Stochastic models for characterisation and prediction of time series with long-range dependence and fractality

Time series from various fields, such as geophysics, meteorology, hydrology, air pollution, often display long-range dependence and small-scale behaviour (fractality). This paper develops a new class of stochastic models to represent such properties. An efficient estimation procedure is described and tested on two concentration time series collected in an environmental wind tunnel. These time series simulate two different types of odour sources and possess quite different statistical properties that are well described by the new model.