Dissolved organic C export is highly dynamic - capturing this variability and challenges in modelling

High resolution, field-deployable sensors offer opportunities to deepen our understanding of natural environmental systems, and measure the ';riverine pulse'. Studies utilising high-resolution equipment have demonstrated that sampling hydrological variables on traditional low frequency rates (such as once a week) creates a simplified picture of conditions that does not capture a true reflection of how fluvial systems operate. Dissolved organic carbon (DOC) represents a large and diverse mixture of compounds (including sugars, amino acids and humic substances) and concentration and composition of this pool varies globally. Understanding transport of this C pool in fluvial systems is important as it 1) represents the lateral export of C no longer sequestered in the terrestrial system, 2) surface water concentrations have been observed to have increased globally and we need to know if this trend is continuing and 3) when water is abstracted the purification processes removing DOC from the water, can create harmful by-products and so prior knowledge of inflow loading is valuable. Traditionally [DOC] has been measured using manual sampling methods, where a water sample would be collected in the field and returned to the lab. This approach can provide reliable data but the resource required to sustain this make it nearly impossible to measure the ';riverine pulse' through the information in long and detailed time series. In recent years new technology designed to estimate [DOC] in-situ has been developed. We have used one of these devices, which measures absorption in both the visible and UV wavelength regions of the electromagnetic spectrum and from this absorbance profile an algorithm estimates [DOC]. We have deployed this system in the field environment and after overcoming initial challenges have an almost continuous time series of [DOC], measured at 30 minute intervals, since May 2012. The logger has been functioning over a temperature range of 0.5 - 23 °C and a [DOC] range of 8 - 55.6 mg/l C, with the greatest shift in a single day being 23.5 mg/l C. We will present this highly dynamic [DOC] time series and also contemporaneous data from an in-situ water chemistry sonde profiling other measures of the ';riverine circulation system': pH, conductivity, temperature and stage height. The challenge now is how to allow data series of ~17,500 measurements per annum to interact to better understand and model drivers of carbon export. We are exploring the application of wavelet analysis to identify periods of coherence between [DOC] and these other variables. Our initial results indicate show that coherence with [DOC] can be intermittent and irregular, and so the challenge sensor technology presents continues.