Abstract:
The goal is to build up an inverse model capable of finding the release history of atmospheric pollution by using measured gas concen¬tration data at just one location on the ground and identify the factors which affects the accuracy of the model predictions. The problem involves a non-steady point source of pollution at a known location in the atmosphere. This problem of finding the release rate is an ill-posed inverse problem and its solution is extremely sensitive to errors in the measurement data. Special regularisation methods, which stabilise the process of the solution, must be used to solve the problem. The method described in this paper is based on linear least-squares regression and Tikhonov regularisation, coupled with the solution of an advection-diffusion equation for a non-steady point source. The accuracy of the method is examined by imposing normally-distributed relative noise into the concentration data generated by the forward model as well as some real experimental data.
