Keywords

integrated modelling; time-scales; river loads; coastal water quality; biogeochemistry

Location

Session B3: Integrated Hydrodynamic, Hydrological, Water Quality, and Ecological Models

Start Date

18-6-2014 2:00 PM

End Date

18-6-2014 3:20 PM

Abstract

Differences in the temporal resolution of catchment models and receiving water models often represent a problem for coupling of these models. Many catchment models are optimised for prediction of event-mean, monthly, or average annual sediment and nutrient loads, while coupled biogeochemical-hydrodynamic models typically run on time-steps measured in seconds, and often aim to predict patterns on a day-to-day or even sub-daily time-scale. Though previous work has shown that low temporal resolution of river boundary conditions set from in situ measurements can compromise the accuracy of a receiving water model, this is in large part because total river loads derived from such data are likely to be inaccurate. The case of a catchment model that accurately calculates total river loads, but at low temporal resolution, is different. To test the effect of varying catchment model temporal resolution, we ran a coupled, three-dimensional hydrodynamic-biogeochemical model of the Fitzroy Estuary and Keppel Bay, Queensland, Australia, using (a) daily concentrations (EMCs) of nitrogen and phosphorus species for the river boundary condition. All three cases used the same, daily-varying river flow and delivered the same total nutrient loads over the simulation period. The model was run for a three-month period during a very dynamic wet season, with two major flood events in different parts of the catchment. The results were compared in terms of simulated water-column water quality, total system primary productivity, and net exports of nitrogen and phosphorus mass to the Great Barrier Reef Lagoon. The results did not support the hypothesis that daily river constituent boundary conditions produce significantly different results than monthly or event-mean concentrations

Share

COinS
 
Jun 18th, 2:00 PM Jun 18th, 3:20 PM

Varying the temporal resolution of river nutrient boundary conditions to a coupled hydrodynamic-biogeochemical model of a coastal system has surprisingly little impact on model results

Session B3: Integrated Hydrodynamic, Hydrological, Water Quality, and Ecological Models

Differences in the temporal resolution of catchment models and receiving water models often represent a problem for coupling of these models. Many catchment models are optimised for prediction of event-mean, monthly, or average annual sediment and nutrient loads, while coupled biogeochemical-hydrodynamic models typically run on time-steps measured in seconds, and often aim to predict patterns on a day-to-day or even sub-daily time-scale. Though previous work has shown that low temporal resolution of river boundary conditions set from in situ measurements can compromise the accuracy of a receiving water model, this is in large part because total river loads derived from such data are likely to be inaccurate. The case of a catchment model that accurately calculates total river loads, but at low temporal resolution, is different. To test the effect of varying catchment model temporal resolution, we ran a coupled, three-dimensional hydrodynamic-biogeochemical model of the Fitzroy Estuary and Keppel Bay, Queensland, Australia, using (a) daily concentrations (EMCs) of nitrogen and phosphorus species for the river boundary condition. All three cases used the same, daily-varying river flow and delivered the same total nutrient loads over the simulation period. The model was run for a three-month period during a very dynamic wet season, with two major flood events in different parts of the catchment. The results were compared in terms of simulated water-column water quality, total system primary productivity, and net exports of nitrogen and phosphorus mass to the Great Barrier Reef Lagoon. The results did not support the hypothesis that daily river constituent boundary conditions produce significantly different results than monthly or event-mean concentrations