Publications

Karmakar, S; Srinivas, CV; Rakesh, PT; Venkatesan, R; Venkatraman, B (2022). A WRF-FLEXPART simulation study of oil-fire plume dispersion- sensitivity to turbulent diffusion schemes. METEOROLOGY AND ATMOSPHERIC PHYSICS, 134(2), 32.

Abstract
In this work, we simulate the dispersion of oil-fire aerosols from the accident at Abqaiq oil processing facility in Saudi Arabia during 14-16 September 2019 using the Lagrangian dispersion model FLEXPART-WRF (version 3.1) to investigate the sensitivity of local and nonlocal turbulence parameterization schemes on the dispersion dynamics. The Weather Research and Forecasting (WRF version 4.0) model is used to simulate the meteorological parameters over the 2-day release period at a resolution of 2 km. Two Planetary Boundary Layer (PBL) schemes (MYNN level 2.5 and YSU) are used to simulate the boundary layer structure during the fire episode. Comparison with available surface and upper-air observations indicate fairly similar results, showing good agreement between simulation and observations. The FLEXPART-WRF v3.1 model is run with two turbulence diffusion schemes, Hanna and Hanna-TKE Hybrid. Meteorological predictions of WRF-YSU are coupled with Hanna diffusion scheme (hereafter YSU-Hanna) and predictions of MYNN2.5 are coupled with Hanna-TKE Hybrid scheme (hereafter MYNN-TKE). Simulated plume dispersion patterns are compared with Moderate Resolution Imaging Spectroradiometer (MODIS)/Terra and LANDVIWER Earth Observing System imagery. Though, in general the simulated plumes compare well with the satellite observed plume, MYNN-TKE simulates a wider plume beyond 20 km and YSU-Hanna produces more accurate plume pattern and plume width. Both schemes could reproduce the downwind variation of concentration. While YSU-Hanna slightly overestimates the concentration at all distances, MYNN-TKE underestimates for an initial 5 km and then closely follows observation derived data. While both schemes produces nearly similar behaviour of vertical concentration variation in both daytime and nighttime conditions, YSU-Hanna overestimates the concentration profile in the lower level region by approximately 25% compared to MYNN-TKE. Overall, YSU-Hanna performs slightly better than MYNN-TKE in reproducing the observed plume behaviour.

DOI:
10.1007/s00703-022-00866-w

ISSN:
1436-5065