The major cause of property loss and death from natural disasters is associated with floods resulting from severe weather.

Having considered the long-term properties of the climate system, and the gradual changes that have affected the Earth, we will now focus on the short-term processes associated with weather. Driven by temperature-induced air pressure variations, winds blow on the surface of the Earth, with a great variation in force. These winds transport both dust particles, providing a source of both damaging dust storms and beneficial windblown sedimentary deposits, and water, resulting in storms and hurricanes as well as beneficial rainfall. In a sense, we must take the bad with the good, as weather circulation plays a key role in the critical water cycle at the surface, without which we could not exist.

First we begin by considering what is wind? The working definition for us is horizontal motions of the atmosphere driven by atmospheric pressure differences. Winds blow from high pressure regions to low pressure regions. The strength of the pressure differences determines the wind speeds, which have a variety of categories:

2 1 light breeze

10 5 moderate breeze

20 10 fresh wind

30 15 gale

40 20 strong gale

50 25 storm

70 >33 hurricane

The familiar names reflect the fact that wind is highly variable, or gusty, with turbulent flow. This is very common experience for all of us. Some of the types of winds include daily sea breezes, driven by differential pressures over land and sea associated with daily heating and evening cooling. There are storm updrafts and downdrafts, notably associated with thunder showers. Tornadoes are intense localized wind cylinders with winds of 150-300 mph. There are drainage winds, or catabatic winds, which involve heavy cool air driven down off of mountain sides or glaciers. In the Alps these are called Foehn, while in the Rockies these are Chinook winds. Winds driven in off of desert regions are called Santa Annas, haboobs and harmatans. The diversity of wind forms reflects the complex factors influencing wind pressures such as soil types, geography, diurnal heating patterns etc.

Wind varies continuously at any one location, as the complex dynamic system of the atmosphere is rich in all scales of motion from very local to global patterns. The variability of winds is recorded by anemometers, or wind meters in various locations. Prevailing wind patterns associated with large-scale atmospheric motions interact with local conditions to drive the specific wind history at any point, such as the fluctuating sea breezes along the Monterey Bay. This is also true of regions with recurring weather patterns, such as tornadoes.

Tornadoes arise from east ward moving high winds overriding northwestward moving lower winds, which sets up strong spinning circulations that can spin-up and establish stable cyclonic motions. The central midwest experiences many tornadoes in a tornado alley associated with the jet stream interactions with winds off of the Gulf of Mexico. While individual tornadoes are not uncommon, large outbreaks of hoards of tornadoes occur about every 50 years in the mid-west to eastern U.S., strung out along trajectories that track the jet stream. While the conditions favoring the origin of tornadoes are well understood, the specific behavior of localized wind conditions is difficult to predict accurately, and great emphasis is on rapid detection by Doppler radar of the strong wind shears that accompany tornado conditions.

Winds and pressure differences can be directly destructive as in the case of tornadoes, but the power of wind is greatly enhanced by the presence of dust and water suspended by the atmospheric turbulence. Sediment is transported by wind with a weight dependence due to the resisting effects of gravity. Wind drags along sediments, either by saltation or bouncing along the surface or by suspension. Big particles, such as sand grains tend to stay low in the atmosphere, driven by drag of the passing wind. Turbulence of the wind is not very important, and the sand grains transport only moderate distances of 1 cm to 1 m, mainly by saltation. When there are large concentrations of sand, wind can sweep into formations called dunes, with a elongate side in the upwind direction called the stoss, and a steep downwind side called the lee. Sand particles move up the stoss by saltation, reaching the brink, and then avalanching down the lee face. The angle of repose is about 31 degrees for typical sands, and dipping layers build outward on the lee side. Effectively, the dune moves downwind, with the cross sets parallel to the lee face defining rock layers if the dune is buried and cemented. Dunes are often important in the rock record because they indicate past wind directions and climatic conditions, they often are sources of water or oil, and modern dunes are great places to play on. Small particles, such as silt and clay are able to transport much longer distances, effectively being suspended in the air by turbulence, which overcomes the resisting gravity. When the winds die, the fine materials settle out and leave deposits called Loess, which is often very rich agriculturally.

The processes of moving sediment by wind have thresholds, where wind speeds must exceed certain levels in order to get particles of a given size to move. This places great importance on rare large wind events, as these are the cases where nonlinear thresholds trigger substantial events. An example is the December 20, 1977 San Joaquin valley dust storm, which was prompted by a very high pressure system sitting over Idaho and Utah, driving winds westward toward and off-shore low pressure region. The winds blew down valleys trending westward from the Sierra Nevada, and achieved speeds of 200 mph. This sufficed to entrain huge amounts of sand and dust into the wind, causing loss of visibility on Interstate 5 and huge car accidents and fatalities. Massive erosion of top soil resulted as well, due to the lack of ground cover resulting from agricultural practices.

While wind-borne dust constitutes an important and sometimes deadly geological agent, far more important is the water carried by wind. Transport through the atmosphere is one of the most important stages of the water cycle. If we consider where water is at any given time in the water cycle, some 97.6% is in the oceans, 1.9% is ephemerally locked up in ice caps and glaciers, 0.5% is in ground and soil water, 0.02% is in rivers and lakes, and 0.0001% is in the atmosphere. But large quantities cycle through the atmosphere, feeding into each of the other reservoirs. The water transported by the atmosphere is of great importance to humans, and when there are excess rainfalls that cause floods, this water becomes one of the deadliest of natural catastrophes. The conditions at the surface of the Earth are such that water can exist in all three phases (gas vapor, liquid and frozen solid), and the fluid is quite dense and chemically reactive, which makes it an effective erosional agent both mechanically and chemically. Water has a low viscosity, which enables its rapid transport on the surface as well, which enhances its role in erosion, sediment deposition, and ground penetration.

The water cycle involves evaporation, precipitation, runoff, ice storage, groundwater, and transpiration from plants. Long-term climatic conditions can greatly modify the water cycle acting at any given time, as do global atmospheric patterns. For example, the latitudinal variation in annual precipitation shows an equatorial peak of about 80 inches/yr, minima of 40 inches/yr at 30 degrees north and south (the limits of the Hadley cells at low latitudes) and little rainfall at the poles. Within the latitudinal structure there is great variability of weather and rainfall, strongly influenced by geography and wind patterns. In the U.S., some of the most common weather patterns include thunderstorms, which pose hazards due to lightning (responsible for 100-150 deaths/yr, about the same as tornadoes), hail and flashfloods. An example of the latter is the Big Thompson Flood in the Colorado Rockies, which took place on July 31, 1976. Beginning at 8:30 PM, there was 10 inches of rain in 90 minutes, which lead to a swollen river that flooded highway 134, drowning 139 people. Flashfloods tend to be most intense in arid regions where runoff quickly gathers into canyons and other drainages. The primary region of U.S. thunderstorms is in the southeast, where 50-100 occur each year, while fewer than 5 occur annually in California. Hailstorms are concentrated in the Rock Mountains, where strong upwelling currents interact with cold air to produce large frozen aggregates. 4-8 occur annually in Colorado and nearby states.

Even larger storm systems tend to spawn in low latitudes where warm tropical oceans provide large water evaporation. These include hurricanes, also known as cyclones, typhoons, and willywillies. Hurricanes originate where surface ocean temperatures are more than 27 degrees C, which tends to limit them to the late summer season. These storms are large enough that corriolis forces play a significant role, so they 'spin' clockwise in the northern hemisphere (tracking a squall line in toward the low pressure center) and counterclockwise in the southern hemisphere. Winds in excess of 70 mph define a hurricane, with speeds of 150 mph characterizing the largest storms. Hurricanes are notable for their intense, damaging winds, the high rate and duration of rainfall that accompanies them, and the intense storm surge that is caused by the low pressure region at the eye. Yearly storms generate near the equatorial Atlantic, feeding into the Caribbean and southeastern U.S., in the northern Indian Ocean feeding into southern Asia, and in the far western Pacific, feeding into southeast Asia, Philippines, China and Japan.

Massive U.S. losses have resulted from hurricanes. Hurricane Andrew caused about $30 billion in damage, while Hugo caused over $10 billion. This is comparable to the damage from large earthquakes such as the 1989 Loma Prieta event ($9 billion), and the 1994 Northridge event ($30 billion). The 1900 hurricane which hit Galveston caused 6000 deaths, but warning systems have kept loss of like down to several hundred for each of the more recent events this century.

The large storm 'footprint' associated with a hurricane, involves the large area over which rain is delivered to the surface. Hurricanes can deposit huge volumes of water in short times, which quickly overload the intrinsic drainage capabilities of the landscape, resulting in flooding. Basically, there is a balance between rainfall input, infiltration as groundwater and runoff. Runoff occurs whenever rainfall rate exceeds infiltration rate, which is common for hurricanes. For example, hurricane Agnes in 1972 dumped 28 cm of rain in 18 hours over 93,000 square kilometers. The water dumped on a landscape transports through a network of streams, feeding from first order streams to higher and higher order streams connecting throughout a given drainage basin. The distance and character of each hillslope length, leading to the nearest stream channel influences how quickly runoff gets into the stream, where it can transport quite efficiently. For example, a paved slope causes very rapid transport of water over the surface, while a vegetated slope allows water to move only slowly through a tortuous path. The response of a drainage system is characterized by a flood hydrograph, which is a plot of time history over which a given input from a storm drains out of a system. This will give a spread out curve for the entire drainage basin, the shape of which determines whether flooding takes place.

For flooding, there are important characteristics of the storm as well as of the landscape: Storm Characteristics

1. rainfall intensity

2. rainfall duration

3. footprint of the storm (how much/many drainage basins are covered)

Landscape Characteristics

1. vegetation

2. infiltration rate (soil type)

3. pore space available; soil thickness, antecedent rain

4. hillslope lengths (distance for overland transport to stream)

Every channel experiences fluctuations in water level, which results in some characteristic stream morphologies. The active river channel is the low region in a river bed, usually banked by course levees, and surrounded by a river floodplain. If the region is uplifting, or has previously had higher water levels there may be fluvial terraces of higher level floodplains that are now abandoned. It is important to recognize that the floodplain is intrinsically part of the river during floods that overbank the levees, which may happen as often as every 2-3 years. Sediment is transported in the river, with analogous behavior to sediment in the atmosphere. Finer silts are suspended and travel in the turbulent flow for long distances, while coarser materials travel as bedload, over short distances with threshold behavior. When floods occur, coarse materials are dropped out near the levees, while silt and clay are deposited out on the floodplain, which makes these rich agricultural areas.

We can try to characterize river flood behavior based on the time between different levels of peak water discharge. This reflects the regional storm statistics, and gives general probabilities upon which to base mitigation efforts. If we plot peak annual discharge versus the logarithm of time between recurring levels of comparable height, a linear curve is found for most streams, with the bankfull stage (maximum channel capacity) typically being achieved every 2-3 years. Higher levels occur less frequently, and we can extrapolate the curve to estimate how large the flow will be for rare events like the 100 year flow. Regions with different flood hydrographs display different behavior for the recurrence times.

Some regions, such as Bangladesh are particularly prone to flooding disasters. This country is very low elevation, and encompasses the Ganghes-Brahmaputra delta. It is densely populated, with about 100 million people, and there are massive cyclones every other year (60 killer events over the past 120 years). The lack of high ground, high population density, and lack of warning systems add to the human toll. On November 13, 1970 1 million people died in a massive flood from monsoon rains. This recurs time and again. In the U.S. flooding disasters have been far less fatal, with the highest toll being 600 lives lost in the 1938 floods in New England. Modern warning systems, more favorable geography, and less common massive storm occurrence reduce the risk in the U.S..

Floods arising from heavy rainfall are regular events, but the greatest floods have involved collapses of natural dams and the draining of lakes in truly catastrophic floods unlike any witnessed in modern human history.

Natural dam failures: Natural dams arise due to blockage of rivers by rock slides, glacial ice, and geomorphology. Accumulation of water behind the dam in lakes can lead to failure of the dam, resulting in sudden release of the water. The character of the resulting flood is quantified by the hydrograph of the flow (the plot of discharge as a function of time), just as for other river floods. The U.S. has been the site of several great flooding events in recent geological history (the last 20,000 years), involving the abrupt drainage of lakes.

The first case history we will consider is the Great Salt Lake. This lake is in a totally enclosed basin (all rivers lead in, none lead out), which is the primary reason for the high salinity. The regional climate has lead to substantial changes in lake level over the past 200 years. In particular, from 1860-1870 there was a 10 foot increase in lake level that coincided with the Mormon settlements in the area. This extended the lake significantly, as is part of the enduring myth of the westward expansion that "water follows the plot", which erroneously attributed cultivation to increases in water abundance. In fact, climate fluctuations can be influenced by agricultural practices, but there are many other causes of local variations. Recent wet years have led to repeated flooding of the lake, and plans to develop a drainage by tunneling through surrounding hillsides into lower valleys.

The Great Salt Lake is a small residual of a much larger lake that once covered about 1/3 of the state of Utah. That lake was called Lake Bonneville, and it resided in the eastern portion of the Great Basin. Lake Bonneville had a surface area of 51,000 square kilometers at its peak, 300 km by 165 km in size. It too was a trapped lake, with all drainages leading in. How do we know that this lake existed? The evidence is provided by the bathtub ring of shorelines that are beveled into the hillsides that once bounded the lake. Shoreline deposits and eroded terraces define the high water level of Lake Bonneville and the lower shoreline of a subsequent lake call Lake Provo. By dating the rock deposits around the lake, using fossils and other means, the time history of the lake elevation has been charted out over the past 30,000 years. The record is somewhat uncertain for times prior to the highest lake level, as those earlier deposits tend to be overridden by the lake maximum, but from 30,000-15,000 years ago there was a gradual 300 m increase in the lake level as Lake Bonneville grew. This corresponds to the time during which the North American ice pack was near its maximum (20,000) years ago, and then receded. The waters for Lake Bonneville were not glacial melt, as the Great Basin was too far south to be fed from the ice pack, but the microclimate was much wetter as a result of the regional weather patterns which were influenced by the ice pack.

Around 14,500 years ago the lake level was at its peak, at an altitude of 1550m, and then it abruptly dropped by 100m, moving from the Bonneville shoreline down to the Provo level, in an event so rapid that there is no intervening shoreline terraced into the hillsides. This event is now recognized as the Bonneville Flood, the result of catastrophic failure of one of the natural dams that had framed the maximum lake level, with the flood waters involving about 5000 cubic kilometers of water! (100m x 50,000 km2) It is estimated that it took about 20 days for most of the water to squirt out, indicating about 1,000,000 cubic m/sec as a flow rate. In contrast, floods in the Colorado river have involved about 1000 cubic m/sec, so this flood was about 1,000 time larger in terms of water flow. Subsequently, more arid climatic conditions have systematically lowered the lake levels, leading to today's residual in the Great Salt Lake.

The point at which the lake walls failed is at the north end of Lake Bonneville, where a saddleback called Red rock pass held in the lake water. This region was overtopped at the high point in the history of Lake Bonneville, and began to erode away, with the geometry allowing very rapid excavation of the retaining wall. About 100 m of elevation was swept away, allowing water levels to drop down to the level of Lake Provo, with a deep gash eroded that allowed water to continue to leak out of Provo until that lake lowered as well, much more gradually. When the wall collapsed, the water raced northward into the Snake River, which flooded throughout Idaho and up to the Washington boundary. All along the Snake River from where Marsh Creek drained into Porneuf River and then into the Snake River, there are geological markers of the massive flood, involving 100-200 m higher water levels in the river system. Large boulders, erosional features, and slosh deposits are traceable for over 1000 km from the rupture point. The flood hydrograph estimated from the various deposits suggests discharge rates of from 0.8 to 1.0 million cubic m/sec, with total flow duration at Red Rock Pass estimated at about 300 days. This remarkable event has parallels in the great lake history of the Rio Grande Rift area, where massive drainage events of much greater age emptied interior lakes into the Rio Grande system, but the Bonneville flood is very well documented due to the relatively young age of the flood.

There was substantial initial resistance by geologists to the evidence favoring a massive flood as Lake Bonneville dropped 100m, but eventually the weight of field evidence was compelling. A similar resistance to the nature of a catastrophic event was encountered in explaining the Channeled Scablands of eastern Washington. The work of J. Harlen Bretz in the 1920's and 1930's involved careful field mapping of the scablands, and he advanced the notion that massive floods of a repeated nature had scalloped the land surface much as a river plain is carved by more modest flows. Indeed along the Columbia river there are great gravel ripples, which appear to be scaled up versions of those normally seen at the bottom of streams. Bretz argued that these must be the product of massive flows, that could activate bedload deposits on a grand scale. This was viewed with great skepticism, as there was no readily apparent explanation for repeated great floods outside modern day experience, and the invocation of any catastrophic floods was deemed too biblical, and at odds with uniformitarianism. However, in the 1960s and 1970s, glaciologists began to understand the nature of floods in glaciated lands, and the specific case of jokulhlaups, or floods from floating of glaciers. In this case the waters are associated with former Lake Missoula, which formed at the southern end of large ice packs over westernmost Canada during the last ice age. The question that arose is why were there intermittent floods caused by spilling out of Lake Missoula?

In part the answer came from studying a very different environment, on the island of Iceland. Iceland has the unusual circumstances of being a glaciated region with active volcanism and hot crustal conditions. The glacier Vatna Jokul is a large ice pack east of the capital, along the southern margin of which there have been repeated massive floods. The explanation for the floods emerged in the 1960s, when it became clear that a subglacial lake (Grimsvotn) accumulates due to melting of the lower surface of the glacier. The water is normally trapped, and water pressure builds up until the glacier actually floats, upon which the water is able to squirt out in a glacial flood (jokulhlaup). This happens quite regularly, about every 20 years, because it takes that length of time for the water pressure to grow to equal the weight of the ice.

The situation in Lake Missoula was somewhat different, as in that case prongs of glacial ice formed boundaries on the glacial Lake Missoula. This was the only outlet of the lake, as on other sides the hills were never overtopped The water level accumulated against the ice dam, eventually leading to pressures sufficient to float the dam, which allowed floodwaters to squirt forth in large volumes, which swept down into the Columbia River drainage as massive floods. These floods formed the scabland deposits. In detail there were multiple sets of ice dams and several blocked lakes. The highest water levels of Lake Missoula reached 500 m against the lobe of ice. The flood deposits down the river drainage are from 150-250 m above the normal river level, indicating flow rates of 13-17 million cubic m/s, or about 10 times more than the huge Lake Bonneville flood. This is considered to be a minimum estimate. The floods moved 15 cubic miles of water/hour! The water levels through the current city of Portland would top the largest building in the city.

Like the repeated floods in Iceland, the lifting and settling of the ice dam at Lake Missoula allowed repeated events. The geological record of sedimentation in the Lake shows alternating cycles of silts and clays, suggesting varying water levels, while deposits in northern Washington and Idaho and in southern Washington show alternating patterns of silts and sands and gravels. The cycles suggest about 50 year intervals between flood events, with more than a dozen floods being recorded. Thus, these amazing floods happened again and again, until the ice receded and the lake levels dropped permanently.

While the outbreak of Lake Bonneville and the periodic floating of the ice dam of Lake Missoula were dramatic natural phenomena, human activity has also resulted in great floods, sometimes intentionally. An example is the history of the Yellow River, which has been used more than once as a massive weapon. The most recent case occurred in 1938, when the Japanese were advancing southwestward into China in their attempt to take over the country. They had overrun the city of Kaifeng, driving Chiang Kai-Shek before them. In an act of desperation, to slow the advancing Japanese, the Chinese leader mined the levees of the Yellow River, which build up high above the surrounding flood plain due to the high sediment load as the river drops out of the highlands to the west. The induced flood, released through a 1/4 mile wide gap in the levees created a massive bog, which actually mired the Japanese, who redirected their advance toward the south. The Yellow River actually changed its course, merging into a Yangtze tributary 400 miles south of its prior track, but near to historic drainages from 1494 and 1854. The stratagem was somewhat successful, but a huge price was paid by the Chinese. The massive flood, advancing at 5 mi/hr immersed 21000 square miles. Over 1 million Chinese perished in the flood, and massive crop losses caused famine. The river returned to its northeastward drainage in 1947. The river normally floods almost every other year, with 1500 floods in the last 2500 years due to natural causes. The previous human-induced flood was in 1642 when General Gao Mingheng also intentionally collapsed the levees in order to suppress a peasant rebellion.

On September 12 of 1717 an avalanche cascaded down the Troilet, Italy glacier, gaining speed on a cushion of air, reaching a falling velocity of 320 km/hr over a 3600 m fall. Two towns were destroyed, with 7 people killed and 120 cows lost. The slosh of the avalanche ran up the far side of the valley at a speed of 125 km/hr.

Snow avalanches, rock avalanches, debris flows, mud flows, and rock falls are failures of the surface under the action of gravity. The basic physics controlling the stability or instability of landforms is relatively simple and well understood, but the hazards are not always recognized, even when geological deposits document past slope failures in the region. In many cases surface instabilities of this type are compound events, associated with earthquake or volcanic processes, which enhance their catastrophic potential. From the surface geology perspective, landslides and debris flows are important landscape modifying agents, and play as large of a role in eroding topography and depositing debris as is played by other mechanisms such as rainfall and runoff.

There are two basic classes of surface failures:

1. Single rock failures

These often are Blockfalls of individual rocks falling from an eroding surface, which is usually very steep, or 'oversteepened' beyond the stable angle of repose for the surface materials. The blocks pile up in an apron of debris, called a talus slope or talus cone.

2. Flows of the surface, involving large volumes of rocks and soil

Slow Flows - [creep, solifluction, earthflows]

Fast Flows - [snow avalanches, landslides, mudflows, debris flows]

Fast flows leave jumbled deposits, often piles upon piles, or hummocky surfaces that can be recognized in the geological record by their shape and the poorly sorted (many rock sizes intermixed) nature of the deposit. The fast flows tend to be fluidized, either by mixture of rock and air or rock and water. The relative components of the mixture determine the nomenclature: a mudflow is water with lots of clay and silt material, which is a runny mud, while a debris flow is a jumble of rock fragments with some water. The fluidized nature of the flow accounts for the fast flow velocities attained, and the associated hazard posed.

All falls are driven by gravity, pulling the surface downward, but the tendency for failure depends on many factors. Basically, the driving forces must exceed the resisting forces for a slope to fail. The driving force acting on a rock or soil mass is the product of the mass of the object and the gravitational acceleration component acting parallel to the surface. If the surface is vertically oriented, this acceleration is maximum, if it is horizontal, the acceleration is zero. Effectively, the driving force is the component of the objects weight acting parallel to the surface.

The resisting force is the friction, which involves cohesion of the object (things like roots, stickiness of the soil, protruding rocks, etc.) as well as the intrinsic resistance to sliding of the material on the contact surface. Effectively, it is the weight component acting perpendicular to the be times the friction coefficient.

Some of the factors influencing the balance of driving and resisting forces are:

1. water: water in the pores of the soil and rock can be under pressure that reduces the weight, allowing the material to slide easier. Rainfall may trigger failures because the pores saturate and reduce the friction by decreasing the effective weight on the surface. This makes surface stability vary with intensity and rate of rainfall, as well as prior history. In effect the infiltration rate controls the buildup of water pressure that bears part of the weight of the surface. Coastal landslides and river bank failures often involve direct undermining effects of water erosion and pounding as well.

2. slope: the slope is the angle that the surface makes with the horizontal. The larger the angle, the larger the component of driving force will be relative to the resisting force.

3. material: the coefficient of friction differs for different materials. Clay, rock , sand, all have different stickiness and sliding properties. The layering and internal structure of the material is also important. If there is bedding or stratification parallel to the surface, it is often easier for the material to break and slide on an internal bedding surface, than to break across bedding surfaces.

4. biology: Trees and grasses have roots that can stabilize soils and rock slopes due to the suturing effects of cross-cutting roots. Humans often modify this aspect by deforestation or other induced changes in the flora that may lead to slope destabilization.

Landslides often have some common features that allow them to be recognized. The actual failure surface is arcuate, tending to dip steeply at the highest level, or the headscarp, and to scallop under the surface on a single or multiple slip surface that emerges under the toe of the slide, which is hummocky, with compression ridges in it. Typically the landslides are lobate, or finger shaped, spreading out laterally in the toe as they expand into flatter surface areas.

The speed of flows is important for understanding their energy and destructive potential. In some places there are regular debris flows and avalanches, so films are made. In other cases, we can assess the flow energy by looking where the debris ends up, on the far wall of a valley, or over humps in the terrain. The flow velocity required to reach that final position can be estimated. For flows in well-defined channels, the angle of slosh, or superelevation of the flow as it goes around bends can be used to estimate the velocity of the flow.

The association of landslides and snow avalanches with precipitation is rather clear, as both a source of ground water pressure, erosion, and snow accumulation. Thus, flooding from heavy rains is often accompanied by landsliding.

Earthquakes and landslides are also commonly associated. This is mainly because the ground acceleration can act to temporarily reduce the weight of the unstable surface mass, either by directly accelerating the ground upward to counteract the force of gravity, or by increasing the water pressure by pumping the groundwater system. Accelerations of the ground can also break up cohesive bonding forces such as roots, or soil cementation. In situations where there is sandy soil, the ground vibrations can cause liquefaction of the subsurface leading to a loss of strength that causes a surface failure.

An example of an earthquake induced slide is the 1964 Sherman Slide, in which a huge rock avalanche was triggered by the March 27, 1964 Alaskan earthquake. The slide spilled out onto the Sherman Glacier, with the rupture surface following dipping bedding planes in the adjacent highlands. That great earthquake also induced much smaller scale slides within the city of Anchorage, which were responsible for much of the destruction and loss of life during the huge earthquake.

The 1903 Frank rockslide, in Alberta Canada also ruptured on dipping bedding planes in the hillside, and geologists can map the bedding geometries to assess the potential for such failures in different regions. The surface geometry is also readily assessed by mapping, and pathways of possible slides and debris flows can be charted out. Still, even in areas of repeated slope failures, the lessons are not always taken to heart. The 1970 Nevados Huascaran debris avalanche in the Andes of Peru followed a course well marked by predecessor events yet it still took a deadly toll. The mudflow on Nevado del Ruiz had a similar recognized history, but no effort had been made to avoid the inevitable recurrence of clearly documented past events.

Volcanoes like Nevado del Ruiz are commonly associated with landslides, mudflows and debris flows because they are steep-sided, active structures, with earthquakes, expanding cones, glaciers and heating events. The Mount St. Helens event involved a massive landslide on the north flank of the mountain due to slope instability, and the debris injected into the Toutle River resulted in a damaging mudflow. The same has happened in the past on Mt. Rainier, and threatens Tacoma for the future because there are past mudflows that penetrate right into town. Mt. Shasta has a massive flow deposit on the north side, which involved 26 cubic kilometers of debris avalanche (26 times larger than for Mt. St. Helens!). On Mt. Fuji, debris flows are seasonal events, and tend to follow regular courses down the hillside. Engineers have designed channels to guide the flows around surrounding towns.

Landslides and floods are often coupled, because the slides may block up rivers, providing temporary dams which fail once the water in the transient lake accumulates enough pressure to wash out the dam. Large flood can then ensue. Intermittent great floods occur on the Indus river as a result of avalanche dams in the Himalayas impounding lakes that fail again and again.

Debris flows are common in the Sierra Nevadas due to the high topography, the snowpack and the steep slopes. Occasionally these flows take a human toll, as in the August 1989 Olancha flow which blocked up the LA aqueduct.

One would expect that there are also large submarine slides, and this is indeed the case, although only a few such events are directly witnessed. One of the most famous cases was the 1929 Grand Banks (offshore Newfoundland) event, which involved a magnitude 7.2 earthquake/slide. The ocean bottom motions triggered a tsunami which took 27 lives, but there was also a massive turbidity current (an underwater debris flow) which cascaded down the continental slope, extending over 1700 km long. It was detected by the fact that the region of failure was criss-crossed with trans-Atlantic phone cables, which suffered 28 ruptures. The slump of material involved 500 billion cubic meters. Later investigations of sediment deposits on the seafloor suggest that such events recur about every 30,000-100,000 years.

Another example involved the 1975 Kalapana, Hawaii slump. This was an earthquake/landslide with a magnitude of 7.1 on the flank of Kilauea volcano. A mass of about 10exp(15) or 10exp(16) kg moved, which is about ten thousand times more mass than in the Mt. St. Helens slide. A local tsunami of 14.6 m height swept up the coast. Similar events appear to have taken place in 1823 and 1868, but it has now been found that vastly greater submarine slides have occurred offshore of Oahu and Maui. These events, perhaps the largest known landslides, are an important part of the erosional process wearing down the huge volcanic edifices built up as the Pacific plate has overridden the Hawaiian hotspot.

Closer to home, landslides in the Santa Cruz Mountains have been responsible for much of the damage in the 1989 Loma Prieta earthquake, as well as the cause of many moderately damaged homes being 'red-tagged' because it was determined that they are vulnerable to future destruction as they are built on active landslides. This has been a difficult social issue, as the potential loss of people's home under the unpredictable future threat of landslide failure is a difficult issue to resolve and to make policies about.