Using Inverse Modelling and Dual Isotopes (δ15N and δ18O of NO3) to Determine Sources of Nitrogen Export from a Complex Land Use Catchment

Attributing nitrogen export to specific land use within heterogeneous catchments remains a challenge due to the spatio-temporal variability in conditions influencing the mobilization and fate of nitrogen species. This study demonstrates that the measurement of dual stable isotopes of nitrate, taken along with routine tributary measurement of nitrogen in nitrate (NO3--N) and ammonium (NH4+-N), aids in apportioning sources of the overall nitrogen load during wet periods. An inverse modelling technique was developed to estimate the land use-specific export rates of NO3--N and NH4+-N from the Caboolture River Catchment in Queensland, Australia. Measurements of nitrogen in streamflow at 51 locations during six sampling campaigns (May 2012 to April 2013) were made along with catchment geospatial data that was used to disaggregate sub-catchments into six land use fractions. A hydrological model was applied to compute the runoff from each fraction and water routing through the stream network. This data was used within a nitrogen mixing model with inclusion δ15NNO3 and δ18ONO3. The land use specific export rate was computed inversely as the posterior of a Bayesian interference applied to the model. During higher rainfall periods when export rates were highest, the main land use exporting nitrogen was wetland (110 g/ha/day NO3--N, 27 g/ha/day NH4+-N) resulted from mineralization and nitrification of organic N, followed by urban (16 g/ha/day NO3--N, 2.3 g/ha/day NH4+-N). The advantage of using the dual isotopes in conjunction with the nitrogen concentration data was demonstrated by reduced uncertainty in the computed rates during the higher rainfall periods, relative to calculations without δ15NNO3 and δ18ONO3. Key Points An inverse modelling technique was developed to estimate the land use-specific export rate of nitrogen using isotopes signatures The model was embedded in within a Bayesian framework to assess the predictive uncertainty with and without inclusion of dual isotopes of nitrate During wet periods, when nitrogen export rates were highest, the predictive uncertainty was reduced when the model includes dual isotopes of nitrate Plain Language Summary Attributing nitrogen export to specific land use types within heterogeneous catchments is difficult. This paper demonstrates the use of an inverse modelling technique, dissolved inorganic nitrogen concentration data, a dual stable isotope measurement data and geospatial data to estimate the land use-specific export rates of dissolved inorganic nitrogen from six land use types in the Caboolture River Catchment in Queensland, Australia. The results show that during higher rainfall periods when export rates were highest, the main source of nitrogen was found to be wetland (110 g/ha/day NO3--N and 27 g/ha/day NH4+-N), followed by urban (16 g/ha/day NO3-N and 2.3 g/ha/day NH4+-N) land uses