Evaluation of the operational atmospheric model used in emergency response system at Kalpakkam on the east coast of India

• Published in: Atmospheric environment (1994) 2011-12, Vol.45 (39), p.7423-7442
• Main Authors: Srinivas, C.V., Venkatesan, R., Yesubabu, V., Nagaraju, C., Venkatraman, B., Chellapandi, P.
• Format: Article
• Language: English
• Subjects: air; Applied sciences; atmospheric chemistry; atmospheric circulation; Atmospheric dispersion; Atmospheric models; Atmospheric pollution; Climatology; coasts; Drying; Earth atmosphere; Errors; Exact sciences and technology; Global Positioning System; global positioning systems; humidity; Mathematical models; MM5; Model performance; model validation; Pollution; prediction; rain; Real-time predictions; Satellite navigation systems; temperature; Weather; weather stations; wind speed
• ISSN: 1352-2310; 1873-2844
• DOI: 10.1016/j.atmosenv.2011.05.047

The performance of a triple-nested mesoscale atmospheric model (MM5) implemented in the Online Nuclear Emergency Response System (ONERS) at Kalpakkam on southeast coast of India is evaluated. Real-time atmospheric model predictions are used to compute radiological plume dispersion in the mesoscale ranges using Lagrangian particle models. About 280 days falling in dry and wet weather and distributed in 2006, 2007, 2008 and 2009 years are considered. About 25 upper air, 100 surface station data including radiosonde, GPS sonde, micrometeorological tower and automated weather stations are used for model evaluation. Results indicate that model could reproduce the synoptic pressure, geopotential heights, winds and precipitation patterns in the coarse domain as well as the fine scale features of the atmospheric circulation in the inner fine domain. Model diagnosis with observations shows correlation (r), mean absolute error (MAE) and bias as (0.685, 1.87 C, 1.28 C) for temperature, (0.93, 1.55 hPa, 0.113 hpa) for pressure, (0.56,15 m, 0.53 m) for geopotential, (0.55, 12%, −10.5%) for humidity and (0.45, 2.3 m s −1, 1.70 m s −1) for wind speed indicating appreciable performance in the lower atmosphere for both dry and wet weather events. Model error in wind speed/direction reduced with height and slightly increased for temperature and humidity. Model performance is relatively better for dry weather cases than for the rainfall events. Also simulations from high resolution domain-3 are found to be better with relatively lower error metrics than those over coarse domains 1 and 2. ► Atmospheric model of emergency response system at Kalpakkam is evaluated. ► Forecasts compared with observations over different seasons. ► Observational comparison indicates mesoscale flows are consistently reproduced. ► Correlations are 0.685, 0.93, 0.56, 0.55, 0.45 for temperature, pressure, gpm, RH, wind. ► Simulations are better for dry weather cases than for rainfall events.