Spectral Analysis of Chemical Time Series from Long-Term Catchment Monitoring Studies: Hydrochemical Insights and Data Requirements

Hydrological and hydrochemical data from long-term monitoring stations are vital for inferring travel times and flowpaths of water, and transport of contaminants through catchments. Spectral analysis is particularly powerful for studying the hydrological and chemical dynamics of catchments across a wide range of time scales. Here, recent work is reviewed that illustrates how spectral analyses of long-term monitoring data can be used to infer the travel-time distribution of water through catchments, and to measure the chemical retardation of reactive solutes at the catchment scale. For spectral analysis, it is desirable to have data sets with high sampling frequency and long periods of coverage. Using two data sets, a 3-yr daily data series and a 17-yr weekly data series from the Hafren catchment at Plynlimon, Wales, we demonstrate that high-frequency sampling (e.g., daily or more frequent) is particularly useful for revealing the short-term chemical dynamics that most clearly reflect the interplay of subsurface chemical and hydrological processes. However, data sets that combine high-frequency sampling during storm events with low-frequency sampling between storms can cause spectral artifacts and must be treated with special care.