Evaluating Air Quality Forecasts

RAP-chem

Jordan Schnell at NOAA GSL is using MELODIES MONET to evaluate the performance of a variety of research and operational air quality forecast models including a newly developed research forecast model called RAP-chem every day in near real time against the AirNow surface observations of ozone, PM2.5, CO, and NO2 and AERONET AOD measurements.

Check out the full analysis for each forecast here.

This includes a new feature developed by NOAA GSL summer student, Mackenzie Arnold, to interactively view plots for individual surface sites online.

The code to produce this analysis using MELODIES MONET is in the examples/forecast_evaluation folder on GitHub.

Example plots for ozone and PM2.5 for the forecast on January 18th, 2022 are below.

../_images/OZONE_EPA_f00_rapchem.png
../_images/PM25_EPA_f00_rapchem.png

UFS-AQM (AQM-Eval)

The UFS-AQM scientific team, in collaboration with NOAA-EPIC, developed a standardized MELODIES MONET model evaluation wrapper optimized for rocoto-based workflows running in HPC environments. The suite comprises packages targeting key measures of atmospheric composition model performance: chemistry, meteorology, particulate matter, and volatile organic compounds (VOCs). Automated workflow generation is currently implemented for the UFS-Short-range Weather App (UFS-SRW) with plans to generalize workflow creation and configuration to connect to a wider variety of modeling systems.

Code and wiki-based documentation are hosted on GitHub in NOAA-EPIC’s AQM-Eval repository.

Below is a sequence diagram providing an overview of major operations in the MM wrapper.

../_images/aqm-mm-eval-sequence.png