Commentary: On the water footprint as an indicator of water use in food production

The water footprint (WF) is defined as the volume of water used per unit of food produced.
It is used in assessments relating food production to water use.

Green Water Footprint: Water from precipitation stored in the soil's root zone, which is then evaporated, transpired, or incorporated by plants.
Blue Water Footprint: Water sourced from surface or groundwater resources that is evaporated, incorporated into a product, or moved and returned at a different time. It applies to irrigated agriculture, industry, and domestic water use.
Grey Water Footprint: The amount of fresh water required to assimilate pollutants to meet water quality standards.

If the baseline is the water use of the natural ecosystem, a crop's “green” footprint might be less than the native vegetation's water consumption, increasing downstream water availability.
Water consumed by an irrigated crop is a complex mix of residual soil moisture from rainfall and irrigation. Determining whether the water consumed is green or blue can be difficult, especially with rainfall occurring around irrigation events.
Hydropower dams significantly consume “blue” water due to altered timing of downstream water delivery. Water passing through canal head powerhouses can be counted as “blue” both when going through the generators and when consumed by the crop.
Total water consumption of a crop is computed as the maximum potential crop ET, often higher than actual crop ET, particularly in low-productivity rainfed systems, leading to inflated WFs.
In areas with severe pollution, the “grey” footprint may exceed available water for dilution but can be reduced to zero with a suitable treatment plant.

Criticisms have been voiced since the WF idea was formulated in 2002.
Perry (2014) and Wichelns (2015) argue against using WF to compare production systems or derive policy implications for water allocation and use.
Perry focuses on technical difficulties in deriving robust WF indicators.
Wichelns focuses on the absence of the local opportunity cost of water in WF formulations.
Despite criticisms, the scientific community continues to use WF in diverse assessments, from agricultural production to diets, and from local to global scales, increasingly oriented towards trade and sustainability analyses.
A shared standard on definitions and calculation methods is seen as crucial for companies and governments formulating sustainable water strategies and policies.
Large figures, such as one ton of water used per kilogram of wheat, can capture public attention but do not provide the whole story.

The WF is sometimes seen as paralleling the carbon footprint, but unlike CO2 emissions which have global effects, water evaporation has only local effects, impacting the basin where it occurs.
Hydrologists distinguish between water use (various dispositions like evaporation, transpiration, runoff, drainage) and water consumption (evaporation losses to the atmosphere, ET).
At the basin scale, runoff and drainage may be recovered, but consumption through ET is largely uncontrollable and constitutes a ‘true’ water loss.
Without accounting for the various dispositions of water and their local significance, the WF ratio lacks general meaning.
The ET process is passive and depends on the evaporative power of the atmosphere, which farmers cannot control.
Comparing WF values across different locations or seasons is like comparing unmanageable geographical features.
Using WF for certification of agricultural production could lead to questionable norms, negatively impacting industries in arid countries (e.g., cotton).
Computing the WF of rainfed crops is meaningless because ET would occur regardless, and might even be reduced when forest is replaced by seasonal agriculture.
Taking the WF indicator to the absurd, the WF of fish can be computed based on global ocean evaporation and fish catch, resulting in a meaningless figure since evaporation would proceed regardless of fishing.
Similarly, computing the WF of meat from animals fed on pastures or rainfed crops is futile.

The global footprint of fish from marine ecosystems may be estimated as the ratio of the mean global evaporation from oceans $(413 \ km^3/year)$ and the global fish catch $(126 \ Mt/year)$ resulting in $3.3 \ m^3$ of water per kg of fish. If one takes into account the inedible parts of fish (e.g. 50%) and the water content of fish flesh (assumed 35%) the WF value becomes $10 \ m^3$ per kg of dry edible fish.
A British citizen eating a sandwich with 75 g of white bread, which has required 73 g of wheat grain for its production. An average wheat field in the UK produces $0.8 \ kg/m^2$ of grain, covering a field for 10 months which receives a total solar radiation load of $3200 \ MJ/m^2$ . A kg of wheat is associated with $4000 \ MJ$ of solar energy and thus the sandwich would have required a solar energy input of $292 \ MJ$ , the same energy that is contained in 6.3 liters of gasoline!

Computing the WF of rainfed systems is largely meaningless.
In irrigated systems, additional water is supplied artificially, so the seasonal ET exceeds that of natural ecosystems.
The contribution of irrigation to enhancing global evaporation is only about 0.4%.
- Rainfed agriculture has the same evaporation as the average land areas $(490 \ mm/year)$
- Evaporation from irrigated areas is equal to the rate of evaporation from the oceans $(1144 \ mm/year)$
- The areas of oceans, total land and irrigated areas are 361, 149 and 3.0 million $km^2$ , respectively
Only 17% of cropped lands are irrigated globally, producing more than 40% of our food.
The impact of irrigation on water demand is substantial locally, emphasizing the irrelevance of WFs in a global context.
Water productivity (WP), defined as the ratio of production to ET, is a popular indicator in irrigated agriculture.

Improvement in WP reflects a positive advance in productivity.
WF is essentially the inverse of WP; a higher WF has negative connotations.
Both indicators are too simplistic and may be misleading.
Improvements in WP are mostly attributed to increased production, with little attributed to decreased ET.
WP or WF cannot offer sufficient information to guide policy decisions or make meaningful comparisons for sustainability.
Comparing production systems differing in ET makes little sense unless normalized by Reference ET (ETo), as arid climates would always exhibit higher WF values.

The success of WF reflects societal concerns about water use in food production and irrigation.
Scrutiny of irrigation is justified due to large water amounts used and associated environmental issues.
Meaningful assessments are needed to provide unbiased information on current practices and recommendations for improvement.
Assessments must be based on proper accounting of water applied (AW) and its disposition, quantifying the water balance.
The fate of AW varies, with some draining below the root zone, running off, or being lost to evaporation.
AW losses are often reused downstream or recharge aquifers.
Robust procedures are available to evaluate irrigation practices by determining the fate and disposition of AW.
Technical recommendations for improvement must be integrated into socio-economic and institutional environments.
Irrigation is complex and not amenable to characterization by a single indicator like WF, which claims to integrate rainfall, irrigation water, downstream flows, and pollution impacts on a globally standardized basis, potentially misleadingly.