Have you ever carefully replanted a beautiful potted flower, only to watch it wilt in your yard because the sun was harsher or the rain heavier than what it was used to inside? Or have you memorized a presentation perfectly, then froze when you stood in front of a room full of people because the pressure changed everything? Or bought a shirt that you loved in the store, but never end up wearing it because it doesn’t actually work with your wardrobe? Transfer failures across contexts are common. We like to think that if something works in one instance, it will work in all instances. But the world holds more counterintuition than we often think.
Water interventions are not saved from these transfer failures either. A technology or strategy can work well in a controlled setting or pilot project, and we assume it will perform just as well when scaled or transferred into the real world. But once that intervention enters a living system shaped by people, incentives, norms and politics, the conditions change. The intervention does not fail because the science is wrong, but because water systems are also social systems, and humans shape how interventions function. To understand the importance of acknowledging transfer failures, let’s take a look at a few examples of what happens when it is not taken into consideration.
Drip irrigation
The problem: Drip irrigation systems are designed to reduce the amount of water required to irrigate crops, but farmers are using the same if not more water after adopting them.
Drip irrigation is an efficient form of irrigation that delivers precise amounts of water directly to the roots of individual crops. Compared to other irrigation methods that flood water across a field, this method minimizes runoff, evaporation and waste between crops. Within experimental conditions, drip irrigation reduced the amount of water consumed by up to 30-50%, while still maintaining happy and healthy crops. Technically speaking, this is a perfected solution in which we can produce the same output with less water.
But when drip irrigation was introduced to farmers in Jordan, total water consumption did not fall. In many cases, it increased. So, what happened?
“Because people are the users of technology, the ability of drip systems to reduce water use will depend on how farmers use them,”cites the opening to a research by IWMI’s Soumya Balasubramanya, David Stifel and Rachael McDonnell. The technology itself worked exactly as intended. What changed was how farmers responded to the new opportunity it created.
Farmers do not prioritize water conservation in isolation; they prioritize reducing risk. Water is often not seen as an end goal, but as an input into income, and therefore, stability. When drip irrigation improved efficiency, Jordanian farmers used those gains to strengthen their economic position. Some expanded irrigated land, while others shifted to higher-value, more water-intensive crops that generated greater returns.
In both cases, farmers’ market trips became more profitable because drip irrigation allowed them to produce more with the same water. Because the technology itself does not reward using less water, the drive to increase income outweighed any implicit expectation that efficiency would lead to reduced water use.
A counterintuitive result like this is important because it introduces us to the concept of rebound effects, where when efficiency lowers the effective cost of using a resource, it can increase demand for that resource rather than decrease it. Like mentioned above, the technology did not fail but performed exactly as designed. But its introduction changed incentives, and those incentives changed behavior. Efficiency became a tool for expansion, not conservation.
This distinction is critical. Technologies operate within economic and social systems that shape how their benefits are used. Without aligning efficiency with incentives for conservation, efficiency alone cannot guarantee reduced water demand.
Restoring aquifers with groundwater
The problem: Managed aquifer recharge is intended to aid groundwater availability, yet aquifers may deplete faster after interventions begin.
Managed aquifer recharge (MAR) interventions channel water, such as treated wastewater, stormwater and diverted surface water, back into underground aquifers to replenish what has been extracted. By restoring groundwater levels, MAR can delay the most severe effects of scarcity and stabilize water access. In Gujarat, India, one of the world’s largest MAR efforts combines large-scale recharge, channeling water from dams and percolation tanks into depleted aquifers, with energy policy reforms that help limit groundwater abstraction. MAR efforts like this have been known to contribute to rising groundwater levels, supporting more reliable irrigation, sustaining livelihoods and reducing seasonal migration.
At the same time, where incentives and governance remain unchanged, additional groundwater availability can coincide with rising demand, negating any improvements. In some contexts, irrigation demand, largely supplied by groundwater, has increased substantially following recharge efforts. In Gujarat, there was a roughly 150% rise in demand after 2002, driven by expanded cropping area and more intensive cultivation of water-dependent crops such as cotton and wheat, with demand growth exceeding gains in recharge. Over time, total withdrawals can once again surpass replenishment, returning aquifers to decline despite the intervention.
This reveals a broader lesson across both examples. Drip irrigation can reduce demand, while managed aquifer recharge can increase supply. Yet in both cases, the technology worked as designed but still failed to produce the intended outcome. In both directions, people adapted to the new conditions by using efficiency to expand production and increased availability to consume more. Without accounting for the human incentives and behaviors that shape how these gains are used, interventions can unintentionally reinforce the very risks they were meant to reduce.
Levee effect
The problem: Levees are built to reduce flood risk, yet they have led to communities facing higher risk of more severe damage.
Levees are raised embankments built along rivers and coastlines to keep floodwaters from pouring into surrounding land. They are relatively inexpensive, quick to construct and effective at reducing the frequency and severity of smaller floods. Levees and flood-control reservoirs can protect homes, farmland and infrastructure, stabilizing livelihoods and creating a sense of security in places that once flooded regularly.
But this protection also changes how people perceive risk. When floods stop occurring, the danger feels distant and areas behind levees begin to be seen as safe, encouraging new housing, businesses and infrastructure in locations that remain inherently flood-prone. What was once recognized as a floodplain becomes redefined as usable and even desirable land. Over time, more people and assets accumulate in these protected areas, increasing exposure to future, more intense floods.
At the same time, levees alter the natural function of floodplains. By constraining rivers, they prevent water from spreading across landscapes where it would normally dissipate energy and be absorbed. This can intensify flooding elsewhere and place greater pressure on the levees themselves. Their effectiveness also depends on consistent maintenance and accurate communication of their limits, which are not always met. When levees are presented as complete protection rather than partial risk reduction, communities and planners may underestimate the residual danger, which can reduce preparedness.
The result is a paradox in which levees reduce frequent, smaller floods but increase exposure to more rare, catastrophic ones. When floods exceed design limits or defenses fail, the consequences can be far more severe than if no levee had existed at all. During the Great Midwest Flood of 1993, approximately 70% of levees in affected areas were damaged or overtopped, contributing to widespread losses across both urban and agricultural regions. An intervention intended to reduce risk had, over time, reshaped the system in ways that ultimately amplified vulnerability.
In each of these three instances, a problem prompts a technical solution designed to reduce risk. Yet these interventions often fall short because they treat risk as purely physical, rather than social. They overlook how incentives shift when conditions improve, how behaviors adapt in response to new information, and how people interpret protection. Without addressing the norms, motivations and institutional realities that shaped the problem in the first place, even technically sound interventions can unravel when introduced into the systems they are meant to stabilize.
This is not a failure of engineering, but a transfer failure of integration. Lasting resilience depends not only on the strength of our infrastructure, but also on the depth of our understanding. Solutions must account for the feedbacks between water, people and institutions, recognizing that every intervention becomes part of the system it seeks to change. To improve the problem, we must embrace adaptive approaches that account for these dynamic relationships and remember that improving outcomes requires evolving the system itself, not just the tools we use.
