Challenge of Forecasting Urban Weather with NWP Models Fei Chen 1 , Yubao Liu 1 , Hiroyuki Kusaka 1 4 , Mukul Tewari 1 , Jian-Wen Bao 2 Chun Fung Lo 3 , and Kai Hon Lau 3 1 National Center for Atmospheric Research, Boulder, CO 2 National Oceanic and Atmospheric Administration (NOAA), Boulder, CO 3 Hong Kong University of Science and Technology, Hong Kong 4 Central Research Institute of Electric Power Industry (CRIEPI), Abiko, Japan 1. INTRODUCTION Mesoscale numerical weather prediction (NWP) models in connection with increasing capacities of computers in the last few years have considerably increased the spatial (vertical and horizontal) resolution. It is uncommon that some NWP models run with a grid-spacing of 0.5-1 km for local and regional weather forecasts. At such fine scales, the role of urban landuse in local and regional weather needs to be represented in these models and it is important for NWP models to capture effects of urban on wind, temperature, and humidity in the boundary layer and their influences on the boundary layer depth. Not only these boundary layer weather variables influence people’s daily life in the urban region, but also they are important input for air dispersion and quality models, which will benefit from improved prediction of the urban meteorological conditions. Having a consistent treatment of the planetary boundary layer structure and evolution in meteorological and air quality models is imperative. Errors in the improper parameterization of urban landuse result in bias in forecasted boundary layer variables and further in predicting the temperature and wind fields. The spatial distribution of urban landuse (e.g., building height, geometry) is highly heterogeneous even across urban scales. * Corresponding author address: Fei Chen, NCAR, PO Box 3000, Boulder, CO 80307; email: feichen@ucar.edu To explicitly solve the motions around an individual building or buildings requires the use of computational fluid dynamics (CFD) models, which are computationally intensive. Hence, in the foreseeable future, NWP models have to parameterize the subgrid- scale urban variability. Even in this context, it is not clear that which degree of complexity of urban landuse treatment should be incorporated in NWP models. For instance, Taha (1999) preferred a simple approach and pointed out that a large amount of detail in complex urban models may be lost when averaging back to a coarse model grid. In this paper, we report recent progress in urban landuse modeling for the community MM5 and WRF models. The goals of these efforts are to 1) capture important aspects of momentum, heat, and moisture transfer mechanisms caused by urban landuse, 2) assess the proper degree of complexity in urban treatment for NWP models applied with a grid spacing of 1-5 km, and 3) improve forecasted boundary layer structures over urban regions. Two different approaches were tested: 1) simply modifying the values of albedo, roughness length, soil thermal properties, and evaporation for the urban landuse in the Unified Noah land surface model (LSM), and 2) coupling a single-layer urban canopy model with the Noah LSM, which considers the 2-D geometry of building and roads to represent the radiation trapping and wind shear in the urban canopy.