Multi-model data fusion for hydrological forecasting

This paper outlines some simple data fusion strategies for continuous river level forecasting where data fusion is defined as the amalgamation of information from different data sources. The objective of data fusion is to provide a better solution than could otherwise be achieved from the use of single-source data alone. In this paper, the simplest data-in/data-out fusion architecture was used to combine neural network, fuzzy logic, statistical, and persistence forecasts using four different experimental strategies to produce a single predicted output. In the first two experiments, mean and median values were calculated from the individual forecasts and used as the final forecasts. These types of approaches can be effective when the individual model residuals follow a consistent pattern of over and under prediction. In the other two experiments, amalgamation was performed with a neural network, which provided a more flexible solution based on function approximation. The four individual model outputs were input to a one hidden layer, feed-forward network that had been trained to produce a single final forecast. The second network was similar to the first, except that differenced values were used as inputs and outputs. These various data fusion strategies were implemented using hydrological data for the River Ouse gauge at Skelton, above York, in Northern England. Neither the mean nor the median produced improved results, whereas the two neural network data fusion approaches produced substantial gains with respect to their single solution components. The potential to obtain more accurate forecasts using data fusion methodologies could therefore have significant implications for the design and construction of automated flood forecasting and flood warning systems.