Data Assimilation Enhances WRF-Chem Performance in Modeling Volcanic Ash Clouds from Hunga Tonga–Hunga Ha’apai Eruption

Volcanic eruptions release large amounts of ash clouds and gas aerosols into the atmosphere, which can be simulated by air quality prediction models. However, the performance of these models remains unsatisfactory, even though both relevant physics and chemistry are considered. Hence, exploring the approaches for improvement such as inclusion of data assimilation is significative. In this study, we depict the modeling of the volcanic ash dispersion from the Hunga Tonga–Hunga Ha’apai underwater volcano, which erupted in a series of large explosions in late December 2021 and early January 2022. On 15 January 2022, a particularly significant explosion sent a massive ash cloud high into the atmosphere. We used the inline Weather Research and Forecasting model coupled with chemistry (WRF-Chem) and incorporated meteorological data assimilation within the Flux Adjusting Surface Data Assimilation System (FASDAS). We compared three forecast scenarios: one with only meteorology and no chemistry (OMET), one with gas and aerosol chemistry and no assimilation (NODA), and one with both chemistry and assimilation (FASDAS). We found that FASDAS resulted in lower planetary boundary layer height (PBLH), downward surface shortwave flux, and 2-m temperature by up to 800 m, 200 W m⁻², and 6°C on the land portion, respectively, while the opposite was observed near the eruption site. We validated the model against the observations and the results showed that FASDAS significantly enhanced the model performance in retrieving meteorological variables. However, the simulations also revealed significant biases in the concentration of volcanic ash around the ash clouds. Data from the Copernicus TROPOspheric Monitoring Instrument Sentinel-5 Precursor (TROPOMI-S5P) showed a westward trend of the total SO₂ emissions. This work demonstrates the significant contribution of data assimilation to the results of operational air quality predictions during violent volcanic eruption events.