WIRADA foundation data and modelling tools

We developed foundation data products and hydrological modelling tools to support WIRADA research and the Bureau of Meteorology's delivery of water information products.

Australian Water Resources Assessment (AWRA) modelling system

The AWRA modelling system underpins the interactive information on daily hydrological fluxes and stores across Australia delivered through the Bureau of Meteorology's Australian Landscape Water Balance website .

Data products used to develop and validate AWRA include:

  • curated and analysed daily streamflow from 700 unimpaired catchments
  • WIRADA gridded evapotranspiration products
  • satellite and in-situ estimates of soil moisture
  • modelled and point estimates of groundwater recharge
  • pedotransfer estimates of soil hydraulic parameters
  • land use types.

Flood and short-term streamflow forecasting

We developed modelling tools for flood and short-term streamflow forecasting. They are used in the Bureau of Meteorology's seven-day streamflow forecasting service .

These tools include:

  • ESDIIM (Ensemble Spatial Data Imputation and Interpolation Model) for generating ensemble climate surfaces from gauged data with non-concurrent and extensively missing data records
  • RPP (Rainfall Post-Processor) for post-processing rainfall forecasts from numerical weather prediction models to produce unbiased and reliable forecasts
  • ERRIS (Error Reduction and Representation In Stages) method for hydrological modelling  
  • SCHEF (System for Continuous Hydrological Ensemble Forecasting) to forecast out to nine days.

Seasonal streamflow forecasting

We developed modelling tools for seasonal streamflow forecasting. They are used in the Bureau of Meteorology's seasonal streamflow forecasting service .

These tools include:

  • BJP (Bayesian Joint Probability) statistical model, which has been the operational model used by the Bureau of Meteorology for seasonal streamflow forecasting since December 2010
  • WAPABA monthly water balance and partition model for rainfall-runoff simulation
  • CBaM (Calibration, Bridging and Merging) method for post-processing outputs from dynamical climate forecasting models, to produce ensemble seasonal forecasts of rainfall and temperature with improved skill and reliable uncertainty spread
  • FoGGS (Forecast Guided Stochastic Scenarios) model for producing ensemble time series forecasts of monthly streamflow out to 12 months
  • BMA (Bayesian Model Averaging) and QMA (Quantile Model Averaging) methods for merging forecasts from multiple models to improve skill.


Our informatics research provides the backbone to many water tools used within the Bureau of Meteorology. This includes:


WIRADA explored the blending of rain gauge and satellite rainfall to improve rainfall information across Australia. Methods for merging the spatially-comprehensive yet less accurate rainfall estimates from satellite sensing systems with the ground-based rain gauge measurements were developed and assessed to generate enhanced gridded daily rainfall products for Australia. The research demonstrated that the inclusion of satellite-derived rainfall can improve rainfall estimates, particularly for the vast expense of the interior of the continent. In addition, satellite-gauge blended daily rainfall is particularly useful for near real-time applications (like short-term hydrological forecast) where the number of reporting gauges is typically low.

Actual Evapotranspiration

WIRADA assessed the range of actual evapotranspiration (ET) products across Australia for water accounting and water resources assessment applications in an ET product Inter-Comparison Experiment (ET-ICE). The experiment found that the remote sensing based actual ET product (known as CMRS-ET) of Guerschman et al. (2009) worked best overall, especially in vegetated and irrigated areas. A CMRS-ET production system was then developed and implemented on the National Computational Infrastructure (NCI), which exploited the massive multiprocessing capacity of the NCI to produce long actual ET time series for the whole country in just a few hours.


Have an enquiry about this page?

Contact us