Vast distances between Australia’s water monitoring networks gauges makes comprehensive national-scale water balance modelling and resource assessments extremely challenging – our research is helping fill the gaps with satellite observations.

Dr Luigi Renzullo, IWF Senior Research Fellow, and Dr Siyuan Tian, Postdoc Fellow from Fenner School of Env & Soc, have led a 2-year project in a Collaborative Research Partnership with the Bureau of Meteorology into the use of satellite observations to improve on the Bureau’s operational water balance modelling capability. Bureau of Meteorology provided the funding for this work.

The Bureau of Meteorology’s current National Water Resources Assessments and Accounts are driven by outputs from the Australian Water Resources Assessment (AWRA) modelling system. The operational AWRA system in the Bureau provides water balance modelling capability for the whole continent every day. However, the model is driven primarily by measurements from a sparse network of rain gauges with often 100’s – 1000’s of km between them.  This makes the AWRA model outputs unreliable for large parts of the country, thus limiting the Bureau’s ability to an accurate, comprehensive assessment of water in the landscape. As such, they limit their mandated reporting (under the Water Act 2007) to those part of the country where they believe their modelling is most reliable. The research explored the potential of satellite data to extend the reliability of the modelling into rural and central Australia, which would benefit many remote communities in the region.

The project’s main aim was to develop methods that could routinely make use of satellite sensing of soil moisture in the operational AWRA modelling system to improve water balance modelling and forecasting capability for the whole country. Three candidate satellite soil moisture data were explored: NASA-JAXA’s Soil Moisture Active Passive (SMAP); ESA’s Soil Moisture and Ocean Salinity (SMOS); and Eumetsat’s Advanced Scatterometer (ASCAT).  However, some satellite technologies work better in different parts of Australia than others. “Any two satellite systems, for example SMAP and SMOS, provides complete coverage of surface soil moisture across Australia every day, but there is lots of overlap”, Dr Renzullo explains. “So we developed a weighting scheme that ensures that the best satellite product dominates AWRA prediction for any given part of the continent.”

Drs Tian and Renzullo employed concepts of data assimilation – methods of combining numerical models with observations based on error statistics – to integrate the satellite data into AWRA water balance estimates for the whole country. In an independent evaluation of model estimates, the assimilation of satellite data was shown to improve the soil water representation in AWRA significantly. “The assimilation of satellite soil moisture was shown to compensate for the errors the rainfall, particularly in the gauge spares part of the country”, said Dr Tian. “More importantly the improvements were observed to persist for several days, even a few weeks for some AWRA model variables. This is a very important result with a positive message for water balance forecasting.”

The method developed is currently being integrated into the Bureau’s official operational AWRA modelling system. “It very is exciting to know that our work is helping to improve water balance modelling for the whole of Australia”, said Dr Tian.

In conjunction with:

NASA – National Aeronautic and Space Administration

JAXA – Japan Aerospace Exploration Agency

ESA – European Space Agency

Eumetsat – European Organisation for the Exploitation of Meteorological Satellites

Publications:

Tian, S., Renzullo, L.J., et al, 2019, https://doi.org/10.36334/modsim.2019.H6.tian

Tian, S., Renzullo, L.J., et al, 2020, Geophysical Research Letters, under review (preprint available here: https://www.essoar.org/doi/10.1002/essoar.10503460.1)