Performance Analysis of Weather Research and Forecasting Chemistry (WRF-Chem) Model in DKI Jakarta Area (Case Study: July 7–9, 2021)

Air quality predictions for operational purposes are still very limited, especially for the DKI Jakarta region. One of the air quality prediction models that is commonly used is WRF-Chem. Before the model is operationally implemented, it is necessary to evaluate performance to represent air quality, such as PM2.5 in the DKI Jakarta. In this study, GFS, MEGAN, and EDGAR data will be used as initial conditions and lateral boundaries in the WRF-Chem model. In this study, there are also two scenarios to test the effect of cumulus radiation feedback on PM2.5 distribution. Hereafter, the simulation result will be compared to meteorological observations using wind and temperature parameters, and air quality from PM2.5 data. The performance of model simulation is going to be analyzed using statistical methods. The simulation from two scenarios shows a visibly different result. Scenario 2 produces an average RMSE PM2.5 value for DKI Jakarta region, which is slightly lower than Scenario 1, 17 μg/m3 and both produce a negative correlation. However, the WRF-Chem model with Scenario 2 tends to be more representative than Scenario 1. This condition is supported by the simulation results on meteorological variables which are in accordance with the observation data and the model error values that tend to be better. In addition to the influence of feedback, the range of PM2.5 values for the WRF-Chem model is also influenced by EDGAR data that has underestimated values and does not match current conditions.