Water footprints - do they mean anything?

Return to water home


Water footprints have become a topic for discussion and debate on matters of sustainability They are often used by policy makers. The Camden Council in north London decided to reduce the meat content of the meals served by the council in its various establishments because of the large water footprint of meat - particularly of ruminant animals such as cattle and sheep. Why did they stop there? If they replaced wheat products with sugar from sugar beet they could reduce the water footprint of the energy component of the diet by ten times.

Even in Australia and New Zealand, both countries with some of the highest levels of water per head of population in the world, there is discussion of water footprints and calls to put water footprint data on food labels.

There are some severe theoretical problems with the water footprint concept that make them unsuitable for policy implementation at this stage.

Below is the water footprint explained in the words of the water footprint network.

"Why a Water Footprint Network?

The interest in the concept of the water footprint and the accompanying methods and tools is overwhelming. This interest is rooted in the recognition that human impacts on freshwater systems can ultimately be linked to human consumption and that issues like water shortages and pollution can be better understood and addressed by considering production and supply chains as a whole. It is increasingly acknowledged that local water depletion and pollution are often closely tied to the structure of the global economy. Many countries have significantly externalised their water footprint, importing water-intensive goods from elsewhere. This puts pressure on the water resources in the exporting regions, where too often mechanisms for wise water governance and conservation are lacking. Not only governments acknowledge their role in achieving a better management of water resources, also businesses and public-service organisations increasingly recognise their role in the interplay of actors involved in water use and management.
The basis for the water footprint concept and methodology has been laid by Prof. Arjen Hoekstra at UNESCO-IHE and further developed at the University of Twente, the Netherlands. The concept and methods have been firmly established in scientific literature. By today, tens of institutions and thousands of individuals have expressed interest in further developing and/or applying the water footprint methodology. The interest focuses on questions such as: How to implement proper water footprint accounting in the context of my country or organisation? How to identify the spots where water footprints have the largest impact? How to reduce and possibly offset those impacts?
A big gap has grown between the demand for support in application and implementation and the capacity to supply this support. Besides, there is a need to bring together in a structural way developers of methods and tools and institutions that seek application. In order to co-ordinate efforts to further develop and disseminate knowledge on water footprint concepts, methods and tools, a number of institutions have taken the initiative to establish a Water Footprint Network. Unique to the network is that it brings together partners from diverse origins: knowledge institutions, non-governmental sector, private sector, governments and UN."

Virtual water

Virtual water is a concept developed by Prof. Tony Allan at King's College, University of London, to compare the water needed to grow food and other agricultural commodities.

Virtual water is a description of the water requirements of food production. For example when I was a farmer in Australia our farm had an average rainfall of about 500 mm and in a year when the rain fell at the right time and other agronomic factors were good we could expect a yield of 5 tonnes of barley per ha.

500 mm of rainfall over an area of one hectare represents 5000 cubic metres of water. That means that 5000 kg of barley in this example required 5000 cu m of water. One kilo of barley requires 1 cu m. or 1000 kg of water. The 1000 kg is the virtual water content of this barley. (1388 kg is the average figure for world production). It is not the actual water content. Barley in our case was harvested with a moisture content of less than 12% so typically a kilo of barley would have an actual water content of 100 gm.

I have painted a somewhat rosy picture of our cereal production. It was only in exceptional years when everything went just right with the rainfall pattern, the fertiliser, the weed control and the harvesting weather that we achieved 10 kg of grain for every mm of rainfall. Of course all these errors and assumptions are magnified when one moves from an individual crop to a region or country. In spite of these assumptions the concept is useful in explaining world trading patterns. Australia, Canada, Europe, Argentina and USA are huge exporters of virtual water. A very large part of this virtual water goes to North Africa and West Asia. The trade in virtual water plays a major but hidden part in the politics of the region.

The next stage is the virtual water embedded in animal products such as meat, butter and cheese. Here we need to make some more assumptions. As far as I can see these assumptions are based on grain feeding. Thus a broiler chicken intensively grown might have a conversion ratio of 2.5 to 1. This means that 1 kilo of meat is produced from 2.5 kilos of grain. Each kilo of grain in turn contains from our example 1000 kilos of virtual water so chicken meat contains 2500 kilos of virtual water. Similar figures can be constructed for other secondary products.

I have no quarrel with this as long as it is seen as an approximation and as long as it does not lead to a water centric view of farming. In my example I mention all the other factors that need to be in place for rainfall to be converted into grain. In some discussions these are ignored and a simple formula of water=crop dominates.

Carbon footprint

The term carbon footprint is increasingly used in discussions on global warming. It is the the carbon used by the people, families, factories or countries for the production of goods and services. That is the consumption of gas, electricity (converted for carbon used in generation), petrol and other fuels for a family is calculated from the fossil carbon content and used as their carbon footprint. Unfortunately the carbon footprint rarely includes the virtual carbon consumed. Virtual carbon is the carbon that was used to produce the food they eat, the cars they drive and so on. Leaving these difficulties aside one can see that there is a major different between a carbon footprint and a water footprint and it is misleading to discuss the two in the same breath. Carbon is consumed. Of course there is a carbon cycle but we are talking about carbon from fossil sources that are outside the carbon cycle not firewood, straw or other renewable carbon.

With the water footprint we are talking about the water cycle not fossil water. Of course there is fossil water being used in Libya, Algeria and Egypt but in world terms it is insignificant and so far we have not started to mine the Greenland or Antarctic glaciers. Transferring the carbon footprint concept to water is misleading.

The water cycle is a process of constant renewal. The rain falls on the ground. Some is used by plants - both natural vegetation and crops. The rest flows into rivers or underground acquifers. Some is used for irrigation, industry or domestic use. The rest flows into the sea. Water is evaporated from the sea. It is carried inland in clouds and falls as rain. It is wrong to say that water evaporated and transpired by crops is consumed. The term "consumption" comes from the hydraulic engineers who provide water for irrigation, industrial and domestic users. Water for rainfed crops is borrowed not consumed. In international trade this does not matter. To be precise the sequence should go like this. Australia borrows water from the water cycle but then returns the actual water. It converts it into virtual water. It sells the virtual water to Egypt and they consume the virtual water. Applying this formula on a small scale throws up anomalies

From the water footprint web site again;-


The mission of the Water Footprint Network is to promote the transition towards sustainable, fair and efficient use of fresh water resources worldwide by:
advancing the concept of the ‘water footprint’, a spatially and temporally explicit indicator of direct and indirect water use of consumers and producers;

increasing the water footprint awareness of communities, government bodies and businesses and their understanding of how consumption of goods and services and production chains relate to water use and impacts on fresh-water systems; and

encouraging forms of water governance that reduce the negative ecological and social impacts of the water footprints of communities, countries and businesses."

The water footprint

The problem with the water footprint is that it does not distinguish the source of the virtual water. Virtual water that comes from irrigation can in some limited ways be consumed by the crop. It has other uses. It can be used for other crops, for industrial and domestic use or it can be left in the river for environmental purposes. The scope for these changes is limited. For example it is not possible to say automatically that wheat should be grown instead of rice. Such a change would reduce the water footprint per tonne from 2291 cu m. to 1334 cu m. but dozens of other factors come in the decision besides the water footprint. If the change was made from wheat to sugar beet - something that is agronomically feasible - the water footprint can be reduced by a factor of ten times but is that the diet we want.

Perhaps 35% of the world's food production comes from irrigated land so this sector of food production is significant but the other 65% comes from rainfed farming. With rainfed farming the water footprint becomes an academic curiosity not a tool for policy makers.

Let us return to the Camden Council and their move to reduce meat consumption because of its high water footprint Obviously they should start by eliminating meat that has a higher footprint. The conversion of grain to chicken meat is comparatively efficient at 2.5 to 1 but Welsh lamb has a much worse conversion ration. Even when one excludes the water that flows off the Welsh hills and just counts the water evaporated and transpired by the pastures on the hills, the production of grass is not efficient. The next stage of conversion to meat is even worse. I do not know if any separate calculations have been made for Welsh lamb but my guess is that would compare badly with cheap industrial chicken in terms of its water footprint.

If however we leave the world of water footprints and the virtual water behind, to the real world of farming and look at the impact of a shift from Welsh lamb to industrial chicken we see that it is quite the opposite to that predicted by the water footprint. Let us say that lamb is no longer produced on the Welsh hills. The grass still grows. It borrows water from the water cycle and uses it in evapo-transpiration but it is not converted into lamb. There is no extra run off into the rivers. The city of Birmingham does not gain any extra water from domestic and industrial uses. Perhaps wild deer take over the grasslands but that is a separate issue. Eliminating the production of Welsh lamb has not produced a single extra drop of useable water. It has not improved the quality of the water. It has had no impact on the water cycle whatsoever.

The Welsh lamb may be an extreme example in the sense that conversion ratios of water to pasture and then from pasture to meat are low but other pasture systems will apparently consume large quantities of water and hence produce large water footprints for the the meat and milk produced from them. The more intensive the system the lower the water footprint. In other words pour on the nitrogen fertiliser for maximum grass production. Evapo-transpiration will hardly change and the water footprint will be reduced. Of course the nitrogen fertiliser has a much greater carbon footprint than a legume based pasture and the pollution of rivers and underground water is much greater due to the leaching of soluble nitrogen compounds. Water footprints as a means of reducing pollution? It does not seem to be the case.

On the other side of the equation the chicken meat is produced in factories using grain. Some of this grain may come from irrigated land where water consumption has greater significance. The water is not so clearly borrowed from the system but taken from other uses such as industrial or domestic consumers. In fact the switch from Welsh lamb to chicken would have the reverse effect to that predicted by the water footprint.

Protagonist of water footprints will claim that I have chosen an extreme example but the idea falls apart quite frequently. Another example is the growing of tea and coffee in the hills of Kerala in south India. Using the figures from Hoekstra, A.Y. and Chapagain, A.K. (2008) Globalisation of water: Sharing the planet's freshwater resources, Blackwell Publishing, Oxford, UK. we get a figure of 4255 kg of water footprint for 1 kg of tea while coffee has a footprint of a staggering 86464 kg of virtual water. If we want to save the planet through saving water we switch from coffee to tea. The coffee plantations of Kerala and elsewhere are pulled up and replaced by tea. A huge disruption for the growers but presumably it is all worth while because of the huge savings in water. The only problem is to find these huge savings. Will the rice growers lower down the slope in Kerala have more water? Will the coastal cities have more water? I doubt it. The evapo-transpiration from complete vegetation cover on the hills is much the same whether it is coffee, tea, forest, spices or grassland and changing the mix of crops will not release any water unless drastic steps are taken to increase runoff. I cannot see this happening. To concrete over the land to increase run off would create so many other problems in addition to the cost.

Fresh vegetables pose another problem. Grains are harvested and dried to a low moisture content to ensure preservation. It is therefore legitimate to compare rice with wheat or barley but vegetables have a high moisture content. Tomatoes for example have a water footprint of 184 kg of water per kg of tomatoes but it certainly is not valid to compare this with 1334 for wheat or even 113 for sugar from beets.

The water footprint does have some limited uses in the quest for better water use. For example the figure quoted for sugar cane is 175 kg of virtual water per kg of sugar while sugar beets are 113 kg of virtual water. Apparently there is a saving of real water in changing from sugar cane to sugar beet if this is climatically feasible and there are no other barriers but these direct substitutions are rare and in most cases abandoning a crop or pasture because of its high water footprint would not save any water unless something was done to artificially increase runoff.

The sugar story raises some other complications, which other people are better qualified to discuss in detail, but if beet sugar has a footprint of 113 kg per kg of sugar and wheat a figure 1334 per kg then the Camden Council should be reducing meat and increasing sugar. Reduced meat consumption would be more healthy but not increased sugar.

Negative aspects of water footprints

The calculation of water footprints follows the current fashion for mono culture. The water footprint of wheat is the water needed to grow a wheat crop in a single year. The same applies to other crops. Animal production comes from grain production and is a simple calculation from conversion ratios.

Much of the farming in the world outside the subsidised regimes of USA, Japan and Europe follows a different pattern that cannot be so easily divided into annual budgets. They follow rotations and integrated farming systems where various crops interact with each other and with animals.

Going back to my own farming experience in Australia we practised a rotation of medic pasture and cereals. The medic pasture provided a protein rich feed for the sheep and nitrogen for the following cereal crop. Unlike nitrogen in a bag the production of nitrogen from medic pasture has a zero carbon footprint. In addition to the nitrogen the medic provided other benefits to the cereals that are still not fully understood. It provided a break from the cycle of disease and the still not fully explained hydrogen release. These effects could increase the yield of the cereal crop by 25% to 50% beyond the effect of the nitrogen.

Using purely water footprint criteria the medic should be eliminated and the cereal crop grown with more artificial nitrogen. This would push up the carbon footprint considerably. Alternatively some of the water footprint of the medic might be transferred to the wheat. If this is done then the straw eaten by the sheep after the cereal harvest should be transferred back. These calculations would make the whole calculation too complex to have any relevance but without them we are in danger of glossing over the reality of farming with some simple figures that do not reflect reality.

Water or carbon?

Carbon footprints and water footprints tread on each others toes.

One of the best ways of reducing the water footprint is to increase yields - the old more crop per drop formula. Intensive cereal production has a lower water footprint. The evapo-transpiration from a high yielding crop is not significantly higher than a low yielding one so the water footprint per kilo of output is lower.

To reduce the carbon footprint we should reduce the intensity. Nitrogen fertiliser has a particularly high carbon footprint. Carbon is used in the production process and when the nitrogen is applied to the soil nitrous oxides are released which are much more effective greenhouse gases than carbon dioxide. If agriculture is to make any contribution to the reduction in greenhouse gases the use of artificial nitrogen must be reduced.

The alternative is to use legumes such as the medic pasture I used in Australia. Production is some cases will be lower but the carbon footprint will be zero or negative as carbon is added to the soil. The water footprint will be higher with less intensive production.

Carbon or food?

Some people claim that the only way to produce more food for the growing world population is too use even more nitrogen from a bag and even more intensive systems. Of course the food requirements for the world are difficult to estimate and those provided by the FAO (most commonly used in these discussions) are particularly flawed. Renewable sources of energy to produce nitrogen for crops and pasture need to be found. Wind, solar and bio fuels have been suggested for producing more fertiliser but surprisingly not the most obvious - legumes. Producing artificial nitrogen from green power sources ignores the release of nitrous oxides from the soil which may be more damaging than the carbon used in fertiliser production.

Can food production, carbon reduction and water use be reconciled or is it a question of constant trade offs?

The following would be a compatible approach:-

* Convert bio fuel production to bio fertiliser production.

Bio fuels have yet to prove themselves as effective means of reducing fossil carbon use. The distillation of sugar cane into alcohol has a positive carbon balance but grains and cellulose produce only marginally more renewable carbon than the amount consumed in their production. The technical optimists say this will change but there are some fundamental laws of thermo dynamics that are difficult to reverse.

Bio fertilisers do not need any technical optimism. They work. We know they work. They have been used since farming began. They do not release the nitrous oxides that are harmful green house gases. They also add carbon to the soil so the positive effects are two fold. They eliminate nitrogen fertiliser and increase soil carbon.

Legumes also produce food. The best bio fertilisers are pastures grazed by animals. We obtain more animal production. Less animal production would be needed from grain. Grain growing could be reduced further and we could enter a cycle of reduced grain production without reducing grain for human consumption. Outside the carbon, food and water triangle this would reduce soil erosion.

I am not sure what the impact would be on water footprints but the quality of the water would be greatly improved if it was not polluted with leached nitrates from fertilisers. The impact of animal manure would be less if they were grazing pastures rather than being fed in feedlots.

* Legumes in dryland farming

World agriculture can be divided into three broad division. Irrigated agriculture accounts for about 22% of the area and about 35% of the production. The balance is rainfed - 78% of the area and 65% of the production. Irrigation provides reliable water throughout the growing season. The so called Green Revolution - that is the package of nitrogen fertiliser, high yielding crop varieties and reliable water - has been successful in terms of grain production alone with irrigation.

Rainfed covers two distinct type of agriculture - or at least there are two poles and a cross-over zone in between.

The most productive rainfed agriculture is in the temperate regions with reliable moisture. Canada, USA, Europe etc. produce grains in this zone. Yields are as high as irrigated crops and in many cases higher because the same package has been used. Converting these area to legume based systems would reduce grain yields. Of course there would be some compensation from the output of the legumes.

Where rainfall is unreliable as in the Mediterranean and drier tropical regions the package fails. In the Curse of Nitrogen and elsewhere on this site I explain why bag nitrogen fails when rainfall is unreliable. There have been many and continuing attempts to transfer the Green Revolution to these areas but the package has consistently failed. The technical optimists keep on trying. They believe that there is always another technical fix around the corner which will make bag nitrogen effective. They persistently ignore legumes which are already effective and have a proven record. Adopting a legume based system in these regions would increase yields. Overall surpluses from temperate agriculture in Europe and USA might be reduced and grain production in Africa would be increased. Perhaps that is what we are looking for anyway?

In Asia the use of irrigation has increased considerably and the switch to more legumes in irrigated systems would not be easy. On the other hand grain legumes are already a larger part of the food culture and could be used in the legume rotations. They are not as effective as pasture because the protein is removed and eaten by humans not returned to the soil through animals.

The water footprint story is not supportive. Grain legumes such as lentils, chick peas and beans have a much greater water footprint than cereal grains but if one see them as a partial substitute for meat their water footprint is good.

* More grain legumes - less meat.

We all know that the Western world eats too much meat. It is bad for our health to eat so much. Meat production consumes vast quantities of grain which in turn has a large water and carbon footprint as well as causing increased soil erosion.

The increase in consumption can be attributed to rising incomes and lower food prices. For the last two hundred years food prices have been falling in real terms. Production has continued to increase in spite of lower prices due in part to the generous subsidies paid to farmers in the developed world. In the developed world the average expenditure on food has fallen. Meat has become cheaper during the last fifty years - particularly chicken meat. It is not surprising that consumption should have increased. Of course there are always exceptions. Fresh sheep meat in North Africa is expensive in money terms and even more so in terms of purchasing power for most of the population.

China, India and other developing countries are now following that path.

Making the change from meat to grain legumes is going to be difficult as meat has become associated with affluence in most parts of the world. India is fortunate to have a very large vegetarian population but the non-vegetarians are expanding their consumption.

Tax on meat.

A tax on food! Politicians will say it is impossible. They prefer subsidies on food. It has been done by stealth in Europe. Subsidies to farmers in Europe were paid by consumers for fifty years. The public was aware of the huge cost of purchasing surplus production to dump on world markets but there was another hidden cost - levies on imports to maintain domestic food prices. The population accepted this because they were told it was essential to maintain food security and avoid the famines that occurred after World War II. A tax on food should be explained as an environmental protection measure. Of course this is happening already with the transfer of subsidies away from production to the land in return for wildlife protection but more direct measures are needed.

A number of options are available:-

A tax on nitrogen fertiliser would be a good starting point.

More environmental taxes based on water pollution by fertilisers.

Pollution taxes for intensive raised chickens, pigs and cattle.

Taxes on unhealthy foods.

But not taxes based on water footprints.