Australia reflects the world water situation well. Australia’s water is already recognised as an activity-limiting resource that is further challenged by population growth and climate change

In the early 1880s when Alfred Deakin argued the case for irrigation in Australia, he observed that irrigation was as old as human records – it is inseparably linked to the dawn of civilisation.

Natural flood irrigation in the Nile Delta sustained ancient Egypt and the extensive irrigation infrastructure in Mesopotamia underpinned the greatness of the Sumerian and Babylonian empires, as epitomised by the hanging gardens of Babylon.

Throughout time, the case for irrigation has always been made on the grounds of productivity and economic gains in agriculture. Deakin observed anecdotally a 10-fold productivity gain of irrigated land over dryland farming, but factors between three and five are more commonly reported. He also observed that water was not in short supply, but that Australia had a water-management problem that could be addressed through an irrigation system.

Essentially, any irrigation infrastructure redistributes, both temporally and spatially, hydro-cycle water so as to protect the crops against the vagaries of natural rainfall. As our understanding of hydraulics has matured this infrastructure has evolved, nevertheless the basic ingredients were already well mastered by the Sumerians.

At some altitude, run-off water is harvested and stored in dams. When water is required it is released from the dams. Under the influence of gravity it runs through a system of channels equipped with flow-regulating structures to irrigate the cultivated land. After its agricultural use most of the water re-enters the natural hydro-cycle, with typically less than one per cent of supplied water stored in plants and animals.

As the plain of Mesopotamia was an arid region, the Sumerians had to administer water allocations and managed their infrastructure carefully so as to achieve water allocation fairness. An extensive set of regulations and experience-based practices implemented by a dedicated temple-dependent labour force, overseen by a special cast of priests, saw to that.

Unfortunately they did not grasp the need for adequate drainage and, as a consequence of the oversupply of irrigation water, they eventually poisoned the soils as mobilised salts rose to the surface.

The world’s present water management is equally considered unsustainable because of the observed ecological deterioration of most of our natural water systems, but actually estimating the water resource that is readily and sustainably available for human activity is difficult. It is severely hampered by the paucity of quality data. Groundwater data in particular are scantily available.

Australia reflects the world water situation well. Australia’s water is already recognised as an activity-limiting resource that is further challenged by population growth and climate change: most of our urban communities are used to water restrictions; our farmers are subject to vastly reduced irrigation allocations; and our main river system – the Murray–Darling – suffers severe environmental stress.

In spite of this, our water infrastructure (like the rest of the world) is still largely fixated in the Deakin era, and our operational practices are definitely Sumerian/Babylonian in nature (although admittedly no longer temple-based) and suffer from low efficiency. This does not augur well for the future.

It is essential to modernise the water infrastructure so as to improve water efficiency and in general achieve a more sustainable socioeconomic and environmental water-management outcome within the ill-understood constraint of the available water resource. Irrigation is not a bad place to start.

Some progress can be made by investing in more civil infrastructure – better leak control, precision irrigation, water recycling, desalination and so on. However, these are inherently very expensive options and many come with a heavy and presently totally unwelcome energy penalty. Although it is clear that some of these investments are necessary and will lead to better water efficiency and water availability, the relative trade-offs are not well understood because of a lack of data.

Ours is a water-management problem, so greater gains can be achieved by improving the water decision-making processes through better information at all levels of decision making. More uncertainty (climate change) and tight constraints (demand pressure) combine to create a very challenging decision-making environment overall.

Australia, and the world, is poorly equipped to deal with this because our water infrastructure (both the natural as well as the engineered water infrastructure) is information-poor. The available data lack both the spatial and the temporal resolution necessary to support appropriate decision making. Neither the water user, nor the policy maker, nor the water infrastructure manager obtains sufficient data. Presently available water consumption data are uninformative about how and when to adjust consumptive behaviour.

The policy maker is in a vacuum as to what the actual trade-offs and constraints are. Water infrastructure managers try to meet demand, not knowing what the actual demand options or supply constraints are. At a point in time where the consequences of our actions are more apparent than ever, water resource management is severely limited by the lack of information.

By contrast, a well-informed water allocation would rely on a relatively long-term (multiple years) and large-scale (catchment size) prediction of the available water resource, taking into account all sources – surface water, run-off, stored capacity (including groundwater), desalination and recycling capacity. Such water allocation would be informed by demand (urban, rural, industrial and environmental) patterns, and trade-offs, which would depend on water pricing and water trading.

Moreover, the water allocation policy would necessarily have to incorporate infrastructure limitations and potentially should allow for the evaluation of infrastructure investments. It should be adaptive and responsive as new information becomes available. This requires the prediction of water supply/demand over some length of time, and this is limited by the chaos-induced complexity inherent in the underlying climate/weather/population dynamics.

In such a situation, even a conservative prediction, enabling resilient decision making, is immediately constrained by accuracy and timeliness of data that must feed computer models. Also the reliability of these models plays an important role, and this too is derived from data.

One could envisage such a water allocation policy as the outcome of a rather large mathematically stated, hybrid and heavily constrained, computational optimisation problem. By comparison, present water allocation policies are merely Sumerian.

The computational and algorithmic techniques to tackle such large-scale problems are within present technological reach, (some of the real recent advances in this area are developed within NICTA and IBM) but the measured data to make this a meaningful endeavour are largely lacking.

On a smaller scale, our own work has focused on improving the efficiency of bulk irrigation distribution, from dam to farm, by upgrading the existing civil infrastructure with an information infrastructure. A typical manually operated system of open channels is managed so as to oversupply water because the penalty of undersupply is crop loss. A water-distribution efficiency of between 50 and 70 per cent is considered good practice – which means less than 70 per cent of the water released from a dam goes onto a farm. Evaporation and seepage account for around a third of these losses, the rest is a combination of measurement errors and unused water.

Researchers at the University of Melbourne and NICTA in partnership with Rubicon Systems Australia have developed a system where the existing in-channel flow regulating structures are replaced by networked sensors and actuators. Using data-mining techniques, the sensor data enable the control of the channel dynamics in real time to achieve near on-demand water distribution that matches supply with demand and maintains channel water levels, which is the main other quality-of-service requirement in a gravity-fed flood irrigation network.

The water flow and level data derived from the sensor network also provide more detail about the non-management losses, evaporation and seepage, and enable more targeted channel maintenance.

The next logical step in this development of information-rich water management is to link water supply directly to crop demand, so as to achieve more crop per drop as advocated in the UNESCO World Water Reports. Our approach is to close an information loop between crops’ water needs, measured and predicted soil moisture deficit and the water supply, so the crop orders precisely its own water needs.

The rationale is based on the essence of timing, linked to crop development, as the main driver for irrigation efficiency. Early results in prototype systems at the University of Melbourne’s Dookie farm indicate that on-farm economic productivity can be significantly improved while greatly reducing water consumption.

There is much more that can be done – similar advances can be made in urban and industrial water settings. Moreover, not all water is created equal, and keeping track of water quality is important. A system-engineering approach that considers the entire water system is overdue.

The present water for growth responses in Victoria for example involve both new civil infrastructure as desalination and the lining or replacement of some open channels but also the implementation of an information-rich infrastructure in irrigation as well as a more comprehensive water trading framework. These are steps in the right direction.

In this line, building on Australia’s unique position as a first world country with a serious climate change issue, we have the opportunity to develop within Australia a world-leading climate change adaptation and sustainability technology with an appropriate regulatory framework of world-wide importance.

A comprehensive local response will be essential to Australia’s own future, and has the capacity to generate significant export income and improve our national GDP considerably. It is time to rebuild our irrigation infrastructure and lift it from its Babylonian heritage into the information age.
 
Iven Mareels FTSE is Dean of the Melbourne School of Engineering, a position he commenced in 2007. In 2008 he was the recipient of a Clunies Ross Award for his work smart irrigation systems. He has published four books, 125 journal publications and 225 conference publications and holds six international patents. His research interests are in large-scale systems, in particular with applications in water systems. He is also a Fellow of the Institute of Electrical and Electronics Engineers (USA), a member of the Society for Industrial and Applied Mathematics, a Fellow of Engineers Australia and a founding member of the Asian Control Association.