Erosion and Sedimentation

Background

Erosion features at the Grand Canyon in northern Arizona

Erosion over geologic time has flattened mountains, and excavated such features as the Grand Canyon. These are just a few examples of the lasting impressions that erosion has left on our landscape. These impressions are even more pronounced in arid and semi-arid areas, such as the southwestern United States, due to the tenuous position our ecosystems are in due to the delicate balance of interactions between climate, geology, and vegetation and the subsequent difficulty encountered in recovering from past abuses (over the last 150 years or more). Today the public sees the cumulative effects of past abuses and feels that these must be the result of current management. There is evidence that much of this abuse is the result of actions that occurred years ago, but because of the arid conditions of the Southwest, it will take many decades to return to conditions present hundreds of years ago.

Erosion processes in the Southwest are often energy and supply limited. Energy limitation develops because of the distribution and variability of precipitation. As previously stated, 80 to 95 % of runoff occurs in the winter season as the result of snowmelt or low intensity rain events (energy limited). Precipitation variability is also a factor here because a large percentage of the available precipitation is being used to recharge or rewet the litter and soil horizons.

Summer rainfall is not energy limited but is much more localized than winter precipitation. Thus effects of erosion on large basins are usually limited.

Erosion can be supply-limited in coarse textured soils with few fines. If the rainfall energy is insufficient to transport coarser materials then erosion will not proceed. This is the case with naturally coarse-textured (lithic) soils. Erosion pavements that typically develop when fine-textured sediments have been removed inhibit further erosion because the surface sediments are too large for rainfall alone to detach and move. Concentrated surface runoff is then the only mechanism that can lead to additional erosion.

Upland Erosion

Soil moves off slopes either by dry creep and mechanical dislodgment, or by overland flow of water. Sometimes overland flows generate sufficient momentum on the slopes to start rilling, but soon dissipate before reaching the channel, or if it reaches the channel where the slope is normally less, they dissipate there. By these processes, headwater channels accumulate sediments until heavy flows flush it further on downstream. When channel gradients flatten, velocity of flow decreases and the coarse sediments are left to form broad alluvial fills.

In coarse textured soils, much of the sediment moved in small headwater streams is classified as bedload (Hibbert et al. 1974). If suspended by turbulent flow, the coarser particles quickly settle out when stream gradients drop and water velocity slows down. Fine silts and clays that stay in suspension for longer time periods account for only a small percentage (about 10%) of the total sediment load.

Channel Erosion

Storage and subsequent movement of sediment through channel systems in response to disturbances is a complex process (Heede, 1980). Factors such as a loss of plant cover by "poor" management practices, fire, or flooding can result in large amounts of surface runoff which are concentrated in the channels and can often move sediment even during smaller storm events (Bleich et al. 2005). Vegetation and topography interact to affect sediment movement and storage on the watershed and within the stream system (Heede, 1985; 1986).

The primary role of vegetation in regulating sediment is dissipation of streamflow energy and stabilization of the stream banks (Brooks et al., 1997; Medina, 1996). This allows more water to infiltrate into the stream bank and to recharge the groundwater. Energy dissipation and increased infiltration also prevents excessive erosion, maintains the physical stability of the riparian system, and encourages the establishment of riparian vegetation (Medina 1996, DeBano and Schmidt, 1989). In this semi-arid environment it is not the small storms (less than 10-year return interval) that move most of the sediment. The large, infrequent storms do most of the sediment transport. Alternate cutting and filling occurs along channel systems as discharge rate and sediment supply fluctuate. Because large storms are relatively rare, the long-term average or 'normal' rate of erosion is difficult to determine. Short-term records of sediments stored at some point in the watershed can be misleading.

Big Storm Events—Episodic transport of sediment is a characteristic of erosion in ecosystems of the Southwest (DeBano and Baker 1992, DeBano et al., 1996b, Hibbert et al. 1974). When addressing long-term erosion rates, the rare climatic event must be considered. However, it is difficult to assess the impact of a 100- or 500-year storm, especially if it should come just after a disturbance like a wildfire or prescribed burn. The Labor Day Storm of September 5, 1970 was classed as a 100-year or greater storm over much of central Arizona (Thorud and Ffolliott 1973). The magnitude of this storm on Three Bar is unknown, but rainfall was 203 to 250 mm (8 to 10 in) (Hibbert et al 1974). This was 30 to 50% of the annual rainfall. The impressive thing about this storm at Three Bar was the lack of erosion. Despite heavy flows, the only watershed to produce sediment at the gaging site was the most recently treated watershed F. The lack of sediment from the other catchments is attributed to lack of overland flow and stable channel conditions, particularly in the lower portions of the catchments. Thus, even the large infrequent storm events may have no detectable effect on sediment yield.

Hydrophobic Soil Layers—One of the main causes of accelerated erosion is the development of hydrophobic layers in the soil. These layers typically form after severe wildfires. However, they may develop in thermic soils without any associated disturbance. Infiltration is restricted when hydrophobic conditions exist since water cannot penetrate nonwettable layers, and that leads to surface runoff and eventually rill development. Scholl (1971) demonstrated how wettability varies under an undisturbed Utah juniper stand at Beaver Creek. Resistance to wetting in the surface soil increased from completely wettable in openings between trees to highly nonwettable in the litter under the juniper canopy. It is noteworthy that one of the most intensely rilled areas found on the Whitespar watersheds with chaparral vegetation near Prescott, Arizona, after an intense July storm was under the canopy of large spreading alligator juniper trees (Hibbert et al. 1974). Given the right storm event, significant erosion can even occur in undisturbed areas.

Wildfire—Wildfire creates a very unstable soil condition, after which water runs off the surface much more readily than before. Slope erosion and scouring of headwater channels are greatly accelerated, and flooding and sedimentation of downstream channels are common.

For example, sediment production during the first 3 years after a wildfire in 1959 at the Three Bar Experimental Watersheds contrasted sharply with production in the years before and after this period, and with other watersheds where no fire occurred (Hibbert et al. 1974). Some of the accelerated sediment production came from the channels where sediments had accumulated for years. However, much of it came directly off the slopes from severe rilling in portions of the catchments where high fire severity produced water repellency problems. Total yield for the first 3 postfire years on Three Bar watersheds D, C, and B combined was 25 times the amount measured during the other 13 years (3 prefire and 10 postfire years). Of significance is the fact that the area was particularly vulnerable to erosion from heavy rainfall for only a few years after the fire.

Prescribed Burning—Control burning in contour strips can also produce sediment, although exposing only a portion of the slope each year to fire reduces the impact of broadcast burning (Hibbert et al. 1974). Three small catchments on the Sierra Ancha Experimental Forest were treated with prescribed fire in early fall for 4 consecutive years beginning in 1961. One-fourth of each catchment was burned each year in varying width strips up to 61 m (200 ft) wide. A fourth catchment was left undisturbed as a control. The objective was to test prescribed burning in strips to temporarily reduce chaparral cover for improving water, forage, and browse production (Pase and Lindenmuth 1971).

Sediment moved off the treated slopes on three occasions during the treatment and evaluation period, twice during periods of heavy summer rains, and once during the wet months of November and December 1965 when rain and snow totaled 437 mm (17.2 in). No erosion or overland flow was observed on the control catchment. Sediment yield did not vary consistently with strip width, probably because of differences in slope, which appeared to be more influential in causing overland flow and rilling on the burned areas (Hibbert et al. 1974). Residual litter and relatively low burning temperatures were important factors in controlling sediment yield. Erosion was less noticable on lightly burned areas where 70 % or more of the litter residue was retained than on areas where less than 60 % was retained.

Watershed-Riparian Linkages

In the arid Southwest environment, erosion and runoff processes are key factors affecting the stability of both riparian areas and the surrounding watersheds. Recognizing the interrelationship between watershed condition and riparian health provides a framework for managers to synchronize the improvement of the riparian areas and surrounding watersheds (Bleich et al. 2005, LaFayette and DeBano, 1990). Sediment movement in and through these riparian systems is controlled by vegetation, topography, and hydrology, along with control exerted by geologic formations (Rosgen 1996). If riparian systems are in dynamic equilibrium, the volumes of incoming sediment equal those of the outgoing sediment (Heede, 1980). Riparian vegetation remains vigorous under these conditions (Medina 1996). Also, streamflow does not expand stream meander-cutting or channel bed erosion (Heede, 1980; 1986). In the long-term, riparian areas reflect both vegetation and physical processes such as rainfall, runoff, and the geomorphological features of the watershed in which they reside (Medina 1996, Rosgen 1996). However, the conditions in downstream riparian areas can lag behind the major active erosional processes on the watershed at any given point in time (Heede et al., 1988).

The climatic, vegetative, and hydrologic processes operating in the Southwest provide a unique setting for evaluating erosion processes in ecosystems of the region (Bleich et al. 2005, Medina 1996). These dryland environments and resulting landscapes engender settings that are different from those found in more humid climates. The transitions from hillslope to riparian and aquatic ecosystems in this region also tend to be more abrupt than in more humid regions.

Most of the streams in the lower elevations of the Southwest region are intermittent, flowing mostly in the winter, or infrequently in response to convectional-type storms in the summer. Potential evapotranspiration generally exceeds precipitation in these ecosystems. Although streamflow is largely discontinuous, riparian vegetation frequently occupies the floodplain. Precipitation at the higher elevations is generally sufficient to sustain longer streamflow periods or, in some systems, perennial flows. As a consequence, a more reliable source of water is available for vegetative development and potential erosion.

Supporting Data From Beaver Creek

Sediment yields—Mean annual sediment yields from the untreated pinyon-juniper watersheds (Vertisols of the Springerville series) sampled on Beaver Creek (13 station years) have varied from 0.02 to 0.60 Mg/ha (0.01 to 0.27 ton/ac) with an average of 0.02 Mg/ha (0.10 ton/ac) over a 9-year period of record (Clary et al. 1974). Sediment yields from undisturbed watersheds in the western USA can range from <0.01 to 6.00 Mg/ha (Robichaud et al. 2000). These yields are at the lower end of that range. Mean winter sediment yields from untreated watersheds have ranged from a trace to 0.60 Mg/ha (0.27 ton/ac), with an average of 0.16 Mg/ha (0.07 ton/ac). Summer yields had the same range, with an average of 0.07 Mg/ha (0.03 ton/ac). These mean values indicate that over 50 % of the sediment is generated during the winter season. This seasonal relationship is even more striking when one considers that eight of nine summer seasons had sediment yields of less than 0.01 Mg/ha (0.005 ton/ac). The remaining summer (1970) had an estimated yield of 0.60 Mg/ha (0.27 ton/ac) from untreated watershed 4 as the result of the Labor Day storm. That event had an estimated recurrence interval of 100 years (Thorud and Ffolliott 1973).

Mean annual sediment yields from treated pinyon-juniper watersheds on Beaver Creek have varied from a trace to 2.46 Mg/ha (1.1 tons/ac) on a cabled watershed, and from a trace to (0.18 Mg/ha (0.08 tons/ac) on a herbicide-treated watershed. Because of the short period of record (13 station years) and confounding due to treatment intensity, storm frequency, and climatic variation, a mean cannot realistically be calculated to generalize the treated condition in pinyon-juniper.

The largest sediment yield of 2.46 Mg/ha (1.1 tons/ac) was produced on a cabled watershed during the Labor Day storm, which produced a peak discharge of 9.17 m3/sec/km2 (800 ft3/sec/mi2) with an estimated recurrence interval of 100 to 150 years. This watershed received a maximum 30-minute precipitation intensity of 55 mm/hr (2.17 in/hr) and a total storm precipitation of 103 mm (4.06 in). The only peak discharge in the pinyon-juniper type on Beaver Creek which exceeded this amount during the period of record was on the same watershed in 1964, 1 year after the its cabling treatment. The discharge of 10.92 m3/sec/km2 (1,000 ft3/sec/mi2) produce a sediment yield of 0.74 Mg/ha (0.33 tons/ac). The storm producing this discharge occurred on August 3, 1964, and had a total precipitation of 40 mm/hr (1.59 in/hr). Apparently the greater total rainfall in 1970 was a factor in producing over three times more sediment than the storm event in 1964.

Based on records obtained during the past 9 years on Beaver Creek, and from knowledge of sediment losses resulting from various treatment intensities and storm frequencies, it appears that sediment yields of 2.2 to 4.5 Mg/ha (1 to 2 tons/ac) are approaching the maximum sediment loss potential for watersheds with similar physical characteristics and climatic regimens in the pinyon-juniper type. It is also concluded that there appears to be no meaningful change in sediment yield after either cabling or applying herbicide in the pinyon-juniper type (Clary et al. 1974). These sediment yields are at worst, only six times background erosion levels. They are still within the range for sediment yield from undisturbed watersheds in the West and well below the range for agricultural areas (Neary and Hornbeck 1994).

Sediment rating curves—Research by Lopes et al. (2001) demonstrated that disturbance from vegetative practices generally increased suspended sediment transport above those of control (reference) watersheds. About 85 % of the data analyzed represents snowmelt-runoff events. Factors controlling sediment generation and export from a watershed include geologic structure, soil properties, topography, vegetation, land use, temporal and spatial distributions of precipitation, and streamflow generation mechanisms. It is difficult to combine all of these factors into one reliable expression for estimating sediment discharge from a watershed or to isolate the individual effects of these factors on sedimentation process (Lopes and Ffolliott 1992a, 1993a). One method of analyzing the effects of land-use practices on sediment discharges is through interpretations of a sediment-rating curve relating sediment concentrations to streamflow discharge (Shen and Li 1976, Lopes and Ffolliott 1993b, Brooks et al. 1997).

A sediment-rating curve reflects the pattern of soil erosion and sediment delivery operating in a watershed, and provides a readily accessible starting point for investigating the impacts of land-use practices on sediment discharge. Sediment-rating curves have been used for estimating sediment discharge from large watersheds (Livesey 1975, Elliott and Defeyter 1986, Hansen and Bray 1993) and small-to-medium-size watersheds (Piest 1963, Sidle and Campbell 1985, Lopes and Ffolliott 1993b). These curves can be used along with streamflow-frequency data (flow duration curves) to calculate sediment yields by the flow duration-sediment-rating curve method (Crawford 1991).

There were differences in sediment rating curves among the treated pinyon-juniper watersheds and the control (Lopes et al. 1996). These rating curves were developed with 500 to 600 sediment samples. The main difference was higher sediment concentrations from the cabled watershed than the control watershed from similar streamflow discharges. Higher concentrations of suspended sediment on the cabled watershed were likely a reflection of the soil disturbances caused by uprooting trees in the cabling treatment. There was also a difference between sediment rating curves derived for the herbicide-treated watershed, which experienced little soil disturbances as a result of treatment, and the control. Table 3.2 shows minimum, mean, and maximum values of streamflow and suspended sediment concentration for pinyon-juniper watersheds on Beaver Creek.

Table 3.2. Minimum, mean, and maximum values of streamflow and suspended sediment concentration for pinyon-juniper watersheds.

While these differences in rating curves are statistically significant, suspended sediment concentrations on the Beaver Creek watersheds are relatively low (Table 3.2). This finding is not surprising because erodibility of the volcanic soils on Beaver Creek is inherently low and, therefore, the sediment supply is limited as well as energy limited (Lopes and Ffolliott 1993a, Baker 1999b). More than 50 % of the ponderosa pine and pinyon-juniper woodlands in the southwestern United States are found on soils of similar parent material.