Modified ensemble Kalman filter for nuclear accident atmospheric dispersion: Prediction improved and source estimated

Optimization of an Urban Monitoring Network for Retrieving an Unknown Point Source Emission

“…In this study, 20 trials in various atmospheric stability conditions are selected and the meteorological variables are taken from an analysis of meteorological and micro-meteorological observations in Yee and Biltoft (2004) (Table 1). It is noted that the errors related to meteorological data can affects the accuracy of the source term 25 estimation (Zhang et al, 2014(Zhang et al, , 2015, although this error is not considered in this study. In each trial, the gas was continuously released for ≈ 15 min, during which the concentration measurements were made.…”

“…), Monte Carlo algorithms using Markov chains (MCMC) (Gamerman and Lopes, 2006;Keats, 2009, etc. ) and various stochastic sampling algorithms (Zhang et al, 2014(Zhang et al, , 2015. Deterministic methods use cost functions to assess the difference between observed and modeled concentrations and are based on an iterative process to minimize this difference (Seibert, 2001;Penenko et al, 2002;Issartel, 2005;Sharan et al, 2009Sharan et al, , 2012Kumar et al, 2015b, etc.).…”

Iterative ensemble Kalman filter for atmospheric dispersion in nuclear accidents: An application to Kincaid tracer experiment

“…M(x) is the dispersion model, Á is the model error, y 0 is the observation vector, m is the number of the observations, H(x) is observation model, and is the measurement error. The dispersion model is Lagrangian puff-model, the same as in [24]. A continuous release is modeled by the release of a discrete set of ellipsoidal clouds with Gaussian density distribution called "puffs".…”

“…Iterations are introduced to improve the estimates in the cases where the observation model is not linear. The new iterative EnKF method has been successfully applied in the data assimilation system for NPP accidents [24]. The results of twin experiments in our previous study indicate that the proposed iterative EnKF scheme effectively calibrates the errors in emission rate, plume rise height, wind speed and wind direction, and time delay of the EnKF is substantially alleviated.…”

“…The objective of this study is to thoroughly evaluate our proposed iterative EnKF data assimilation scheme in [24] with the data from Kincaid tracer experiment [25], which involves highly buoyant plumes. The errors in the horizontal wind will be corrected and the emission rates will be reconstructed using the iterative EnKF method.…”