The Power of the Variogram for Characterising Spatial and Temporal Patterns of Streamflow Variability.

Abstract:

Variograms are widely used in spatial statistics, as a way of examining correlations between points in space, but also have potential for application to temporal data. A variogram provides a robust and flexible way to quantify the temporal dependence (or autocorrelation, i.e. the dependence of flow on a given day with previous days) in daily river flow time series. There are, however, very few published examples of variogram techniques applied to hydrological datasets. The power of the variogram lies in its ability to characterise temporal dependence and, as such, describe the precipitation-to-flow relationship; as this is largely controlled by the catchment characteristics (e.g. elevation, soil type, rock type and land cover), it opens up a range of applications for characterising spatial patterns of streamflow regimes, as well as the dynamics of streamflow over time. This presentation describes a novel variogram-based method for investigating the influence that catchment characteristics have on moderating how streamflow responds to temporal changes in precipitation for a set of 116 catchments from across the UK. In the new approach, catchments are first classified based on the shape of streamflow variograms, with the classes being predictable based on catchment characteristics. The classification method therefore sheds light on how spatial variations in landscape properties influence the precipitation-to-flow relationship, and has significant potential for un-gauged site applications. Temporal variability is then assessed using a moving-window approach to index changes in variogram parameters over time, with a key benefit of the method being that different variogram parameters capture distinct aspects of the changing flow regime. Results demonstrate that precipitation alone cannot explain the variation in flow responses: catchment characteristics have a substantial role in moderating how a river responds to climatic variability, with the findings paving the way for assessments of the varying sensitivity of UK catchments to future streamflow change.