This article summarises Colin Austin’s practical, innovation-driven approach to Australia’s water challenge. It explains why “competence research” often produces incremental gains, while real breakthroughs come from risk, trial, and many failed prototypes. It covers irrigation depth control, self-learning scheduling, micro-flood irrigation, and the unexpected invention pathway that led to wicking beds. The central message is simple: Australia has abundant rainfall in total, but we harvest very little. The solution is to extend the useful life of water with decentralised systems and better on-farm practices.
Why Innovation Matters in a Water Crisis
This paper is an extract from a talk titled “Solving the Water Crisis” (Part 2: Innovative Solutions to the Water Crisis), delivered by Colin Austin in March 2006. It is not a sales pitch and not a single-technology argument. It is a description of how practical innovation actually happens—how ideas are generated, tested, abandoned, reshaped, and sometimes turned into something simple enough to use in the real world.
Colin’s starting point is personal and technical. In the early 1970s, while lecturing at RMIT, he developed software that predicted the way hot plastic flows into a cold mould. At first glance, this seems unrelated to irrigation. The link is the mathematics: it is a “moving front” problem—exactly the same kind of challenge as water moving through soil during irrigation. If you can model the front, you can understand the process and, crucially, control it.
That software became Moldflow, a company that grew into a major Australian exporter of technical software, serving global manufacturers across electronics, automotive, appliances, and aerospace. Large corporations could have developed similar tools, but they did not. A small “mouse among giants” survived by doing something different: staying ahead through continuous innovation rather than head-on competition. Colin argues the same logic applies to water.
Competence Research vs Pioneering Research
Australia already has highly competent water research institutions, including large organisations with extensive resources. Competing directly—trying to do the same type of research, in the same way—is rarely effective for a small group. Instead, the talk contrasts two styles of research.
Competence research is disciplined and step-by-step. Each experiment follows logically from the last. It is essential work, but it tends to produce incremental improvements.
Pioneering research looks messy by comparison. It starts projects without a clear end result. It often cannot promise neat milestones. Most projects fail. But the failures generate new pathways. Resources are cut and redirected. The process “zig-zags” until a usable breakthrough appears. Colin describes this not as recklessness, but as a deliberate engine for producing ideas that conventional research pathways often miss.
Accepting Risk, Failure, and “Crazy” Experiments
Innovation requires accepting that some projects will look ridiculous and will fail. Colin gives two examples from early work in soil and irrigation, both aimed at the same goal: getting the right blend of air and water into soil. Soil needs both; water does not naturally distribute evenly; and oxygen availability strongly affects soil biology.
The Froth Machine: The idea was to pump an air-water froth through soil, so air would carry fine droplets into spaces water alone could not reach. A high-pressure fan blasted air through pipes while water was injected into the stream. In practice, the air simply rose to the surface in jets and the water drained downward. Total failure—though it did provide comic relief watching farm dogs trying to locate countless underground air jets.
The Expanda Pipe: A layflat pipe was ploughed into the ground and pressurised to expand and crack the soil. It was then drained, leaving a channel intended to deliver water directly into surrounding soil. It also failed. Yet both failures mattered: they helped shape later, more practical ways of achieving the “magic” air-water blend.
The point is not that failure is admirable. The point is that, in a complex real-world domain like water, the path to a usable system is rarely linear.
From Flood Irrigation to Subsurface Irrigation
The talk then moves into a practical engineering target: finding alternatives to conventional flood irrigation. Colin explored subsurface irrigation and immediately challenged the conventional approach. Standard subsurface systems tended to push water downward with limited sideways spread. That wastes water and misses roots.
The proposed change was simple in principle: inject water at much higher flow rates for short periods. The high flow forces water sideways by hydraulic action rather than letting it percolate straight down. The barrier was practical delivery: huge pumps and pipes are not viable across large areas.
The solution was to break an irrigated area into small sections and irrigate them in sequence. A simple sequential switch diverted flow automatically from one section to the next once water reached a target depth. This later revealed something bigger: controlling irrigation depth is central to both efficiency and environmental management.
Even where the technical approach worked, cost mattered. The system was not likely to be economically justified at the time. Meanwhile, many farmers were coping reasonably well by better scheduling during dry periods. So the research direction shifted again—another zig-zag.
Scheduling: The Real Irrigation Problem
Irrigation scheduling has two problems: knowing when to irrigate, and knowing how much water to apply. Colin argues that “when” is not the hard part. Farmers often have a good eye for crop stress, and simple instruments can help.
A plant and soil moisture sensing system was developed where the plant sensor signal stayed largely flat (aside from daily variations that could be removed mathematically) until the plant began to run short of water. A dip indicated it was time to irrigate. For practical use, a simple evaporation meter gave a decent indication of timing.
The harder problem is “how much.” If you consistently apply the right amount of water, then timing becomes less critical (provided you irrigate before stress). This leads to the next key idea: irrigation depth.
Irrigation Depth: The Simple Metric That Matters
Conventional scheduling often leans on “water holding capacity” as the guiding concept. Colin argues this can be misleading in practice because water does not spread uniformly through soil. It forms complex, shifting three-dimensional distributions. Plants do not take up water evenly. Rain disrupts everything. Measuring total water content accurately would require an impractical three-dimensional array of sensors outside a research environment.
Instead, after irrigation, the question is not “what is the average moisture?” but “how far did the wetting front travel?” In practical terms, that means: how deep did the water penetrate?
Depth is directly linked to outcomes. Too shallow and water is lost to evaporation. Too deep and water is lost below the root zone—wasting water and potentially driving nutrient leaching or salinity problems. Depth is also easy to explain and easy to measure.
Self-Learning Depth Control
The remaining challenge is that the relationship between “water applied” and “penetration depth” varies with many factors. Some factors change slowly (soil structure, crop stage). Others change quickly (existing moisture content). A fixed rule fails.
The proposed solution is an adaptive, self-learning approach: a predictor–corrector scheme. Data is measured before and after irrigation, analysed across multiple irrigation events, and the relationship is continuously corrected. In practice, the farmer specifies a desired penetration depth, and the system predicts how much water to apply to achieve it.
The talk also notes that irrigation depth should not be identical every time. Often, a series of shallower irrigations followed by a deeper irrigation is more effective than repeating the same depth on every event.
Africa: When “Working” Is Not Enough
A major turning point came when World Vision asked Colin to explore irrigation for sustenance food production in Africa. Low-cost tubing originally developed for subsurface irrigation was repurposed to feed conventional furrow irrigation. Testing in Australia looked promising. Field results in Africa were mixed. The lesson was blunt: it does not matter if a technology works in a controlled environment if local farmers cannot make it work reliably in their conditions, without extensive technical support.
This shift pushed the research back toward improving flood irrigation rather than replacing it—another loop in the zig-zag path.
Simulation and the Birth of Micro-Flood
To learn faster, Colin built a computer simulation of flood irrigation. Simulation allows hundreds of experiments in hours, providing insight that physical trials cannot match in time or cost. The model showed something counterintuitive: flood irrigation can be made highly efficient at a fraction of the cost of subsurface systems—if it is done differently.
The key was short runs and rapid flows. This led to a system called Micro-Flood. To make it practical, small areas were irrigated in sequence, controlled by a simple valve described as a “tilt” or “sloshing” valve. Technically successful, but still not a complete solution in settings with limited skills and limited extension support.
The “Plastic Film” Breakthrough and the Start of Wicking Beds
Some failures in the field were not caused by the concept, but by levelling problems or inadequate flow rates. Water pooled at one end or soaked too deeply at another. As an experiment, a polythene film was placed under a furrow to reduce downward soak. The result was dramatic: water travelled to the end and pooled more predictably.
The next step was bolder: shape the film into a trough under the furrow. Experts expected waterlogging and root rot. The opposite happened. Productivity improved sharply. The mechanism was different from conventional flood irrigation: irrigation became less frequent, creating an underground pool of water that wicked upward to keep soil moist rather than saturated. This avoided the damaging wet–dry cycles common in surface flooding.
From there, it was a small step to replace the furrow with a subsurface pipe, run lines horizontally, and allow water to cascade from bed to bed. This produced a highly effective and productive form of subsurface irrigation: wicking beds. The talk emphasises that wicking beds can be rain-fed or supplemented from external water sources, and they emerged through iteration rather than a single “eureka” moment.
Water Has Two Dimensions: Volume and Usefulness
One of the most important arguments in the talk is that irrigation efficiency is not always the core problem. In Africa, people can face hunger even in years with reasonable rainfall. A short dry spell during a nominal rainy season can prevent seed heads from filling and maturing, even if plants look green.
Engineers use the concept of entropy to describe energy not only by quantity but by usefulness. Colin applies the same idea to water: water has a quantity, but it also has a second dimension—its useful life. Extending the useful life of water can matter more than squeezing out small percentage gains in efficiency.
Wicking beds, in this framing, are not merely efficient irrigation devices. Their larger value is that they harvest and store water, keeping it available through dry intervals.
Implications for Australia: We Are Not Short of Water
Colin then makes a statement designed to challenge assumptions: if you take a rainfall map of Australia, estimate the total rainfall volume, and divide by population, the water received per person per day approaches one million litres. Australia is not short of water in total volume; we are short of harvested, accessible water with a long useful life.
Yes, rainfall is uneven. Some falls in the tropics and flows to sea. Some falls in deserts where few people live. But the talk argues these details do not change the central point: huge quantities of water fall on Australia, yet with conventional dam-and-distribution thinking we harvest only a tiny fraction—about 1 in 2,000.
The talk uses an energy analogy: we do not use premium high-pressure steam to heat houses, because it wastes “high-quality” energy on low-grade tasks. Yet with water, we often do the equivalent—using premium dam water (long useful life) when lower-grade, shorter-life water is frequently available for irrigation.
Rainfall Probabilities and Anticipatory Irrigation
Another practical idea is to treat rainfall as probabilities. In any region and season, there is a probability of receiving certain rainfall amounts. If you know the daily probability, you can estimate the likelihood of rain within a longer window. The talk gives an example: if there is a 1-in-20 chance of rain tomorrow, there is roughly a 64% chance of rain within the next 20 days.
The implication is strategic: if you can manage irrigation so crops can cope for 20 days without needing another irrigation, you can often avoid using premium dam water—on average saving large volumes. Some periods will still require dam water. Other periods will bring enough rain that irrigation can be greatly reduced. This, in turn, would allow dams to fill in wetter periods so they are full when drought arrives—something the talk notes has not been seen for many years because dams are drawn down even when alternative water is available.
Extending Useful Life with Decentralised Storage
Wicking beds are presented as one tool among many for extending useful life. Other available approaches include swales, leaky dams (with percolation holes), and twin dams—generally low-cost and accessible technologies.
Colin describes a practical example from orchard irrigation using a twin-dam system on an ephemeral creek. When rain filled the lower dam, it triggered a switch that filled an upper leaky dam and turned on irrigation—sometimes during the rain itself. The dams held only modest water volume; the bulk of storage was in the soil profile, slowly released as water infiltrated downslope. With one decent rain per month, external irrigation was often unnecessary, though some seasons required supplementation.
Tanks are another conventional option, but storage size becomes prohibitive for meaningful irrigation volumes. The talk gives an example: a tank roughly 5.4 metres in diameter and height stores about 120,000 litres, roughly a 30-day supply in that scenario. For many households or farms, that is expensive and impractical for irrigation as the primary solution.
By contrast, a wicking bed can store around 30 days of irrigation water without taking additional garden space. It can also recycle grey water and be supplemented by smaller tanks, creating a flexible, decentralised storage approach.
Using Plants to Harvest Small Rainfalls
The talk closes with an observation that looks simple, but has big consequences: plants themselves can be water-harvesting systems. An example is given of Egyptian spinach with an impressive root structure—deep roots plus surface hair roots—which “mops up” small rains and then effectively hibernates until the next event. Small rains of around 2 mm are normally lost to evaporation on the next hot day; plants with the right root systems can intercept and store that moisture in biomass and soil structure.
The Paradigm Shift
The conclusion is not that there is a free lunch. Extending useful life of water requires effort: designing systems, building them, maintaining them, and changing habits. The public is used to having large volumes of water delivered at minimal cost. Expanding conventional catchments is possible, but it can bring major environmental costs and may be unacceptable to much of the community.
Therefore, Colin argues that the real requirement is a shift in thinking—a change in expectations about how water is harvested and used, and a willingness to adopt alternative practices. Government leadership is presented as essential for this shift, because it influences infrastructure, incentives, standards, and the confidence required for communities and irrigators to adopt different models.
