climate change and water resources - CiteSeerX

The National Center for Atmospheric Research

CLIMATE CHANGE AND WATER RESOURCES: A Primer for Water Utilities Preliminary Draft Not for citation or reproduction – all figure permissions are pending

Kathleen A. Miller And

David Yates National Center for Atmospheric Research

With Assistance from Conrad Roesch and Jan Stewart

Draft – Not for quotation or reproduction. Figure permissions are pending.

1

CLIMATE CHANGE AND WATER RESOURCES: A Primer for Water Utilities Draft 3/8/2004

The Science of Climate Change What is climate and what does “climate change” mean? We all know that weather varies from day to day, that it changes with the seasons, and that no two years are ever exactly alike. One way to describe the distinction between weather and climate is that “climate is what you expect, and weather is what you get.” In other words, weather describes the evolution of the current state of the atmosphere, while climate is a measure of the expected weather for a particular place, hour of day, and time of year. Climate, as a statistical concept, measures not only expected average conditions, but also the typical range of variability of those conditions. The climate system, as shown in Figure 1, includes the atmosphere, the Sun, oceans, ice, land, vegetation, surface water, and human activities. Interactions among all of these components determine the geographical and seasonal distribution of climates across the surface of the globe. A change in any of these elements can cause changes in global and regional climates. The process may involve a sequence of adjustments and feedbacks in other components of the system. For example, there is a positive feedback between temperature and ice cover. Suppose that a period of increased solar input causes air temperatures to warm. That would tend to reduce the total area covered by ice and snow. Because snow and ice are very bright, they reflect sunlight back into space. With less ice and snow, the surface of the planet would be darker and would absorb more solar radiation. That, in turn, would lead to further warming. The Sun is the source of energy that drives the climate system. Solar radiation heats the atmosphere and the surface of the Earth. To balance the amount of energy coming in from the sun, the Earth must radiate the same amount of energy back to space − primarily in the form of infrared radiation. Greenhouse gases, which include water vapor, carbon dioxide, methane, nitrous oxide and a variety of human-made chemical compounds, trap some of the outgoing infrared radiation. If the energy balance should be upset by, for example, a change in the amount

Draft – Not for quotation or reproduction. Figure permissions are pending. of solar radiation reaching the Earth, or by a change in the amount of greenhouse gases in the atmosphere, then the Earth either warms or cools until a new balance is established.

Figure 1: An idealized graphic of the climate system (from Bureau of Meteorology, Australia).

Figure 2 provides a globally averaged view of the Earth’s energy budget. The term “greenhouse effect” refers to the fact that the atmosphere absorbs most of the infrared radiation leaving the surface of the Earth and re-emits part of that energy back downwards. Increased concentrations of greenhouse gases in the atmosphere would increase the “back radiation” term on the righthand side of the figure. That, in turn, would cause warming of the surface. Increased loss of energy through sensible heat, infrared radiation, and the release of latent heat through evapotranspiration and precipitation would restore the energy balance.

Figure 2: Global heat flows (Kiehl and Trenberth, 1997).

2


3

The global water cycle plays an important role in the global energy balance because evaporation and cloud formation help to regulate both incoming and outgoing radiation. Water vapor is itself one of the most important greenhouse gases, and because a warmer atmosphere can hold more water vapor, it provides a powerful positive feedback to other sources of warming. Clouds, in particular, play a complicated role in the energy balance. They act as a blanket − warming the Earth’s surface by absorbing and emitting thermal radiation. On the other hand, they also act to cool the surface of the Earth by reflecting incoming sunlight back into space. These opposing effects almost cancel each other out, but in our current climate, clouds appear to have a slight net cooling effect. Because the Earth is a sphere, the Sun’s heating is uneven (Figure 3). There is an energy surplus near the equator and a deficit near the poles. This imbalance works with the rotation of the Earth to drive the circulation of the atmosphere and oceans. Those movements transport heat from the tropics toward the poles, making the Earth’s tropical regions cooler, and its polar regions warmer, than they would be if the Earth had no atmosphere or oceans.

Figure 3. Heating dynamics of the Earth.

The atmosphere and oceans are constantly in motion. That motion displays some stable patterns, which define contrasting climatic zones. For example, the Intertropical Convergence Zone (ITCZ) is a broad band that girdles the equator and is characterized by rising air, frequent convective storms, and high annual precipitation. Just north and south of the ITCZ, centered at latitudes around 30 degrees north and south, are bands of hot, dry, descending air that create deserts in the world’s subtropical regions. In the temperate regions, storms are steered by broad wind


4

bands, called jet streams, that flow from west to east. The position of each jet stream migrates with the changing seasons, and planetary waves of high- and low-pressure regions develop within the jet streams. These vary over time, but they are anchored, to some extent, on underlying geographical features such as mountains and boundaries between oceans and land. That anchoring results in semi-permanent predominant storm tracks that help to define the characteristics of regional climates. Figure 4 is a generalized picture of these circulation patterns. Seasons are caused by the fact that the axis of the Earth is tilted. As the Earth orbits around the Sun, the northern and southern hemispheres alternate between pointing slightly toward or away from the sun, so they experience opposite periods of winter and summer. During each hemisphere’s winter season, there is a greater imbalance between the energy deficit at that pole and the energy surplus in the tropics. This contributes to the formation of water storm fronts, as the poleward transport of heat intensifies. Figure 4: Generalized representation of circulation patterns.

In addition to these broad global climate patterns, local climates are heavily influenced by proximity to large bodies of water and by the location of mountain ranges. For example, the windward side of a mountain range may receive considerably more precipitation than nearby locations on the downwind side. What one calls a climate change depends on the time period being considered. Climate varies naturally from one year to the next, and over longer time periods as well, so the distinction between climate variability and climate change is somewhat fuzzy. Basically, any trend or persistent change in the statistical distribution of climate variables (temperature, precipitation, humidity, wind speed, etc.) constitutes a climate change. Regional climate changes may be caused by persistent changes in the details of oceanic and atmospheric circulation. For example, the El Niño-Southern Oscillation (ENSO) phenomenon causes changes in the distribution of heat within the Pacific Ocean and the surrounding atmosphere. That, in turn, leads to changes in predominant storm tracks. The effects on local climates can be striking, with some areas receiving much heavier than normal precipitation, while other areas experience severe drought.


5

Figure 5 demonstrates that the effects of El Niño episodes (warm events) and the cool events, known as La Niña, are felt across the entire globe.

Figure 5: Expected seasonal effects of El Niño (warm episodes) across the globe during December−February (top) and expected seasonal effects of La Niña (cold episodes) during the same time period (from Climate Diagnostics Center, NOAA).

There are also longer-term changes in ocean-atmosphere circulation − marked by shifts in the location and/or intensity of the semi-permanent high- and low-pressure cells. These changes can persist for several decades. For example, temperature and circulation patterns in the North Pacific appear to get “stuck” in one of two modes for long periods of time. Various indices are used to measure this tendency, but they are all strongly linked to the intensity and position of the winter Aleutian low-pressure system. Figure 6 displays one such index: the Pacific Decadal Oscillation (PDO) Index. When the PDO is in its positive coastal warm phase, as it was for most of the period from 1977 through the mid-1990s, sea surface temperatures along the west coast of North America are unusually warm, the winter Aleutian low intensifies, and the Gulf of Alaska is unusually stormy.


Coastal Warm Phase

Temperature Anomaly o -C

6

Coastal Cool Phase

Figure 6: Pacific Decadal Oscillation. Upper panel: sea surface temperature and wind stress anomalies. Lower panel: Monthly values of PDO Index. Red is coastal warm phase; blue is coastal cool phase (courtesy of Dr. Nathan Mantua, JISAO, University of Washington).

The slowly evolving state of the ocean, as measured by the PDO, interacts with the more rapid ENSO-related changes to influence storm tracks and, thus, the likelihood of unusually heavy or light seasonal precipitation. For example, a positive PDO appears to reinforce the effects of an El Niño, making wet winter conditions in the southwestern United States and dry conditions in the Pacific Northwest more likely than would be the case if the PDO were in the negative (coastal cool) phase. A similar pattern of multi-year variability occurs in the Atlantic basin as well. The North Atlantic Oscillation (NAO) measures swings in the relative intensity of the winter low-press cell centered


7

over Iceland, and the high-pressure cell centered over the Azores. The NAO is defined as being positive when that pressure difference is larger than normal. A positive NAO pattern drives strong, westerly winds over northern Europe, bringing warm stormy winter weather, while southern Europe, the Mediterranean and Western Asia experience unusually cool and dry conditions (Figure 7a). Also in the positive phase, northeastern Canada is more likely to experience unusually cold winter conditions. In the negative phase, the pressure differential is smaller than average and winter conditions are unusually cold over northern Europe and milder than normal over Greenland, northeastern Canada, and the Northwest Atlantic. There have been long periods during which the NAO has tended to be either unusually low or unusually high. In particular, it was generally low throughout the 1950s and 1960s, and then abruptly switched to a positive state for most of the period from 1970 to the present (Figure 7b). Figure 7a: Schematic of the positive index phase of the NAO during the Northern Hemisphere winter (courtesy of Dr. James Hurrell, CGD/NCAR).

Figure 7b: NAO Index 1864-2003 (courtesy of Dr. James Hurrell, CGD/NCAR).


8

ENSO, the PDO, and the NAO are all natural modes of climate variability, but any change in global climate is also likely to affect these processes. At the global scale, climate changes are linked to changes in the Earth’s energy budget. Global warming, in particular, refers to the rise in the average surface temperature of the Earth that is expected to result from increased concentrations of greenhouse gases, such as carbon dioxide, in the atmosphere. The Earth’s climate has changed throughout geologic time − why did those changes occur? There is good evidence that the Earth has experienced long periods during which average global temperature were much colder and much warm than today (Figure 8). Changes in the Earth’s climate system throughout geologic time can be linked to changes in the components of the climate system, including changes in the Earth itself, the composition of the atmosphere, and the seasonal distribution and total amount of incoming solar energy.

Warm period of dinosaurs ◄

Extensive glaciation of supercontinent of Pangea ◄

Figure 8: Estimated mean global temperatures over the last 570 million years.


9

There have been enormous changes in the surface of the Earth − with continents moving, mountain ranges growing and eroding away, and the area covered by oceans and by ice growing or shrinking. The composition of the atmosphere has also changed as a result of biological and geophysical processes, including storage of carbon in the ocean and its subsequent release, volcanic eruptions, and the occasional sudden release of methane from sediments on the ocean floor. In addition, there have been changes in solar output, and in the Earth’s orbit, and Earth-Sun geometry. All of these changes affect climate at both the global and regional scale. Consider, for example, the effects of slow changes in the Earth’s orbit around the Sun. Over the course of approximately 100,000 years, the Earth’s orbit around the Sun changes shape from a thin oval to a circle, and back again. At present, the shape of the Earth’s orbit is almost a perfect circle. There is only a small difference in our distance from the Sun at the time when we are closest to it (the perihelion, currently in January), and when are farthest away (the aphelion, currently in July). The fact that the Earth is now closest to the Sun during the northern hemisphere winter is just a coincidence, because the date of the perihelion slowly moves to come later in the year, following a 21,000-year cycle. In other words, 10,000 years from now, the perihelion will occur in the northern hemisphere summer, causing Northern Hemisphere seasonal contrasts to be somewhat more pronounced than at Figure 9: Graphic illustration of the Earth’s orbit and average solar radiation.

present (Figure 9).

Even such subtle differences can have profound impacts on regional climates. When the perihelion last occurred in northern hemisphere summer, the Sahara was much wetter than it is now, and was covered with savanna-like vegetation. As the seasonal distribution of solar


10

radiation gradually changed to modern conditions, the Sahara dried out. Its transformation to the present day desert accelerated dramatically about 5,500 years ago. The abruptness of the change suggests that the climate system crossed a threshold, triggering a series of biophysical feedbacks that amplified the trend toward regional drying (IGBP, 2001). Seasonal contrasts would also tend to be more extreme when the shape of the Earth’s orbit is more elliptical than it is at present. In addition, the Earth wobbles slightly on its axis, so that the angle of the tilt varies over a 41,000-year cycle. Recall that it is the tilt that causes seasons in the first place. So, the greater the angle of tilt, the stronger the seasonal contrasts. These astronomical Milankovich cycles are believed to have played a significant role in the timing of ice ages and interglacial periods in the recent past, but they clearly cannot explain all of the Earth’s climate history.

Figure 10: Four glacial cycles are recorded in Vostok ice core. The graphic represents thousands of years before the present. Purple is carbon dioxide, red is temperature, green is methane.

Changes in the seasonal distribution of incoming solar energy may have triggered the beginning and end of previous ice ages, but the solar impacts were greatly amplified by positive feedbacks within the climate system, including changes in the reflection of sunlight back into space by icecovered areas, changes in ocean circulation, and dramatic changes in atmospheric concentrations of greenhouse gases, especially carbon dioxide and methane. Over the past 400,000 years, the record of temperatures in the world’s high-latitude regions followed a sawtoothed pattern. Global concentrations of carbon dioxide and methane followed a nearly identical


11

pattern (Figure 10). There were four long but erratic periods of cooling, each followed by a dramatic warm-up. The reasons for this pattern are not fully understood, but changes in the ocean’s thermohaline circulation (Figure 11) and changes in the release of carbon dioxide from the oceans, and the release of methane from wetlands, appear to have played important roles. As can be seen in Figure 10, rapid warming and increases in atmospheric carbon dioxide and methane occurred nearly simultaneously. This suggests a positive feedback loop, with initial warming causing the greenhouse gas concentrations to rise, and rising concentrations promoting further warming.

Figure 11: The Great Ocean Conveyor: global thermohaline circulation.

Figure 11 depicts the pattern of thermohaline circulation in the World’s oceans − that is, the connection between the movement of cold, salty water in the oceans’ depths and the movement of warm, less saline water at the surface (Broecker, 1997). Warm, low-salinity water from the tropical Pacific and Indian Oceans flows around the tip of South Africa and ultimately joins the Gulf Stream to transport heat from the Caribbean to Western Europe. As the water moves northward, evaporative heat loss cools the water and leaves it saltier and more dense. The cold, salty water sinks in the North Atlantic and flows back toward Antarctica, thus pushing the conveyor along. It is hypothesized that inflow of fresh water into the North Atlantic during warm periods can cause this conveyor to dramatically slow down or even collapse. Such a mechanism could explain the sudden reversals of warming that appear in the geologic record.


12

It is possible that increased high-latitude runoff and ice-melt caused by human-induced global warming also could slow the thermohaline circulation. However, we do not know how much that would reduce projected temperature increases for Europe and the northern latitudes, because the mechanisms of human-induced global warming are different from the mechanisms of previous natural warming episodes (IPCC, WG I, 2001). This is an area of active research.

Why should I believe that emissions of carbon dioxide and other greenhouse gases will cause global warming? The major greenhouse gases, carbon dioxide, methane, nitrous oxide and water vapor, occur naturally in the atmosphere. Without them, the Earth would be too cold to support life, as we know it. The basic science of the greenhouse effect is well understood and easily reproduced in the laboratory. There is no controversy about the fact that these gases are transparent to incoming short-wave solar radiation, and that they tend to absorb outgoing long-wave radiation and re-emit part of that radiation back down to the Earth’s surface. In effect, they act as a blanket to warm the surface of the Earth. Table 1 Selected chemically reactive greenhouse gases and their precursors: abundances, trends, budgets, lifetimes, and GWPs.

Chemical species

Carbon dioxide

Formula

CO2

Trend – annual % change

Abundance

Annual emission

100-yr c GWP

Lifetime

(units)

2002

1750

1990s

late 1990s

(yr)

(ppm)

372

280

0.4 %

6.3 +/- 0.4 PgC

~5 to 200

1 23

Methane

CH4

(ppb)

1729 – d 1843

700

0.4 %

600 Tg

12

a

Nitrous oxide

N2O

(ppb)

314

270

0.3 %

16.4 TgN

114

a

296

Perfluoromethane

CF4

(ppt)

80

40

1.3 %

~15 Gg

>50000

5700

Perfluoroethane

C2F6

(ppt)

3.0

0

2.7 %

~2 Gg

10000

11900

Sulphur hexafluoride SF6

(ppt)

4.2

0

5.7 %

~6 Gg

3200

22200

HFC-23

~7 Gg

260

12000

CHF3

(ppt)

14

0

3.9 %

CFC-11

b

CFCl3

(ppt)

268

0

-0.5 %

45

4600

CFC-12

b

CF2Cl2

(ppt)

533

0

0.8 %

100

10600

Sources: IPCC WGI, 2001; Blasing and Jones, 2003. a

Species with chemical feedbacks that affect the duration of atmospheric response – values are perturbation lifetimes Regulated under Montreal Protocol c Global Warming Potential (GWP) is an index describing the relative effectiveness of well-mixed greenhouse gases in absorbing outgoing infrared radiation. The index approximates the time-integrated warming effect of a unit mass of a given greenhouse gas relative to that of carbon dioxide. d As measured at Cape Grim, Tasmania and Mace Head, Ireland, respectively (Blasing and Jones, 2003). b


13

Concern about global warming arises from the fact that human activities are releasing large quantities of these substances − and other even more powerful manufactured greenhouse gases such as halocarbons − into the atmosphere (Table 1). Because carbon dioxide and many of the halocarbons have very long atmospheric lifetimes, the increased concentrations are expected to result in an enhanced greenhouse effect for many years to come. We are also loading the atmosphere with other types of pollutants. Some of these tend to produce cooling by reflecting incoming sunlight. Dust from disturbed soil surfaces and other tiny particles from combustion, including soot and sulphate aerosols, act in this way. Unlike carbon dioxide and many other greenhouse gases, however, these aerosols only stay in the atmosphere a very short time. So, although they may temporarily mask the warming effects of the greenhouse gases, warming will eventually dominate. Figure 12 depicts the estimated relative impacts of greenhouse gases, aerosols, and other factors on global temperatures from pre-industrial times (circa 1750) to the present (circa 2000).

Figure 12: Many external factors force climate change.

Over the past 400,000 years, atmospheric carbon dioxide concentrations varied from about 180 parts per million (ppmv) at the height of each glaciation to about 310 ppmv at the peak of each


14

warming. Similarly, methane concentrations varied from approximately 350 to 800 parts per billion (ppbv). Since the beginning of the Industrial Revolution, burning of fossil fuels, deforestation, expanding agriculture, and other human activities have contributed to rapid increases in CO2 and methane concentrations. In the mid-eighteenth century, the estimated atmospheric concentration of CO2 stood at 280 ppmv. As of the year 2002, it had risen to approximately 372 ppmv. Similarly, methane concentrations increased from approximately 700 ppbv at the beginning of the Industrial Revolution to current levels between 1,729−1,843 ppbv, as measured at different locations. These modern levels are, thus, well above the range of natural variability in the recent geologic past. Future emissions are expected to further increase these concentrations (Figure 13).

Figure 13: Carbon dioxide (CO2) and methane (CH4) concentrations: past, present, and future. Compiled by K. Alverson, PAGES IPO.


15

Why is there so much uncertainty about future climate changes? There are really two big questions here: 1) How much warming is likely to result from humancaused increases in greenhouse gas concentrations? And 2) What will that do to local and regional climates? The recent Third Assessment Report of the Intergovernmental Panel on Climate Change (IPCC) estimates that global average temperature will rise by between 1.4o to 5.8o C by the year 2100. This rather wide range of uncertainty results primarily from the fact that it is difficult to forecast future emissions, and also from the fact that the ultimate warming will depend on the size and direction of many feedback processes in the climate system that cannot be precisely estimated. Changes in atmospheric water vapor and cloud formation are two of the most important processes in this regard. The warming from increased CO2 will tend to be amplified by increases in atmospheric water vapor, while changes in the extent of cloud cover and the characteristics of clouds may either enhance or diminish the initial warming. Attempts to account for the range of uncertainty in these feedbacks results in a range of possible changes in global average temperatures for any given change in CO2, or its equivalent in other greenhouse gases. Future emissions are the real wild-card because they depend on how fast the world economy grows, how fast world population increases, how quickly our energy technology evolves, how much our land uses change and what policies we put in place to guide these processes.


16

Figure 14 presents a range of possible future paths for CO2 emissions along with two different estimates of the resulting changes in the atmospheric concentrations of CO2.1 One of the important things to note is that atmospheric CO2 concentrations will be higher in the year 2100 than they are now, even in the scenarios in which emission rates eventually decline significantly relative to present rates. This suggests that some global warming will be inevitable. Recall that CO2 has a long atmospheric lifetime, and that emissions cannot be avoided completely – even under the most optimistic assumptions about future innovations in energy technology. The other important thing to notice is that there are huge differences in projected CO2 concentrations at the end of the century, depending on the development path Figure 14: A possible range of carbon dioxide emissions and the resulting atmospheric changes.

followed by the world economy. Also note that there is uncertainty about the eventual

CO2 concentrations that would result from any given emission scenario – arising, largely, from our incomplete understanding of possible changes in the uptake and release of carbon by biological processes on both land surfaces and in the ocean.

1

The SRES emissions scenarios pictured here were developed as part of the IPCC 2001 assessment process. They represent a wide range of possible futures, as follows: A1Fl = rapid economic growth, continued reliance on fossil fuels, converging world living standards, world population peaking in mid century and declining thereafter. A1T = Same as above except with increasing reliance on new technologies using renewable energy rather than fossil fuels A1B = Same as above except with a balance of fossil and non-fossil fuel sources A2 = regionally divergent economic growth, continuing population growth, slower and more fragmented technological change B2 = emphasis on local solutions to economic, social and environmental sustainability, intermediate technological change, economic growth and population growth B1 = population as in A1, rapid change toward service and information economy, emphasis on clean, highly resource-efficient technologies.


17

Scientific understanding of the sensitivity of the climate system to changes in the concentrations of CO2 and other trace gases is also imperfect. Different climate models will produce different projected temperature changes because they incorporate different estimates of the parameters that describe the behavior of the climate system. The range of temperature changes projected by the IPCC reflects the combined effects of all of these sources of uncertainty. Figure 15 compares the range of IPCC temperature projections over the coming century with the estimated record of global average temperature changes over the past 1000 years. The gray shading represents the range of uncertainty in both the projections and the record of past variation.

Figure 15: Variations in the Earth’s surface temperature from the year 1000 projected to the year 2100.

Of course, global average temperature is a very crude metric of climate change. Nobody lives at the global average. What we really care about is what will happen to climate at particular places, and temperature is only one of several important variables. Water supplies, for example, will be affected by changes in temperature, precipitation (including changes in timing and intensity), insolation, humidity, and wind-speed, among other factors. In addition, many human and natural


18

systems are likely to be sensitive to changes in extremes (e.g. the frequency and severity of floods and droughts). Unfortunately, the details of climate changes at particular places and times in the future cannot be reliably predicted at this time – even if we could reliably predict the change in global average temperature. We will discuss this lack of predictability, and its implications for water utility planning below. Here, it is important to emphasize that while the details of local climate changes cannot be predicted, we are beginning to accumulate some evidence on the likely characteristics of climate changes on gross regional scales. This body of evidence is sufficient to allow utilities to explore the implications of a range of potential local climate changes that are consistent with projections of global warming.

Is global warming really likely to happen on a time scale relevant to water utilities?

Figure 16: Variations of the Earth’s surface temperature for the past 140 years (1860 to 2000).

First, it is important to understand that global warming is already happening. Over the past century, global average surface temperature increased by approximately 0.6o C (Figure 16). Warming is expected to accelerate during the current century. The modest warming, to date, has not been evenly distributed over the surface of the globe. In particular, arctic areas have warmed more rapidly than other areas. Climate model simulations also suggest that future warming will tend to be most pronounced near the poles (Figure 17). That picture might change if there is a


19

significant slowing of the oceanic thermohaline circulation, because that would reduce the poleward transport of heat through the ocean.

Figure 17: The multiple model ensemble map for the end of the 21st century predicts that most warming will occur over the Arctic and land areas, when compared with the 1961-1990 Normals.

As for impacts on water resources, warming over the past half century appears to be associated with reduced spring snow packs in some of the mountainous areas of the western United States. In California’s Sacramento River Basin, for example, spring runoff has been peaking earlier, and there has been a century-long downward trend in late spring and early summer flow as a proportion of total annual flow. (Dettinger and Cayan, 1995).

Warming could have some benefits, couldn’t it? Why is it usually portrayed as some sort of catastrophe? Global warming will certainly produce a mixture of both beneficial and harmful impacts. Beneficial impacts might include reduced winter heating demands and longer growing seasons in some areas, while harmful impacts will include the health and energy demand impacts of more frequent summer heat waves, and increased stress on natural ecosystems. Here, we are particularly interested in the possible impacts of climate change on the water utility industry. As will be described below, large changes are possible in total available water supplies, in the seasonal distribution of surface flows, in water quality and in the frequency and severity of flood and drought events. However, the details of how these changes will unfold at any given location are likely to remain highly uncertain. Water demands, particularly for irrigation, are likely to change


20

as well. Effective adaptation to such changes may require careful evaluation of the implications of a wide range of possible future climate scenarios.

If we institute policies to slow down or stop the growth in emissions, how long would it take for climate to stop warming? Global warming cannot be “turned off” immediately. The concentrations of greenhouse gases in the atmosphere, rather than current emissions are what determine how warm the climate will be. To draw upon a water resources analogy, atmospheric concentrations of CO2 can be thought of as a large reservoir, while current emissions are like a small stream entering that reservoir. The level of the reservoir rises or falls depending upon whether the natural draw-down processes (e.g. evaporation in the case of water) are larger or smaller than the current rate of inflow. There are natural processes by which CO2 and other greenhouse gases are removed from the atmosphere (including uptake and storage of CO2 in biota, soils or ocean sediments), but we are currently adding these gases much faster than they are being removed. In addition, these natural sinks may become saturated as CO2 concentrations rise. On average, natural sinks currently remove over half of the carbon emitted by fossil fuel use each year, but these processes could become less effective in a warmer world. For example, the solubility of CO2 in seawater declines as the water warms (IGBP, 2001), which would reduce the effectiveness of the ocean sink.

Figure 18: Possible time paths for selected stabilization targets, and the estimated range of global temperature changes that would result from stabilization of carbon dioxide at each level.


21

Stopping global warming would require stabilizing atmospheric concentrations of greenhouse gases2. The reductions in emissions that would be required to do that, depend on the levels at which CO2 and other greenhouse gas concentrations are to be stabilized, and the target date for that stabilization. Figure 18 depicts possible time paths for selected stabilization targets, and the estimated range of equilibrium global temperature changes that would result from stabilization of CO2 (or its equivalent in other greenhouse gases) at each level.

Impacts on Hydrology Global warming will likely alter the hydrologic cycle in ways that may have substantial impacts on water resource availability and water quality. Some aspects of these changes, particularly at large scales, are understood with a high level of confidence, but our understanding generally diminishes as we move from these large-scale changes down to the more relevant regional and local scale changes. The discussion below describes the climatic changes in which the scientific community has a greater degree of confidence, and those in which there is less confidence primarily due to the complexity of our climate system.

If the earth gets warmer, can we predict how much total global precipitation will change? The one change that appears most likely is that global average annual precipitation will increase as global average temperature rises. Evaporation will increase with warming and a warmer atmosphere can hold more moisture. A simple-minded explanation for the resulting intensification of the hydrologic cycle is that “what goes up, must come down.” Of course, it really isn’t quite that simple, but the overall scientific consensus is that globally the earth will be warmer with higher globally averaged precipitation. Exactly how much global average precipitation will increase is somewhat less certain. The overall precipitation change will depend on how much warming occurs, and as we have discussed, that is highly uncertain. However, there is uncertainty even if we focus on the narrower question of how much global average precipitation will increase for each degree of warming, with a high-end estimate of a little over 6% per degree Celsius by applying well-known atmospheric principles. However, most current climate models suggest a far smaller increase, averaging about 3% per degree Celsius for warming forced by CO2. The difference results from the dynamics of the

2

Note that it is actually the radiative forcing of the entire suite of greenhouse gases that would have to be stabilized, so there may be several ways to achieve a mix of emission reductions to meet any specific stabilization target.


22

atmospheric energy budget, as simulated in the models. Recent research suggests that the nature of the “forcing” also matters. In other words, warming caused by increased solar radiation may entail a greater increase in precipitation than would arise from an equivalent temperature change caused by increased CO2 concentrations (Allen and Ingram 2002). There is reason to believe that, on average, precipitation will tend to be less frequent, but more intense when it does occur (Trenberth et al., 2003), implying greater incidence of more extreme floods and droughts, with somewhat obvious implications. Most precipitation comes from moisture already in the atmosphere at the time the storm begins, and transport of heat and moisture by the storm-scale circulation into the storm often creates greater storm vigor. Precipitation amount is governed by the surface energy and water budgets through evaporation, implying that an increase in intensity is offset by a decrease in duration or frequency of events. This is depicted graphically in the simple figure below, where the time increments are represented on the horizontal axis, considering both “current” climate and a “warmer” climate.

Current atmosphere

Warmer atmosphere

time Figure 20. A highly stylized representation of changes in precipitation frequency and intensity under current (top) and a warmer (bottom) atmosphere. The blue dots represent precipitation, while all others are water vapor. The red dots are used to simply indicate atmospheric water vapor attributable to evaporation.


23

Can climate models tell us where it will get wetter or drier? An increase in global average precipitation does NOT mean that it will get wetter everywhere and in all seasons. In fact, all model simulations show complex patterns of precipitation change, with some regions receiving less precipitation in warming scenarios than they do now, while other regions receive more. The major difficulty is that different model simulations show slightly different regional temperature changes, and often very different regional precipitation changes. To understand why this occurs and what it implies for the usefulness of climate model projections, it is useful to begin with an explanation of what climate models are, and how they are used to simulate present and future climates. Coupled Atmosphere-Ocean General Circulation Models (AOGCMs) are currently the primary tool used to analyze the potential impacts of increased greenhouse gases, aerosols and other factors on global climate. The atmospheric part of a climate model is a mathematical representation of the behavior of the atmosphere. A horizontal and vertical grid structure (as depicted in figure 21) is used to track the movement of air parcels and the exchange of energy and moisture between parcels.

Figure 21 The structure of an atmospheric GCM Source: Henderson-Sellers and McGuffie 1987.

To be useful for the analysis of climate change, the atmospheric model must be coupled to models of other components of the climate system, such as the oceans and sea ice. The major climate models currently in use may include several vertical layers in the atmosphere and the oceans, as well as a dynamic sea-ice model. State-of-the-art climate models also include the


24

effects of changes in vegetation and other land surface characteristics. Climate models differ in the degree of complexity with which the various components of the climate system are modeled and in the way in which the coupling in carried out. While greater detail in the processes represented by a model may provide greater realism, computational costs and uncertainties in process representation also increase. There is no agreement as to a single best approach to climate modeling, but, in general, model development has proceeded from the early, simple atmosphere-only models to present-day coupled models designed to capture responses and feedbacks involving several components of the climate system (Gates et al., 1996; Washington, 1996). Despite improvements in the ability of these climate models to simulate large-scale climate processes, there continue to be differences across models in their representation of large-scale patterns of moisture transport. For example, simulations run on different models may project slightly different storm tracks and patterns of change in sea surface temperatures – leading to different conclusions about the location and seasonal timing of precipitation changes over land surfaces. Nevertheless, when one compares climate change simulations produced by several different climate models broad similarities in the patterns of regional precipitation change are evident. In general, the models agree in projecting precipitation increases over high latitude land areas, smaller and less certain increases over the equatorial regions, and decreases over some subtropical areas. Elsewhere, precipitation changes are more variable across models (Carter and Hulme 1999; IPPC 2001). Wigley (2002) has developed a statistical summary of the spatial distribution of precipitation change, and Figure 22 displays these results in the form of normalized signal to noise ratios – with the noise representing the scatter among model projections. 90 70 50 Latitude_pr

30 10 -10 -30 -50 -70 -90 -180 -2

-120 -1

-60 0

0 Longitude_pr 1

60

120 2

3 o

Figure 22 – Normalized signal to noise ratio – percentage change in precipitation per 1 C warming. Source: Wigley, 2003


25

Figure 23 represents these findings in a different way. Here, the values on the map are measures of the modeled likelihood of an increase in winter precipitation for a 1 degree Celsius warming, based on the output of nine AOGCMs. The light magenta values at the top of the scale indicate that 95% or more of the scenario simulations show increased winter precipitation. At the opposite end of the scale, the brown values indicate that the simulations suggest a less than 5% chance of increased winter precipitation.

Figure 23 – scenario frequency of increased winter precipitation per 1 C Source: Wigley 2003

o

warming

It is important to remember, however, that these simulation frequencies are not the same thing as true probabilities because they are dependent on the models incorporated in the analysis, which may not fully capture the true dynamics of the climate system. So far, we have discussed only changes in average annual temperatures and precipitation at the global scale, but it will be the changes in runoff that determine changes in water availability. Runoff changes will depend on changes in temperatures and precipitation, among other variables. A study by Arnell (1999) used two climate model simulations to examine changes in annual average land surface runoff by 2050 (Figure 24). The striking thing about this figure is the fact that both simulations yield a global average increase in precipitation (not shown), but likewise exhibit substantial areas where there are large decreases in runoff. Thus, the global message of increased precipitation clearly does not translate into regional increases in water availability. Also, one gets a feeling for the uncertainty related to climate projections, as these different simulations produce quite different regional impacts. North America in particular shows nearly opposite results between the two model runs, and there are other important differences.


Figure 24 – Change in average annual runoff by 2050 relative to average runoff for 1961-1990 Panel (a) HadCM2 ensemble mean; Panel (b) HadCM3

26


27

Can climate change projections help us to identify future changes in regional runoff and groundwater recharge? From the above discussion, one should not expect any single climate model projection to yield a reliable prediction of regional hydrologic changes. In fact, one should be cautious about referring to regional hydrologic changes derived from AOGCMs as projections. This suggests that it would be prudent to consider a range of changes projected by different models. However, before launching into such an analysis, it is important to understand that the ability of AOGCMs to accurately represent regional and local climates is limited by the coarse horizontal scale at which the models resolve the earth, ocean, and atmospheric systems. A typical AOGCM divides these systems into 3-dimensional space, represented by a set of grid points at which coupled equations of mass, energy, and momentum are written. When a climate model is run to simulate present or future climate (a “model integration”), these equations are iteratively solved over long-time horizons of perhaps hundreds of years at time-steps as short as several minutes, and with horizontal resolutions of perhaps 10’s to 100’s of kilometers. The temporal scale is necessarily short since the underlying atmospheric equations are based on first-principles, whose dynamic processes must be resolved at these short timescales at which the processes operate. Despite tremendous technological advances in computing capability, model integrations with this type of configuration still take a very long time to complete. One of the most important compromises for achieving model integrations in a reasonable amount of time is to increase the model’s horizontal resolution. This limitation means that it is prohibitively costly to run full coupled climate models at a spatial resolution that would be sufficiently fine to accurately depict the effects of mountains and other complex surface features on regional climates. The problem with such a coarse horizontal resolution is that important processes that occur at finer scales are not well resolved. Topography, for example, is very important in determining the location of precipitation. As moist air rises over mountains or hills, the moisture condenses, producing clouds and if conditions are right, precipitation occurs. The coarse horizontal resolution of typical climate models essentially smoothes out these landscape effects. At the resolution of most AOGCMs the models see the mountains of the western United States as a smooth ridge with a high point in the vicinity of central Utah. Clearly, that spatial resolution is too coarse to reproduce the effects of topography on the region's precipitation and runoff patterns (Grotch and MacCracken, 1991; Giorgi and Mearns, 1991). For example, the Great Basin area, which is a desert, would be predicted to be wet because it would be seen as located on an upslope. The global-scale models cannot capture the actual rain-shadowing effect of the Sierra Nevada Mountain Range. In short, raw AOGCM output will put the precipitation in the wrong places and perhaps at the wrong time.


28

The recognition of this deficiency has led to the application of downscaling as a means of trying to understand how local scale processes might respond to larger-scale weather and climate changes as represented by AOGCMs. Downscaling is also used to derive finer scale datasets that are used to force hydrology and land-surface related models. A key reason for developing downscaled datasets is the hypothesis that the statistical characteristics of the downscaled data will differ from that of the historic record, not only in the mean condition, but also in other statistical attributes such as the sequences of storms and dry periods. Climate scenarios generated through downscaling techniques involve the development of statistical relationships between historic meteorological observations and outputs from regional and/or global climate models [Wilks 1992; Robock et al. 1993; Easterling 1999; Hewitson and Crane 1996; Semenov 1997; Wilby et al. 1998; Sailor and Li 1999a,b; Mearns et al. 1999; Wood 1997. These methods include 1) purely statistical methods (refs, Yates et al. 2003); 2) stochastic methods based on the relationship between GCM atmospheric circulation patterns and those same surface variables (Refs); and 3) statistical-dynamical downscaling which utilizes regional weather models driven by coarser scale GCM boundary and initial conditions to resolve the finer scale atmospheric processes, which are then related to surface variables. This method captures the stochastic characteristics of large area circulation patterns, which are arguably better represented by GCMS than are surface processes, most notably precipitation (Lettenmaier 1995, Wood 1997, others). While these approaches for generating climate scenarios for impact analysis are useful, they do have limitations. For example, a climate change scenario might be generated from large-scale GCM features such as pressure patterns that are then related to storm tracks and thus surface precipitation. However, the GCM’s simulation of past and current climate might show biases in the current pressure patterns, which would then be propagated to the downscaled precipitation estimate. Thus, this technique in producing either current or future climate sequences for impact analysis can be problematic. So, if an AOGCM does not adequately replicate the historic climate of a region, its utility in helping to generate scenarios of the impacts increasing CO2 and aerosol changes is questionable. Stochastic weather generators have also been used to develop climate datasets for impact analysis. These can address some of the issues just raised with their ability to simulate plausible climate scenarios, and have themselves been used as downscaling techniques in global change studies [Wilks, 1992]. Typically, a stochastic weather generator is developed based on the historically observed data at a location, and can then be used to simulate climate scenarios consistent with the global change scenarios. However, Katz [1996] points out that modifying the


29

parameters of a stochastic model can lead to unanticipated effects. For example, modifying the probability of daily precipitation occurrence using a stochastic weather generator [Richardson 1981] changed not only the mean of daily temperature, but its variance and autocorrelation as well. Doing such downscaled analyses based on the output of multiple climate model simulations can be a very laborious and time-consuming task. That daunting prospect has led a number of researchers to use simpler, almost back-of-the-envelope approaches to explore the possible implications of climate change for water resources. Several analyses have used hypothetical changes in temperature and precipitation amounts by simply scaling a historic record by some predefined amount, essentially amounting to a sensitivity analysis to a climate perturbation. A drawback of this approach is that the hypothetical scenarios may not be internally consistent (Nemec and Schaake, 1982; Flashcka et al., 1987; Schaake, 1990; Duell, 1992; 1994; Nash and Gleick, 1993). Others have derived alternative climate scenarios by using differences between coarse resolution 1xCO2 and 2xCO2 AOGCM simulations imposed on top of historical climate traces. However, it is difficult to assess the plausibility of these scenarios. Despite those drawbacks, systematic analysis of such scenarios can be useful for delineating the relative importance of changes in temperature and precipitation and can provide a fairly inexpensive way to explore vulnerabilities of water supply systems, water quality, and instream resources. One example of such an analysis is a study by Nash and Gleick (1993) that examined the impacts of several climate change scenarios on the flow of the Upper Colorado River. In the hypothetical scenarios, temperature increases of 2 oC and 4 oC were considered, along with precipitation changes ranging from a 20-percent increase to a 20-percent decrease. The study concluded that temperature increases would reduce Lake Powell inflows unless offset by significant precipitation increases. For example, inflows would decline by almost 21% if precipitation remained unchanged while temperatures increased by 4 oC. For a warming of that magnitude, the study concludes that precipitation would need to increase by nearly 20% to avoid reduced reservoir inflows. To put these modeled changes in context, the authors note that treering evidence suggests that a 20% reduction in natural flow from the recent historical record is conceivable even without greenhouse warming. They argue that if such natural variations were coupled with reduced flows arising from greenhouse warming, the impacts would be far more severe than any of the scenarios that they present. A more complete assessment of the possible impacts of climate change on the hydrologic balance of a drainage basin must account for the fact that several of the elements in the balance are likely to be affected. For example, the amount, intensity, and temporal distribution of


30

precipitation are likely to change. Warmer temperatures will affect the proportion of winter precipitation falling as rain or snow, and will increase evaporation. Changes in soil moisture availability, together with changes in plant characteristics, will lead to changes in plant transpiration. Finally, changes in the quantity of water percolating to groundwater storage will result in changes in aquifer levels, in base flows entering surface streams, and in seepage losses from surface water bodies to the groundwater system. To estimate the impacts of climate change on water resource availability, one must, in some way, account for the effects of changes in each of these elements. Several approaches to modeling the relationship between climate and hydrology are available; each differing in the amount of detail used to represent these various processes. Although different hydrologic models can yield different values in terms of streamflow, groundwater recharge, water quality results, etc (Boorman and Sefton 1997; Beven 2001), their differences are usually small in comparison to the uncertainities due to the effect of climate. However, the chain of effects from climate, to hydrologic response, to water resource systems, to the actual impacts on water supply, power generation, navigation, water quality, etc. will depend on many factors, each with a different level of uncertainty. Simulations of the water cycle at the land surface have long been used to assess the potential impacts of climate as well as other forms of change such as land use and land cover conversion on watershed response. It is worth distinguishing between the simulation of physical processes (e.g. the hydrologic cycle), and water resource planning processes and decisions. The former typically focuses on natural processes, while the later has addressed the managed distribution of water through the watershed as it propagates through management elements such as dams, diversions, penstocks, spillways, treatment facilities, canals, etc. Physical simulations have ranged from more stylized water balance models at monthly or longer time-scales to daily and sub-daily models of the rainfall-runoff process that are chiefly aimed at understanding the runoff response of watersheds to climatic forcing of various natures. Advanced physical models attempt to rigorously describe the landscape process by simulating the potential biosphere impacts and biosphere-atmosphere feedbacks, including consideration of ecosystem nutrient cycling and the dynamics of vegetation at the landscape (Neilson and Drapek 1998, Aber et al., 2001). Because these models describe the hydrologic process, they can be used to look at watershed response not only to changes in precipitation forcing but also to landuse/land-cover modification and progression and the subsequent impact on runoff, water quality and water supply.

What do we know with high confidence and what other changes appear likely? •

There is a very high level of confidence in projections of warmer temperatures over most land surface areas. As a result, river basins in which the hydrology currently is


31

dominated by annual snowpack accumulation and melting are likely to experience smaller total snowpack accumulations and earlier melt-off. This melt-off will cause increases in winter or spring flows and tend to reduce summer flows. The altered pattern of seasonal flow results primarily from the effects of warmer temperatures on the form of precipitation and the rate of melt. •

The effects of given temperature and precipitation changes will vary across catchments, depending on the physical characteristics and climatic regime of the catchment (Lettenmaier and Gan, 1990; Schaake, 1990; Arnell, et al., 1996). For example, in an analysis of four sub-basins of the Sacramento, Lettenmaier and Gan (1990) found that the area-elevation distribution of the catchment would determine the extent to which snowmelt remained an important component of seasonal runoff. In addition, differences in soil and geologic characteristics and, therefore, in the moisture storage capacity of the watershed, can affect the maintenance of base flows through the summer.

•

In basins that are currently glaciated, declining glacier reservoir capacity may eventually lead to an earlier peak in the annual hydrograph and reduced late summer streamflows. In the near term, increased melting of glacial ice can, in some cases, sustain summer streamflows (Pelto, 1993).

•

Warmer temperatures could increase the number of rain on snow events in some river basins, increasing the risk of winter and spring floods (Lettenmaier and Gan, 1990; Hughes et al., 1993).

•

In the absence of offsetting increases in precipitation, warmer temperatures could lead to reduced annual streamflows. Warmer temperatures may cause runoff to decline even where precipitation increases. For the Colorado River, Nash and Gleick (1993) estimate that a 4oC temperature increase would cause inflows into Lake Powell to decline unless annual precipitation increased by nearly 20 percent. However, projected declines in annual flows are much less certain than the changes in the seasonal distribution of flows noted above.

•

Water resources in arid and semi-arid environments could be inherently more vulnerable to the effects of global warming depending on the direction and character of change, particularly precipitation. Often, these environments are characterized by highly nonlinear relationships between precipitation and runoff. So, if precipitation becomes more intense and less frequent, and if there is inadequate storage capacity to contain runoff, supplies could decline, even with increased precipitation.

Draft – Not for quotation or reproduction. Figure permissions are pending. •

32

Reservoir evaporation could increase under warmer conditions (Callaway and Currie, 1985). However, while warmer temperatures would tend to increase evaporative losses, the losses would diminish if the surface area of the reservoir shrinks. Nash and Gleick (1993) found that a 10-percent reduction in average annual inflow into Lake Powell could cause evaporative and bank storage losses to decline by as much as 500,000 acre-feet per year as a result of the smaller areal extent of the lake. The net impact on reservoir losses is, thus, uncertain.

•

Water quality may be adversely affected by the impacts of warmer temperatures, an increased frequency of low flow conditions, and possible increases in the intensity of episodic high precipitation events. The location of water infrastructure, including both intakes and pipe distribution networks could be increasingly vulnerable to precipitation extremes. Water quality impacts are, therefore, likely to be rather complex.

How might water utilities in coastal zones be affected by climate change? The IPPC Working Group II Third Assessment Report identifies sea-level rise as one of the most important aspects of climate change at the coast, and identified several key impacts. A number of these are particularly relevant for water utilities located in coastal areas, including: 1) lowland inundation and wetland displacement; 2) altered tidal range in rivers and bays; 3) changes in sedimentation patterns; 4) severe storm-surge flooding; 5) saltwater intrusion into estuaries and freshwater aquifers. Water utility infrastructure is the most likely place where these impacts could be felt. For example, intakes located in transition areas between freshwater and saltwater interfaces of both surface and sub-surface systems could be affected. Sedimentation patterns in estuaries and deltas have been shown to be strongly tied to tidal patterns, storm surges and flow conditions, whose changes could affect utility supplies. Saltwater intrusion into freshwater aquifers that are a source of supply is already a problem in many coastal communities, primarily due to overdrafting of those groundwater supplies. Because of the higher density of saltwater, a rise in sea-level could result in a disproportionate loss of freshwater reservoirs in coastal zones due to the intrusion of the saltwater wedge.

What about natural variability? Relatively short instrumental records may not provide an adequate picture of the full range of natural climatic variability. A longer-term view is provided by the work of several researchers who


33

have developed proxy records for precipitation and streamflow based on tree rings and geological evidence ( e.g. Meko et al., 1991). Figure 25 provides examples of such proxy records for the reconstructed streamflow of the Colorado River at Lee Ferry and for the Four Rivers Index in northern California (Sacramento, American, Yuba, and Feather). These 20-year moving averages indicate that both regions have experienced extended periods of drought as well as periods of sustained above average flow. While there is a very weak positive correlation between annual flows in the two regions, there is no consistent pattern of association for the longer-term fluctuations between wet and dry conditions. For example, northern California experienced an extended dry period from 1918-37, during which time the Four Rivers Index dropped to 13.55 million acre-feet (Maf) from its longterm mean of 17.4 Maf. At the same time, conditions in the Upper Colorado River were much wetter than the long term mean of 13.5 Maf. On the other hand, the most severe extended drought in the Upper Colorado River Basin occurred during the period 1579-98, when average annual flow was only 10.95 Maf. That same period was among the driest in the northern California tree ring record (Meko et al., 1991).

Figure 25 – Time series plots of 20-year running means of reconstructed flows for the Colorado River at Lee’s Ferry (lower line) and for the Four Rivers Index, northern California (upper line). Source: Meko et al., 1991 (courtesy of the National Academy Press, Washington DC)

These records suggest that water supplies can change dramatically, and for extended periods of time, even without anthropogenic climate change. Where they are available, such reconstructions of past variability could be useful for examining the vulnerability of a water system to conditions outside of the range of recent experience.


34

Implications for Water Utilities Supply and Infrastructure Deciding how to evaluate system reliability given future uncertainty is a major challenge for water resource managers. Projected changes in water demand and variation in water supply are already considered in current water resource planning methods, but often based upon limited historical datasets. Will existing strategies be adequate to address climate change, or will the effects of climate change on water resources vary fundamentally from other anticipated changes? Currently infrastructure investments and long-term management strategies are based on the assumption that precipitation and runoff will follow past trends. Mounting evidence for climate change makes this an increasingly tenuous assumption. Some utilities and municipalities charged with securing water supplies are turning to interesting approaches to assess system security and reliability. For example, the City of Boulder, Colorado completed a study that evaluated 12 potential water supply/demand ‘futures’ for the city. These combined four alternative projected future water demands with three hypothetical climate scenarios. The study made use of a 300-year tree-ring hydrologic reconstruction to derive alternative hydrologic traces with the different changes in mean flow an annual variability. Thus, the study took a sensitivity approach to investigating the vulnerability of their system to climate variability, and made reference to climate change studies as providing bounds for their stylized scenarios: “While some research has suggested that climate change may result in earlier runoff and lower late summer stream flows due to more rain and less snow, we did not attempt to redistribute seasonal stream flows in this scenario. This decision was made for the sake of simplicity and conservatism. The degree of shift in seasonal runoff patterns has not been suggested by research to date. Earlier runoff is likely to increase the yield of Boulder’s water supply because Boulder’s reservoirs would be able to store more water before the onset of the irrigation season.” (Used with permission of the City of Boulder, CO USA). Based on the scenarios examined, the study concluded if climate change results in significantly reduced stream flows, Boulder’s water supply system will not be able to meet the future water demands at an acceptable level of reliability and with a reasonable margin of safety. This would be the case unless additional water supplies are acquired or developed, or additional reductions in per capita water use are achieved beyond the levels anticipated in Boulder’s comprehensive water conservation program.


35

Incorporating climate change into water management decisions is challenging. As discussed previously, there are several layers of uncertainty inherent in assessing climate change impacts. For example, uncertainty in projected greenhouse gas emissions, limitations of climate models, loss of accuracy when climate forecasts are downscaled to watershed resolution and imperfections in hydrological models are all relevant. Perhaps even more frustrating is the fact that there is no universally accepted standard for quantifying, or even defining, these uncertainties. This means that it is difficult to define a meaningful confidence level for these forecasts. Given the indefinite nature of climate impact analysis it may be tempting to disregard climate change in decision analysis. However, it is vital to understand that not everything regarding climate change is unknown, and uncertainty should not be used as an excuse to dismiss all aspects of climate change from water resource planning. Rather, the uncertainty introduced by climate emphasizes the importance of incorporating flexibility in long-term water resource planning. Intensification of the hydrological cycle is an effect of climate change that is quite probable. Therefore, there could be periods of runoff that are both significantly above and below current averages. This can make reservoir management more challenging, since there is often a tradeoff between storing water for dry-period use and evacuating reservoirs prior to the onset of the flood season to protect downstream communities. It may become more difficult to meet delivery requirements during prolonged periods between reservoir refilling without also increasing the risk of flooding. Earlier spring runoff from snowmelt is a likely repercussion of global warming. Much of Europe and the western United States depend on snowmelt as a water source for most of the year, so earlier runoff clearly affects water storage on a broad scale. Again these repercussions may be mitigated by the re-operation of reservoirs. To evaluate the impact of hydrological intensification and changes in snowmelt patterns on water storage, we must calculate the sensitivity of reservoir yields to changes in inflows. Unfortunately, a number of papers have demonstrated that the reliability of water yields can vary dramatically with only a small change in reservoir inflows (Nemec and Schaake 1982; USEPA 1989; Lettenmaier and Sheer 1991; McMahon et al. 1989; Mimikou et al. 1991; Nash and Gleick 1991,1993). These conclusions are based on the yields that various reservoirs would deliver under contemporary operating procedures given inflow scenarios that vary slightly from current patterns. An obvious question is whether reservoir operating rules could be adapted to maintain deliveries with current infrastructure given a change in inflows. Lettenmaier and Sheer (1991) demonstrated that this is indeed possible, but perhaps at the cost of increasing flood risk.


36

Conjunctive use – an innovative strategy that optimally uses surface and groundwater supplies in coordinated fashion is a promising approach. Construction of new reservoirs is another alternative, but there are few new good sites, and potentially high economic, environmental, and social costs. Increasing storage capacity of existing reservoirs also could be useful, but as the surface area increases and the climate warms there would potentially be increased evaporative losses. Also, an increase in runoff could make additional storage superfluous, while if net inflows decrease the additional capacity provided may go unused. The ecological cost of diverting water from streams can also be prohibitively expensive from a social standpoint. Therefore, more flexible strategies for water planning may be preferable, including increased reliance on a variety of water market options.

Water quality Intensification of the hydrological cycle may lead to more intense but less frequent precipitation. Possible increases in both flooding and drought pose potential threats to water quality. An increased incidence of heavy precipitation events may result in increased leaching and sediment transport, causing greater sediment and non-point source pollutant loadings to watercourses. This may make water treatment more difficult. Physical damage to water storage and treatment facilities is another possible consequence of severe floods. Regions where sewage and storm runoff systems are coupled will have further sanitary considerations regarding floods. Where stream flows and lake levels decline, water quality deterioration is likely as nutrients and contaminants become more concentrated in reduced volumes of carrying water. Warmer water temperatures may have further direct impacts on water quality, for example by reducing dissolved oxygen concentrations. In addition the salinity of surface waters, especially lakes and reservoirs with high residence times, could be increased by evaporative water losses. These stresses on water quality will be exacerbated periodically if climate warming leads to longer dry spells.


37

Case Study The impact of heavy runoff on drinking water: Milwaukee, 1993 In the spring of 1993 a waterborne gastrointestinal illness afflicted an estimated 403,000 residents of Milwaukee, Wisconsin and was responsible for over 100 deaths. The illness was caused by oocysts of the protozoa Cryptosporidium , which were introduced to the city’s drinking water after heavy spring runoff. Prior to the outbreak, the watershed that supplied Milwaukee’s drinking water experienced heavy rain and high runoff from snowmelt. The Cryptosporidium outbreak in Milwaukee demonstrated two threats that flooding poses to water treatment. The first is an increase in contaminants carried into the watershed by high flows. Cryptosporidium was likely introduced into the water supply by manure runoff from dairy farms. The parasite thrives in cattle intestines, and therefore is present in their manure. Hundreds of dairy farms are located near rivers and streams that empty into Lake Michigan and Wisconsin had a history of manure runoff problems before the 1993 epidemic. Waste from other agriculture and human waste are also possible sources of the parasite. The second problem was caused by the impact of heavy streamflow on the treatment process itself. Cryptosporidium bypassed the filtering process because Milwaukee’s south water treatment plant was not prepared to handle the high turbidity caused by flooding at the water source. The treatment plant failed to remove sufficient particulates, which allowed the parasite passage into the city’s drinking water. Turbidity at the source of Milwaukee’s other treatment plant at the north end of the city was not affected by the floods, and the quality of water it supplied was not compromised. This provides further evidence that the failure of the south treatment plant to respond to heavy runoff was responsible for the contagion. Runoff that is higher than historical trends, like that experienced by Milwaukee in the spring of 1993, should be expected since both precipitation and early snowmelt generally intensify as climate warms. The Milwaukee Cryptosporidium outbreak also demonstrates why water utilities should consider weather events that are beyond conventional expected variances.


38

How should water utilities respond to the prospect of climate change and plan for adaptation? Given our current level of uncertainty regarding regional changes in climate, it is difficult to give definitive advice as to how water utilities should respond to the generic issue of climate change. Climate is one among many sources of uncertainty affecting water utilities. Some would argue that the impacts of climate change are so uncertain and so far in the future that they pale in significance relative to more pressing concerns. However, the risk should not be ignored. Climate change is a new source of uncertainty. As such, it provides further reason to take actions that will improve resilience to the droughts and floods that arise from ongoing climate variability. Perhaps the best answer is to rethink traditional approaches to the planning process, historically grounded in assumptions such as the stationarity of climate. To date, future scenarios of how climate might change may be too general and aggregated and perhaps do not encourage more precautionary approaches to planning (Subak, 2000). However, the prospect of climate change adds to the future supply and demand uncertainties and reinforces the need for utilities to promote more efficient management of supplies and the infrastructure they manage (Frederick 1997).

References Aber, J.; Neilson, R. P.; McNulty, Steve; Lenihan, J. M.; Bachelet, D., and Drapek, R. J. Forest, 2001. processes and global environmental change: predicting the effects of individual and multiple stressors. BioScience. 51(9):735-751. Beven, K. Rainfall-Runoff Modeling- The Primer. Wiley: Chichester, UK. Boorman, D. and C. Sefton 1997, Recognising the uncertainty in the quantification of the effects of climate change on hydrological response, Climatic Change, 35, (4), 415-434. Allen, M.R., and W. J. Ingram, 2002. Constraints on future changes in climate and the hydrologic cycle. Nature, 419: 224-232. Arnell, N., et al., 1996. "Hydrology and Freshwater Ecology," in IPCC, 1996b, Climate Change 1995: Impacts, Adaptations and Mitigation of Climate Change: Scientific-Technical Analyses, Contribution of Working Group II to the Second Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge. Arnell, N.W. 1999. “Climate Change and Global Water Resources,” Global Environmental Change. 9: S31-S49. Blasing, T.J. and S. Jones, 2003: Current Greenhouse Gas Concentrations, Carbon Dioxide Information and Analysis Center (CDIAC) http://cdiac.esd.ornl.gov/pns/current_ghg.html Broecker, W.S., 1997: Thermohaline circulation, the Achilles heel of our climate system: Will man-made CO2 upset the current balance? Science, 278, 1582-1588.


39

Changnon, S.A. and K.E. Kunkel. 1999. “Rapidly Expanding Uses of Climate Data and Information in Agriculture and Water Resources: Causes and Characteristics of New Applications,” Illinois State Water Survey, Champaign, Illinois. Dettinger, M. and D. Cayan, 1995: Large-scale atmospheric forcing of recent trends toward early snowmelt in California, Journal of Climate 8: 606-623. Duell, L. F. W. Jr., 1992. "Use of Regression Models to Estimate Effects of Climate Change on Seasonal Streamflow in the American and Carson River Basins, California-Nevada," in Managing Water Resources During Global Change, American Water Resources Association 28th Annual Conference Proceedings, AWRA, Bethesda, MD. Duell, L. F. W. Jr., 1994. "The Sensitivity of northern Sierra Nevada streamflow to climate change. Water Resources Bulletin, 30(5): 841-859. Easterling, D., 1999. Development of regional climate scenarios using a downscaling approach, Climatic Change, 41(3), 615-634. Flashcka, I. M., C. W. Stockton and W. R. Boggess, 1987. Climatic Variation and Surface Water Resources in the Great Basin Region, in Water Resources Bulletin, 3(1): 47-57. Fredrick, K. 1997, Adapting to climate impacts on the supply and demand for water, Climatic Change, 37: 141-156. Gates, W. L., et al., 1996. "Climate Models - Evaluation," in IPCC, 1996a. Climate Change 1995 - The Science of Climate Change. Contribution of Working Group I to the Second Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge. Giorgi, Filippo and Linda O. Mearns, 1991. "Approaches to the simulation of regional climate change: a review," Reviews of Geophysics, 29(2): 191-216. Georgakakos, K.P. 2003. “Probabilistic Climate-Model Diagnostics for Hydrologic and Water Resources Impact Studies,” Journal of Hydrometeorology. 4: 92-105. Gleick, P.H. 2000a. “A Look at Twenty-first Century Water Resources Development,” Water International. 25(1): 127-138. Gleick, P.H. 2000b. Water: The Potential Consequences of Climate Variability and Change for the Water Resources of the United States. The Report of the Water Sector Assessment Team of the National Assessment of the Potential Consequences of Climate Variability and Change, The Pacific Institute. Grotch, S. L. and M. C. MacCracken, 1991. "The use of general circulation models to predict regional climatic change," Journal of Climate, 4: 286-303. Hewitson, B.C. and Crane, R.G., 1996. Climate downscaling: techniques and application, Climate Research, 7, 85-95. Hobbs, Benjamin F. and Philip T. Chao. 1997. “Using Decision Analysis to Include Climate Change in Water Resources Decision Making.” Climatic Change. 37: 188-202. Hughes, J. P., D. P. Lettenmaier and E. F. Wood, 1993. "An approach for assessing the sensitivity of floods to regional climate change," in Rafael Bras (ed.), The World at Risk: Natural Hazards and Climate Change. American Institute of Physics Conference Proceedings 277, AIP, New York. IGBP (International Geosphere Biosphere Program) 2001: Global Change and the Earth System: A Planet Under Pressure, IGBP Science No. 4. Steffen, W. and P. Tyson, (eds.). Stockholm, Sweden: IGBP.


40

IPCC (Intergovernmental Panel for Climate Change), WG I (Working Group I), 2001: Climate Change 2001: The Scientific Basis. Contribution of Working Group I to the Third Assessment Report of the IPCC. Cambridge, UK: Cambridge University Press. Katz, R., 1996. Use of conditional stochastic models to generate climate change scenarios, Climate Change, 32, 237–255. Kiehl, J., and K. Trenberth, 1997: Earth’s annual global mean energy budget. Bull. Amer. Meteor. Soc., 78, 197-208. Lettenmaier, D. P. and T. Y. Gan, 1990. "Hydrologic Sensitivities of the Sacramento-San Joaquin River Basin, California, to Global Warming," Water Resources Research, 26(1): 69-86. Lettenmaier, D.P. and D.P. Sheer. 1991. “Climate sensitivity of California water resources,” Journal of Water Resources Planning and Management. 117: 108-125. Lettenmaier, D.P., A.W. Wood, R.N. Palmer, E.F. Wood and E.Z. Stakhiv. 1999. “Water Resources Implications of Global Warming: A U.S. Regional Perspective,” Climatic Change. 43: 537–579. Mearns, L.O., I. Bogardi, F. Giorgi, I.Matyasovszky and M. Palecki, 1999. Comparison of climate change scenarios generated from regional climate model experiments and statistical downscaling, J. of Geophysical Research, 104, 6603-6621. Meko, D., M. Hughes and C. Stockton, 1991. "Climate Change and Climate Variability: The Paleo Record," in National Research Council, Managing Water Resources in the West Under Conditions of Climate Uncertainty. Proceedings of a Colloquium November 14-16, 1990, Scottsdale, AZ, National Academy Press, Washington, D.C. McMahon, M., Y. Kouvopoulos, G. Cavadias, and N. Vayianos, 1989: effects of climate change, J. of Hydrology, 123, (1-2), 119-146.

Regional hydrological

Nash, L. L. and P. H. Gleick, 1993. The Colorado River Basin and Climatic Change: The Sensitivity of Streamflow and Water Supply to Variations in Temperature and Precipitation. Report prepared for the U.S. Environmental Protection Agency, Office of Policy, Planning and Evaluation - Climate Change Division, EPA 230-R-93-009, Pacific Institute for Studies in Development, Environment, and Security, Oakland, CA. Neilson, R. and R. Drapek, 1998, Potentially complex biosphere responses to transient global warming, Global Change Biology, 4(5): 505-521. 1998. Nemec J. and J. Schaake, 1982. "Sensitivity of Water Resource Systems to Climate Variation," in Hydrological Sciences, 27(3): 327-343. Pelto, Mauri S., 1993. "Changes in Water supply in alpine regions due to glacier retreat", in Rafael Bras (ed.), The World at Risk: Natural Hazards and Climate Change, 61-67. American Institute of Physics Conference Proceedings 277, AIP, New York. Robock, A., R.P. Turco, M.A. Harwell, T.P. Ackerman, R. Anderssen, H.S. Chang, and M. Sivakumar, 1993. Use of general circulation model output in the creation of climate change scenarios for impact analysis, Climatic Change, 23 293-335. Sailor, D.J. and X. Li, 1999a. Tree-structured regression downscaling for regional climate change predictions of precipitation, World Resource Review, 11, 64-75. Sailor, D.J. and X. Li, 1999b. A semi-empirical downscaling approach for predicting regional temperature impacts associated with climatic change, Journal of Climate, 12, 103-114. Schaake, J. C., 1990. "From Climate to Flow," in P. E. Waggoner, ed., Climate Change and U. S. Water Resources, Wiley, New York. Semenov, M.A., 1997. Use of a stochastic weather generator in the development of climate change scenarios, Climatic Change, 35(4): 397-414.


41

Subak, S. 2000. “Climate Change Adaptation in the U.K. Water Industry: Managers’ Perceptions of Past Variability and Future Scenarios,” Water Resources Management 14: 137–156. Trenberth, K.E., A. Dai, R.M. Rasmussen, DB. Parsons. 2003. “The Changing Character of Precipitation,” Bulletin of the American Meteorological Society: 84(9): 1205–1217. Washington, W., 1996. "An Overview of Climate Modeling," in K. A. Miller and R. K. Parkin, eds., Final Report: An Institute on the Economics of the Climate Resource, University Corporation for Atmospheric Research, Boulder, CO. Wigley, T. M. L., 2003. “Assigning Probabilities to Climate Change” talk presented at the Workshop on Scientific and Applied Perspectives on Climate and Weather Related Risks, NCAR, Boulder, CO, October 9, 2003. Wilby, R.L., Wigley, T.M.L., Conway, D., Jones, P.D., Hewitson, B.C., Main, J. and Wilks, D.S., 1998. Statistical downscaling of general circulation model output: A comparison of methods, Water Resour. Res., (42) 2995-3008. Wilks, D.S, 1992. Adapting stochastic weather generation algorithms for climate change studies, Climatic Change, 22, 67-84. WMO (World Meteorological Organization), 2003: Climate: Into the 21st Century. Edited by William Burroughs, (ed.). Cambridge, UK: Cambridge University Press. Wood, A., D. Lettenmaier, and R. Palmer, 1997, Assessing climate change implications for water resources planning, Climatic Change, 37, 203-228. Yates, D., S. Gangopadhyay, B. Rajagopalan, and K. Strzepek 2003, A technique for generating regional climate scenarios using a nearest-neighbor algorithm, Water Resour. Res.,39, 7.