Evaluation of numerical weather prediction models commonly considers the surface and near surface layers where more observations with greater detail are often available. For larger-scale models, publically available comparisons mostly consist of certain standard levels at set time intervals. Comprehensive and sophisticated statistical packages are available that can generate a variety of statistical measures, although they mostly consider the surface, near surface, and a limited number of higher levels, mostly within or near the boundary layer or at standard pressure levels. Evaluation of vertical soundings from model output at user-defined intervals vertically through the atmosphere up or near to the uppermost model output level would allow a more complete evaluation of model performance. An ability to readily compare profiles from two or more sources of data, such as a model and a radiosonde, will assist evaluation over those levels or layers. This report describes a series of methods for extraction of vertical soundings from the output of several representative models, including the Global Forecast System, the Global Air Land Weather Exploitation Model, and the Weather Research and Forecast model, as well as methods to evaluate single or multiple soundings from model output.
展开▼
