Estimating Volcanic Sulfur Dioxide to Sulfate Aerosol Conversion Rates in Hawai‘i

Abstract

Volcanic smog, known as vog, has been a persistent issue on the island of Hawai‘i since the eruption of Kīlauea Volcano began in 1983. Vog, made up of sulfur dioxide (SO2) and sulfate aerosols (collectively referred to as SO4) poses a significant health risk to communities surrounding the volcano and on the leeward coast of Hawai‘i. The Vog Measurement and Prediction Project (VMAP) was launched by the University of Hawai‘i at Mānoa (UHM) and United States Geologic Survey in 2010 as a feasibility study to evaluate whether vog forecasts are achievable and usable. The UHM Vog Model has been operational since 2010 producing forecasts of SO2 and SO4 for the entire state of Hawai‘i. The UHM Vog Model’s purpose is to warn the public, particularly sensitive groups including the elderly, children, and those with respiratory problems, to avoid high concentrations of vog. More recently, the Hawai‘i Department of Health has expressed interest in using Vog Model forecasts to issue air quality alerts when projected concentrations exceed certain thresholds. An early evaluation of the model performance suggested that forecasts are poor for high-end concentrations of both SO2 and SO4, bringing the reliability of the model into question for issuing warnings. To address this shortcoming, a number of improvements to the model were considered. The most promising improvement was including a more comprehensive sulfur chemistry scheme to represent the conversion of SO2 to SO4. A new sulfur chemistry scheme was constructed for inclusion into the UHM Vog Model. This scheme was based on theory and past research of sulfur chemistry. To provide a baseline for the new scheme a brief field experiment was conducted in July 2015 to estimate the conversion rate of SO2 to SO4. The new sulfur chemistry scheme, along with other available schemes, were evaluated against observations of SO2 and PM2.5 around the island of Hawai‘i during November 2015. The results of this evaluation revealed the new sulfur chemistry scheme improved forecasts for SO2 and SO4. Also, as forecast windows extended from one hour to six hours the probability of detection for SO2 increased from 10-30% to 50-70%.