Title: Aermod modelling of hydrogen sulfide (H2S) concentration from geothermal power plants in Ulubelu, Indonesia and Hellisheidi-Nesjavellir, Iceland

Abstract

The AERMOD model was evaluated with the aim to assess the applicability of the software to give
reasonable results, in estimating H₂S concentration from two geothermal fields affected by
different weather conditions. The study cases were geothermal emissions from the Ulubelu power
plants in Indonesia, and the emissions from the Hellisheidi and Nesjavellir power plants in
Iceland. The modelled H₂S distribution was also compared to observation H₂S values with periods of
up to one-year data. AERMOD was used to calculate the maximum concentration of 1-hour (odour
standard), 8-hour (occupational health standard), 24-hour and annual time averages (public health
standard). The test cases included different model setup of elevated and flat terrain options, as
well as various meteorological data (e.g. onsite and offsite). Overall, the model performed better
for a long-term period (annual) than a short-term period (1-hour and 24-hour), except for the
Ulubelu case, where the model at 24-hour period agreed well with the measurement data sample points
taken from up to 3 km from the source. In contrast, for the Hellisheidi and Nesjavellir case, the
models had difficulty in predicting the concentration at receptors within 25 km from the sources.
When evaluating the level of H₂S concentration based on seasons, the results of the model showed
higher concentrations during the winter season than summer season for the Hellisheidi and
Nesjavellir case. For the Ulubelu case, the predicted H₂S concentration during the dry season was
estimated to be higher than during the wet season. The study highlighted the influence of weather
conditions (i.e., wind stability in a tropical climate compared to cold weather) on the dispersion
of geothermal emissions, as well as the effect distance of meteorological stations, receptor´s and
source’s location, and terrain height have on the results of model simulations. The study shows
that the model simulation does not work well when the source is far away, the weather changes
rapidly and the terrain is complex. However, for stable weather conditions, it provides valuable
ations measures decisions, for instance, to define H₂S monitoring
station points at receptors, which indicates high concentration of H₂S.