Pol(F)lux software, a dedicated tool to stream nutrient fluxes and uncertainties calculations for survey optimization

Abstract:

Data on stream material fluxes are essential for calculating element cycles (carbon, nutrients, and pollutants) and erosion rates from local to global scales. In most water-quality stations throughout the world stream fluxes are calculated from daily flow data (Q) and discrete concentration data (C), the latter being often the main cause of large uncertainties. This paper present the Pol(F)lux software tool, which addresses with two major issues: i) the selection of the optimal (minimal uncertainties) flux calculation method among 8 methods based on the flux variability matrix. ii) for the the discharge-weighted concentration method (the most commonly used method and recommended in the international convention for the protection of the North Sea and the Northeast Atlantic, OSPAR Convention), sampling frequency can be predicted to achieve a specified level of precision from the flux variability indicator (M2%, cumulative material fluxes discharged during the upper 2% of highest daily fluxes) through a nomograph for sampling intervals of 3 to 60 days. The software was validated for water-quality stations in medium to large basins (basin area>500 km²).

The flux variability matrix, the cornerstone of the Pol(F)lux software, is based on two indicators: (a) cumulative flow volume discharged during the upper 2% of highest daily flow, W2%, which characterizes the hydrological reactivity of the catchment during highest flow, and (b) the truncated b50sup exponent, calculated as the exponent of the relationship between concentration and discharge (in logarithmic scale) at the high-water stages (discharges greater than median flow), which characterize the behaviour of stream material. We postulate that performance is similar for stream materials found in the same flux variability class, composed of 4 classes of hydrological reactivity (W2%) and 5 classes of biogeochemical behavior (b50sup), defining 20 potential variability classes.