Increases in average atmospheric temperature accelerate the rate of evaporation and demand for cooling water in human settlements, thereby increasing overall water demand, while simultaneously either increasing or decreasing water supplies (depending on whether precipitation increases or decreases and whether additional supply, if any, can be captured or simply runs off and is lost). Shimizu (1993, quoted in Mimura et al., 1998) showed that daily water demand in Nagoya, Japan, would increase by 10% as the highest daily temperature rose from 25 to 30°C. Boland (1997) looked at several climate transient forecasts for the Washington, DC, metropolitan area for the year 2030 and estimated increases in summertime use of 13 to 19% and annual use of –8 to +11% relative to a future increase from 1990 without climate change of approximately 100%. In China, using four GCMs for the year 2030, water deficiency under normal (50%) and extreme dry (95%) hydrological conditions for various basins was predicted as -1.6 x 10⁸ to 1.43 x 10⁹ m³ in the Beijing-Tianjin-Tangshan area and -1.1 x 10⁸ to 121.2 x 10⁸ m³ in the Yellow River Basin (China Country Study Team, 1999).

Estimates of effects on water supply mostly have dealt with linking atmospheric and hydrologic models in an attempt to produce more plausible forecasts with statistical variability. Although they do not directly forecast water supplies to human settlements, Kwadijk and Rotmans (1995) do show an increase in variability of supply in the Rhine River with climate change, but little change in estimated annual flow. The study is unusual in that it attempts to directly link impacts to mitigation policy. A more conventional study that still links water supply (directly for a municipal water system) and climate scenarios is Wood et al. (1997).

Additional research since the SAR seems to have added to concerns about increased intensity of rainfall and urban flooding (e.g., for the highly urbanized northeast United States—Rosensenzweig and Solecki, 2000). Increases in intensity are projected by Fowler and Hennessy (1995) and Hennessy et al. (1997). Smith et al. (1999) have performed a series of four case studies that combine climate and hydrological modeling to directly model flood frequency and magnitude and economic losses under enhanced greenhouse rainfall intensities. The study concludes that there would be little change in forecast flood damage by the year 2030 but that there would be substantial increase in flood risk (shortened average return interval) and flood damage (as a result of building inundation and failure) by the year 2070. The estimation technique was an improvement over earlier studies that assumed that changes in intensities of rainfall events would be associated with changes in flood frequency (e.g., Minnery and Smith, 1996).

Generally speaking, climate change will change the level and type of climatic effects that need to be covered by infrastructure design codes. This could affect infrastructure durability and energy usage. Potential changes in humidity and climate may change distribution in factors such as termites, for example—with potential degradation of structures and more serious impacts from given extreme events. There also could be adverse effects of storms, heat, and humidity on walls and insulation, though perhaps less winter damage.

There still are relatively few reports that estimate the impact of global warming on the value of economic losses resulting from effects on infrastructure. However, Smith et al. (1999) and Penning-Powswell et al. (1996) provide estimates that show that increases in damages from urban riverine flooding probably would be substantial. See also the discussion on sea-level rise in Section 7.2.2.2.

IPCC Homepage