Article prepared for FAO by Hubert H.G. Savenije, IHE, Delft, The Netherlands

[S-1] Of all water resources, green water is probably the most under-valued resource. Yet it is responsible for by far the largest part of the world's food and biomass production. The concept of green water was first introduced by Falkenmark (1995), to distinguish it from blue water, which is the water that occurs in rivers, lakes and aquifers. The storage medium for green water is the unsaturated soil. The process through which green water is consumed is transpiration. Hence the total amount of green water resources available over a given period of time equals the accumulated amount of transpiration over that period. In this definition irrigation is not taken into account. Green water is transpiration resulting directly from rainfall, hence we are talking about rainfed agriculture, pasture, forestry, etc. The average residence time of green water in the unsaturated zone is the ratio of the storage to the flux (the transpiration). At a global scale the storage in the unsaturated zone is about 500 mm, whereas the average global transpiration is 100 mm/month (see Table 1). The average residence time of green water is hence approximately 5 months. At a local scale, depending on climate, soils and topography, these numbers can vary significantly.

[S-2] Green water is a very important resource for global food production. About 60% of the world staple food production relies on rainfed irrigation, and hence green water. The entire meat production from grazing relies on green water, and so does the production of wood from forestry. In Sub-Saharan Africa almost the entire food production depends on green water (the relative importance of irrigation is minor) and most of the industrial products, such as cotton, tobacco, wood, etc.

[S-3] There is no green water without blue water, as their processes of origin are closely related. Blue water is the sum of the water that recharges the groundwater and the water that runs-off over the surface. Blue water occurs as renewable groundwater in aquifers and as surface water in water bodies. These two resources can not simply be added, since the recharge of the renewable groundwater eventually ends up in the surface water system. Adding them up often implies double counting. Depending on the climate, topography and geology, the ratio of groundwater recharge to total blue water varies. In some parts the contribution of the groundwater to the blue water can be as high as 70-80%, in some parts (on solid rock surface), it can be negligible. Generally the groundwater contribution to the blue water is larger than one thinks intuitively. The reason that rivers run dry is more often related to groundwater withdrawals, than to surface water consumption.

[S-4] Engineers always have had a preference for blue water. For food production, engineers have concentrated on irrigation and neglected rainfed agriculture, which does not require impressive engineering works. Irrigation is a way of turning blue water into green water. Drainage is a way of turning green water into blue water.

[S-5] Water has many colours. To complete the full picture of the water resources, besides green water and blue water, there is white water. White water is the part of the rainfall that feeds back directly to the atmosphere through evaporation from interception and bare soil. Some people consider the white water as part of the green water, but that adds to confusion since green water is a productive use of water whereas the white water is non-productive. The white and green water together form the vertical component of the water cycle, as opposed to the blue water, which is essentially horizontal. In addition, the term white water can be used to describe the rainfall which is intercepted for human use, e.g. through rainwater harvesting. Figure 1 gives a schematic representation of these three colours. The groundwater is part of the blue water and may be painted "deep blue". The fossil water does not enter into the picture, since it is on-renewable and not related to rainfall.

note:

*) indicates rough estimates

[S-6] Table 1 presents the quantities of fluxes and stocks of these water resources, and the resulting average residence times, at a global scale. For catchments and sub-systems similar computations can be made. The relative size of the fluxes and stocks can vary considerably between catchments. Not much information on these resources exist at sub-catchment scale. The study of the Mupfure catchment in Zimbabwe is an exception, see Table 2.

Table 2. Water resources partitioning and variability in the Mupfure River Basin, Zimbabwe

[S-7] Fig. 2, based on 20 years of records (1969-1989) in the Mupfure basin in Zimbabwe (1.2 Gm²), shows the separation of rainfall into interception (White), Green and Blue water. The model used for this separation is described by Savenije (1997). It can be seen that there is considerably more green water than blue water available in the catchment. Table 2 shows the average values over the 20 years of records. Moreover, the model showed that more than 60% of the blue water resulted from groundwater, a resource until recently neglected in Zimbabwe.

[S-8] Finally, the last colour of the rainbow is the ultra-violet water, the invisible water, or the virtual water. Virtual water is the amount of water required to produce a certain good. In agriculture, the concept of virtual water is used to express a product in the amount of water required for its production. The production of grains typically requires 2-3 m³/kg, depending on the efficiency of the production process. Trading grains, implies the trade of virtual water.

[S-9] The data on the annual renewable per capita water availability, that are widely used as indicators of water stress, are deceptive. as they concentrate on blue water and disregard green water. In temperate climates food production is primarily from green water. The quantities of food produced and the further potential in North America, South America and Europe are large. Yet this resource is fully disregarded in the data. The inclusion of green water in the data on national water resources would paint a substantially less gloomy picture of global food security.

[S-10] Although the potential for rainfed food production is enormous, rainfed agriculture has been marginalised by water resources planners, who are mostly engineers. This does not do justice to the fact that more than 60% of all food in the world is produced under rainfed conditions (Lundqvist & Sandström, 1997). And this figure does not even include meat production from grazing. The trade of this food (virtual green water) is an important mechanism for food security. Or as Allan (1997, unpublished Internet report) puts it: More water flows into the Middle East each year in its virtual form than is used for annual crop production in Egypt. In addition, the recycling of grey water, or the harvesting of white water is not taken into consideration.

[S-11] A distinction should be made between rainfed agriculture in temperate zones, the wet tropics and the semi-arid tropics. The rainfed agriculture in temperate zones has been highly mechanised and production is highly efficient, at least financially. However it is very energy demanding and some people claim that it uses more energy than the (solar) energy it absorbs (Trevor Graham in: McDonald, 1998). In the wet tropics rainfed agriculture is mostly done by small farmers on small scale schemes (e.g. Bangladesh). These schemes can be very efficient as long as the risk emanating from the occurrence of dry spells can be minimised by supplementary irrigation or rain water harvesting. In the wet tropics, in addition, there is often enough water to allow a second or a third crop during the dry season through irrigation. Indonesia, Bangladesh and Taiwan (to name a few) have shown that the Malthusian precipice can be avoided by intensifying small scale agriculture.

[S-12] In the semi-arid tropics, however, the situation is completely different. Except at a number of well-known sites along major rivers (Nile, Niger, Zambezi, Euphrates, etc.) there is not much scope for large scale irrigation. Probably more than 90% of the food consumed in Sub-Saharan Africa is from small scale rainfed agriculture. Most of this agriculture is subsistence farming with a very low efficiency. For that reason, the potential to increase food production from rainfed agriculture is often disregarded. However, this is precisely the wrong conclusion. If this rainfed agriculture is inefficient, then there is an enormous scope for increasing food production by bringing it up to standards. By improving rainfed agriculture we short-cut the food cycle, producing the food were the consumers are.

[S-13] One of the reasons why rainfed agriculture is deficient is the poverty trap. Poor people living in a marginal environment try to survive by avoiding damage resulting from hazards. Risk avoidance generally results in maximizing the use of labour while minimizing the use of capital intensive resources. Often the poor can not afford to invest sufficiently in their crops or in their natural resource base. This behaviour leads on the one hand to economic inefficiency but it also leads to exploitation and degradation of the resource base, both of which, in turn, sustain their poverty.

[S-14] Moreover, an increase in rural income from rainfed agriculture and subsequently an increase in capacity to invest in the resource base is only possible if there is sufficient demand for the products that can be grown. This may require the creation of markets at the appropriate scale and adequate access to these markets.

[S-15] The challenge is to break the poverty trap by reducing the risks of crop failure and by stimulating demand. With regard to water resources, the prime factor is securing water to bridge dry spells during the rainy season. If that can be done, by applying relatively small amounts of blue water (supplementary irrigation), then high value crops can be grown on mainly green water. What we are looking at are innovative ways of combining blue and green water (Falkenmark et al., 1998). The most probable way forward is not the building of gravity irrigation schemes (involving a number of farmers), but the prudent use of groundwater or rainwater harvesting in combination with soil and water conservation (Rockström, 1997).

[S-16] The management of groundwater is much simpler than surface water. One could think of a single well per farmer of a limited capacity. Wells can have relatively small capacities since quantities required by small farmers to bridge dry spells are small. For supplementary irrigation one can temporarily exceed the safe yield from a well and pump it dry, only to bridge a dry spell. Small capacities would prevent irrigation in the dry season (which is not desirable in water-scarce areas) but would be more than sufficient for household supply, which allows farmers (and particulalry women) to dedicate time to other economic activities.

[S-17] Consequently, securing blue (ground)water to bridge dry spells allows farmers to invest in seeds and fertilizers, to use their labour more efficiently and to conserve their resources. Under such improved conditions, the economic returns of external fertilizers and of additional (blue) water will be high. The supply of modest credit facilities, such as exist in Bangladesh, could make a major difference, particularly if they are given to women even in the absence of collateral.

[S-18] The economic feasibility of small scale supplementary irrigation has to be seriously analysed in a broad socio-economic context. For this analysis one should consider the satisfaction of multiple objectives including: reduction of poverty, access to safe water, conserving the environment, prevention of erosion and land degradation, and provision of food security. Since population growth is closely linked to poverty, breaking the poverty trap would also be instrumental in attaining a stable population.

[S-19] Water is the key to restoring the balance between crop production and the carrying capacity of the natural system. Closely related to it is the nutrient balance. Investments in (appropriate use of) fertilisers, improved seeds and sustainable land use practises, however, require that the risk of crop failure due to the occurrence of dry spells is minimized and the poverty trap is broken. In the semi-arid tropics, supplementary irrigation from groundwater could be the key to breaking the poverty trap, by optimizing the use of green water for crop production. Relatively small quantities of blue water can safeguard crop production and by doing so have very high yields per cubic metre of blue water, much higher than can be attained from full scale (dry season) irrigation. Addressing poverty, water resources, nutrients and the environment at the same time is both the biggest challenge and the biggest opportunity to warrant food production for the coming generations.

This article makes use of the information provided in the key note paper presented by Savenije (1998) at the 8th Stockholm Water Symposium, entitled: How do we feed a growing world population in a situation of water scarcity?.

Falkenmark, M., 1995. Coping with water scarcity under rapid population growth. Conference of SADC Ministers, Pretoria 23-24 November 1995.

Falkenmark, M., W. Klohn, J. Lundqvist, S. Postel, J. Rockstrom, D. Seckler, H. Shuval and J. Wallace, 1998. Water scarcity as a key factor behind global food insecurity: Round Table Discussion. Ambio, Vol. 27 No. 2. Royal Swedisch Academy of Sciences.

Lundqvist, J. and K. Sandstrom. 1997. Most worthwhile use of water: efficiency, equity and ecologically sound use; pre-requisites for 21st century management. Publications on water resources Nr. 7, Dep. for Natural resources and the environment, SIDA, Stockholm, Sweden.

McDonald, Frank, (ed.), 1998. The Ecological Footprint of Cities. The International Institute for the Urban Environment, Delft, The Netherlands.

Rockström, J., 1997. On-farm agrohydrological analysis of the Sahelian yield crisis: rainfall partitioning, soil nutrient and water use efficiency of pearl millet. Natural Resources Management, Department of Systems Ecology, Stockholm University, Sweden.

Savenije, H.H.G., 1997. Determination of evaporation from a catchment water balance at a monthly time scale. Hydrology and Earth System Sciences, Vol. 1, pp. 93-100, EGS, Katlenburg-Lindau, Germany.

Savenije, H.H.G., 1998. How do we feed a growing world population in a situation of water scarcity?, Key note paper presented at the 8th Stockholm Symposium, 10-13 August, 1998, Published in Water-The Key to Socio-economic Development and Quality of Life (pp.49-58), SIWI, Stockholm, 1998.