Time series analysis of the world's longest fluvial nitrate record : evidence for changing states of catchment saturation.

Abstract

Processes that drive the occurrence of nitrate concentrations in surface waters are known to operate over many decades longer than the available observations. This study considers the world's longest water quality record of nitrate concentrations in the River Thames (1868–2009) in order to understand whether the nature of the time series has changed with time and such external drivers as climate change and land use of hydrology. The study considers the linear trend, the seasonality, the memory and the impulsivity relative to river flow of the time series for moving windows of 6 years in length. The study can show that: Time series analysis proved effective at discriminating controls upon the nitrate concentration in the long term as different components of the record respond to different drivers in different ways. There was decoupling of the annual minimum, annual maximum and the amplitude of the seasonal cycle. The nature of the time series is dominantly controlled by changes in source of nitrate and not by climate change. That even similar increases in nitrate concentration in surface waters can have distinct character that illustrates that they are the result of different sources of nitrate. Changes in the impulsivity of the record show that the study catchment has recovered from a state of saturation, but the memory effect shows that there is an increased contribution from a shallow groundwater.