| Abstract: | Assessment of man's impact on fluvial systems frequently involves extrapolation in time and space from data collected as part of a water quality network. Extrapolation represents information transfer and is used for scenario building and risk assessment. Transfer functions for non-conservative quality variables are poorly understood. Metals, phosphorus, and organic contaminants, which are preferentially transported in association with suspended sediment, are especially susceptible to extrapolation errors. Information loss, which is inherent in extrapolation procedures, is characterized for sediment-associated substances by examining the assumptions and limitations of the sediment delivery ratio. A practical demonstration of scale effects and interpretive difficulties is provided using data from the Bow River, Alberta. Guidelines for data assessment and extrapolation are summarized for temporal effects and for three different spatial scales. |
