New insights into biogeochemical processing gained from sub-daily river monitoring

This talk will focus on the insights obtained from sub-daily hydrochemical monitoring for a sustained time periods (> 1 year), at multiple sites within a catchment and across different catchment types. Sub-daily instream hydrochemical dynamics were investigated, using non-stationary time-series analysis techniques, for two catchments representative of upland and lowland UK. The River Hafren at Plynlimon, mid-Wales drains an upland catchment where half the land cover is unmanaged moorland and the other half is first generation plantation forestry. The Hafren was monitored at two sites on a 7-hourly basis, between March 2007 and January 2009, using a Xian automatic sampler. The River Enborne, Berkshire, southeast England, is a rural lowland catchment, impacted by agricultural runoff, and septic tank and sewage treatment works discharges. The Enborne was monitored on an hourly basis between November 2009 and February 2012, using in situ field deployable analytical equipment to measure: Total Reactive Phosphorus (TRP: Systea Micromac C), Nitrate (Hach-Lange Nitratax), pH, dissolved oxygen, conductivity and water temperature (YSI 6600 Multi-parameter sonde). The results reveal complex diurnal patterns which exhibit seasonal changes in phase and amplitude, and are influenced by both flow conditions and nutrient sources. The comparison of the upland and lowland nitrate time series highlights how the different nitrogen sources within each system results in marked differences in the seasonal and diurnal dynamics, with a seasonal maximum in winter and a single peak diurnal cycle in the upland system, compared to a summer maximum and a two peak diurnal cycle in the lowland system. The analysis of TRP and nitrate concentrations in the Enborne catchment, in combination with flow, pH, dissolved oxygen, conductivity and water temperature, allowed the main processes controlling the observed sub-daily nutrient dynamics to be investigated. The different monitoring approaches adopted revealed the complexities involved in the accurate extraction of diurnal dynamics under lower frequency sampling, and the inherent issues of aliasing. Monitoring for 2 years also allowed an initial assessment of the inter-annual variability in the observed dynamics.