Daniel Hillel
Center for Environmental Studies, Karkur, Israel

The task of managing water equitably and sustainably requires balancing supply and demand, with due regard to ever-changing environmental, social, political, and - of course - economic conditions. Greatest attention is typically devoted to ensuring the quantity of water supply, which involves the harnessing and diversion of water resources, generally by means of extensive engineering works. Somewhat less attention is paid to safeguarding water quality, and still less is given to managing the demand for water so as to avoid waste and derive the maximal benefit from the limited supply. No less important is the oft-neglected provision of sufficient water (quantity as well as quality) for natural ecosystems, particularly aquantic ecosystems such as wetlands and riverine esturaries. Within this complex of issues, we shall address some of the problems encountered in the task of balancing supply and demand in the face of a changing climate.

Both the supply and the demand side of water management are greatly affected by climate. The primary source of supply is precipitation (in warm regions consisting mainly of rainfall), which tends to vary both spatially and temporally. On the demand side, the major consumer of water in natural ecosystems as well as in agriculture is the process of evapotranspiration. The problem is especially acute in arid regions, where the supply is meager and intermittent, even while the demand (imposed by the intense influx of solar radiation and by the warm, dry, and windy atmosphere) is high and continuous. Hence plants growing in the field are highly vulnerable to water deficits that inhibit growth and reduce yields.

Irrigation, being the supply of water to agricultural crops by artificial means, is designed to permit farming in arid regions and to offset drought in semi-arid regions. As such, it can help to ensure stable production and to provide food security for an expanding population. In many semi-arid and arid countries, irrigation is the major consumer of water supplies. (In Israel, for example, it still receives over 60% of the state's freshwater supplies, as well as a growing portion of the recycled wastewater.) To be sustainable, irrigation must be practiced in a manner that prevents the degradation of land (e.g., by waterlogging and salination) and it must be made ever more efficient as the competition for water from the domestic, industrial, and environmental sectors also increases.

Enter the prospect of global warming, caused by the enhanced atmospheric "greenhouse effect." Some benefits are expected to result from the warming of cool regions and from the enhancing effect of carbon-dioxide enrichment on photosynthesis and on water-use efficiency. However, the problems to be encountered are likely to override the beneficial effects in many cases (Rosenzweig and Hillel, 1998). In a warmer climate, both rainfall and evaporation are expected to be greater than at present, but the two will not be commensurate everywhere. Some places will be more humid, while other places will become more arid. All climatic processes are likely to intensify. Not only average conditions will change but also their variability, with increasing frequency and severity of extreme events such as floods, heat waves, and droughts.

Rainfall in arid regions is generally not normally distributed. The occurrence of just a few seasons with anomalously high rainfall, however infrequent and unpredictable, tends to increase the average and skew the distribution. Consequently, the average (mean) rainfall does not represent the most frequent (mode) or the median rainfall. The greater number of seasons in fact have less than average rainfall. This affects farming, which cannot be carried out "on average" but is an annual enterprise, the failure of which in any given season (and even more so in successive seasons) can cause widespread famine.

The interannual distributions of both surface-water and groundwater supplies are further skewed by the fact that neither the process of runoff (feeding streams) nor the process of deep percolation (replenishing groundwater aquifers) is simply proportional to rainfall. Though both processes are functions of rainfall, their dependence is nonlinear. There appears to be a threshold amount (and intensity) of rainfall below which practically no runoff occurs, and a different threshold below which practically no groundwater recharge occurs. As rainfall amounts exceeding those thresholds, each of the two processes, separately, appears to respond to increasing rainfall with an exponent greater than unity.

Under the circumstances, the task of sustaining food security and environmental integrity seems bound to become ever more difficult.