Present computing power allows fine grid models to resolve small mesoscale flows within individual valleys. This capability is of utmost importance for mountainous British Columbia because the valleys contain most of the population centers, industries, and transportation routes. Accurate fine-resolution forecasts depend on accurate and representative initial fields from which to start.

Dense surface observations are valuable data sources for mesoscale data assimilation and forecasting. How to best assimilate surface observations (particularly in complex terrain) into numerical models remains a challenging problem.

Built upon the recent success of objective analyses of potential temperatures in complex terrain using a mother-daughter approach, experiments are conducted to evaluate the impacts of assimilating surface observations on subsequent forecasts. The mother-daughter approach spreads surface valley observations along circuitous valleys in complex terrain, while reducing spread over mountain ridges. This work is to further refine the mother-daughter approach for mountain top observations, and to combine detailed surface analyses with coarse-resolution Eta analyses. To assess the impact on mesoscale forecasts of surface data assimilation, the assimilation runs started from the combined analyses are compared with control runs driven by unmodified Eta initial conditions. The preliminary results will be presented.