The traditional hydrological modelling workflow is to first calibrate a hydrological model against streamflow observations at each streamflow gauging station, then regionalise the model parameters from a nearest streamflow gauging station to a target ungauged area for streamflow prediction (Chiew et al., 2009; Zhang and Chiew, 2009). As a result, model calibration performance is noticeably better than model predictions.
The workflow used here, however, follows a regional calibration for the reasons outlined in companion submethodology M06 (as listed in Table 1) for modelling (Viney, 2016). Here, a regional calibration is first undertaken across multiple streamflow gauging stations in the broader area, and then this calibration is applied to all locations of interest (including ungauged). The regional model calibration results (Table 4 and Figure 8) suggest that the Australian Water Resources Assessment landscape model (AWRA-L) performs acceptably well at estimating streamflow in the Richmond river basin and its surrounding area when it is calibrated against in situ high streamflow and low streamflow, respectively.
It is noted that when the regional model is calibrated against observations from the nine streamflow gauging stations it does not generate a uniform model performance. Though the AWRA‑L model performs well overall, it performs poorly at some streamflow gauging stations. For instance, the high-streamflow model calibration generates a poor model performance at streamflow gauge 203030 and 204043 and the low-streamflow model calibration exhibits a poor model performance at streamflow gauge 204043 (Table 4).
Compared to the local model calibration, the regional model calibration performs similarly between the streamflow gauging stations used for calibrations and those used for predictions (Viney et al., 2014; Zhang et al., 2011). The nine calibration streamflow gauging stations cover a wide range of climate and topographic conditions, where mean annual streamflow varies from 201 mm/year at streamflow gauging station 203030 to 905 mm/year at streamflow gauging station 203002 (Table 4).
Furthermore, the performance of the high-streamflow calibration for the nine streamflow gauging stations does not appear to be significantly affected by catchment wetness, although it does perform slightly better with a wetter climate. The performance of low-streamflow calibration appears less sensitive to catchment wetness, with the exception of three of the drier areas where model performance is lower.
Product Finalisation date
- 184.108.40.206 Methods
- 220.127.116.11 Review of existing models
- 18.104.22.168 Model development
- 22.214.171.124 Calibration
- 126.96.36.199 Uncertainty
- 188.8.131.52 Prediction
- 184.108.40.206.1 Annual flow (AF)
- 220.127.116.11.2 Interquartile range (IQR)
- 18.104.22.168.3 Daily streamflow at the 99th percentile (P99)
- 22.214.171.124.4 Flood (high-flow) days (FD)
- 126.96.36.199.5 Daily streamflow at the 1st percentile (P01)
- 188.8.131.52.6 Low-flow days (LFD)
- 184.108.40.206.7 Low-flow spells (LFS)
- 220.127.116.11.8 Longest low-flow spell (LLFS)
- 18.104.22.168.9 Zero-flow days (ZFD)
- 22.214.171.124.10 Summary and conclusions
- Contributors to the Technical Programme
- About this technical product