erosion

erosion Erosion is the process that moves material resulting from the breakdown, or weathering, of bedrock. Weathering itself may be distinguished into two broad categories of process: physical and chemical (see weathering). Physical weathering results from the mechanical action of environmental agents on bedrock, including wind, moving water, freezing and thawing of water, crystallization and dissolution of salts, expansion of roots, digging by animals, and fire. Glacial erosion is perhaps the most powerful manifestation of the mechanical breakdown of bedrock. Sometimes, physical erosion acts alone, for example during the erosion of barely consolidated bedrock. Chemical weathering entails the chemical breakdown of bedrock minerals by water; it can also operate alone, as when carbonate or evaporite deposits dissolve. For much of the Earth's surface, however, physical and chemical weathering act in concert. When liquid water interacts with siliceous bedrock, chemical weathering breaks down some of the silicate minerals in the bedrock, converting these to solutes plus clays and aluminium and iron sesquioxides.

Measuring rates

Most weathered materials are removed and transported by water and wind. The discharge of material transported by rivers is normally measured at fixed gauging sites. The material discharge divided by the area of the gauged catchment gives the yield. The yield does not equate directly to erosion or weathering rates because of mass storage during transport downslope and through river channels. Instead, yields are said to equate to rates of denudation, a measure of the removal of material, thus generally leading to a reduction of elevation and relief. Possible sites of storage include (1) colluvium or sediment deposited at the base of hillslopes before even entering a channel, (2) alluvium or sediment deposited by fluvial activity within a channel or on adjacent floodplains, and (3) lacustrine sediments deposited in lakes. Various techniques are used to measure local rates of physical erosion, such as erosion pins, erosional troughs, and isotopic techniques. Particles moved by wind do not conveniently pass by a single gauging point, and intermediate storage in dunes and sediment sheets is almost always important. Thus aeolian denudation rates must be synthesized from many local measurements.

In rivers, the concentrations of individual solutes normally decrease with increasing discharge, whereas suspended sediment concentrations and bedload movement normally increase. Consequently, solute transport is dominated by typical flow conditions, and solid transport by discharge events. Good estimates of sediment yield from a watershed, or catchment area, thus require years of measurement. Many rigorous studies have been undertaken to determine sedi-ment yields from large watersheds, including those on sedimentary substrates. The mass of sediment in motion is large, 30–100 × 10¹⁵ g year⁻¹, with approximately 20 × 10¹⁵ g year⁻¹ being transferred to the ocean (Fig.1). Still larger quantities of sediment are probably deposited on land. Storage of solutes on land appears minimal, and about 9 × 10¹⁵ g year⁻¹ are transferred to the ocean.

Natural erosion in tropical to temperate settings

Under natural conditions, the supply of terrigenous sediments from a watershed is largely controlled by its topographic relief, its bedrock geology, and its Pleistocene climatic history. The fundamental division is between steepland and flatland terrains. Runoff follows in importance, whereas temperature appears less significant. The erosion of Pleistocene glacial and periglacial deposits continues to affect the solute and solid loads of many rivers. Finally, given the same general rock type, topography, runoff, and temperature, rates of chemical weathering of younger igneous rocks, such as those in island arcs, are roughly twice rates in old cratonic settings.

Feedback among the chemical and mechanical processes that erode developing soils strongly influences both the rate of erosion and the composition of solid and dissolved erosion products. This interaction can be examined in a simple framework, the division of erosional regimes into two types: weathering-limited and transport-limited. Typically, soils develop from the weathered and partially weathered mineral grains, while solutes are transported away by surface and ground waters. Loose soil material can then be eroded by physical processes. For weathering-limited erosion, the supply of loose weathering products is controlled by weathering rate. Soils tend to be thin. Most rock types are more vulnerable than young crystalline igneous bedrock to chemical and physical erosion, and weakly cemented sediments, such as some shales, and unconsolidated glacial deposits, especially loess, contribute exceptional quantities of sediments to rivers. For transport-limited erosion, weathering supplies loose solid materials at rates that exceed the capacity of transport processes. As soil thickens, it develops an internal structure which in turn reduces the rate of interaction between water and fresh minerals, thereby reducing weathering rates. Rates may drop exponentially with soil thickness. Weathering rates can be low in the absence of soil, because water fails to make prolonged contact with fresh minerals. Millions of years may be required to reach equilibrium.

Vegetation has two primary roles in weathering and erosion. First, it reduces the power of physical erosion processes. Plants protect and anchor loose material with their roots, while litter and low ground cover protect soil from raindrop impact and surface wash. This allows thicker soils to develop and stabilizes vulnerable bedrock. Physical erosion usually accelerates after loss of vegetation by fire, cutting, cropping, or grazing. Second, plants produce chemicals that promote the breakdown of bedrock and soil minerals. Some are released through roots, perhaps with the help of symbiotic root fungi (mycorrhyzae). Chemicals include an array of complexing agents, organic acids, and carbonic acid. Other chemicals, such as carbonic and nitric acids, are generated during the decay of organic matter. Experiments indicate substantially greater rates of chemical weathering under well-developed plant cover. Soils with plant cover and active soil organisms appear to have greater infiltration rates of water, perhaps allowing more interaction with bedrock minerals. On slopes where soils tend to be thin owing to weathering-limited erosion, plants probably greatly increase erosion rates because of their chemical effects. In transport-limited settings, with thick soils, the influence of plants may be less; carbon dioxide, however, is mobile throughout the soil profile.

The interaction of vegetation with erosional processes promotes styles of erosion where mass movement is episodic. Two types of cycles seem to be especially important. One is a landslide (soil–avalanche) cycle that may be a dominant form of mass wasting in many humid regions. The other is a fire–flood erosional cycle. In the landslide cycle, soil is anchored by vegetation. The soil profile eventually thickens to the point where heavy rainfall or an earthquake destabilizes the slope and causes soil avalanches. This normally involves the entire profile above a hard saprolite (partially weathered rock). Vegetation re-establishes, and the soil profile begins to thicken, starting the cycle once again. In eastern Puerto Rico, the recurrence interval is roughly 10 000 years. In a fire–flood sequence, vegetation anchors loose soil. After sufficient burnable material accumulates in and upon the soil, a hot fire can occur. Such burns commonly transform soils chemically into a hydrophobic form. Instead of infiltrating, heavy rains run off, thereby eroding the soil and scouring channels. Debris torrents and landslides develop. Vegetation re-establishes and the cycle starts once again. The fire–flood sequence requires a threshold accumulation of soil. Thus, the recurrence intervals of the fire–flood sequence are much shorter, probably by several hundred years. Because these processes involve loose material and act minimally on hard saprolite, they cause a characteristic partitioning of major cations in stream waters. Sodium and calcium are preferentially leached from the hard saprolite, and the sodium-to-potassium ratio in water (after correcting for sea salt) exceeds the bedrock ratio, as does the calcium to magnesium ratio.

For landscapes in which chemical weathering represents the rate-limiting step for physical erosion, the rates of physical erosion (solid yields) can be predicted from chemical erosion (Fig.2). A hypothetical starting point for analysing physical erosion in a watershed is to assume that the watershed is in geomorphic equilibrium. This is a state in which the statistical measures of a landscape (hypsography, river-channel density, soil types and thicknesses, vegetation types, disturbance types and frequencies) do not change through time. Most importantly, equilibrium requires that there is no net thickening of soils or accumulation of sediments within the landscape. In this context, we can define trends for equilibrium erosion for geologically young, crystalline, igneous bedrock and carbonate-cemented bedrock. When sediment yields for watersheds are compared with run-off, the predicted equilibrium–erosion trends separate high-sediment-yield rivers that drain steeplands from low-sediment-yield rivers that drain flatlands. Thus, these trends approximate to the transition between weathering-limited erosion in steeplands and transport-limited erosion in flatlands.

Sediment yields of flatland watersheds do not seem to depend on watershed area, whereas yields in mountainous watersheds generally decrease with increasing watershed area. The yields of the world's largest rivers are only a few-fold greater than those of flatland watersheds. Several factors are involved in the apparent decrease in yields with increasing catchment area. In uplands, sediments accumulate as colluvium at the base of hillslopes or in other areas of reduced steepness, or they deposit as alluvium in lower-order streams, without having ever entered a large river channel. Alluvial accumulation is especially significant in flatter regions. Large watersheds tend to have extensive flatland areas. These settings are transport-limited and either fail to contribute sediment through physical erosion or act as depositional areas, where sediments accumulate as alluvium in river channels and adjacent floodways, and as a mix of clastic and organic deposits in lakes and wetlands.

Glacial and cold climate weathering and erosion

Glacial erosion is the most powerful form of physical erosion. Perhaps hundreds of metres of rock were removed during the Pleistocene by continental ice sheets from the Laurentide region and smaller ice sheets in Asia and Europe. Glaciers have also deeply sculpted mountainous areas worldwide. Ice caps still cover much of Antarctica and Greenland. The intensity of glacial erosion is much greater when liquid water is present at the ice base, a condition referred to as a warm-based glacier. Glaciers lacking liquid water near the base, cold-based glaciers, sometimes erode rather little. For some regions with extensive Pleistocene glacial sediments, sediment yields of watersheds increase with increasing area, an indication that Pleistocene sediments have been eroded out of the smaller basins but are still moving downstream in larger watersheds.

Chemical weathering in cold, wet environments contrasts in notable ways from that of warmer settings. Glacierization and freeze-thaw processes fracture bedrock, exposing fresh mineral surfaces. In effect, the bedrock-to-water ratio in these settings is much greater than for warmer settings. Because of this, the weathering of less abundant but easily weathered minerals in siliceous bedrock, especially carbonates and sulphides, contributes disproportionally to solutes in surface waters. Subglacial waters are especially rich in potassium relative to sodium, perhaps because the physical breakdown of micas releases large quantities of potassium.

Human effects

Much of the Earth's land surface is now occupied by people and has been cleared for timber, agriculture, and urbanization. Although land clearing increases rates of physical erosion, the resulting changes in rates of physical erosion have, in general, not been well documented. Agriculture, civil engineering, and mining mobilize vast quantities of soils, unconsolidated sediment, and bedrock. Engineering activities, such as road building and building construction, may, in concert, move more solids over the Earth's surface than all ordinary natural processes combined. The hydraulic architecture of the land has been transformed to serve agriculture, water supplies, land reclamation, navigation, and power generation. Few rivers lack dams and reservoirs; large rivers are confined to their channels by levees; normal floodplains are converted to paddylands and aquiculture; water is mined; irrigation raises soil moisture and creates artificial distributary systems; and so forth. These changes confound the quantification of sediment dispersal from the uplands to the ocean.

Human activities have accelerated physical erosion over equilibrium levels. In flat coastal plains and piedmonts, yields appear to increase five- to ten-fold with agriculture, whereas in loess terrains 100-fold increases are typical. The greatest yields are in the smallest watersheds. In tiny watersheds in the upper Mississippi River valley, yields approaching 10 000 g m⁻² year⁻¹ have been reported.

Research directions

Much of the process-level research relating biogeochemical processes to erosion is routinely undertaken in small research watersheds. Most of these are on igneous and metamorphic rock to minimize groundwater losses. In contrast, most of the world's chemical and physical erosion is from sedimentary rocks, and these are mostly shales. Moreover, sampling for most small watersheds has been from periodic (interval-based) sampling, thereby ignoring a major characteristic of such watersheds, that sediment moves during discharge events. Finally, most research watersheds are not large enough to examine the effects of sediment storage during transport. Thus, many opportunities remain for well-designed process-level studies of weathering and erosion.

Perhaps one of the most fascinating research directions in the study of local erosion is the analysis of cosmogenic radionuclides produced in situ. These are created when cosmic-ray neutrons and muons strike atoms within mineral grains in a soil profile. Some of the important isotopes are ¹⁰Be, ²⁶Al, and ¹⁴C. The concentration of these isotopes depends on erosion rate and irradiation intensity. The latter depends on altitude and latitude and depth below the ground surface. Two situations are easily modelled. One is the erosion event, where it is assumed that a surface was formed at some time in the past, and that there was no subsequent erosion. Isotope concentrations are used to date this event. Settings locally dominated by transport-limited erosion would be suited to this approach. The second situation is to assume that erosion is steady, as might be expected in an erosion-limited setting. The ground-surface isotope concentration equates, for a broad range of erosion rates, to the erosion rate. These methods average erosion over thousands of years.

R. F. Stallard

Bibliography

Milliman, J. D. and and Syvitski, J. P. M. (1992) Geomorphic tectonic control of sediment discharge to the ocean—the importance of small mountainous rivers. Journal of Geology, 100, 525–44.
Stallard, R. F. (1995a) Relating chemical and physical erosion. Reviews in Mineralogy, 31, 543–64.
Stallard, R. F. (1995b) Tectonic, environmental, and human aspects of weathering and erosion—a global review using a steady-state perspective. Annual Review of Earth and Planetary Sciences, 23, 11–39.