The Regional Impacts of Climate Change

Other reports in this collection

8.3.3. Hydrology and Water Resources Hydrological Trends and Variability

Several reports of recent trends in precipitation and streamflow have shown generally increasing values throughout much of the United States; in Canada, total precipitation trends indicate an increase, but monthly streamflow analyses show varying seasonal changes. Lettenmaier et al. (1994) analyzed data over the period 1948-88 and found generally increasing trends in precipitation during the months of September to December and increasing trends in streamflow during the months of November to April, particularly in the central and north-central portions of the United States. Similarly, Lins and Michaels (1994) report that streamflow has increased throughout much of the conterminous United States since the early 1940s, with the increases occurring primarily in autumn and winter. In Louisiana, precipitation and simulated runoff (streamflow per unit drainage area) have increased significantly over the past 100 years (Keim et al., 1995).

Mekis and Hogg (1997) analyzed annual and seasonal precipitation (total, rain and snow) trends for periods from 1948-96 to 1895-1996 for regions of Canada and noted significant increases in total annual precipitation and snow for most regions. In Ontario, 41 hydrometric stations with a minimum of 30 years of data ending in 1990 were analyzed by Ashfield et al. (1991). Mean monthly flows increased for the period September to January in more than 50% of the stations; approximately 25% of the stations show a downward trend in flow for the April to September period. Anderson et al. (1991) analyzed low-, average-, and maximum-flow time series for 27 stations (unregulated flow) across Canada; the data show a decrease in summer low flows and an increase in winter average and low flows but little trend in seasonal maximum flows. Burn (1994) analyzed the long-term record of 84 unregulated river basins from northWestern Ontario to Alberta for changes in the timing of peak spring runoff. In the sample, the more northerly rivers exhibited a trend to earlier spring snowmelt runoff; the observed impacts on timing were more prevalent in the recent portion of data. These trends generally are consistent with climate models that produce an enhanced hydrological cycle with increasing atmospheric CO2 and warmer air temperatures, although some of the streamflow trends also may be the result of water-management or land-use changes that reduce surface infiltration and storage.

Recent investigations have shown how natural modes of variability at scales from seasons to years (e.g., ENSO, Pacific Decadal Oscillation) affect hydrological variability in different regions of North America and thereby have underlined the importance of increasing our understanding of the roles these features play in influencing hydrological characteristics. The ENSO phenomenon, a predictable climate signal, affects precipitation and streamflow in the northWestern, north-central, northeastern, and Gulf coast regions of the United States (Kahya and Dracup, 1993; Dracup and Kahya, 1994). For example, La Ni�a events (the cold phase of the ENSO phenomenon) produce higher than normal precipitation in winter in the northWestern United States, whereas El Ni�o events (the warm phase of the ENSO phenomenon) cause drier winters in the NorthWest on roughly a bidecadal time scale. Precipitation over a large region of southern Canada extending from British Columbia through the prairies and into the Great Lakes shows a distinct pattern of negative precipitation anomalies during the first winter following the onset of El Ni�o events; positive anomalies occur in this region with La Ni�a events. On the other hand, the northern prairies and southeastern NorthWest Territories show significant positive precipitation anomalies with El Ni�o events (Shabbar et al., 1997). Variability in ENSO phenomena contributes natural variations in hydrology at decadal and longer time scales that are problematic for CO2 climate change analysis. Changes in ENSO behavior related to increasing CO2 are highly uncertain but could produce enhanced variability in precipitation and streamflow for the regions most sensitive to ENSO fluctuations (IPCC 1996, WG I, Section 6.4.4).

Wetlands in North America traditionally have been viewed as wasted land available for conversion to more productive use. This opinion has contributed to the loss of millions of wetland hectares that have been drained or filled for agriculture, highways, housing, and industry. In Canada, where wetlands occupy an estimated 14% of the landscape, 65-80% of Atlantic coastal marshes, southern Ontario wetlands, prairie potholes, and the Fraser River delta have been lost-largely to agriculture (Environment Canada, 1986, 1988). Figures for the United States indicate that approximately 53% of the original wetland area in the lower 48 states has been lost, mostly (87% of this figure) to agriculture (Maltby, 1986). These losses are accompanied by the loss of ecological, hydrological, and cultural functions wetlands provide, including water purification, groundwater recharge/discharge, stormwater storage/flood control, sediment and pollutant sequestering, carbon storage, cycling of sulfur, and wildlife habitat (Mitsch and Gosselink, 1986; IPCC 1996, WG II, Chapter 6).

Socioeconomically, wetlands provide direct benefits through the harvesting of timber, wild rice, cranberries, and horticultural peat-as well as through recreational activities such as hunting, fishing, and bird watching. The cultures and spiritual values of many First Nation peoples are linked to the health of wetlands.

Other reports in this collection