroneous results.