|
Abstract:
|
Behavior of air pollution is governed by the complex dynamics, in which air quality of a site is affected by the pollutants transported from neighboring locations via phys- ical processes. To estimate the source of observed pollution, it is crucial to take the atmospheric condition account. Traditional approach to build empirical models uses observations, but is not able to incorporate the physical knowledge. This drawback becomes particularly severe for the situations where a near-real time source estimation is needed. In this paper, we propose a Bayesian method to estimate the pollution sources, by exploiting both the physical knowledge and observed data. The proposed method uses a flexible approach to utilize the large scale data from the numerical weather prediction model while incorporating the physical knowledge into the model.
|
