AQEx evaluation package

AQEx is an automated evaluation package developed in R to assess the performance of the AQFx air quality forecasting system. It uses a suite of continuous and categorical verification statistics, leveraging the OpenAir package (Carslaw, 2015), to provide detailed insights into model accuracy and reliability.

Key Features:

• Visual Diagnostics:
  • Time series plots of observed vs. modelled pollutant concentrations
  • Scatter plots for 1-hour and 24-hour PM₂.₅ comparisons
  • Quantile-Quantile (Q-Q) plots to assess distributional fit
  • Diurnal variation plots
  • Taylor diagrams for model skill assessment
  • Timeseries heatmap
• Statistical Evaluation:
  • Common model performance metrics (e.g., bias, RMSE)
  • Categorical scores including:
    • Probability of Detection (POD)
    • False Alarm Ratio (FAR)
    • Critical Success Index (CSI)
    • Success Ratio (SR)

These tools help evaluate how well AQFx captures the onset, duration, and magnitude of smoke plume events across different locations.