Datasets:

izumi-lab

/

open-text-books

like 14

A body of rock that is under stress becomes deformed. When the rock can no longer withstand the deformation, it breaks and the two sides slide past each other. Most earthquakes take place near plate boundaries, but not necessarily right on a boundary, and not necessarily even on a pre-existing fault. The engineering principle of elastic deformation, which can be used to understand earthquakes, is illustrated in Figure 11.2. The stress applied to a rock — typically because of ongoing plate movement — results in strain or deformation of the rock (Figure 11.2b). Because most rock is strong (unlike loose sand, for example), it can withstand a significant amount of deformation without breaking. But every rock has a deformation limit and will rupture (break) once that limit is reached. At that point, in the case of rocks within the crust, the rock breaks and there is displacement along the rupture surface (Figure 11.2c). The magnitude of the earthquake depends on the extent of the area that breaks (the area of the rupture surface) and the average amount of displacement (sliding).

The concept of a rupture surface, which is critical to understanding earthquakes, is illustrated in Figure 11.3. An earthquake does not happen at a point, it happens over an area within a plane, although not necessarily a flat plane. Within the area of the rupture surface, the amount of displacement is variable (Figure 11.3), and, by definition, it decreases to zero at the edges of the rupture surface because the rock beyond that point isn’t displaced at all. The extent of a rupture surface and the amount of displacement will depend on a number of factors, including the type and strength of the rock, and the degree to which it was stressed beforehand.

Figure 11.3 A rupture surface (dark pink), on a steeply dipping fault plane (light pink). The diagram represents a part of the crust that may be up to tens or hundreds of kilometres long. The rupture surface is the part of the fault plane along which displacement occurred. In this example, the near side of the fault is moving to the left, and the lengths of the arrows within the rupture surface represent relative amounts of displacement. [SE] Earthquake rupture doesn’t happen all at once; it starts at a single point and spreads rapidly from there. Depending on the extent of the rupture surface, the propagation of failures out from the point of initiation is typically completed within seconds to several tens of seconds (Figure 11.4). The initiation point isn’t necessarily in the centre of the rupture surface; it may be close to one end, near the top, or near the bottom. Figure 11.4 Propagation of failure on a rupture surface. In this case, the failure starts at the dark blue heavy arrow and propagates outward, reaching the left side first (green arrows) and the right side last (yellow arrows). [SE] Figure 11.5 shows the distribution of immediate aftershocks associated with the 1989 Loma Prieta earthquake. Panel (b) is a section along the San Andreas Fault; this view is equivalent to what is shown in Figures 11.3 and 11.4. The area of red dots is the rupture surface; each red dot is a specific aftershock that was recorded on a seismometer. The hexagon labelled “main earthquake” represents the first or main shock. When that happened, the rock at that location broke and was displaced. That released the stress Physical Geology 300 on that particular part of the fault, but it resulted in an increase of the stress on other nearby parts of the fault, and contributed to a cascade of smaller ruptures (aftershocks), in this case, over an area about 60 km long and 15 km wide. Figure 11.5 Distribution of the aftershocks of the 1989 M 6.9 Loma Prieta earthquake (a: plan view, b: section along the fault, c: section across the fault.) [from Open University under CC Sharealike, So, what exactly is an aftershock then? An aftershock is an earthquake just like any other, but it is one that can be shown to have been triggered by stress transfer from a preceding earthquake. Within a few tens of seconds of the main Loma Prieta earthquake, there were hundreds of smaller aftershocks; their distribution defines the area of the rupture surface. Aftershocks can be of any magnitude. Most are smaller than the earthquake that triggered them, but they can be bigger. The aftershocks shown in Figure 11.5 all happened within seconds or minutes of the main shock, but aftershocks can be delayed for hours, days, weeks, or even years. As already noted, aftershocks are related to stress transfer. For example, the main shock of the Loma Prieta earthquake triggered aftershocks in the immediate area, which triggered more in the surrounding area, eventually extending for 30 km along the fault in each direction and for 15 km toward the surface. But the earthquake as a whole also changed the stress on adjacent parts of the San Andreas Fault. This effect, which has been modelled for numerous earthquakes and active faults around the world, is depicted in Figure 11.6. Stress was reduced in the area of the rupture (blue), but was increased at either end of the rupture surface (red and yellow).

Figure 11.6 Depiction of stress changes related to an earthquake. Stress decreases in the area of the rupture surface, but increases on adjacent parts of the fault. [by SE based on data from 2010 Laguna Salada earthquake by Stein and Toda at: Stress transfer isn’t necessarily restricted to the fault along which an earthquake happened. It will affect the rocks in general around the site of the earthquake and may lead to increased stress on other faults in the region. The effects of stress transfer don’t necessarily show up right away. Segments of faults are typically in some state of stress, and the transfer of stress from another area is only rarely enough to push a fault segment beyond its limits to the point of rupture. The stress that is added by stress transfer accumulates along with the ongoing buildup of stress from plate motion and eventually leads to another earthquake.

Episodic tremor and slip (ETS) is periodic slow sliding along part of a subduction boundary. It does not produce recognizable earthquakes, but does produce seismic tremor (rapid seismic vibrations on a seismometer). It was first discovered on the Vancouver Island part of the Cascadia subduction zone by Geological Survey of Canada geologists Herb Dragert and Gary Rogers.* The boundary between the subducting Juan de Fuca Plate and the North America Plate can be divided into three segments, as shown below. The cold upper part of the boundary is locked. The plates are stuck and don’t move, except with very large earthquakes that happen approximately every 500 years (the last one was M8.5+ in January 26, 1700). The warm lower part of the boundary is sliding continuously because the warm rock is weaker. The central part of the boundary isn’t cold enough to be stuck, but isn’t warm enough to slide continuously. Instead it slips episodically, approximately every 14 months for about 2 weeks, moving a few centimetres each time.

You might be inclined to think that it’s a good thing that there is periodic slip on this part of the plate because it releases some of the tension and reduces the risk of a large earthquake. In fact, the opposite is likely the case. The movement along the ETS part of the plate boundary acts like a medium-sized earthquake and leads to stress transfer to the adjacent locked part of the plate. Approximately every 14 months, during the two-week ETS period, there is a transfer of stress to the shallow locked part of the Cascadia subduction zone, and therefore an increased chance of a large earthquake.

The distribution of earthquakes across the globe is shown in Figure 11.7. It is relatively easy to see the relationships between earthquakes and the plate boundaries. Along divergent boundaries like the mid- Atlantic ridge and the East Pacific Rise, earthquakes are common, but restricted to a narrow zone close to the ridge, and consistently at less than 30 km depth. Shallow earthquakes are also common along transform faults, such as the San Andreas Fault. Along subduction zones, as we saw in Chapter 10, earthquakes are very abundant, and they are increasingly deep on the landward side of the subduction zone. Figure 11.7 General distribution of global earthquakes of magnitude 4 and greater from 2004 to 2011, colour coded by depth (red: 0-33 km, orange 33-70 km, green: 70-300 km, blue: 300-700 km) [from Dale Sawyer, Rice University, ,used with permission] Earthquakes are also relatively common at a few intraplate locations. Some are related to the buildup of stress due to continental rifting or the transfer of stress from other regions, and some are not well understood. Examples of intraplate earthquake regions include the Great Rift Valley area of Africa, the Tibet region of China, and the Lake Baikal area of Russia.

Figure 11.8 provides a closer look at magnitude (M) 4 and larger earthquakes in an area of divergent boundaries in the mid-Atlantic region near the equator. Here, as we saw in Chapter 10, the segments of the mid-Atlantic ridge are offset by some long transform faults. Most of the earthquakes are located along the transform faults, rather than along the spreading segments, although there are clusters of earthquakes at some of the ridge-transform boundaries. Some earthquakes do occur on spreading ridges,

Figure 11.8 Distribution of earthquakes of M4 and greater in the area of the mid-Atlantic ridge near the equator from 1990 to 1996. All are at a depth of 0 to 33 km [SE after Dale Sawyer, Rice University,

The distribution and depths of earthquakes in the Caribbean and Central America area are shown in Figure 11.9. In this region, the Cocos Plate is subducting beneath the North America and Caribbean Plates (ocean-continent convergence), and the South and North America Plates are subducting beneath the Caribbean Plate (ocean-ocean convergence). In both cases, the earthquakes get deeper with distance from the trench. In Figure 11.9, the South America Plate is shown as being subducted beneath the Caribbean Plate in the area north of Colombia, but since there is almost no earthquake activity along this zone, it is questionable whether subduction is actually taking place. Figure 11.9 Distribution of earthquakes of M4 and greater in the Central America region from 1990 to 1996 (red: 0-33 km, orange: 33-70 km, green: 70-300 km, blue: 300-700 km) (Spreading ridges are heavy lines, subduction zones are toothed lines, and transform faults are light lines.) [SE after Dale Sawyer, Rice University, 305 Chapter 11 Earthquakes There are also various divergent and transform boundaries in the area shown in Figure 11.9, and as we’ve seen in the mid-Atlantic area, most of these earthquakes occur along the transform faults. The distribution of earthquakes with depth in the Kuril Islands of Russia in the northwest Pacific is shown in Figure 11.10. This is an ocean-ocean convergent boundary. The small red and yellow dots show background seismicity over a number of years, while the larger white dots are individual shocks associated with a M6.9 earthquake in April 2009. The relatively large earthquake took place on the upper part of the plate boundary between 60 km and 140 km inland from the trench. As we saw for the Cascadia subduction zone, this is where large subduction earthquakes are expected to occur. In fact, all of the very large earthquakes — M9 or higher — take place at subduction boundaries because there is the potential for a greater width of rupture zone on a gently dipping boundary than on a steep transform boundary. The largest earthquakes on transform boundaries are in the order of M8. Figure 11.10 Distribution of earthquakes in the area of the Kuril Islands, Russia (just north of Japan) (White dots represent the April 2009 M6.9 earthquake. Red and yellow dots are from background seismicity over several years prior to 2009.) [SE after Gavin Hayes, from data at subduction_zone/us2009fdak/szgc/ku6_trench.pdf] The background seismicity at this convergent boundary, and on other similar ones, is predominantly near the upper side of the subducting plate. The frequency of earthquakes is greatest near the surface and especially around the area where large subduction quakes happen, but it extends to at least 400 km depth. There is also significant seismic activity in the overriding North America Plate, again most commonly near the region of large quakes, but also extending for a few hundred kilometres away from the plate boundary. The distribution of earthquakes in the area of the India-Eurasia plate boundary is shown in Figure 11.11. This is a continent-continent convergent boundary, and it is generally assumed that although the India

Figure 11.11 Distribution of earthquakes in the area where the India Plate is converging with the Asia Plate (data from 1990 to 1996, red: 0-33 km, orange: 33-70 km, green: 70-300 km). (Spreading ridges are heavy lines, subduction zones are toothed lines, and transform faults are light lines. The double line along the northern edge of the India Plate indicates convergence, but not subduction. Plate motions are shown in mm/y.) [SE after Dale Sawyer, Rice University, The entire northern India and southern Asia region is very seismically active. Earthquakes are common in northern India, Nepal, Bhutan, Bangladesh and adjacent parts of China, and throughout Pakistan and Afghanistan. Many of the earthquakes are related to the transform faults on either side of the India Plate, and most of the others are related to the significant tectonic squeezing caused by the continued convergence of the India and Asia Plates. That squeezing has caused the Asia Plate to be thrust over top of the India Plate, building the Himalayas and the Tibet Plateau to enormous heights. Most of the earthquakes of Figure 11.11 are related to the thrust faults shown in Figure 11.12 (and to hundreds of other similar ones that cannot be shown at this scale). The southernmost thrust fault in Figure 11.12 is equivalent to the Main Boundary Fault in Figure 11.11.

Figure 11.12 Schematic diagram of the India-Asia convergent boundary, showing examples of the types of faults along which earthquakes are focussed. The devastating Nepal earthquake of May 2015 took place along one of these thrust faults. [SE after D. Vouichard, from a United Nations University document at: unupbooks/80a02e/80A02E05.htm] There is a very significant concentration of both shallow and deep (greater than 70 km) earthquakes in the northwestern part of Figure 11.11. This is northern Afghanistan, and at depths of more than 70 km, many of these earthquakes are within the mantle as opposed to the crust. It is interpreted that these deep earthquakes are caused by northwestward subduction of part of the India Plate beneath the Asia Plate in this area.

There are two main ways to measure earthquakes. The first of these is an estimate of the energy released, and the value is referred to as magnitude. This is the number that is typically used by the press when a big earthquake happens. It is often referred to as “Richter magnitude,” but that is a misnomer, and it should be just “magnitude.” There are many ways to measure magnitude — including Charles Richter’s method developed in 1935 — but they are all ways to estimate the same number: the amount of energy released. The other way of assessing the impact of an earthquake is to assess what people felt and how much damage was done. This is known as intensity. Intensity values are assigned to locations, rather than to the earthquake itself, and therefore intensity can vary widely, depending on the proximity to the earthquake and the types of materials and conditions of the subsurface.

Before we look more closely at magnitude we need to review what we know about body waves, and look at surface waves. Body waves are of two types, P-waves, or primary or compression waves (like the compression of the coils of a spring), and S-waves, or secondary or shear waves (like the flick of a rope). An example of P and S seismic wave records is shown in Figure 11.13. The critical parameters for the measurement of Richter magnitude are labelled, including the time interval between the arrival of the P- and S-waves — which is used to determine the distance from the earthquake to the seismic station, and the amplitude of the S waves — which is used to estimate the magnitude of the earthquake.

When body waves (P or S) reach Earth’s surface, some of their energy is transformed into surface waves, of which there are two main types, as illustrated in Figure 11.14. Rayleigh waves are characterized by vertical motion of the ground surface, like waves on water, while Love waves are characterized by horizontal motion. Both Rayleigh and Love waves are about 10% slower than S-waves (so they arrive

Figure 11.14 Depiction of seismic surface waves [SE after: https://en.wikipedia.org/wiki/Rayleigh_wave#/media/ File:Rayleigh_wave.jpg and https://en.wikipedia.org/wiki/ Love_wave#/media/File:Love_wave.jpg] Other important terms for describing earthquakes are hypocentre (or focus) and epicentre. The hypocentre is the actual location of an individual earthquake shock at depth in the ground, and the epicentre is the point on the land surface directly above the hypocentre (Figure 11.15).

A number of methods for estimating magnitude are listed in Table 11.1. Local magnitude (ML) was widely used until late in the 20th century, but moment magnitude (MW) is now more commonly used because it gives more accurate estimates (especially with larger earthquakes) and can be applied to earthquakes at any distance from a seismometer. Surface-wave magnitudes can also be applied to measure distant large earthquakes. Because of the increasing size of cities in earthquake-prone areas (e.g., China, Japan, California) and the increasing sophistication of infrastructure, it is becoming important to have very rapid warnings and magnitude estimates of earthquakes that have already happened. This can be achieved by using P-wave data to determine magnitude because P-waves arrive first at seismic stations, in many cases several seconds ahead of the more damaging S-waves and surface waves. Operators of electrical grids, pipelines, trains, and other infrastructure can use the information to automatically shut down systems so that damage and casualties can be limited.

Physical Geology 312 32 times higher than that released by one of M3, and this ratio applies to all intervals in the scale. If we assign an arbitrary energy level of 1 unit to a M1 earthquake the energy for quakes up to M8 will be as shown on the following chart: In any given year, when there is a large earthquake on Earth (M8 or M9), the amount of energy released by that one event will likely exceed the energy released by all smaller earthquake events combined.

The intensity of earthquake shaking at any location is determined not only by the magnitude of the earthquake and its distance, but also by the type of underlying rock or unconsolidated materials. If buildings are present, the size and type of buildings (and their inherent natural vibrations) are also important. Intensity scales were first used in the late 19th century, and then adapted in the early 20th century by Giuseppe Mercalli and modified later by others to form what we know call the modified Mercalli intensity scale (Table 11.2). Intensity estimates are important because they allow us to characterize parts of any region into areas that are especially prone to strong shaking versus those that are not. The key factor in this regard is the nature of the underlying geological materials, and the weaker those are, the more likely it is that there will be strong shaking. Areas underlain by strong solid bedrock tend to experience much less shaking than those underlain by unconsolidated river or lake sediments.

An example of this effect is the 1985 M8 earthquake that struck the Michoacán region of western Mexico, southwest of Mexico City. There was relatively little damage in the area around the epicentre, but there was tremendous damage and about 5,000 deaths in heavily populated Mexico City some 350 km from the epicentre. The key reason for this is that Mexico City was built largely on the unconsolidated and water-saturated sediment of former Lake Texcoco. These sediments resonate at a frequency of about two seconds, which was similar to the frequency of the body waves that reached the city. For the same reason that a powerful opera singer can break a wine glass by singing the right note, the amplitude of the seismic waves was amplified by the lake sediments. Survivors of the disaster recounted that the ground in some areas moved up and down by about 20 cm every two seconds for over two minutes. Damage was greatest to buildings between 5 and 15 storeys tall, because they also resonated at around two seconds, which amplified the shaking.

An intensity map for the 1946 M7.3 Vancouver Island earthquake is shown in Figure 11.16. The intensity was greatest in the central island region where, in some communities, chimneys were damaged on more than 75% of buildings, some roads were made impassable, and a major rock slide occurred. The earthquake was felt as far north as Prince Rupert, as far south as Portland Oregon, and as far east as the Rockies

Some of the common impacts of earthquakes include structural damage to buildings, fires, damage to bridges and highways, initiation of slope failures, liquefaction, and tsunami. The types of impacts depend to a large degree on where the earthquake is located: whether it is predominantly urban or rural, densely or sparsely populated, highly developed or underdeveloped, and of course on the ability of the infrastructure to withstand shaking. As we’ve seen from the example of the 1985 Mexico earthquake, the geological foundations on which structures are built can have a significant impact on earthquake shaking. When an earthquake happens, the seismic waves produced have a wide range of frequencies. The energy of the higher frequency waves tends to be absorbed by solid rock, while the lower frequency waves (with periods slower than one second) pass through the solid rock without being absorbed, but are eventually absorbed and amplified by soft sediments. It is therefore very common to see much worse earthquake damage in areas underlain by soft sediments than in areas of solid rock. A good example of this is in the Oakland area near San Francisco, where parts of a two-layer highway built on soft sediments collapsed during the 1989 Loma Prieta earthquake (Figure 11.17).

Building damage is also greatest in areas of soft sediments, and multi-storey buildings tend to be more seriously damaged than smaller ones. Buildings can be designed to withstand most earthquakes, and this practice is increasingly applied in earthquake-prone regions. Turkey is one such region, and even though Turkey had a relatively strong building code in the 1990s, adherence to the code was poor, as builders did whatever they could to save costs, including using inappropriate materials in concrete and reducing

317 Chapter 11 Earthquakes the amount of steel reinforcing. The result was that there were over 17,000 deaths in the 1999 M7.6 Izmit earthquake (Figure 11.18). After two devastating earthquakes that year, Turkish authorities strengthened the building code further, but the new code has been applied only in a few regions, and enforcement of the code is still weak, as revealed by the amount of damage from a M7.1 earthquake in eastern Turkey in 2011.

Fires are commonly associated with earthquakes because fuel pipelines rupture and electrical lines are damaged when the ground shakes (Figure 11.19). Most of the damage in the great 1906 San Francisco earthquake was caused by massive fires in the downtown area of the city (Figure 11.20). Some 25,000 buildings were destroyed by those fires, which were fuelled by broken gas pipes. Fighting the fires was difficult because water mains had also ruptured. The risk of fires can be reduced through P-wave early warning systems if utility operators can reduce pipeline pressure and close electrical circuits. Figure 11.19 Some of the effects of the 2011 Tohoku earthquake in the Sendai area of Japan. An oil refinery is on fire, and a vast area has been flooded by a tsunami. [from: 2011_T%C5%8Dhoku_earthquake_and_tsunami#/media/ File:SH-60B_helicopter_flies_over_Sendai.jpg]

Earthquakes are important triggers for failures on slopes that are already weak. An example is the Las Colinas slide in the city of Santa Tecla, El Salvador, which was triggered by a M7.6 offshore earthquake in January 2001 (Figure 11.21).

Figure 11.21 The Las Colinas debris flow at Santa Tecla (a suburb of the capital San Salvador) triggered by the January 2001 El Salvador earthquake. This is just one of many hundreds of slope failures that resulted from that earthquake. Over 500 people died in the area affected by this slide. [from: images/foreign/ElSalvadorslide.jpg] Ground shaking during an earthquake can be enough to weaken rock and unconsolidated materials to the point of failure, but in many cases the shaking also contributes to a process known as liquefaction, in which an otherwise solid body of sediment is transformed into a liquid mass that can flow. When water-saturated sediments are shaken, the grains become rearranged to the point where they are no longer supporting one another. Instead, the water between the grains is holding them apart and the material can flow. Liquefaction can lead to the collapse of buildings and other structures that might be otherwise undamaged. A good example is the collapse of apartment buildings during the 1964 Niigata earthquake (M7.6) in Japan (Figure 11.22). Liquefaction can also contribute to slope failures and to fountains of sandy mud (sand volcanoes) in areas where there is loose saturated sand beneath a layer of more cohesive clay.

Figure 11.22 Collapsed apartment buildings in the Niigata area of Japan. The material beneath the buildings was liquefied to varying degrees by the 1964 earthquake. Niigata_earthquake#/media/File: Liquefaction_at_Niigata.JPG Parts of the Fraser River delta are prone to liquefaction-related damage because the region is characterized by a 2 m to 3 m thick layer of fluvial silt and clay over top of at least 10 m of water- saturated fluvial sand (Figure 11.23). Under these conditions, it can be expected that seismic shaking will be amplified and, the sandy sediments will liquefy. This could lead to subsidence and tilting of buildings, and to failure and sliding of the silt and clay layer. Current building-code regulations in the Fraser delta area require that measures be taken to strengthen the ground underneath multi-storey buildings prior to construction.

Exercise 11.4 Creating Liquefaction and Discovering the Harmonic Frequency There are a few ways that you can demonstrate the process of liquefaction for yourself. The simplest is to go to a sandy beach (lake, ocean, or river) and find a place near the water’s edge where the sand is wet. This is best done with your shoes off, so let’s hope it’s not too cold! While standing in one place on a wet part of the beach, start moving your feet up and down at a frequency of about once per second. Within a few seconds the previously firm sand will start to lose strength, and you’ll gradually sink in up to your ankles. If you can’t get to a beach, or if the weather isn’t cooperating, put some sand (sandbox sand will do) into a small container, saturate it with water, and then pour the excess water off. You can shake it gently to get the water to separate and then pour the excess water away, and you may have to do that more than once. Place a small rock on the surface of the sand; it should sit there for hours without sinking in. Now, holding the container in one hand gently thump the side or the bottom with your other hand, about twice a second. The rock should gradually sink in as the sand around it becomes liquefied. As you were moving your feet up and down or thumping the pot, it’s likely that you soon discovered the most effective rate for getting the sand to liquefy; this would have been close to the natural harmonic frequency for that body of material. Stepping up and down as fast as you can (several times per second) on the wet beach would not have been effective, nor would you have achieved much by stepping once every several seconds. The body of sand vibrates most readily in response to shaking that is close to its natural harmonic frequency, and liquefaction is also most likely to occur at that frequency. Earthquakes that take place beneath the ocean have the potential to generate tsunami.[footnote]Tsunami is the Japanese word for harbour wave. It is the same in both singular and plural.[/footnote] The most likely situation for a significant tsunami is a large (M7 or greater) subduction-related earthquake. As shown in Figure 11.24, during the time between earthquakes the overriding plate becomes distorted by elastic deformation; it is squeezed laterally (Figure 11.24B) and pushed up. When an earthquake happens (Figure 11.24C), the plate rebounds and there is both uplift and subsidence on the sea floor, in some cases by as much as several metres vertically over an area of thousands of square kilometres. This vertical motion is transmitted through the water column where it generates a wave that then spreads across the ocean.

Figure 11.24 Elastic deformation and rebound of overriding plate at a subduction setting (B). The release of the locked zone during an earthquake (C) results in both uplift and subsidence on the sea floor, and this is transmitted to the water overhead, resulting in a tsunami. [SE] Subduction earthquakes with magnitude less than 7 do not typically generate significant tsunami because the amount of vertical displacement of the sea floor is minimal. Sea-floor transform earthquakes, even large ones (M7 to M8), don’t typically generate tsunami either, because the motion is mostly side to side, not vertical. Tsunami waves travel at velocities of several hundred kilometres per hour and easily make it to the far side of an ocean in about the same time as a passenger jet. The simulated one shown in Figure 11.25 is similar to that created by the 1700 Cascadia earthquake off the coast of British Columbia, Washington, and Oregon, which was recorded in Japan nine hours later.

Figure 11.25 Model of the tsunami from the 1700 Cascadia earthquake (~M9) showing open-ocean wave heights (colours) and travel time contours. Tsunami wave amplitudes typically increase in shallow water. [from NOAA/PMEL/Center for Tsunami Research, at:

It has long been a dream of seismologists, geologists, and public safety officials to be able to accurately predict the location, magnitude, and timing of earthquakes on time scales that would be useful for minimizing danger to the public and damage to infrastructure (e.g., weeks, days, hours). Many different avenues of prediction have been explored, such as using observations of warning foreshocks, changes in magnetic fields, seismic tremor, changing groundwater levels, strange animal behaviour, observed periodicity, stress transfer considerations, and others. So far, none of the research into earthquake prediction has provided a reliable method. Although there are some reports of successful earthquake predictions, they are rare, and many are surrounded by doubtful circumstances. The problem with earthquake predictions, as with any other type of prediction, is that they have to be accurate most of the time, not just some of the time. We have come to rely on weather predictions because they are generally (and increasingly) accurate. But if we try to predict earthquakes and are only accurate 10% of the time (and even that isn’t possible with the current state of knowledge), the public will lose faith in the process very quickly, and then will ignore all of the predictions. Efforts are currently focused on forecasting earthquake probabilities, rather than predicting their occurrence. There was great hope for earthquake predictions late in the 1980s when attention was focused on part of the San Andreas Fault at Parkfield, about 200 km south of San Francisco. Between 1881 and 1965 there were five earthquakes at Parkfield, most spaced at approximately 20-year intervals, all confined to the same 20 km-long segment of the fault, and all very close to M6 (Figure 11.26). Both the 1934 and 1966 earthquakes were preceded by small foreshocks exactly 17 minutes before the main quake.

The U.S. Geological Survey recognized this as an excellent opportunity to understand earthquakes and earthquake prediction, so they armed the Parkfield area with a huge array of geophysical instruments and waited for the next quake, which was expected to happen around 1987. Nothing happened! The “1987 Parkfield earthquake” finally struck in September 2004. Fortunately all of the equipment was still there, but it was no help from the perspective of earthquake prediction. There were no significant precursors to the 2004 Parkfield earthquake in any of the parameters measured, including seismicity, harmonic

325 Chapter 11 Earthquakes tremor, strain (rock deformation), magnetic field, the conductivity of the rock, or creep, and there was no foreshock. In other words, even though every available technique was used to monitor it, the 2004 earthquake came as a complete surprise, with no warning whatsoever. The hope for earthquake prediction is not dead, but it was hit hard by the Parkfield experiment. The current focus in earthquake-prone regions is to provide forecasts of the probability of an earthquake of a certain magnitude within a certain time period — typically a number of decades — while officials focus on ensuring that the population is educated about earthquake risks and that buildings and other infrastructure are as safe as can be. An example of this approach for the San Francisco Bay region of California is shown in Figure 11.27. Based on a wide range of information, including past earthquake history, accumulated stress from plate movement, and known stress transfer, seismologists and geologists have predicted the likelihood of a M6.7 or greater earthquake on each of eight major faults that cut through the region. The greatest probabilities are on the San Andreas, Rogers Creek, and Hayward Faults. As shown in Figure 11.27, there is a 63% chance that a major and damaging earthquake will take place somewhere in the region prior to 2036.

As we’ve discussed already, it’s not sufficient to have strong building codes, they have to be enforced. Building code compliance is quite robust in most developed countries, but is sadly inadequate in many developing countries. It’s also not enough just to focus on new buildings; we have to make sure that existing buildings — especially schools and hospitals — and other structures such as bridges and dams, are as safe as they can be. An example of how this is applied to schools in B.C. is described in Box 11.2.

British Columbia is in the middle of a multi-billion-dollar program to make schools safer for students. The program is focused on older schools, because, according to the government, those built since 1992 already comply with modern seismic codes. Some schools would require too much work to make

An example is Sangster Elementary in Colwood on southern Vancouver Island. The school was originally built in 1957, with a major addition in 1973. Ironically, the newer part of the school, built of concrete blocks, required strengthening with the addition of a steel framework, while the 1957 part, which is a wood-frame building, did not require seismic upgrading. The work was completed in 2014.

As of January 2015, upgrades had been completed at 145 B.C. schools, 11 were underway, and an additional 57 were ready to proceed with funding identified.* Another 129 schools were listed as needing upgrades. In May 2015, the provincial government announced that the target date for completion of the upgrades, originally set for 2020, had been delayed to 2030.* The final part of earthquake preparedness involves the formulation of public emergency plans, including escape routes, medical facilities, shelters, and food and water supplies. It also includes personal planning, such as emergency supplies (food, water, shelter, and warmth), escape routes from houses and offices, and communication strategies (with a focus on ones that don’t involve the cellular network).

What is an aftershock and what is the relationship between aftershocks and stress transfer? Episodic slip on the middle part of the Cascadia subduction zone is thought to result in an increase in the stress on the upper part where large earthquakes take place. Why? Explain the difference between magnitude and intensity as expressions of the size of an earthquake. How much more energy is released by an M7.3 earthquake compared with an M5.3 earthquake? The map shows earthquake locations with the depths coded according the colour scheme used in Figure 11.11. What type of plate boundary is this? Draw a line on the map to show approximately where the plate boundary is situated. In which directions are the plates moving, and where in the world might this be? Earthquakes are relatively common along the mid-ocean ridges. At what type of plate boundary do most such quakes occur? The northward motion of the Pacific Plate relative to the North America Plate takes place along two major transform faults. What are they called? Why is earthquake damage likely to be more severe for buildings built on unconsolidated sediments as opposed to solid rock? Why are fires common during earthquakes? What type of earthquake is likely to lead to a tsunami? What did we learn about earthquake prediction from the 2004 Parkfield earthquake? What are some of the things we should know about an area in order to help minimize the impacts of an earthquake? What is the difference between earthquake prediction and forecasting?

Observing and understanding geological structures helps us to determine the kinds of stresses that have existed within Earth in the past. This type of information is critical to our understanding of plate tectonics, earthquakes, the formation of mountains, metamorphism, and Earth resources. Some of the 331 Steven Earle types of geological structures that are important to study include fractures, faults, and folds. Structural geologists make careful observations of the orientations of these structures and the amount and direction of offset along faults.

Rocks are subject to stress —mostly related to plate tectonics but also to the weight of overlying rocks—and their response to that stress is strain (deformation). In regions close to where plates are converging stress is typically compressive—the rocks are being squeezed. Where plates are diverging the stress is extensive—rocks are being pulled apart. At transform plate boundaries, where plates are moving side by side there is sideways or shear stress—meaning that there are forces in opposite directions parallel to a plane. Rocks have highly varying strain responses to stress because of their different compositions and physical properties, and because temperature is a big factor and rock temperatures within the crust can vary greatly. We can describe the stress applied to a rock by breaking it down into three dimensions—all at right angles to one-another (Figure 12.2). If the rock is subject only to the pressure of burial, the stresses in all three directions will likely be the same. If it is subject to both burial and tectonic forces, the pressures will be different in different directions. Figure 12.2 Depiction of the stress applied to rocks within the crust. The stress can be broken down into three components. Assuming that we’re looking down in this case, the green arrows represent north-south stress, the red arrows represent east-west stress, and the blue arrows (the one underneath is not visible) represent up-down stress. On the left, all of the stress components are the same. On the right, the north-south stress is least and the up-down stress is greatest. [SE] Rock can respond to stress in three ways: it can deform elastically, it can deform plastically, and it can break or fracture. Elastic strain is reversible; if the stress is removed, the rock will return to its original shape just like a rubber band that is stretched and released. Plastic strain is not reversible. As already noted, different rocks at different temperatures will behave in different ways to stress. Higher temperatures lead to more plastic behaviour. Some rocks or sediments are also more plastic when they are wet. Another factor is the rate at which the stress is applied. If the stress is applied quickly (for example, because of an extraterrestrial impact or an earthquake), there will be an increased tendency for the rock to fracture. Some different types of strain response are illustrated in Figure 12.3.

Figure 12.3 The varying types of response of geological materials to stress. The straight dashed parts are elastic strain and the curved parts are plastic strain. In each case the X marks where the material fractures. A, the strongest material, deforms relatively little and breaks at a high stress level. B, strong but brittle, shows no plastic deformation and breaks after relatively little elastic deformation. C, the most deformable, breaks only after significant elastic and plastic strain. The three deformation diagrams on the right show A and C before breaking and B after breaking. [SE] The outcomes of placing rock under stress are highly variable, but they include fracturing, tilting and folding, stretching and squeezing, and faulting. A fracture is a simple break that does not involve significant movement of the rock on either side. Fracturing is particularly common in volcanic rock, which shrinks as it cools. The basalt columns in Figure 12.4a are a good example of fracture. Beds are sometimes tilted by tectonic forces, as shown in Figure 12.4b, or folded as shown in Figure 12.1.

When a body of rock is compressed in one direction it is typically extended (or stretched) in another. This is an important concept because some geological structures only form under compression, while others only form under tension. Most of the rock in Figure 12.4c is limestone, which is relatively easily deformed when heated. The dark rock is chert, which remains brittle. As the limestone stretched (parallel to the hammer handle) the brittle chert was forced to break into fragments to accommodate the change in shape of the body of rock. A fault is a rock boundary along which the rocks on either side have been displaced relative to each other (Figure 12.4d).

When a body of rock, especially sedimentary rock, is squeezed from the sides by tectonic forces, it is likely to fracture and/or become faulted if it is cold and brittle, or become folded if it is warm enough to behave in a plastic manner. The nomenclature and geometry of folds are summarized on Figure 12.5. An upward fold is called an anticline, while a downward fold is called a syncline. In many areas it’s common to find a series of anticlines and synclines (as in Figure 12.5), although some sequences of rocks are folded into a single anticline or syncline. A plane drawn through the crest of a fold in a series of beds is called the axial plane of the fold. The sloping beds on either side of an axial plane are limbs. An anticline or syncline is described as symmetrical if the angles between each of limb and the axial plane are generally similar, and asymmetrical if they are not. If the axial plane is sufficiently tilted that the beds on one side have been tilted past vertical, the fold is known as an overturned anticline or syncline.

A very tight fold, in which the limbs are parallel or nearly parallel to one another is called an isoclinal fold (Figure 12.6). Isoclinal folds that have been overturned to the extent that their limbs are nearly horizontal are called recumbent folds.

Physical Geology 336 Large folds can have wavelengths of tens of kilometres, and very small ones might be visible only under a microscope. Anticlines are not necessarily, or even typically, expressed as ridges in the terrain, nor synclines as valleys. Folded rocks get eroded just like all other rocks and the topography that results is typically controlled mostly by the resistance of different layers to erosion (Figure 12.8).

A body of rock that is brittle—either because it is cold or because of its composition, or both— is likely to break rather than fold when subjected to stress, and the result is fracturing or faulting.

Fracturing is common in rocks near the surface, either in volcanic rocks that have shrunk on cooling (Figure 12.4a), or in other rocks that have been exposed by erosion and have expanded (Figure 12.9).

A fracture in a rock is also called a joint. There is no side-to-side movement of the rock on either side of a joint. Most joints form where a body of rock is expanding because of reduced pressure, as shown by the two examples in Figure 12.9, or where the rock itself is contracting but the body of rock remains the same size (the cooling volcanic rock in Figure 12.4a). In all of these cases, the pressure regime is one of tension as opposed to compression. Joints can also develop where rock is being folded because, while folding typically happens during compression, there may be some parts of the fold that are in tension (Figure 12.10).

Finally joints can also develop when rock is under compression as shown on Figure 12.11, where there is differential stress on the rock, and joint sets develop at angles to the compression directions.

A fault is boundary between two bodies of rock along which there has been relative motion (Figure 12.4d). As we discussed in Chapter 11, an earthquake involves the sliding of one body of rock past another. Earthquakes don’t necessarily happen on existing faults, but once an earthquake takes place a fault will exist in the rock at that location. Some large faults, like the San Andreas Fault in California or the Tintina Fault, which extends from northern B.C. through central Yukon and into Alaska, show evidence of hundreds of kilometres of motion, while others show less than a millimetre. In order to

Figure 12.12 A fault (white dashed line) in intrusive rocks on Quadra Island, B.C. The pink dyke has been offset by the fault and the extent of the offset is shown by the white arrow (approximately 10 cm). Because the far side of the fault has moved to the right, this is a right-lateral fault. If the photo were taken from the other side, the fault would still appear to have a right-lateral offset. [SE] There are several kinds of faults, as illustrated on Figure 12.13, and they develop under different stress conditions. The terms hanging wall and footwall in the diagrams apply to situations where the fault is not vertical. The body of rock above the fault is called the hanging wall, and the body of rock below it is called the footwall. If the fault develops in a situation of compression, then it will be a reverse fault because the compression causes the hanging wall to be pushed up relative to the footwall. If the fault develops in a situation of extension, then it will be a normal fault, because the extension allows the hanging wall to slide down relative to the footwall in response to gravity. The third situation is where the bodies of rock are sliding sideways with respect to each other, as is the case along a transform fault (see Chapter 10). This is known as a strike-slip fault because the displacement is along the “strike” or the length of the fault. On strike-slip faults the motion is typically only horizontal, or with a very small vertical component, and as discussed above the sense of motion can be right lateral (the far side moves to the right), as in Figures 12.12 and 12.13, or it can be left lateral (the far side moves to the left). Transform faults are strike-slip faults.

Figure 12.13 Depiction of reverse, normal, and strike-slip faults. Reverse faults happen during compression while normal faults happen during extension. Most strike-slip faults are related to transform boundaries. [SE after: education/images/GRAPHICS/fault_types_2.jpg] In areas that are characterized by extensional tectonics, it is not uncommon for a part of the upper crust to subside with respect to neighbouring parts. This is typical along areas of continental rifting, such as the Great Rift Valley of East Africa or in parts of Iceland, but it is also seen elsewhere. In such situations a down-dropped block is known as a graben (German for ditch), while an adjacent block that doesn’t subside is called a horst (German for heap) (Figure 12.14). There are many horsts and grabens in the Basin and Range area of the western United States, especially in Nevada. Part of the Fraser Valley region of B.C., in the area around Sumas Prairie is a graben.

A special type of reverse fault, with a very low-angle fault plane, is known as a thrust fault. Thrust faults are relatively common in areas where fold-belt mountains have been created during continent- continent collision. Some represent tens of kilometres of thrusting, where thick sheets of sedimentary rock have been pushed up and over top of other rock (Figure 12.15).

There are numerous thrust faults in the Rocky Mountains, and a well-known example is the McConnell Thrust, along which a sequence of sedimentary rocks about 800 m thick has been pushed for about 40 km from west to east (Figure 12.16). The thrusted rocks range in age from Cambrian to Cretaceous, so in the area around Mt. Yamnuska Cambrian-aged rock (around 500 Ma) has been thrust over, and now lies on top of Cretaceous-aged rock (around 75 Ma) (Figure 12.17).

Exercise 12.2 Types of Faults The four images are faults that formed in different tectonic settings. Identifying the type of fault allows us to determine if the body of rock was under compression or extension at the time of faulting. Complete the table below the images, identifying the types of faults (normal or reversed) and whether each one formed under

Geologists take great pains to measure and record geological structures because they are critically important to understanding the geological history of a region. One of the key features to measure is the orientation, or attitude, of bedding. We know that sedimentary beds are deposited in horizontal layers, so if the layers are no longer horizontal, then we can infer that they have been affected by tectonic forces and have become either tilted, or folded. We can express the orientation of a bed (or any other planar feature) with two values: first, the compass orientation of a horizontal line on the surface—the strike—and second, the angle at which the surface dips from the horizontal, (perpendicular to the strike)—the dip (Figure 12.18). It may help to imagine a vertical surface, such as a wall in your house. The strike is the compass orientation of the wall and the dip is 90˚ from horizontal. If you could push the wall so it’s leaning over, but still attached to the floor, the strike direction would be the same, but the dip angle would be less than 90˚. If you pushed the wall over completely so it was lying on the floor, it would no longer have a strike direction and its dip would be 0˚. When describing the dip it is important to include the direction. In other words. if the strike is 0˚ (i.e., north) and the dip is 30˚, it would be necessary to say “to the west” or “to the east.” Similarly if the strike is 45˚ (i.e., northeast) and the dip is 60˚, it would be necessary to say “to the northwest” or “to the southeast.” Measurement of geological features is done with a special compass that has a built-in clinometer, which is a device for measuring vertical angles. An example of how this is done is shown on Figure 12.19.

Strike and dip are also used to describe any other planar features, including joints, faults, dykes, sills, and even the foliation planes in metamorphic rocks. Figure 12.20 shows an example of how we would depict the beds that make up an anticline on a map.

The beds on the west (left) side of the map are dipping at various angles to the west. The beds on the east side are dipping to the east. The middle bed (light grey) is horizontal; this is denoted by a cross within a circle. The dyke is dipping at 80˚ to the west. The hinge of the fold is denoted with a dashed line with two arrows point away from it. If it were a syncline, the arrows would point towards the line.

Explain why fractures are common in volcanic rocks. What is the difference between a normal fault and a reverse fault, and under what circumstances would you expect these to form? What type of fault would you expect to see near to a transform plate boundary? This diagram is a plan view (map) of the geology of a region. The coloured areas represent sedimentary beds. (i) Describe in words the general attitude (strike and dip) of these beds. ( ii) Which of these beds is the oldest? What is “a” and what is its attitude? What is “b” and what is its attitude? Which of these terms applies to “b”: “left lateral” or “right lateral”?

Explain the hydrological cycle and its relevance to streams and what residence time means in this context Describe a drainage basin and explain the origins of different types of drainage patterns Explain how streams become graded and how certain geological and anthropogenic changes can result in a stream losing its gradation Describe the formation of stream terraces Describe the processes by which sediments are moved by streams and the flow velocities that are necessary to erode them from the stream bed and keep them suspended in the water Explain the origins of natural stream levées Describe the process of stream evolution and the types of environments where one would expect to find straight-channel, braided, and meandering streams Describe the annual flow characteristics of typical streams in Canada and the processes that lead to flooding Describe some of the important historical floods in Canada Determine the probability of a flood of a particular size based on the flood history of a stream Explain some of the steps that we can take to limit the damage from flooding

Streams are the most important agents of erosion and transportation of sediments on Earth’s surface. They are responsible for the creation of much of the topography that we see around us. They are also places of great beauty and tranquility, and of course, they provide much of the water that is essential to our existence. But streams are not always peaceful and soothing. During large storms and rapid snowmelts, they can become raging torrents capable of moving cars and houses and destroying roads and bridges. When they spill over their banks, they can flood huge areas, devastating populations and infrastructure. Over the past century, many of the most damaging natural disasters in Canada have been floods, and we can expect them to become even more severe as the climate changes.

Water is constantly on the move. It is evaporated from the oceans, lakes, streams, the surface of the land, and plants (transpiration) by solar energy (Figure 13.2). It is moved through the atmosphere by winds and condenses to form clouds of water droplets or ice crystals. It comes back down as rain or snow and then flows through streams, into lakes, and eventually back to the oceans. Water on the surface and in streams and lakes infiltrates the ground to become groundwater. Groundwater slowly moves through the rock and surficial materials. Some groundwater returns to other streams and lakes, and some goes directly back to the oceans. Figure 13.2 The various components of the water cycle. Black or white text indicates the movement or transfer of water from one reservoir to another. Yellow text indicates the storage of water. [SE after Wikipedia: Water_cycle_blank.svg] Even while it’s moving around, water is stored in various reservoirs. The largest, by far, is the oceans, accounting for 97% of the volume (Figure 13.3). Of course, that water is salty. The remaining 3% is fresh water. Two-thirds of our fresh water is stored in ice and one-third is stored in the ground. The remaining fresh water — about 0.03% of the total — is stored in lakes, streams, vegetation, and the atmosphere.

Figure 13.3a The storage reservoirs for water on Earth. Glacial ice is represented by the white band, groundwater the red band, and surface water the very thin blue band at the top. The 0.001% stored in the atmosphere is not shown. [SE using data from: https://web.archive.org/web/ 20180518215745/https://water.usgs.gov/edu/ watercyclefreshstorage.html] To put that in perspective, let’s think about putting all of Earth’s water into a 1 L jug (Figure 13.3b). We start by almost filling the jug with 970 ml of water and 34 g of salt. Then we add one regular-sized (~20 mL) ice cube (representing glacial ice) and two teaspoons (~10 mL) of groundwater. All of the water that we see around us in lakes and streams and up in the sky can be represented by adding three more drops from an eyedropper.

Although the proportion of Earth’s water that is in the atmosphere is tiny, the actual volume is huge. At any given time, there is the equivalent of approximately 13,000 km3 of water in the air in the form of water vapour and water droplets in clouds. Water is evaporated from the oceans, vegetation, and lakes at a rate of 1,580 km3 per day, and just about exactly the same volume falls as rain and snow every day — over both the oceans and land. The precipitation that falls on land goes back to the ocean in the form of stream flow (117 km3/day) and groundwater flow (6 km3/day). Most of the rest of this chapter is about that 117 km3/day of streamflow. The average discharge of the Fraser River into the ocean is approximately 0.31 km3/day, or 0.26% of the world’s total.

A stream is a body of flowing surface water of any size, ranging from a tiny trickle to a mighty river. The area from which the water flows to form a stream is known as its drainage basin. All of the precipitation (rain or snow) that falls within a drainage basin eventually flows into its stream, unless some of that water is able to cross into an adjacent drainage basin via groundwater flow. An example of a drainage basin is shown in Figure 13.4.

Figure 13.5 Profile of the main stem of Cawston Creek near Keremeos, B.C. The maximum elevation of the drainage basin is about 1,840 m, near Mount Kobau. The base level is 275 m, at the Similkameen River. As shown, the gradient of the stream can be determined by dividing the change in elevation between any two points (rise) by the distance between those two points (run). [SE] Cawston Creek is a typical small drainage basin (approximately 25 km2) within a very steep glaciated valley. As shown in Figure 13.5, the upper and middle parts of the creek have steep gradients (averaging about 200 m/km but ranging from 100 to 350 m/km), and the lower part, within the valley of the 355 Physical Geology 356 Similkameen River, is relatively flat (<5 m/km). The shape of the valley has been controlled first by tectonic uplift (related to plate convergence), then by pre-glacial stream erosion and mass wasting, then by several episodes of glacial erosion, and finally by post-glacial stream erosion. The lowest elevation of Cawston Creek (275 m at the Similkameen River) is its base level. Cawston Creek cannot erode below that level unless the Similkameen River erodes deeper into its flood plain (the area that is inundated during a flood). Metro Vancouver’s water supply comes from three large drainage basins on the north shore of Burrard Inlet, as shown in Figure 13.6. This map illustrates the concept of a drainage basin divide. The boundary between two drainage basins is the height of land between them. A drop of water falling on the boundary between the Capilano and Seymour drainage basins (a.k.a., watersheds), for example, could flow into either one of them.

The pattern of tributaries within a drainage basin depends largely on the type of rock beneath, and on structures within that rock (folds, fractures, faults, etc.). The three main types of drainage patterns are illustrated in Figure 13.7. Dendritic patterns, which are by far the most common, develop in areas where the rock (or unconsolidated material) beneath the stream has no particular fabric or structure and can be eroded equally easily in all directions. Examples would be granite, gneiss, volcanic rock, and sedimentary rock that has not been folded. Most areas of British Columbia have dendritic patterns, as do most areas of the prairies and the Canadian Shield. Trellis drainage patterns typically develop where sedimentary rocks have been folded or tilted and then eroded to varying degrees depending on their strength. The Rocky Mountains of B.C. and Alberta are a good example of this, and many of the drainage systems within the Rockies have trellis patterns. Rectangular patterns develop in areas that have very little topography and a system of bedding planes, fractures, or faults that form a rectangular network. Rectangular drainage patterns are rare in Canada. In many parts of Canada, especially relatively flat areas with thick glacial sediments, and throughout much of Canadian Shield in eastern and central Canada, drainage patterns are chaotic, or what is known as deranged (Figure 13.8, left). Lakes and wetlands are common in this type of environment.

A fourth type of drainage pattern, which is not specific to a drainage basin, is known as radial (Figure 13.8, right). Radial patterns form around isolated mountains (such as volcanoes) or hills, and the individual streams typically have dendritic drainage patterns.

Over geological time, a stream will erode its drainage basin into a smooth profile similar to that shown in Figure 13.9. If we compare this with an ungraded stream like Cawston Creek (Figure 13.5), we can see that graded streams are steepest in their headwaters and their gradient gradually decreases toward their mouths. Ungraded streams have steep sections at various points, and typically have rapids and waterfalls at numerous locations along their lengths.

A graded stream can become ungraded if there is renewed tectonic uplift, or if there is a change in the base level, either because of tectonic uplift or some other reason. As stated earlier, the base level of Cawston Creek is defined by the level of the Similkameen River, but this can change, and has done so in the past. Figure 13.10 shows the valley of the Similkameen River in the Keremeos area. The river channel is just beyond the row of trees. The green field in the distance is underlain by material eroded Physical Geology 358 from the hills behind and deposited by a small creek (not Cawston Creek) adjacent to the Similkameen River when its level was higher than it is now. Sometime in the past several centuries, the Similkameen River eroded down through these deposits (forming the steep bank on the other side of the river), and the base level of the small creek was lowered by about 10 m. Over the next few centuries, this creek will seek to become graded again by eroding down through its own alluvial fan. Figure 13.10 An example of a change in the base level of a small stream that flows into the Similkameen river near Keremeos. The previous base level was near the top of the sandy bank. The current base level is the river. [SE] Another example of a change in base level can be seen along the Juan de Fuca Trail on southwestern Vancouver Island. As shown in Figure 13.11, many of the small streams along this part of the coast flow into the ocean as waterfalls. It is evident that the land in this area has risen by about 5 m in the past few thousand years, probably in response to deglaciation. The streams that used to flow directly into the ocean now have a lot of down-cutting to do to become regraded.

Sediments accumulate within the flood plain of a stream, and then, if the base level changes, or if there is less sediment to deposit, the stream may cut down through those existing sediments to form terraces. A terrace on the Similkameen River is shown in Figure 13.10 and some on the Fraser River are shown in Figure 13.12. The Fraser River photo shows at least two levels of terraces.

In the late 19th century, American geologist William Davis proposed that streams and the surrounding terrain develop in a cycle of erosion (Figure 13.13). Following tectonic uplift, streams erode quickly, Physical Geology 360 developing deep V-shaped valleys that tend to follow relatively straight paths. Gradients are high, and profiles are ungraded. Rapids and waterfalls are common. During the mature stage, streams erode wider valleys and start to deposit thick sediment layers. Gradients are slowly reduced and grading increases. In old age, streams are surrounded by rolling hills, and they occupy wide sediment-filled valleys. Meandering patterns are common. Davis’s work was done long before the idea of plate tectonics, and he was not familiar with the impacts of glacial erosion on streams and their environments. While some parts of his theory are out of date, it is still a useful way to understand streams and their evolution.

As we discussed in Chapter 6, flowing water is a very important mechanism for both erosion and deposition. Water flow in a stream is primarily related to the stream’s gradient, but it is also controlled by the geometry of the stream channel. As shown in Figure 13.14, water flow velocity is decreased by friction along the stream bed, so it is slowest at the bottom and edges and fastest near the surface and in the middle. In fact, the velocity just below the surface is typically a little higher than right at the surface because of friction between the water and the air. On a curved section of a stream, flow is fastest on the outside and slowest on the inside.

Other factors that affect stream-water velocity are the size of sediments on the stream bed — because large particles tend to slow the flow more than small ones — and the discharge, or volume of water passing a point in a unit of time (e.g., m3/second). During a flood, the water level always rises, so there is more cross-sectional area for the water to flow in; however, as long as a river remains confined to its channel, the velocity of the water flow also increases. Figure 13.15 shows the nature of sediment transportation in a stream. Large particles rest on the bottom — bedload — and may only be moved during rapid flows under flood conditions. They can be moved by saltation (bouncing) and by traction (being pushed along by the force of the flow). Smaller particles may rest on the bottom some of the time, where they can be moved by saltation and traction, but they can also be held in suspension in the flowing water, especially at higher velocities. As you know from intuition and from experience, streams that flow fast tend to be turbulent (flow paths are chaotic and the water surface appears rough) and the water may be muddy, while those that flow more slowly tend to have laminar flow (straight-line flow and a smooth water surface) and clear water. Turbulent flow is more effective than laminar flow at keeping sediments in suspension. Stream water also has a dissolved load, which represents (on average) about 15% of the mass of material transported, and includes ions such as calcium (Ca+2) and chloride (Cl-) in solution. The solubility of these ions is not affected by flow velocity.

The faster the water is flowing, the larger the particles that can be kept in suspension and transported within the flowing water. However, as Swedish geographer Filip Hjulström discovered in the 1940s, the relationship between grain size and the likelihood of a grain being eroded, transported, or deposited is not as simple as one might imagine (Figure 13.16). Consider, for example, a 1 mm grain of sand. If it is resting on the bottom, it will remain there until the velocity is high enough to erode it, around 20 cm/s. But once it is in suspension, that same 1 mm particle will remain in suspension as long as the velocity doesn’t drop below 10 cm/s. For a 10 mm gravel grain, the velocity is 105 cm/s to be eroded from the bed but only 80 cm/s to remain in suspension.

On the other hand, a 0.01 mm silt particle only needs a velocity of 0.1 cm/s to remain in suspension, but requires 60 cm/s to be eroded. In other words, a tiny silt grain requires a greater velocity to be eroded than a grain of sand that is 100 times larger! For clay-sized particles, the discrepancy is even greater. In a stream, the most easily eroded particles are small sand grains between 0.2 mm and 0.5 mm. Anything smaller or larger requires a higher water velocity to be eroded and entrained in the flow. The main reason for this is that small particles, and especially the tiny grains of clay, have a strong tendency to stick together, and so are difficult to erode from the stream bed. 363 Chapter 13 Streams and Floods It is important to be aware that a stream can both erode and deposit sediments at the same time. At 100 cm/s, for example, silt, sand, and medium gravel will be eroded from the stream bed and transported in suspension, coarse gravel will be held in suspension, pebbles will be both transported and deposited, and cobbles and boulders will remain stationary on the stream bed. A stream typically reaches its greatest velocity when it is close to flooding over its banks. This is known as the bank-full stage, as shown in Figure 13.17. As soon as the flooding stream overtops its banks and occupies the wide area of its flood plain, the water has a much larger area to flow through and the velocity drops significantly. At this point, sediment that was being carried by the high-velocity water is deposited near the edge of the channel, forming a natural bank or levée.

Figure 13.17 The development of natural levées during flooding of a stream. The sediments of the levée become increasingly fine away from the stream channel, and even finer sediments — clay, silt, and fine sand — are deposited across most of the flood plain. [SE]

Stream channels can be straight or curved, deep and slow, or rapid and choked with coarse sediments. The cycle of erosion has some influence on the nature of a stream, but there are several other factors that are important. Youthful streams that are actively down-cutting their channels tend to be relatively straight and are typically ungraded (meaning that rapids and falls are common). As shown in Figures 13.1 and 13.18, youthful streams commonly have a step-pool morphology, meaning that the stream consists of a series of pools connected by rapids and waterfalls. They also have steep gradients and steep and narrow V- shaped valleys — in some cases steep enough to be called canyons.

In mountainous terrain, such as that in western Alberta and B.C., steep youthful streams typically flow into wide and relatively low-gradient U-shaped glaciated valleys. The youthful streams have high sediment loads, and when they flow into the lower-gradient glacial valleys where the velocity isn’t high enough to carry all of the sediment, braided patterns develop, characterized by a series of narrow channels separated by gravel bars (Figure 13.19).

Braided streams can develop anywhere there is more sediment than a stream is able to transport. One such environment is in volcanic regions, where explosive eruptions produce large amounts of unconsolidated material that gets washed into streams. The Coldwater River next to Mt. St. Helens in Washington State is a good example of this (Figure 13.20).

A stream that occupies a wide, flat flood plain with a low gradient typically carries only sand-sized and finer sediments and develops a sinuous flow pattern. As you saw in Figure 13.14, when a stream flows around a corner, the water on the outside has farther to go and tends to flow faster. This leads to erosion of the banks on the outside of the curve, deposition on the inside, and formation of a point bar (Figure 13.21). Over time, the sinuosity of the stream becomes increasingly exaggerated, and the channel migrates around within its flood plain, forming a meandering pattern.

Figure 13.21 The meandering channel of the Bonnell Creek, Nanoose, B.C. The stream is flowing toward the viewer. The sand and gravel point bar must have formed when the creek was higher and the flow faster than it was when the photo was taken. [SE] A well-developed meandering river is shown in Figure 13.22. The meander in the middle of the photo has reached the point where the thin neck of land between two parts of the channel is about to be eroded through. When this happens, another oxbow lake will form like the others in the photo. Figure 13.22 The meandering channel of the Nowitna River, Alaska. Numerous oxbow lakes are present and another meander cutoff will soon take place. [Oliver Kumis, http://commons.wikimedia.org/wiki/File:Nowitna_river.jpg]

At the point where a stream enters a still body of water — a lake or the ocean — sediment is deposited and a delta forms. The Fraser River has created a large delta, which extends out into the Strait of Georgia (Figure 13.23). Much of the Fraser delta is very young in geological terms. Shortly after the end of the last glaciation (10,000 years ago), the delta did not extend past New Westminster. Since that time, all of the land that makes up Richmond, Delta, and parts of New Westminster and south Surrey has formed from sediment from the Fraser River. (You can see this in more detail at Geoscape Vancouver .)

Figure 13.23 The delta of the Fraser River and the plume of sediment that extends across the Strait of Georgia. The land outlined in red has formed over the past 10,000 years. [September 2011, SE after NASA:

The discharge levels of streams are highly variable depending on the time of year and on specific variations in the weather from one year to the next. In Canada, most streams show discharge variability similar to that of the Stikine River in northwestern B.C., as illustrated in Figure 13.24. The Stikine River has its lowest discharge levels in the depths of winter when freezing conditions persist throughout most of its drainage basin. Discharge starts to rise slowly in May, and then rises dramatically through the late spring and early summer as a winter’s worth of snow melts. For the year shown, the minimum discharge on the Stikine River was 56 m3/s in March, and the maximum was 37 times higher, 2,470 m3/s, in May.

Streams in coastal areas of southern British Columbia show a very different pattern from those in most of the rest of the country because their drainage basins do not remain entirely frozen and because they receive a lot of rain (rather than snow) during the winter. The Qualicum River on Vancouver Island typically has its highest discharge levels in January or February and its lowest levels in late summer (Figure 13.25). In 2013, the minimum discharge was 1.6 m3/s, in August, and the maximum was 34 times higher, 53 m3/s, in March.

When a stream’s discharge increases, both the water level (stage) and the velocity increase as well. Rapidly flowing streams become muddy and large volumes of sediment are transported both in suspension and along the stream bed. In extreme situations, the water level reaches the top of the stream’s banks (the bank-full stage, see Figure 13.17), and if it rises any more, it floods the surrounding 370 371 Chapter 13 Streams and Floods terrain. In the case of mature or old-age streams, this could include a vast area of relatively flat ground known as a flood plain, which is the area that is typically covered with water during a major flood. Because fine river sediments are deposited on flood plains, they are ideally suited for agriculture, and thus are typically occupied by farms and residences, and in many cases, by towns or cities. Such infrastructure is highly vulnerable to damage from flooding, and the people that live and work there are at risk. Most streams in Canada have the greatest risk of flooding in the late spring and early summer when stream discharges rise in response to melting snow. In some cases, this is exacerbated by spring storms. In years when melting is especially fast and/or spring storms are particularly intense, flooding can be very severe. One of the worst floods in Canadian history took place in the Fraser Valley in late May and early June of 1948. The early spring of that year had been cold, and a large snow pack in the interior was slow to melt. In mid-May, temperatures rose quickly and melting was accelerated by rainfall. Fraser River discharge levels rose rapidly over several days during late May, and the dykes built to protect the valley were breached in a dozen places. Approximately one-third of the flood plain was inundated and many homes and other buildings were destroyed, but there were no deaths. The Fraser River flood of 1948, which was the highest in the past century, was followed by very high river levels in 1950 and 1972 and by relatively high levels several times since then, the most recent being 2007 (Table 13.1). In the years following 1948, millions of dollars were spent repairing and raising the existing dykes and building new ones; since then damage from flooding in the Fraser Valley has been relatively limited.

Physical Geology 372 Table13.1 Ranking of the maximum stage and discharge values for the Fraser River at Hope between 1948 and 2008. Typical discharge levels are around 1,000 m3/s. [From date in Mannerstrom, 20082] Serious flooding happened in July in 1996 in the Saguenay-Lac St. Jean region of Quebec. In this case, the floods were caused by two weeks of heavy rainfall followed by one day of exceptional rainfall. July 19 saw 270 mm of rain, equivalent to the region’s normal rainfall for the entire month of July. Ten deaths were attributed to the Saguenay floods, and the economic toll was estimated at $1.5 billion. Just a year after the Saguenay floods, the Red River in Minnesota, North Dakota, and Manitoba reached its highest level since 1826. As is typical for the Red River, the 1997 flooding was due to rapid snowmelt. Because of the south to north flow of the river, the flooding starts in Minnesota and North Dakota, where melting starts earlier, and builds toward the north. The residents of Manitoba had plenty of warning that the 1997 flood was coming because there was severe flooding at several locations on the U.S. side of the border. After the 1950 Red River flood, the Manitoba government built a channel around the city of Winnipeg to reduce the potential of flooding in the city (Figure 13.26). Known as the Red River Floodway, the channel was completed in 1964 at a cost of $63 million. Since then it has been used many times to alleviate flooding in Winnipeg, and is estimated to have saved many billions of dollars in flood damage. The massive 1997 flood was almost too much for the floodway; in fact the amount of water diverted was greater than the designed capacity. The floodway has recently been expanded so that it can be used to divert more of the Red River’s flow away from Winnipeg. Figure 13.26 Map of the Red River Floodway around Winnipeg, Manitoba (left), and aerial view of the southern (inlet) end of the floodway (right). [Map from 1997_Red_River_Flood#/media/ File:Rednorthfloodwaymap.png and photo from Natural Resources Canada 2012, courtesy of the Geological Survey of Canada (Photo 2000-118 by G.R. Brooks).] Canada’s most costly flood ever was the June 2013 flood in southern Alberta. The flooding was initiated by snowmelt and worsened by heavy rains in the Rockies due to an anomalous flow of moist air from the Pacific and the Caribbean. At Canmore, rainfall amounts exceeded 200 mm in 36 hours, and at High River, 325 mm of rain fell in 48 hours. Mannerström, M, 2008, Comprehensive Review of Fraser River at Hope Flood Hydrology and Flows Scoping Study, Report prepared for the B.C. Ministry of the Environment. Available at: review_fraser_flood_flows_hope.pdf

In late June and early July, the discharges of several rivers in the area, including the Bow River in Banff, Canmore, and Exshaw, the Bow and Elbow Rivers in Calgary, the Sheep River in Okotoks, and the Highwood River in High River, reached levels that were 5 to 10 times higher than normal for the time of year (see Exercise 13.5). Large areas of Calgary, Okotoks, and High River were flooded and five people died (see Figures 13.27 and 13.28). The cost of the 2013 flood is estimated to be approximately $5 billion. For more about Alberta’s flood of the century, visit: . Figure 13.28 Flooding in Calgary (June 21, left) and Okotoks (June 20, right) during the 2013 southern Alberta flood wikipedia/commons/6/6a/ Riverfront_Ave_Calgary_Flood_2013.jpg Okotoks_-_June_20%2C_2013_-_Flood_waters_in_ local_campground_playground-03.JPG]

The largest flood recorded on the Bow River over that period was the one in 2013, 1,840 m3/s on June 21. Ri for that flood is (95+1)/1 = 96 years. The probability of such a flood in any future year is 1/Ri, which is 1%. The fifth largest flood was just a few years earlier in 2005, at 791 m3/s. Ri for that flood is (95+1)/5 = 19.2 years. The recurrence probability is 5%. Calculate the recurrence interval for the second largest flood (1932, 1,520 m3/s). What is the probability that a flood of 1,520 m3/s will happen next year? Examine the 100-year trend for floods on the Bow River. If you ignore the major floods (the labelled ones), what is the general trend of peak discharges over that time?

One of the things that the 2013 flood on the Bow River teaches us is that we can’t predict when a flood will occur or how big it will be, so in order to minimize damage and casualties we need to be prepared. Some of the ways of doing that are as follows: Mapping flood plains and not building within them Building dykes or dams where necessary Monitoring the winter snowpack, the weather, and stream discharges Creating emergency plans Educating the public

Why do many of the streams in the southwestern part of Vancouver Island flow to the ocean as waterfalls? Where would you expect to find the fastest water flow on a straight stretch of a stream? Sand grains can be moved by traction and saltation. What minimum stream velocities might be required to move 1 mm sand grains? If the flow velocity of a stream is 1 cm/s, what sizes of particles can be eroded, what sizes can be transported if they are already in suspension, and what sizes of particles cannot be moved at all? Under what circumstances might a braided stream develop? How would the gradient of a stream be affected if a meander is cut off? The elevation of the Fraser River at Hope is 41 m. From there it flows approximately 147 km to the sea. What is the average gradient of the river (m/km) over that distance? How do B.C.’s coastal streams differ from most of the rest of the streams in Canada in terms of their annual flow patterns? Why? Why do most serious floods in Canada happen in late May, June, or early July? There is a 65-year record of peak annual discharges on the Ashnola River near Princeton, B.C. During this time, the second highest discharge was 175 m3/s. Based on this information, what is the recurrence interval (Ri) for that discharge level, and what is the probability that there will be a similar peak discharge next year?

Explain the concepts of porosity and permeability and the importance of these to groundwater storage and movement Describe the relative porosities and permeabilities of some common geological materials Define aquifers, aquitards, confining layers, and the differences between confined and unconfined aquifers Explain the concepts of hydraulic head, the water table, potentiometric surface, and hydraulic gradient, and apply the Darcy equation for estimating groundwater flow Describe the flow of groundwater from recharge areas to discharge areas Describe the nature of groundwater flow in karst systems Explain how wells are used to extract groundwater and the implications of over-pumping a well Describe how observation wells are used to monitor groundwater levels and the importance of protecting groundwater resources Distinguish between natural and anthropogenic contamination of groundwater Describe some of the ways that groundwater can become contaminated, and how contamination can be minimized

Physical Geology 378 As we saw in Chapter 13, fresh water makes up only 3% of the water on Earth. Approximately two- thirds of that is glacial ice and most of the rest is groundwater. We can’t live without water, and it’s easy to see that groundwater represents a critically important component of our water supply. Groundwater is not as easily accessed as surface water, but it is also not as easily contaminated as surface water. If more than 7 billion of us want to continue living comfortably here on Earth, we have to take great care of our groundwater and learn how to use it sustainably.

Groundwater is stored in the open spaces within rocks and within unconsolidated sediments. Rocks and sediments near the surface are under less pressure than those at significant depth and therefore tend to have more open space. For this reason, and because it’s expensive to drill deep wells, most of the groundwater that is accessed by individual users is within the first 100 m of the surface. Some municipal, agricultural, and industrial groundwater users get their water from greater depth, but deeper groundwater tends to be of lower quality than shallow groundwater, so there is a limit as to how deep we can go. Porosity is the percentage of open space within an unconsolidated sediment or a rock. Primary porosity is represented by the spaces between grains in a sediment or sedimentary rock. Secondary porosity is porosity that has developed after the rock has formed. It can include fracture porosity — space within fractures in any kind of rock. Some volcanic rock has a special type of porosity related to vesicles, and some limestone has extra porosity related to cavities within fossils. Porosity is expressed as a percentage calculated from the volume of open space in a rock compared with the total volume of rock. The typical ranges in porosity of a number of different geological materials are shown in Figure 14.2. Unconsolidated sediments tend to have higher porosity than consolidated ones because they have no cement, and most have not been strongly compressed. Finer-grained materials (e.g., silt and clay) tend to have greater porosity — some as high as 70% — than coarser materials (e.g., gravel). Primary porosity tends to be higher in well-sorted sediments compared to poorly sorted sediments, where there is a range of smaller particles to fill the spaces made by the larger particles. Glacial till, which has a wide range of grain sizes and is typically formed under compression beneath glacial ice, has relatively low porosity. Consolidation and cementation during the process of lithification of unconsolidated sediments into sedimentary rocks reduces primary porosity. Sedimentary rocks generally have porosities in the range of 10% to 30%, some of which may be secondary (fracture) porosity. The grain size, sorting, compaction, and degree of cementation of the rocks all influence primary porosity. For example, poorly sorted and well-cemented sandstone and well-compressed mudstone can have very low porosity. Igneous or metamorphic rocks have the lowest primary porosity because they commonly form at depth and have interlocking crystals. Most of their porosity comes in the form of secondary porosity in fractures. Of the consolidated rocks, well-fractured volcanic rocks and limestone that has cavernous openings produced by dissolution have the highest potential porosity, while intrusive igneous and metamorphic rocks, which formed under great pressure, have the lowest.

Porosity is a measure of how much water can be stored in geological materials. Almost all rocks contain some porosity and therefore contain groundwater. Groundwater is found under your feet and everywhere on the planet. Considering that sedimentary rocks and unconsolidated sediments cover about 75% of the continental crust with an average thickness of a few hundred metres, and that they are likely to have around 20% porosity on average, it is easy to see that a huge volume of water can be stored in the ground. Porosity is a description of how much space there could be to hold water under the ground, and permeability describes how those pores are shaped and interconnected. This determines how easy it is for water to flow from one pore to the next. Larger pores mean there is less friction between flowing water and the sides of the pores. Smaller pores mean more friction along pore walls, but also more twists and turns for the water to have to flow-through. A permeable material has a greater number of larger, well-connected pores spaces, whereas an impermeable material has fewer, smaller pores that are poorly connected. Permeability is the most important variable in groundwater. Permeability describes how easily water can flow through the rock or unconsolidated sediment and how easy it will be to extract the water for our purposes. The characteristic of permeability of a geological material is quantified by geoscientists and engineers using a number of different units, but the most common is the hydraulic conductivity. The symbol used for hydraulic conductivity is K. Although hydraulic conductivity can be expressed in a range of different units, in this book, we will always use m/s. The materials in Figure 14.3 show that there is a wide range of permeability in geological materials from 10-12 m/s (0.000000000001 m/s) to around 1 m/s. Unconsolidated materials are generally more permeable than the corresponding rocks (compare sand with sandstone, for example), and the coarser materials are much more permeable than the finer ones. The least permeable rocks are unfractured intrusive igneous and metamorphic rocks, followed by unfractured mudstone, sandstone, and limestone. The permeability of sandstone can vary widely depending on the degree of sorting and the amount of cement that is present. Fractured igneous and metamorphic rocks, and especially fractured volcanic rocks, can be highly permeable, as can limestone that has been dissolved along fractures and bedding planes to create solutional openings.

The surface of most silicate mineral grains has a slight negative charge due to imperfections in the mineral structure. Water (H2O) is a polar molecule. This means that while it has no overall electrical charge, one side of the molecule has a slight positive charge (the side with the two hydrogens), compared to a slight negative charge on the other side. Water is strongly attracted to all mineral grains and water within that bound water layer (a few microns around each grain) is not able to move and flow along with the rest of the groundwater. In the lower diagrams shown here, the bound water is represented by dark blue lines around each grain and the water that can move is light blue. In the sand, there is still a lot of water that is able to move through the sediment, but in the clay/silt almost all of the water is held tightly to the grains and this reduces the permeability. [SE] We have now seen that there is a wide range of porosity in geological materials and an even wider range of permeability. Groundwater exists everywhere there is porosity. However, whether that groundwater Physical Geology 382 is able to flow in significant quantities depends on the permeability. An aquifer is defined as a body of rock or unconsolidated sediment that has sufficient permeability to allow water to flow through it. Unconsolidated materials like gravel, sand, and even silt make relatively good aquifers, as do rocks like sandstone. Other rocks can be good aquifers if they are well fractured. An aquitard is a body that does not allow transmission of a significant amount of water, such as a clay, a till, or a poorly fractured igneous or metamorphic rock. These are relative terms, not absolute, and are usually defined based on someone’s desire to pump groundwater; what is an aquifer to someone who does not need a lot of water, may be an aquitard to someone else who does. An aquifer that is exposed at the ground surface is called an unconfined aquifer. An aquifer where there is a lower permeability material between the aquifer and the ground surface is known as a confined aquifer, and the aquitard separating ground surface and the aquifer is known as the confining layer. Figure 14.4 shows a cross-section of a series of rocks and unconsolidated materials, some of which might serve as aquifers and others as aquitards or confining layers. The granite is much less permeable than the other materials, and so is an aquitard in this context. The yellow layer is very permeable and would make an ideal aquifer. The overlying grey layer is a confining layer. The upper buff-coloured layer (K = 10-2 m/s) does not have a confining layer and is an unconfined aquifer. The yellow layer (K = 10-1 m/s) is “confined” by the confining layer (K = 10-4 m/s), and is a confined aquifer. The confined aquifer gets most of its water from the upper part of the hill where it is exposed at the surface, and relatively little by seepage through the fine silt layer. Figure 14.4 A cross-section showing materials that might serve as aquifers and confining layers. The relative permeabilities are denoted by hydraulic conductivity (K = m/s). The pink rock is granite; the other layers are various sedimentary layers. [SE]

If you go out into your garden or into a forest or a park and start digging, you will find that the soil is moist (unless you’re in a desert), but it’s not saturated with water. This means that some of the pore space in the soil is occupied by water, and some of the pore space is occupied by air (unless you’re in a swamp). This is known as the unsaturated zone. If you could dig down far enough, you would get to the point where all of the pore spaces are 100% filled with water (saturated) and the bottom of your hole would fill up with water. The level of water in the hole represents the water table, which is the surface of the saturated zone. In most parts of British Columbia, the water table is several metres below the surface. Water falling on the ground surface as precipitation (rain, snow, hail, fog, etc.) may flow off a hill slope directly to a stream in the form of runoff, or it may infiltrate the ground, where it is stored in the unsaturated zone. The water in the unsaturated zone may be used by plants (transpiration), evaporate from the soil (evaporation), or continue past the root zone and flow downward to the water table, where it recharges the groundwater. A cross-section of a typical hillside with an unconfined aquifer is illustrated in Figure 14.5. In areas with topographic relief, the water table generally follows the land surface, but tends to come closer to surface in valleys, and intersects the surface where there are streams or lakes. The water table can be determined from the depth of water in a well that isn’t being pumped, although, as described below, that only applies if the well is within an unconfined aquifer. In this case, most of the hillside forms the recharge area, where water from precipitation flows downward through the unsaturated zone to reach the water table. The area at the stream or lake to which the groundwater is flowing is a discharge area. What makes water flow from the recharge areas to the discharge areas? Recall that water is flowing in pores where there is friction, which means it takes work to move the water. There is also some friction between water molecules themselves, which is determined by the viscosity. Water has a low viscosity, but friction is still a factor. All flowing fluids are always losing energy to friction with their surroundings. Water will flow from areas with high energy to those with low energy. Recharge areas are at higher elevations, where the water has high gravitational energy. It was energy from the sun that evaporated the water into the atmosphere and lifted it up to the recharge area. The water loses this gravitational energy as it flows from the recharge area to the discharge area. In Figure 14.5, the water table is sloping; that slope represents the change in gravitational potential energy of the water at the water table. The water table is higher under the recharge area (90 m) and lower at the discharge area (82 m). Imagine how much work it would be to lift water 8 m high in the air. That is the energy that was lost to friction as the groundwater flowed from the top of the hill to the stream.

The situation gets a lot more complicated in the case of confined aquifers, but they are important sources of water so we need to understand how they work. As shown in Figure 14.6, there is always a water table, and that applies even if the geological materials at the surface have very low permeability. Where there is a confined aquifer — meaning one that is separated from the surface by a confining layer — this aquifer will have its own “water table,” which is actually called a potentiometric surface, as it is a measure of the total potential energy of the water. The red dashed line in Figure 14.6 is the potentiometric surface for the confined aquifer, and it describes the total energy that water is under within the confined aquifer. If we drill a well into the unconfined aquifer, the water will rise to the level of the water table (well A in Figure 14.6). But if we drill a well through both the unconfined aquifer and the confining layer and into the confined aquifer, the water will rise above the top of the confined aquifer to the level of its potentiometric surface (well B in Figure 14.6). This is known as an artesian well, because the water rises above the top of the aquifer. In some situations, the potentiometric surface may be above the ground level. The water in a well drilled into the confined aquifer in this situation would rise above ground level, and flow out, if it’s not capped (well C in Figure 14.6). This is known as a flowing artesian well.