An extreme and persistent heat wave event hit Missouri in summer 2012. Current operational forecast models failed to predict such an event at a long lead. The objective of the current study is to simulate this extreme event using a high-resolution Weather Research and Forecasting (WRF) model with eight combined physical (including longwave/shortwave radiation, microphysics, and planetary boundary layer) parameterization packages. Integrated for one month, the model successfully simulates the spatial pattern and temporal evolution of surface air temperature, compared to in-situ observations. The interesting feature is an oscillatory development of the surface air temperature, with a pronounced synoptic timescale. Such a temperature evolution is consistent with the local zonal wind fluctuation, implying the important role of anomalous temperature advection.
An overall skill score that combines the performance of 2-m air temperature, relative humidity, and precipitation fields is defined. The result shows that the combination of Thompson, Rapid Radiative TransferModel for GCMs (RRTMG), and Mellor–Yamada–Nakanishi–Niino level-3 (MYNN3) schemes presents the best WRF simulation. A further analysis of this best simulation shows that the model successfully reproduces the urban heat island (UHI) effect in the Kansas City Metropolitan Area with realistic diurnal variation of 2-m air temperature in the urban and non-urban areas with a larger UHI effect at nighttime.