| << | Integrated systems and Synergies/Assimilation | >> |
Abstract
Simulations of lower atmospheric flow patterns on scales of order 1 day with resolution of a few kilometers are commonly used for interpretation of field measurements, prediction of air quality, wind energy availability, or electricity demand. Measured wind profiles can be assimilated into retrospective simulations to improve their fidelity to reality, or to provide better initial conditions for predictions. It is vital to evaluate the impact of assimilation, which may not always be helpful. We will present results from WRF simulations of Houston and southeast Texas for 10 weeks of 2006. Data from three boundary layer wind profilers was assimilated by four-dimensional data assimilation (FDDA or nudging). Traditional statistics do not crisply display differences between runs. A new metric of sea breeze correspondence shows improved model performance at 7 surface sites with FDDA. The average wind over the middle of the day is a very good predictor of maximum ozone, and runs with FDDA are clearly better than without by this metric.