Virtual weather observations, forecasting and validation

点击翻译按钮获取中文摘要

Pelmorex Media Inc. via The Weather Network provide high quality weather information to consumers and commercial clients across Canada and in other parts of the world. Forecasts for approximately 20,000 sites in Canada are distributed via TV, internet and mobile phones with an emphasis on timely delivery. While current observations are available for about 1,000 of these forecast locations there are many more without in-situ observations. The aim of our research is to provide a methodology to generate best estimates with a "Virtual Observation Engine" (VOE). Pelmorex have a prototype engine using interpolations from available data (including about 1,000 weather stations in Canada and 7,000 world wide) coupled with numerical weather prediction forecasts. We will aim to improve the virtual observation engine through the use of data assimilation techniques, and careful down-scaling using simple guidelines or high resolution boundary-layer flow models. In addition York University (Dr Chen's group) are currently using the WRF (Weather Research and Forecasting) model and assimilating a range of data, including radar wind profiler data, in a joint, MITACS supported, project with Pelmorex. This is currently focused on Southern Ontario but will be expanded to test the use of this approach as a VOE for critical locations or areas across the country where other techniques have difficulty in predicting reliable virtual observations. Dr Taylor's group have worked with Pelmorex to assess the performance of their prototype VOE. In areas of even moderately complex terrain the effects of local topography and variations in surface properties (such as roughness and albedo) can play a significant role in determining local wind and temperature. The aim is to develop approaches which take better account of these factors within the Virtual Observation Engine approach.