Linking monitoring and modelling: can long-term datasets be used more effectively as a basis for large-scale prediction?

摘要: Data from long-term monitoring sites are vital for biogeochemical process understanding, and model development. Implicitly or explicitly, information provided by both modelling must be extrapolated in order to have wider scientific policy utility. In many cases, large-scale utilises little of the data available monitoring, instead relying on simplified models limited, often highly uncertain, parameterisation. Here, we propose a new approach whereby outputs applications upscaled landscape using simple statistical method. For 22 lakes streams UK Acid Waters Monitoring Network (AWMN), standardised concentrations (Z scores) Neutralising Capacity (ANC), dissolved organic carbon, nitrate sulphate show high temporal coherence among sites. This permits annual mean solute at site predicted back-transforming Z scores derived observations other The requires limited observational site, such as estimates two synoptic surveys. Several illustrative method suggest that it is effective predicting ANC change upland surface waters, may application. Because possible parameterise constrain more sophisticated with intensively monitored sites, extrapolation relevant scales this could provide robust, less computationally demanding, alternative application generalised input data.