Stage 3 – Foundation Research

Sensors and Data Integration | Modelling and Forecasting | Spatio-Temporal Reporting and Visualisation

Early in 2006 CSIRO made a $9M commitment to develop technologies for the monitoring and management of water resources through the Water Resources Observation Network (WRON), which by 2010 will entail:

• A robust National Water Account available to all water users
• A National Water Sensor Network for monitoring water use and ecosystem health
• An accredited National Water Forecasting System (including a National Hydrological Model)
• A modular and flexible National Water Reporting System.

Research towards WRON will be provided by the Water for a Healthy Country National Research Flagship, with input from at least three different divisions of CSIRO (Land and Water, the ICT Centre and Mathematical and Information Sciences).

At this stage the research portfolio has been structured around three major streams:

1. Sensors and Data Integration
2. Modelling and Forecasting
3. Spatio-Temporal Reporting and Visualisation

Sensors and Data Integration

Next generation sensor networks could provide an order of magnitude improvement in scale, resolution and frequency of water data sampling at a fraction of the cost of current infrastructure.

Research in this area will develop new sensors to measure parameters such as water quality, levels, volume, velocity, or flow (for example).

Streaming real-time information will then be integrated with existing online historical datasets and models, providing interactive data analysis and information services to Australia’s water industry managers, operators, farmers and researchers.

Data integration requires standards for data exchange and communication, so this is a key element of the foundation research program.

Research will also focus on the development of new data capture techniques, such as the assimilation of remote sensing and in situ measurements. New algorithms, architectures and software will be required to integrate uncertain streaming and historical data, enabling real-time modelling, decision-making and reporting.

See also: The WRON vision for Data Integration and Sensing.

Modelling and Forecasting

We have the technology to vastly improve the power of predictive models. By linking sensor webs, remote sensing receiving stations, and distributed data to decision-support tools and grid computing technologies we can boost forecasting speed and currency. With the development of advanced statistical methods we can also make forecasts with more certainty.

At the present time, models are typically delivered to the desktop as distinct, stand-alone applications. This limits the computational power that can be brought to bear and makes auditing, archiving and analysing the results of model runs difficult.

Furthermore, models developed for a specific set of questions in a particular catchment do not readily translate to another catchment.

WRON will bring the delivery of models in the water resource domain up to speed with the latest advances in technology for research and management.

The approach is modular, where model components are selected in a web services framework. Calculations can be performed on a remote server farm or supercomputing facility, rather than on the desktop. This will have many advantages, including speed of model calculation, transparent auditing and archiving and flexibility of model delivery. The increased speed of calculation will enable many more scenarios to be tested and facilitate sensitivity and uncertainty analysis as a routine component of model delivery.

Research projects will:

• Develop new model/data assimilation techniques
• Create innovative applications of dynamic spatial simulations
• Minimise error propagation and quantify uncertainty
• Investigate distributed model calculation.

See also: The WRON vision for Forecasting.

Spatio-Temporal Reporting and Visualisation

A national water resource reporting system needs to be modular and flexible to enable more automated, cost-effective and responsive reporting.

Such a system would allow reports to be tailored to suit the individual needs of government or industry, whether at a national, state, regional or local council level. Reporting standards also make it possible to consider information on a comparative basis.

The first step for WRON research is to address standards and potential tools for reporting and visualisation, to:

• Work with vast quantities of data
• Provide confidence in the quality of the data or models
• Support a range of users (from operation managers, making decisions in near real-time, to regulatory authorities and policy-makers)
• Automate interpretation of information (to the level required by the user)
• Generate scenarios to enable ‘what if’ and ‘actual versus forecast’ advice to decision-makers
• Address the need for privacy
• Allow reports to be compared, combined and otherwise integrated
• Enable feedback to inform modification and improvement
• Archive all data, model runs and reports for future reference

There will be a tremendous amount of foundation research activity behind the scenes. New algorithms, architectures, standards and software will concentrate on the interpretation of data, models and forecasts.

With WRON it will be possible to visualise the status of Australia’s water resources, including trends in key areas of interest, and present a variety of scenarios to assist with management decisions.

See also: The WRON vision for Reporting.