Modelling Rating Curves to Manage Uncertainty in the Fitzroy River Catchment in North West Western Australia

Lauren Greening, Leith Bowyer and Dean Pegoraro - Department of Water and Environmental Regulation, Perth, WA , 29 July, 2019

Seventeen streamflow gauging stations have been commissioned in the Fitzroy River catchment from 1955 to 1997, thirteen of which are currently operating. Stage-flow relationships at streamflow gauging stations are dependent on discharge measurements (gaugings) to relate the water level to a flow rate. Effort to record high flow discharge measurements has occurred when possible and continues to be a priority, however the majority of stations don’t have the number of measurements required to generate a reliable stage-flow relationship (rating curve). Hydraulic modelling to understand stage-flow relationships for high to mid flows has been adopted in the absence of field measured discharge at the majority of gauging stations in the Fitzroy River catchment to enable a time series of streamflow to be generated for use in hydrologic analysis.

Hydraulic models of the Looma (802007) and Fitzroy Barrage (802003) gauging stations in the Lower Fitzroy catchment have been used to develop the stage flow relationship. Both gauges were modelled in one dimension using HEC RAS in the late 2000s and a rating was developed and adopted although it was acknowledged that a two dimensional model would be preferential to capture flow through multiple braided channels across the 10 km wide floodplain. The same high flow discharge measurements were used to calibrate both models that were taken at a gauged cross section in between the two sites. A reliable relationship exists to transpose the discharge measurements to the Looma gauging station, however no relationship is available to transpose the measurements to the Fitzroy Barrage gauge and numerous hydrologic interactions occur between the two gauges which are difficult to define. Both gauges have been re-modelled in two dimensions using HEC RAS as the available input data, software and computing power are now accessible. The same discharge measurements are used for calibration with the same uncertainty in relation to the Fitzroy Barrage gauge. Near the completion of this project, an additional high flow discharge measurement was captured in 2017, used to validate the modelled rating and also to verify the earlier
discharge measurements. The two dimensional modelled rating at the Looma station is similar to the previous rating up to the highest discharge measurement, however flow is greater for the same water depth above this height, up to 18% greater at the high end of the rating. The two dimensional modelled rating is more accurate at this location, because of the more detailed representation of the geometry, and has been adopted. The two dimensional modelled rating at the Fitzroy Barrage gauge is the same as the current rating which confirms the one dimensional model, however uncertainty remains in the calibration data. Sensitivity analysis of the calibration variable Manning’s n shows that up to 25% uncertainty exists in the applied rating.

Modelled stage-flow rating analysis means the number of discharge measurements required at a gauging station can be minimised, but over time the number of gaugings across a network of gauges can be maximised. Field measurements can be targeted and fewer in number, whilst enabling streamflow data to be generated within a quantified level of uncertainty. Quantifiable uncertainty allows the data to be used in management decisions.