Chris Perry (former Editor in Chief, Agricultural Water Management) & Quentin Grafton (Australian National University, Australia)
According to the World Meteorological Organization’s just released State of Global Water Resources report, five billion people, or around two-thirds of the world’s population, will face at least one month of water shortage by 2050. Another way of saying this is that in a few decades, the majority of humanity will be facing profound water insecurity. A common response to the challenge of water insecurity are cries for greater efficiency, especially in the irrigation sector through the use of ‘high-tech’ solutions that mean more water gets to the crop, less is ‘lost’. Here, Chris Perry and Quentin Grafton challenge this belief demonstrating the unintended consequences that ill-considered efforts to improve irrigation efficiency can have and undermine efforts to restrain water consumption. Their key message is that where water is scarce, and water access is uncontrolled, ‘efficient’ irrigation technologies tend to increase consumption by irrigation. Alternatively, with appropriate water governance including controls on water consumption, farmers are incentivised to adopt irrigation technologies that maximise net return to their resources, within the constraint of sustainable water supplies.
‘Efficiency’ is a value-laden term: increased efficiency is typically presumed to be a good objective. In the water sector, ‘efficiency’ is frequently a misleading shorthand for a variety of different indicators, including productivity, water-use efficiency, irrigation efficiency, and allocative efficiency. Each of these indicators has a different meaning because increasing ‘efficiency’ in any these various forms of efficiency will result in different and potentially contradictory outcomes.
An engineer’s approach to water scarcity
Irrigation engineers define ‘irrigation efficiency’ as the ratio of water that supports plant growth to the total water available at various scales (abstracted from an aquifer, released from a dam delivered to a field, etc.). The higher this dimensionless ratio, the greater the area that can be irrigated from the water source and the more crops that can be grown—entirely appropriate objectives in designing and operating a project. Irrigation efficiency is maximised when beneficial local water consumption (water that returns to the atmosphere through crop transpiration) is maximised, and as little water as possible is ‘lost’ from the local system as drainage and infiltration.
Their success in devising technologies and management systems that maximised local production and minimised these ‘losses’ gave the engineers a strong voice as water became scarce at larger scales. Reducing ‘losses’ was their trade, so who better to advise on water scarcity? But the concept of irrigation efficiency that is so valuable at the project or farm scale translates poorly when responding to water scarcity at the basin scale. Tackling water scarcity at larger scales depends on reduced water consumption. This is because local ‘losses’ rarely disappear when viewed from this wider perspective: they may return to rivers as drainage, or recharge aquifers for subsequent abstraction. The automatic assumption that higher irrigation efficiency is ‘good’ loses relevance when water consumption is the driver of water scarcity beyond the local scale.
The critical added dimension is return flows. Higher irrigation efficiency that increases local consumption at a farm and project level must mean lower return flows to streams, rivers, and aquifers. Some of these return flows are reusable, some end up in saline sinks and are indeed lost to further reuse—but until we identify these alternative flow paths, we cannot claim to understand the implications of ‘improved’ technologies. These different flows are summarised in Figure 1.
Applying this simple framework can produce essential insights that decision-makers frequently miss. Rice, for example, is widely assumed to require double or triple the water consumed per hectare when compared to maize or other grains. But much of the extra water applied to rice either runs off to the next field or drain, or recharges aquifers. A carefully calibrated water balance of the California Delta summarising water consumption for 25 crops and other land uses (Anderson et al, 2018) reports that rice consumes about 12% more water than maize.
Incorporating return flows
Until we assess the extent to which reusable return flows are impacted, the full hydrological and economic impact of increased irrigation efficiency is uncertain. Return flows are often valuable, serving other farmers and other sectors including drinking water, fisheries, navigation, and—often most significantly—support for ecosystems services.
A recent review of some 230 reports and papers where sufficient data were available confirmed that in more than 90% of cases, return flows (ie, ‘losses’) were potentially reusable; in more than 70% of cases, increases in local irrigation efficiency reduced availability elsewhere; and actual savings were only achieved in 11% of cases, either because controls on access to water were introduced, or because return flows were not reusable (Pérez-Blanco et al, 2021). Locally, in most cases, farmers were able to increase their incomes and, hence, many farmers welcome subsidies and grants to increase irrigation efficiency. Politicians also support interventions that benefit their local constituents. Yet, as shown in the figure below, the most common outcome of these ‘high-tech’ irrigation solutions is higher (not lower) water consumption.
Irrigation efficiency and the ‘rebound’ effect
Improving irrigation efficiency increases the amount of water consumed by crops (which is the driver of crop production). It has the direct, physical result of increasing consumption per unit of water delivered to the farm and supporting higher yields and/or increased irrigated area. A proportion of this incremental local production is offset by reduced economic activity elsewhere where that water may have been used for other purposes.
Beyond this essentially physical impact, there is an additional rebound effect because the economic demand for an input into a production process (such as water to irrigate crops) increases as the input becomes more productive (all else being equal). In the case of water and irrigation efficiency, this occurs because the production per unit of water delivered to the farm increases in parallel with irrigation efficiency. Thus, demand for irrigation water will tend to increase and, in the absence of constraints, will result in an even larger increase in water consumption—an unintended consequence of an increase in irrigation efficiency—as farmers are incentivised to consume more (by, for example, pumping for more hours or digging deeper wells or installing larger pumps).
A detailed statistical analysis of data from China shows that the introduction of measures intended to save water in various districts in China, compared over time with areas without such interventions, resulted in increased abstraction from aquifers, and increased water consumption (Xu & Yang, 2022). The data were sufficiently detailed to allow differentiation between the physical effect of increased irrigation efficiency (~30% of overall impact) and the economic response to the increased value of water as an input (~70%). These figures are location specific, but the existence of the separate components of the rebound effect in the case of water is an important policy distinction.
‘Blue’ water, ‘green’ water, and productivity
First, it’s useful to reflect on where that water comes from, where it lies in the overall landscape.
Water that is managed (stored or extracted from a dam, river diversion, or aquifer) is often characterised as ‘blue’ water. Rainfall, by contrast is uncontrolled. Rainfall adds to soil moisture and supports crop growth and is labelled as ‘green’ water. This shorthand is helpful in thinking about what we can and what we cannot influence, and in formulating management strategies for blue water that maximise its potential, as well as the potential contribution of green water to agricultural production.
This classification by colour is, however, more nuanced than it appears. A change in land use that affects runoff and recharge, will have impacts elsewhere over time and by place. For example, intensive tree plantings in some upper catchments in China appear to be reducing downstream water availability because the trees retain and consume more green water than the prior land us. Thus, downstream availability of blue water is diminished. Similarly, contour ploughing can substantially increase crop transpiration (more green water consumption) and local recharge to aquifers (more blue water) at the expense of runoff to local streams that might have been usable elsewhere (less blue water). Importantly, blue and green water are not independent components of the water supply.
Given these caveats, there are important opportunities to maximise the benefits that can be derived from erratic green rainfall by judicious management of blue water resources. The classifications set out in Figure 1 are equally applicable to green water. Rainfall arriving on to a landscape either evaporates locally (productively or unproductively), runs off to a stream, or infiltrates to groundwater. In some scenarios, unproductive evaporation that does not contribute to crop production or maintenance of environmental assets, is substantial.
Researchers have estimated that in semi-arid areas of sub-Saharan Africa, unproductive evaporation typically accounts for twice as much water consumption as productive transpiration by rain-fed crops (Rockström et al, 2002). In such cases, limited quantities of blue water applied to seedlings can result in large increases in production because a well-established crop will transpire previously unproductive (for crops) evaporation. This makes rainfall more productive in crop production, and encourages improved husbandry (denser planting, more fertilisers, etc.). In such cases, the impact of limited, additional blue water effectively multiplies the potential benefits of the rainfall (green water). The marginal return to these small quantities of blue water over a large area can be very high (2-4 times the average) such that total production, where rainfall is low and unreliable, can be higher than the intensive application of blue water to a smaller cropping and intensively irrigated area in pursuit of maximum yield.
When considering an intervention in water resources management—increasing irrigation efficiency or changing land use—all consequences (including the unintended) must be carefully evaluated. Physically, will the intervention increase local water consumption at the expense of other existing users (including nature)? Economically, will the intervention tend to increase water demand (the rebound effect)?
Both physical and economic impacts are important. The physical consequences arise because higher irrigation efficiency increases water consumption and reduces return flows—this is the definition of an increase in irrigation efficiency. The economic consequences arise because the local impact of higher irrigation efficiency is to increase the productivity and value of water delivered to the farm, thereby increasing the incentives to acquire greater volumes of water (a rebound effect).
In all scenarios, whether blue water is the sole source of water (such as in a desert), or primarily rain-fed farming is supplemented by small quantities of precisely targeted irrigation, the critical management obligation is to ensure that water supplies are constrained to sustainable levels (the hydrological dimension). Further, that all farmers are incentivised to respond to the opportunity cost of scarce water (the economic dimension).
In our view, the key water governance dimension is sequencing. Where water is scarce, and water access and extraction are uncontrolled, the introduction of more efficient irrigation technologies makes water more, not less,scarce. But with appropriate governance in place, and water access and extraction are effectively controlled, farmers are incentivised to adopt irrigation technologies that maximise net returns to their land, labour, and other resources within the constraint of sustainable water supplies.
Anderson M, F Gao, K Knipper, C Hain, W Dulaney, D Baldocchi, E Eichelmann, K Hemes, Y Yang, J Medellin-Azuara & W Kustas (2018). Field-scale assessment of land and water use change over the California delta using remote sensing. Remote Sensing 10. doi:10.3390/rs10060889
Pérez-Blanco CD, A Loch, F Ward, C Perry & A Adamson (2021). Agricultural water saving through technologies: a zombie idea. Environ. Res. Lett. 16 (2021) 114032
Rockström J, J Baron & J Fox (2002). Rainwater management for increased productivity among small-holder farmers in drought prone environments. Physics and Chemistry of the Earth 27: 949-959
Xu H & R Yang (2022). Does agricultural water conservation policy necessarily reduce agricultural water extraction? Evidence from China. Agricultural Water Management,
Dr Chris Perry is an economist specialising is water resources management. He worked for the World Bank for more than twenty years. Chris was subsequently head of research at the International Water Management Institute and after retiring was a Visiting Professor at Cranfield University, and Editor in Chief of Agricultural Water Management.
Professor R. Quentin Grafton is Professor of Economics at the ANU, Convenor of the Water Justice Hub, and Lead Expert and Commissioner of the Global Commission on the Economics of Water. He is also the Executive Editor of the Global Water Forum.
The views expressed in this article belong to the individual authors and do not represent the views of the Global Water Forum, the Water Justice Hub, the UNESCO Chair in Water Economics and Transboundary Water Governance, UNESCO, the Australian National University, Oxford University, or any of the institutions to which the authors are associated. Please see the Global Water Forum terms and conditions here.