Water and the Hydrosphere

Patricia C. Henshaw, Robert J. Charlson and Stephen J. Burges

6.1 Introduction

It is hard to imagine any part of the Earth system that is more essential than or that has as many different functions as the water of the hydrosphere. In particular, the presence of a mobile liquid phase, with its long list of special chemical and physical properties, must be clearly identified as the main feature of Earth that separates it from the other terrestrial planets or from any known astronomical object. Close to home, the “terrestrial planets,” Earth, Mars, and Venus are presumed to have accreted similar abundances of “excess volatiles” – H₂O, CO₂, etc. – but evolved very differently. Even in the earliest stages of planetary evolution, liquid water provided a medium in which chemical reactions occurred between atmospheric CO₂ and the minerals in primitive igneous rocks to allow the precipitation of carbonate minerals and to prevent a runaway greenhouse effect. While no exact chronology or quantification of this early chemical event can be given, it seems clear that some such process prevented the accumulation of all of the Earth’s CO₂ and H₂O (as a vapor) in the atmosphere at the same time. This would have caused the Earth to be more or less like Venus – a condition from which there would appear to be no return to our present state. Before embarking on a description of this most important reservoir, it is useful – perhaps necessary – to reflect on the special properties of water itself. We can then proceed to a discussion of how the hydrosphere works.

is a bent molecule with a very strong permanent dipole moment. This dipole is the result of the negatively charged O atom and the two positively charged H atoms (the whole molecule being neutral). The existence of this charge separation arises due to the near orthogonality of the orbitals of the bonding electrons of the central O atom, while its large magnitude arises comes from the lack of shielding of the bonding electron and the small size of the O atom.

The permanent dipole moment is so strong that it permits the function of what are called hydrogen bonds between the highly electronegative O atom of one molecule and a nearby hydrogen atom of another molecule (see Fig. 6-1). The hydrogen bond is not a chemical bond in the ordinary sense of the forces that hold molecules together, which can be deduced from its strength of ca. 20 kJ/mol. Ordinary molecular bonds have typical strengths (energy required to break them) of a few hundred kJ/mol.

Fig. 6-1 Hydrogen bonds in liquid water.

It is the hydrogen bonds of water that give it unique physical and chemical properties, characteristics that set it apart from all of the other molecules formed from elements near the top of the periodic table. Table 6-1 compares several key properties of water to selected simple compounds that might be expected to have similar properties to water but do not. In regard to melting and boiling point, water behaves like a much larger molecule, but it has low density like the low atomic number compound that it really is.

Table 6-1

Anomalous properties of water

Property	Water	Comparison species
Boiling point	373 K	CH₄: 112 K (comparable size molecule)
		NH₃: 240 K
		H₂S: 211 K (dihydride)
		H₂Se: 231 K (dihydride)
		Ar: 87 K
Melting point	273 K	CH₄: 89 K
		NH₃: 195 K
		H₂S: 190 K
		H₂Se: 209 K
		Ar: 84 K
Heat capacity of liquid	4218 J/(K kg)	CH₄:2170 J/K kg
Latent heat of vaporization (0°C)	2.5 × 10⁶ J/kg
Latent heat of freezing	3.3 × 10⁵ J/kg
Ratio of density frozen/density liquid (0°C)	0.92
Surface tension	73 dyn/cm	CCl₄: 27 (nonpolar liquids)C₆H₆: 29

Likewise, liquid water has anomalously high molar heat capacity (75 J/mol K), meaning that liquid water can absorb relatively large amounts of heat from the sun by day and release it at night without much change of temperature. Owing to the large amount of liquid water at the surface of Earth, this large heat capacity is important in mediating temperatures and therefore climate.

Still further, water has large latent heats of evaporation and freezing (J/mol), all because of the same hydrogen bonds. As one result, the solar heating of the planet (largely in the tropics and subtropics), which results primarily in evaporation, transfers latent heat to the atmosphere in the form of water vapor. Subsequent precipitation at colder temperatures (higher latitudes or altitudes) releases the latent heat, making water vapor an important heat-transport vehicle.

This latent heat of evaporation, L_e, also appears in the fundamental description of the dependence of the vapor pressure of water, p, on temperature, T – the Clausius–Clapeyron equation:

(1)

or in integral form, between locations with p₁, T₁ and p₂, T₂:

(2)

where R is the universal gas constant in appropriate units.

The large value of L_e results in a very strong dependence of vapor pressure on temperature. As a result, the water vapor content of the air is extremely variable, from parts per million by volume in the coldest parts of the atmosphere to several percent in the warmest and wettest parts. The large latent heat of freezing of liquid water imposes a requirement for large transport of heat from bodies of water before the temperature can drop very much below 0°C, yet another type of thermostat.

A further unusual property of water is that it has a maximum density at around 4°C and expands upon freezing, again because of hydrogen bonds. There are more of these bonds in ice than in liquid water, creating a relatively open crystal structure in the solid phase. When ice melts (requiring the addition of a large amount of latent heat of freezing or fusion, L_f) some of the hydrogen bonds are broken, and a tighter packing of H₂O molecule results in the denser liquid.

The high latent heat of evaporation or vaporization – due to the hydrogen bonds causing attraction of water molecules to each other – also causes molecules at the surface of water to have cohesive forces. This results in water having anomalously high surface tension. In turn, this property plays a very strong role in the process of nucleation of cloud droplets, as one of the key factors involved in determining cloud droplet sizes and growth and coalescence rates. The latter is a significant factor in delivery of water to the continents by rain.

Very short wavelengths (λ < 186 nm) of UV radiation are required to dissociate the very strong O–H in water (bond strength = 456 kJ/mol). The large concentrations of O₂ and O₃ in air absorb incoming solar UV radiation high in the atmosphere, preventing much of this photodissociation. The strong bonds and the small size and mass of H₂O also give it a very complex infrared absorption spectrum that extends to shorter wavelengths (i.e. higher frequencies, v, and higher energies, hv) than many other simple molecules (CO₂ for example). One absorption feature, the 6.3 µm band, is extremely strong, as can be seen in Fig. 6-2. The result of this strong IR absorption, the large amount of water vapor in the atmosphere, and the proximity of the 6.3 µm absorption band of water vapor to the peak of the Earth’s black-body emission is that water vapor is by far the dominant greenhouse gas.

Fig. 6-2 Comparison of infrared absorbance of a vertical column of atmospheric CO₂ and H₂O vapor. The nearly total absorbance by H₂O between 5 and 7 µm, nearly coinciding with the peak of the wavelength-dependent emission of the surface, make H₂O a much more effective greenhouse gas. Liquid water (not shown) in clouds adds still more absorbance.

We can see the importance of water vapor as a greenhouse gas by comparing the greenhouse effect on Earth, a relatively humid planet, with that on Mars, which is arid. Mars has an atmosphere that is ca. 95% CO₂ (by mass, about 50 times more than on Earth). The greenhouse effect of Martian CO₂ causes a temperature increase of only a few degrees. In contrast, the total natural greenhouse effect on Earth is ca. 33 K, the majority of which is due to water vapor and water clouds. Nonetheless, as will be seen in Chapter 17, the anthropogenic greenhouse effect due to enhanced CO₂, CH₄, N₂O etc. cannot be dismissed, as changes in the Earth’s average temperature of even 1 K are significant.

Besides these special physical properties, hydrogen-bonded liquid water also has unique solvent and solution properties. One feature is high proton (H⁺) mobility due to the ability of individual hydrogen nuclei to jump from one water molecule to the next. Recalling that at temperatures of about 300 K, the molar concentration in pure water of H₃O⁺ ions is ca. 10⁻⁷M, the “extra” proton can come from either of two water molecules. This freedom of H⁺ to transfer from one to an adjacent “parent” molecule allows relatively high electrical conductivity. A proton added at one point in an aqueous solution causes a domino effect, because the initiating proton has only a short distance to travel to cause one to pop out somewhere else.

The existence of strongly polar water molecules and mobile protons also makes H₂O an excellent and almost universal solvent for ionic compounds and polar organic species. Compounds that will not significantly dissolve in water (i.e. saturated solutions with concentrations less than ca. 10⁻⁵ M) include aliphatic and aromatic hydrocarbons, as well as plastics and many other polymers.

6.1.2 The Right Abundance of Water to Support Life

As can be seen in Fig. 2-1 (abundance of elements), hydrogen and oxygen (along with carbon, magnesium, silicon, sulfur, and iron) are particularly abundant in the solar system, probably because the common isotopic forms of the latter six elements have nuclear masses that are multiples of the helium (He) nucleus. Oxygen is present in the Earth’s crust in an abundance that exceeds the amount required to form oxides of silicon, sulfur, and iron in the crust; the excess oxygen occurs mostly as the volatiles CO₂ and H₂O. The CO₂ now resides primarily in carbonate rocks whereas the H₂O is almost all in the oceans.

While it is clear that the hydrosphere is a significant portion of the planet’s mass, there is not, at least currently, so much water that the continents are submerged. Conversely, the oceans are large enough that their surface area would never become an important limiting factor in the hydrologic cycle. Although there have been many shifts in the balance of the hydrosphere, this condition has prevailed since the biosphere began to evolve. The presence of liquid water allowed the Earth to become and remain a living planet, and by the astronomical coincidence of the Earth’s location, the planet received just the proper abundance to sustain and recycle this all-important resource, almost in perpetuity.

6.2 Global Water Balance

While the hydrosphere has long been appreciated as essential to life on Earth, only in the past couple of decades have scientists expanded their exploration of the global hydrologic cycle and its roles across the spectrum of Earth science disciplines. The Earth and its atmosphere, in the broadest view, are a complex, intimately coupled system of chemical, physical, and biological cycles, and water, with its myriad unique chemical and physical properties, plays a part in almost all of them.

To understand the role water plays in global cycles, it is necessary to first understand the mechanics of the water cycle. The hydrologic cycle is driven by solar radiation, which provides the energy necessary to overcome latent heat capacities involved in phase changes. Gravity plays a key role in returning condensed water to the surface as precipitation, and via runoff from the continents to the oceans. At the simplest level, the global water cycle results from imbalances between precipitation and evapotranspiration (ET) at the ocean and land surfaces. Globally, the oceans lose more water by evaporation than they gain by precipitation, whereas the land surface receives more precipitation than is lost through ET; runoff from the land surfaces then balances the ocean-atmosphere water deficit. The hydrologic cycle is significantly more complex than this simple description would suggest. In addition to the atmosphere, oceans, and rivers, significant amounts of water are stored in groundwater, glaciers and ice sheets, soil moisture, and, to a smaller extent, biomass. Figure 6-3 shows a schematic of the global hydrologic cycle, with storages in km³ and fluxes in km³/yr.

Fig. 6-3 Global water balance. Storages in km³, fluxes in km³/yr. Turnover times calculated as storage divided by total annual inflow. (Data from Shiklomanov and Sokolov, 1983.)

Calculation of the global water balance is a nontrivial problem. Gross storage volumes are calculated predominantly by multiplying surface areas by estimated average depths (UNESCO, 1978). While modern remote sensing technology has made it possible to determine areal extents of surface reservoirs, such as oceans, lakes, and ice sheets quite accurately, estimations of depths or thicknesses are still highly uncertain and often subjective. For example, the wide variations in groundwater estimates can often be attributed to differing interpretations of the extent of groundwater into the Earth’s crust. Similarly, soil moisture depths vary from less than a meter to tens of meters or more, so global soil moisture averages are by necessity highly subjective. Even more well-defined reservoirs, such as oceans or lakes, cannot be accurately quantified without complete knowledge of their bathymetry and vertical temperature profiles, an impossible requirement on a global scale.

Although fluxes of precipitation and river discharge can be quite accurately determined on a local scale, large portions of the globe, especially the oceans and Antarctica, are essentially ungauged, requiring extensive extrapolation of existing data. Evaporation fluxes are even less well known, since calculation requires extensive knowledge of hydrologic and meteorologic parameters. Consequently, numerous estimates of storage and flux volumes exist for the hydrologic cycle; the representation in Fig. 6-3 was selected for its completeness. Some of the variability in estimates of the global water balance is reflected in the reservoir storage values shown in Table 6-2.

Table 6-2

Reservoir storage

6.2.1 Reservoirs

Figure 6-3 shows the hydrologic cycle as seven primary reservoirs interconnected by a number of water fluxes. The role of each reservoir in the hydrologic cycle and its connections with other cycles is briefly summarized below, in order of storage volume.

Table 6-3

Reservoir turnover times

Reservoir	Volume (km³)	Avg. turnover time
Oceans	1.338 × 10⁹	2640 yrs
Cryosphere	24.1 × 10⁶	8900 yrs
Groundwater/permafrost	23.7 × 10⁶	515 yrs
Lakes/rivers	189 990	4.3 yrs
Soil moisture	16 500	52 days
Atmosphere	12 900	8.2 days
Biomass	1120	5.6 days

Many hydrologic reservoirs can be further subdivided into smaller reservoirs, each with a characteristic turnover time. For example, water resides in the Pacific Ocean longer than in the Atlantic, and the oceans’ surface waters cycle much more quickly than the deep ocean. Similarly, groundwater near the surface is much more active than deep reservoirs, which may cycle over thousands or millions of years, and water frozen in the soil as permafrost. Typical range in turnover times for hydrospheric reservoirs on a hillslope scale (10–10³ m) are shown in Table 6-4 (estimates from Falkenmark and Chapman, 1989). Depths are estimated as typical volume averaged over the watershed area.

Table 6-4

Hillslope scale turnover times

Reservoir	Depth (mm)	Turnover times
Atmosphere	25 ^a	8–10 days
Plants	5–50	Hours-days
Streams/rivers	3	Weeks
Lakes/reservoirs	—	Months-years
Soil moisture	10–10⁴	Years
Groundwater	10⁴–10⁵	Days-10⁶ years ^b

^a Global average.

^b Longer turnovers associated with large watershed areas.

Global freshwater reserves (discounting pollution) are a small percentage of global water, accounting for only 35 × 10⁶ km of the total 1.386 × 10⁹ km³ global water supply (UNESCO,1978). Assuming an input flux equal to oceanic evaporation, this would give a turnover time of about 750 years. The turnover time analysis is not strictly correct since freshwater resides in a number of interconnected reservoirs; however, since freshwater volume is essentially equal to non-marine storage and net evaporation is the only output flux from the oceans, this may be taken as a reasonable estimation. For practical purposes, freshwater resources available to humans cycle more rapidly, but since 97% of freshwater is stored in ice, the global average turnover is much longer.

6.2.3 Fluxes

Robert Horton, an influential pioneer in the field of hydrology, developed one of the first comprehensive representations of the hydrologic cycle in 1931. His original diagram, Fig. 6-4, illustrates the processes by which water moves between the Earth’s hydrologic reservoirs. Hydrologic fluxes can be summed up in four processes, shown around the outside of Horton’s wheel – precipitation, surface dissipation of precipitation, evaporation, and atmospheric moisture transport. This section will discuss precipitation, evaporation (and closely related transpiration), and runoff – the processes that link the oceans, atmosphere, and land surface. Atmospheric moisture is addressed in Chapter 7. These fluxes are highly variable over the Earth’s surface in both space and time, which has extremely important implications for water resources; spatial and temporal variability is discussed in Section 6.3.

Fig. 6-4 The fluxes of the hydrologic cycle, developed by Robert Horton (1931).

6.2.3.1 Precipitation

The flow of water from the atmosphere to the ocean and land surfaces as rain, snow, and ice constitutes the atmospheric efflux in the hydrologic cycle. Although most precipitation falls on the oceans (ca. 79% of the global total), precipitation onto land is much more hydrologically significant. On a global scale, nearly two-thirds of the land portion returns to the atmosphere via evapotranspiration (see below), while the remaining one-third contributes to groundwater and surface runoff. Precipitation is highly variable over the globe, with atmospheric circulation patterns concentrating it in the tropics and mid-latitudes.

An important component of precipitation on a regional scale comes from precipitation recycling; that is, a portion of the precipitation in a region comes from water vapor evaporated from within that region, with the remainder composed of atmospheric moisture advected into the region. The precipitation recycling ratio, the ratio of recycled precipitation to total precipitation, is then a function of evaporation and internal and external atmospheric moisture fluxes (Budyko, 1974; Eltahir and Bras, 1996). The settling of the Great Plains in the late 19th and early 20th centuries was in fact spurred by early (and largely unfounded) concepts of precipitation recycling. S. Aughey (cited in Holzman, 1937) wrote of Nebraska in 1880 that increased evaporation from cultivated land would increase moisture and rainfall – i.e., that “rain follows the plow.”

6.2.3.2 Evapotranspiration

Water returns to the atmosphere via evaporation from the oceans and evapotranspiration from the land surface. Like precipitation, evaporation is largest over the oceans (88% of total) and is distributed non-uniformly around the globe. Evaporation requires a large input of energy to overcome the latent heat of vaporization, so global patterns are similar to radiation balance and temperature distributions, though anomalous local maxima and minima occur due to the effects of wind and water availability.

Evapotranspiration (ET) is the collective term for land surface evaporation and plant transpiration, which are difficult to isolate in practice. Transpiration refers to the process in which water is transported through plants and returned to the atmosphere through pores in the leaves called stomata, and is distinct from direct evaporation of intercepted precipitation from leaf surfaces. Some land surface processes and the roles of vegetation in the water and energy balances are illustrated in Fig. 6-5. Due to the number of variables involved, ET can be extremely difficult to measure and is often determined by closing the water or energy balance calculated from better-known components.

Fig. 6-5 Evaporation and transpiration from vegetation are among the complex land surface interactions in the hydrologic cycle. (From Dickinson, 1984.)

6.2.3.3 Runoff

The excess of evaporation from the oceans is made up for by runoff from the land. Although this flux is much smaller than precipitation and ET, it is a major link in many cycles and is of particular importance to humans in terms of water supply. Runoff can be broadly categorized into subsurface, or groundwater, flow and surface flow, consisting of overland runoff and river discharge.

6.2.3.3.1 Subsurface runoff.

When precipitation hits the land surface, the vast majority does not go directly into the network of streams and rivers; in fact, it may be cycled several times before ever reaching a river and the ocean. Instead, most precipitation that is not intercepted by the vegetation canopy and re-evaporated infiltrates into the soil, where it may reside as soil moisture, percolate down to groundwater, or be transpired by plants.

Very little groundwater is discharged directly to the oceans, but groundwater does provide a significant contribution to stream discharge in most areas. Subsurface flow is generally much slower than surface runoff, allowing groundwater to provide perennial baseflow to streams far into a dry season, long after surface storm runoff has been discharged. Figure 6-6 shows a typical storm hydrograph, with baseflow and stormflow components indicated. Groundwater flow velocities have been found to follow Darcy’s law:

(3)

where v is velocity, K is soil hydraulic conductivity with units of (length/time), n is the dynamic (or actively available) porosity, and dh/dL is the hydraulic gradient.

Fig. 6-6 Hydrograph showing the rapid contribution of surface runoff and more steady baseflow. Runoff in cubic feet per second, precipitaion in inches. (From Langbein and Wells, 1955.)

Hydraulic conductivities vary over a range of 10⁻¹² cm/s for unfractured igneous and metamorphic rocks to 2 or 3 cm/s for porous (karst) limestone and gravel; surface flow is typically on the order of a meter per second.

6.2.3.3.2 Surface runoff.

Hydrologists have identified two processes for generating surface runoff over land. The first, saturated overland flow (SOF), is generated when precipitation (or snowmelt) occurs over a saturated soil; since water has nowhere to infiltrate, it then runs off over land. SOF typically occurs only in humid environments or where the water table rises to intersect with a stream. Horton overland flow (HOF or infiltration-limited overland flow) occurs when precipitation intensity exceeds the infiltration capacity of the soil in a non-saturated environment. In this case, only the excess precipitation (that exceeding the infiltration capacity) runs off over the surface. Both types of overland runoff generate relatively rapid flows that constitute the surface water contribution to the hydrograph (Fig. 6-6).

Figure 6-7 illustrates the runoff paths for HOF and SOF, as well as for subsurface stormflow and groundwater flow. Subsurface stormflow is a moderately rapid runoff process in which water flows to a stream through highly permeable surface soil layers (without reaching the water table). Note in Fig. 6-7 that while HOF and subsurface stormflow may occur over a large fraction of an infiltration-limited hillslope, SOF occurs over a smaller portion adjacent to the stream.

Fig. 6-7 Vertical cross-section showing pathways for surface and subsurface runoff. Path 1: HOF; path 2: groundwater flow; path 3: subsurface stormflow; path 4: SOF. (From Dunne and Leopold, 1978.)

Hillslopes occupy about 99% of the landscape and provide stream channels with water supply, making hillslope processes extremely important on a local scale. However, the much more visible component of surface runoff comes from river discharge. Globally, rivers discharge roughly 45 000 km³ per year to the oceans (Shiklomanov and Sokolov, 1983). The 16 largest rivers account for more than one-third of total discharge, and over half of that contribution comes from the three largest. Table 6-5 lists the 10 largest rivers in the world in terms of average discharge rate (m³/s) and annual discharge volume (km³/yr) (Dingman, 1994).

Table 6-5

World’s largest rivers

River	Discharge (m³/s)	Discharge(km³/yr)
Amazon	190 000	6000
Congo	42 000	1330
Yangtze	35 000	1100
Orinoco	29 000	915
Brahmaputra	20 000	630
La Plata	19 500	615
Yenesei	17 800	565
Mississippi	17 700	560
Lena	16 300	515
Mekong	15 900	500

While river discharge is the primary means of transferring water from the land to the oceans, its magnitude pales in comparison to circulation within the oceans themselves. The total average discharge of the world’s rivers is about 10⁶ m³/s or 1 Sverdrup (Sv); by comparison, the oceans’ thermohaline circulation transports about 15 Sv (Broecker, 1997).

In addition to runoff, rivers transport products of upland weathering to the oceans, forming a key link in the tectonic cycle of uplift and erosion. This interaction will be explored further in Section 6.6.

6.3 Hydrologic Variability

Part of the difficulty in studying the hydrologic cycle arises from the huge variability in fluxes over time and space. Most of us have experienced the differences in rain and snowfall associated with changing seasons and observed the different precipitation patterns experienced by other areas of the country and the world. This hydrologic variability is present across virtually all spatial and temporal scales, from the smallest hillslope over a period of minutes to the entire globe over the geologic history of the Earth. Understanding hydrologic variability is particularly important for management of water resources. In that context, continental scale variations in precipitation and runoff within a year and over several years to decades are of the most interest.

Variability has traditionally been accounted for in hydrologic models in one of two ways. Stochastic models attempt to preserve statistical relationships between significant variables determined from historical records, while physically based models are designed to represent natural processes based on values of known variables and empirical parameters. While stochastic models tend to be simpler and less data intensive, they require long historical records, which are unavailable in many areas, and cannot represent conditions outside the range of historic values. Physically based models require extensive input data, but because they explicitly represent physical processes, they are more appropriate for studying effects of and responses to global change and can be used for limited extrapolation beyond the range of data used for their calibration and testing.

6.3.1 Precipitation

On a global scale, and as discussed in Chapter 7, precipitation patterns clearly reflect the convergence and divergence zones in the general atmospheric circulation. Rainfall peaks over the tropics, decreases in the subtropical latitudes, and exhibits more modest peaks at mid-latitudes before going to essentially zero at the poles. Figure 7-7 in the next chapter provides a more complete description of the overall latitude dependence. At the regional or continental scale, the precipitation patterns are complicated by many other factors, including mesoscale atmospheric circulations and orographic effects.

Figure 6-8 shows the average monthly precipitation for selected locations in the United States, illustrating both the spatial variability in total precipitation and the marked differences in distribution of precipitation through the year. The US exhibits four basic annual patterns for precipitation distribution, which can be used to classify climate (Thornthwaite, 1948). The West Coast experiences dry summers and winter precipitation maximums from storms coming in off the Pacific, while the interior of the continent tends to have late spring or early summer peaks associated with peaks in the soil moisture cycles. The northeast receives moderate precipitation throughout the year, while the southeast receives large amounts of rain from late summer thunderstorms fueled by the surrounding warm oceans. The rain shadow effects of the Sierra Nevada, Cascades, and Rocky Mountains, as well as the aridity of the desert southwest, can also be observed in Fig. 6-8.

Fig. 6-8 Annual precipitation patterns for the United States. Precipitation in inches (1 inch = 25.4 mm). (From Linsley et al., 1982.)

6.3.2 Runoff

Because it depends on a number of conditions that are themselves inherently variable, runoff tends to vary even more than precipitation, particularly over time. Seasonal runoff patterns depend largely on latitude and altitude of the watershed, due to the importance of snowmelt in runoff peaks. In high-latitude basins or those with significant high-altitude contributing areas, peaks occur later than in low-latitude or low-altitude areas. Figure 6-9 shows the relative amount of runoff in each month for selected US rivers. While average flow in the southeastern rivers is nearly constant (approximately 8% per month), northern rivers and those fed by mountain snowpack show distinct peaks, with peak timing dependent on when snowmelt occurs.

Fig. 6-9 Annual runoff patterns for the United States. Percent of normalized annual runoff in each month. (From Langbein and Wells, 1955.)

Water resources decision making in many areas, particularly arid and semi-arid climates such as the American West, depends on interannual to decadal variations in surface water availability. In addition to more predictable seasonal differences, runoff tends to exhibit long-term trends alternating between flood and drought periods. Fig. 6-10 shows historical wet and dry periods based on streamflow records for 50 world rivers. For the most part, these periods are consistent on a regional basis, though they appear to alternate on a hemispheric scale.

Fig. 6-10 Long-term global streamflow trends. Wet and dry periods from the historical record of 50 major rivers. (From Probst and Tardy, 1987.)

The importance of these long-term trends for water resources is illustrated by the case of the Colorado River Compact, which allocates the water of the Colorado River to the seven states in its drainage area. The agreement, signed in 1922, based allocations on annual discharge from the preceding decade, which remains the wettest period on record; subsequent drier years have given rise to numerous water rights disputes among the seven states and between the United States and Mexico.

6.3.3 Hydrologic Sensitivity

The ability to predict runoff and water availability is critical to water resources planners. However, the complex non-linearities of the hydrologic cycle make this an extremely difficult process. Even where precipitation is fairly well known, runoff prediction is a non-trivial problem, as land surface response depends as much (or more) on precipitation patterns and timing as on precipitation amount. The historical record of monthly rainfall and inflow at the Serpentine Dam, near Perth, Western Australia, provides an illustration of this sensitivity (Fig. 6-11a and b).

Fig. 6-11 (a) Monthly rainfall record. Serpentine dam, Perth, Western Australia. (b) Monthly inflow. Serpentine Dam, Western Australia. (From Burges, 1998.) (From Burges, 1998.)

The precipitation record shows a mild decrease in rainfall for May, June, and July over the last 20 years of record. However, runoff decreased to less than half the historical average for May and June over the same period, and reduced runoff persisted into December, despite a return to normal or above normal precipitation levels over the latter half of the year.

Runoff sensitivity, particularly in arid and semi-arid climates, is largely a result of sensitivity in soil moisture response. If rainfall amount and frequency decrease, more soil moisture is lost to evapotranspiration, creating a soil moisture deficit that must be replaced before surface runoff or significant groundwater flow returns. The converse also tends to be true: frequent storms can saturate the soil, generating more surface runoff and enhancing subsurface flow.

The complexities of land surface response and runoff generation have also presented a major obstacle to global climate modelers. Hydrologic response is linked to several important climate feedbacks (see Section 6.4.2), so until the hydrologic cycle, and in particular its land surface component, can be accurately represented, there is little hope for accurate assessments of global change.

6.4 Water and Climate

In addition to biogeochemical cycles (discussed in Section 6.5), the hydrosphere is a major component of many physical cycles, with climate among the most prominent. Water affects the solar radiation budget through albedo (primarily clouds and ice/snow), the terrestrial radiation budget as a strong absorber of terrestrial emissions, and global temperature distribution as the primary transporter of heat in the ocean and atmosphere.

6.4.1 Water and the Energy Balance

Water plays a crucial role in the redistribution of heat from the tropics to the high-latitude polar regions. Water transports heat in two forms. Sensible heat refers to the portion of the radiant energy budget that changes the temperature of the surface of the atmosphere. Sensible heat flux produces changes in temperature proportional to the product of density and specific heat (Shuttleworth, 1993). Since liquid water has one of the highest specific heats of any substance (4218 J/(kg K) at 0°C), the oceans can store and release vast amounts of sensible heat without large changes in their own temperatures. Surface ocean currents transport warm tropical water poleward and recirculate cooler surface waters toward the equator. These currents are complemented by the global thermohaline circulation, which circulates water through the depths of the oceans. Ice and water vapor can also store significant amounts of sensible heat, though their specific heat capacities are only one-half to one-third that of liquid water, respectively.

Latent heat is the energy associated with phase changes. Evaporation of water requires an energy input of 2.5 × 10⁶ J per kilogram of water at 0°C, almost 600 times the specific heat. When water vapor is transported via atmospheric circulation and recondensed, latent heat energy is released at the new location. Atmospheric transport of water vapor thus transfers both latent and sensible heat from low to high latitudes.

Sensible and latent heat can be related through the Bowen ratio, which is the ratio of sensible heat to latent heat flux at the surface. The Bowen ratio, R, can be estimated from atmospheric properties as follows:

(4)

where SH = sensible heat (Energy/unit time and unit area); LE = latent energy flux (Energy/unit time and unit area); T₀ = surface temperature (¯C); T_a = reference level air temperature (¯C); e₀ = saturation vapor pressure (mbar); e_a = atmospheric vapor pressure (mbar); p = atmospheric pressure (mbar).

For saturated surfaces, the Bowen ratio can then be used to calculate evapotranspiration as a residual of the surface energy balance (Penman, 1948). Since direct measurement of ET is difficult and expensive, the energy balance method is fairly common.

6.4.2 Climate Feedbacks and Response to Global Warming

Five components of the hydrosphere play major roles in climate feedbacks – atmospheric moisture, clouds, snow and ice, land surface, and oceans. Changes to the hydrologic cycle, among other things, as a result of altered climate conditions are then referred to as responses. Interactions with climate can best be explored by examining potential response to a climate perturbation, in this case, predicted global warming.

Current debate regarding the role of the hydrologic cycle in climate focuses on potential responses to anthropogenically induced global warming through an enhanced greenhouse effect due to increased atmospheric CO₂ and other gases. Based on GCM simulation results, the Intergovernmental Panel on Climate Change (IPCC) has concluded that global mean surface atmospheric temperature will increase by 1.0 to 3.5°C in the next century, though this will be distributed unevenly over the Earth, with the most significant warming expected at high latitudes. Warmer temperatures are expected to accelerate the hydrologic cycle, though the magnitude and distribution of hydrologic changes, particularly at the regional scale important for water resources, are much more speculative (IPCC, 1996a).

6.4.2.1 Climate feedbacks

In its assessment of climate change, the IPCC (1990) identified five hydrosphere-related feedback mechanisms in the climate system likely to be activated by increased greenhouse gas concentrations in the atmosphere. These feedbacks are briefly described below; for more detailed discussion of the climate system, refer to Chapter 17.