Wind and Water Processes

C H A P T E R

4 Wind and Water Processes O U T L I N E 4.1 Wind Erosion 4.1.1 Saltation 4.1.2 Vegetation 4.1.3 Wind and Dead Plant Materials

73 74 75 78

4.2 Water—Redistribution of Rainfall 4.2.1 Interception and Stemflow 4.2.2 Throughfall

80 81 84

4.3 Splash Erosion—Kinetic Energy of Raindrops

85

4.4 Infiltration and Run-Off 4.4.1 Vegetation Effects 4.4.2 Run-Off 4.4.3 Biological Soil Crusts 4.4.4 Physical Crusts 4.4.5 Animals 4.4.6 Rocks and Stones

86 87 88 89 90 91 94

4.5 Exchanges Among Landscape Units 94 4.5.1 Ephemeral Streams (Arroyos, Dry Washes, Wadis) 95 4.6 Exchanges Within Landscape Units 4.6.1 Rills 4.6.2 Steep Hillslopes

96 97 98

4.7 Modeling Overland Flow and Ephemeral Channel Flow

98

4.8 Episodic Events

99

4.9 Ephemeral Ponds and Lakes

101

Summary

102

References

103

Further Reading

107

4.1  WIND EROSION Visitors to arid regions are frequently impressed by the swirling, funnel-shaped winds carrying dust, sand, and plant litter across the surface (dust devils in the United States and willy-willys in Australia). Dust devils or dust whirlwinds are produced by thermals that suck up soil and plant fragments and transport these materials for various distances across the landscape. Dust devils are an interesting manifestation of an important and pervasive abiotic process. Aeolian (wind) erosion is a feature of arid regions not shared with more mesic environments. The probability of wind erosion affecting soil loss decreases exponentially as

Ecology of Desert Systems https://doi.org/10.1016/B978-0-12-815055-9.00004-7

73

© 2020 Elsevier Ltd. All rights reserved.

74

4.  Wind and Water Processes

rainfall increases. Most dryland ecosystems have characteristic windy seasons during which high velocity wind “storms” persist for hours and occur at high frequency. While some undetermined fraction of atmospheric dust comes from uninhabited desert, most dust storms originate in marginally arable semiarid areas due to erosion of surfaces no longer protected by vegetation, lag gravels, or stones. The largest emitters of dust are dry washes (13.8–0.007 mg m−2 s−1), dunes, playa lake margins, distal alluvial fans, and lacustrine beaches (Sweeney et  al., 2011). Low dust emissions features include salt-crusted playas, silt-clay-crusted playas, and desert pavements. Emission rates can vary by three orders of magnitude depending upon variability in soil texture, vegetation height and density, and presence of surface crusts. Unmeasured contributors to wind erosion include accumulations of soil from digging by animals, deposits of soil around ant nest entrances, and termite foraging galleries and sheeting which are easily entrained and moved by wind. Soil ejected by burrowing animals and digging by animals searching for food dries rapidly and loses aggregate stability, which contributes to entrainment of the effected soil at low wind velocities. In arid and semiarid regions the relatively large volumes of soil moved by vertebrates and invertebrates (Whitford and Kay, 1999a, b; Whitford, 2000a, b) are probably significant sources of saltating sand grains and subsequent wind erosion.

4.1.1 Saltation Aeolian processes are initiated when the erosion power of the wind exceeds the erodibility of the surface. Erodibility of the soil surface is generally related to the threshold shear velocity defined as the velocity below which saltation (transport flux of sand particles) will not occur. Soil erosion and dust generations are a function of the areal extent of bare soil, soil moisture, and extent and type of vegetation. Threshold shear velocity for soils is a function of soil texture and coverage of protective crusts and soil moisture (Okin et al., 2006). They report that particles of 80–120 μm are effective as saltating particles with low threshold shear velocity. Some soils may be supply limited in that the quantity of saltation size particles is insufficient to sustain the maximum theoretical flux (Gillette and Chen, 2001). Atmospheric relative humidity controls the moisture of soil surfaces especially in arid or semiarid regions during time periods of no rain. Soil moisture affects the adhesive forces among particles which affect the threshold velocity required to dislodge soil particles. Thus there is a complex relationship between air relative humidity, particle size, and soil erodibility (Ravi et al., 2004). Wind transports soil particles by three mechanisms that are partially related to soil particle diameters but the particle sizes involved in the mechanisms overlap. Surface creep occurs when soil particle diameters are >500 μm and move short distances in the direction of the wind. Saltation is generally the mechanism of wind transport for soil particle diameters ranging from 20 to 500 μm. Saltation occurs when the shear velocity of the wind exceeds the threshold shear velocity of the soil. This results in soil particles bouncing across the soil surface. When bouncing particles impact the soil, the energy is transferred to other soil particles dislodging the particles and either adding to the volume of saltating particles or pushing small soil particles into the wind stream where they are subject to vertical or horizontal transport. Surface creep and saltation contribute to most of the mass movement of soil on a local scale (less than several meters). Small soil particles <20 μm in diameter are generally suspended in the wind stream and may be transported over long distances on a regional, continental, or



75

4.1  Wind Erosion

Windborne dust particles

Accretion mound

Saltating sand grains

FIG. 4.1  Saltating sand grains accumulate under shrubs producing shrub mounds.

even global scale (Field et al., 2010). Soil nutrients like nitrogen and phosphorus and organic matter are frequently associated with smaller soil particles, soil fertility in dust source areas becomes depleted while fertility in sink areas is enriched (Field et al., 2010) (Fig. 4.1). Cohesive resistance to wind erosion varies directly with the percentage of the soil volume made up of aggregates (Leys and Raupach, 1991). Chepil (1951) indicated that as the fraction of nonerodable aggregates increased, erodibility decreased. The clay fraction of the soil contributes directly to the formation of aggregates (Tisdall and Oades, 1982). Erodibility of soils is greatest in soils with less than 10% clay content, minimal erodibility in soils with clay contents of 20%–30%, and increasing erodibility of soils with clay contents >30% (Chepil, 1953). In the Mojave Desert (U.S.A.) the largest producers of dust are dry washes (13.8–0.007 mg m−2 s−1) followed by dunes, playa margins, distal alluvial fans, and lacustrine beaches. Low dust emitters include salt-crusted playas (0.7–0.002 m−2 s−1), silt-crusted playas, and desert pavements (Sweeney et al., 2011). High dust emissions are primarily a function of saltating sand that bombards the surface producing dust-sized particles for entrainment in the wind stream. Low dust emissions are primarily a function of surface crusting, surface gravels and small stones, and vegetation density (Sweeney et al., 2011).

4.1.2 Vegetation Sand moved by saltation is frequently deposited in dunes or dune-like structures around plants or other physical objects. In undisturbed deserts, even strong winds in excess of 30–50 km h−1 will not usually generate air-borne dust because of vegetative cover (Musick et al., 1996). Even sparse vegetation reduces wind erosion by its effect as a roughness factor that reduces velocity by generating turbulence. Van de Ven et al. (1989) developed an equation to predict soil loss as a function of wind velocity, soil type, and vegetation characteristics. They found that if any of the vegetation parameters, density (N), height (H), and diameter (D) were increased, soil loss decreased by the square root of that factor. Their generalized equation is as follows: SL = K (Uh – Ut ) / √ ( NDH )

76

4.  Wind and Water Processes

where SL is soil loss, K is an empirically derived constant, and Uh – Ut is a velocity parameter specific for a soil’s resistance to erosion. The more resistant a soil is to erosion, the higher UT becomes for a particular free-stream wind velocity. N is the number of stems per unit area, D is average stem diameter, and H is the average height of stems above the soil surface. Even dead plants provide a stabilizing feature of soils in arid regions. Grazing or browsing by domestic or native herbivores that reduce the height and/or density of grasses or shrubs will increase soil loss by wind erosion in dryland areas. A variable not accounted for in this model is the porosity of vegetation cover. A sparse multistemmed mesquite plant provides less wind resistance and less protection than a dense multistemmed shrub of the same height and diameter. Compact desert shrubs and grass tussocks provide more wind resistance than large multistemmed shrubs with open canopies. Removal of vegetation by burning reduces the aerodynamic roughness by two orders of magnitude and reduces the average shear velocity by about 60% (Wiggs et al., 1994). They concluded that the relictual nature of the Kalahari dunes was attributable to the extensive cover by grasses and other vegetation. Movement of Kalahari dune sands could only occur immediately after loss of vegetation by fire, drought, or overgrazing (Wiggs et al., 1994). Reduction in canopy height and cover of vegetation plus disruption of the soil surface by grazers especially during droughts and periods of reduced rainfall can create small erosion cells (modified from Pickup, 1985), that is, unvegetated patches of sufficient fetch length for wind to exceed threshold velocities of a particular soil. Continued movement (saltation) of soil particles across an erosion cell will bury or abrade vegetation on the downwind side of the barren patch. Barren patches in desert landscapes are also subject to water erosion, which exacerbates the increase in size of the cell. Eventually erosion cells will coalesce into a large barren area unless colonized by vegetation. Erosion cells that pass some size threshold behave as unvegetated cultivated fields with respect to wind erosion. Many of the present desert landscapes developed as a result of erosion cells that rapidly increased in size and coalesced as a result of intense grazing during periods of drought. Well-developed biological soil crusts composed of cyanobacteria, green algae, lichens, and mosses reduce wind erosion by increasing the threshold velocity required to dislodge soil particles. Foot and vehicle traffic and livestock break up soil crusts. Areas with broken-up soil crusts are frequently exposed to wind speeds that exceed the velocity thresholds of damaged crusts (Belnap and Gillette, 1998). Because biological soil crust organisms are concentrated in the top 3 mm of soil, sand blasting by wind can strip the crusts from the surface. Loss of soil crusts opens up sites to wind and water erosion and subsequent loss of productivity. The most important variable affecting wind erosion in noncultivated deserts is the fetch length of unvegetated patches. A study of wind velocities and fetch lengths in different vegetation assemblages that resulted in dust generation revealed that in a tussock grassland and low height, stoloniferous grassland (burrograss) with patches of bunch grass (tobosa) on fine-textured soils, the minimum fetch length for detachment was approximately 50 cm and the threshold velocities for fetch lengths between 50 and 300 cm ranged between 25 and 52 km h−1 (Herrick, unpublished data). In mesquite coppice sand dunes (nebkhas) with a high degree of surface roughness due to variations in dune heights and coppiced stems, the minimum fetch length for saltation was approximately 500 cm. Threshold velocities ranged between 18 and 35 km h−1. In these vegetation assemblages with different height, cover, and soil characteristics, the relationship between fetch length and threshold velocity appeared to be dependent upon soil characteristics in addition to fetch length.



4.1  Wind Erosion

77

A study examining the effects of crusting, gravel-stony surface, and vegetative cover on wind-blown sediments with collectors at heights of 5, 10, 20, 50, and 100 cm above the soil surface found no difference in amounts of sediments collected at 5 and 10 cm above the soil surface. The total mass of aeolian sediments in collectors decreased with increasing height above the surface. Particle size distribution also decreased with smaller particles in the higher collectors. Where physical crusts were abraded by saltating sand grains, sediment yields were equal to a coarse sand site with minimal crusting. In areas with vegetative cover greater than 20% there were no differences in the total amount of sediments collected. At a site with gravel- rock fragments and sparse shrubs wind-blown sediment yields were similar to areas with high vegetative cover (Hupy, 2004). When an aeolian transport model was modified by empirical data to account for how vegetation affects the distribution of shear velocity on the surface rather than calculating the average effect of vegetation on shear velocity, the model performance predicted horizontal aeolian flux with a low error. This study showed that the impact of vegetation on aeolian transport depends on the distribution of the vegetation rather than average lateral cover. Vegetation impacts surface shear stress locally by reducing shear in the immediate lee (protected opposite side from wind direction) of plants rather than by changing the threshold shear velocity of the bulk surface (Li et al., 2013). Their results also suggest that threshold shear velocity is exceeded more than estimated by single measurements of threshold shear velocity and roughness length commonly associated with vegetated surfaces (Li et al., 2013) Resistance to wind erosion is a function of the adhesive strength of the soil and the presence of surface protectors such as rocks, plant litter, physical crusts, and biological crusts. Accumulations of plant litter (leaves and stems) under shrubs or behind obstructions on the landscape are generally like mats of material held together by fungal hyphae or microbial secretions. Rocks and plant litter masses too large to be moved by wind offer the best soil protection (Okin et al., 2006). Physical soil crusts are formed by the binding together of silt and clay particles which are wetted and then dried to form a crust. Physical soil crusts prove higher resistance to erosion than soils dominated by coarse sand particles. These crusts protect soils unless disturbed by animals or saltating particles dislodged from an unprotected soil. Biological soil crusts are composed of cyanobacteria, lichens, and mosses which stabilize soils by the excretion of mucilaginous materials that bind surficial soil particles together. Bound soil particles increase soil aggregate size and increase soil resistance to wind (Field et al., 2010). An Australian study using a portable wind tunnel on two soil types, a loamy soil with 90% coverage by cryptogamic crusts and a sandy soil with 76% coverage by biological soil crusts, found that a crust cover of approximately 20% is needed to keep sediment transport below a defined erosion target (Eldridge and Leys, 2003). Vegetation affects erodibility of soil surfaces in three ways: (1) directly shelter soil from the force of wind by covering some part of the surface and by producing a leeside wake in which wind speed is reduced, (2) vegetation can reduce the velocity wind by the canopies of shrubs thereby reducing the erosive power of the wind, and (3) vegetation can trap wind-borne particles thereby reducing the total horizontal and vertical flux and providing loci for sediment deposition (Okin et al., 2006). Because of turbulent flow in and around plant canopies, the ability of wind to carry sediments is reduced and wind-blown materials are deposited under plant canopies. The ability of plant canopy to trap wind-borne sediments is in part a function of its porosity. Wind tunnel and theoretical studies have suggested that vegetation with

78

4.  Wind and Water Processes

­ orosities of 20%–40% have the maximum effect (Okin et al., 2006). Wind flow around grasses p and shrubs is an important feature of wind-driven processes in arid ecosystems. A study of turbulent wind flow around grass clumps and shrubs found that wind velocity was reduced by 70% in the immediate lee of grass clumps and 40% in the lee of shrubs. However, velocity recovered to the same equilibrium over the same relative distance with both shrubs and grass clumps at approximately nine element heights downwind. The difference in reduction of wind velocity was attributed to differences in porosity of the vegetation, 53% porosity for the grass and 69% for the shrub. Therefore air bleed through the shrub was responsible for the velocity reduction differences (Mayaud et al., 2016). They conclude that grasses have a more sheltering effect than shrubs. The importance of different life forms of vegetation resisting wind erosion and the effects of wind erosion on downwind vegetation were elucidated in an experimental study with replicated 25 × 50 m plots where all grasses, half-shrubs, and perennial forbs were hand removed (Alvarez et al., 2012). These plots were compared with control plots with no vegetation removal or manipulation. There were plots immediately downwind of the grass removal plots and the controls (Alvarez et al., 2012). Sediment fluxes increased on the grass removal areas plus soil nutrients and seed bank were depleted. Grass cover on plots downwind of vegetation removal decreased over the 4-year study despite above-average rainfall during the term of the study. Grass cover increased in the plots downwind of the control areas (Alvarez et al., 2012). When the effects of enhanced wind erosion on surface particle size distribution and carbon and nitrogen contents of wind-blown sediments were examined, the soils of the grass removal plots were found to be coarser (Li et al., 2009). The content of coarse (250–500 μm) soil particles increased but the content of soil particles in the 50–125 μm and <50 μm was significantly depleted. There was enrichment of carbon and nitrogen in wind-blown sediments and fine particles (<50 μm) were enriched in sediment collected at 50 and 100 cm above the surface to a higher degree than carbon and nitrogen. They found that nearly 12% of the total organic carbon and 9% of the total nitrogen were found in the particles with diameters <50 μm (Li et al., 2009). This study demonstrated that the loss of fine soil particles by wind erosion in arid areas with no grass cover and sparse shrub cover may occur over a short period of a few years resulting in loss of soil N and organic C. Sand burial and wind erosion are important stresses on plants in arid regions. Sand burial and wind erosion had significant effects on the growth and regeneration of the clonal shrub, Calligonum arborescens (Luo and Zhao, 2015). The number of ramets and biomass production decreased with burial depth and with severe erosion. Number of ramets and biomass production increased with moderate erosion and burial. Severe erosion and sand burial had negative effects on C. arborescens while moderate wind erosion had positive effects (Luo and Zhao, 2015).

4.1.3  Wind and Dead Plant Materials In terms of ecosystem function, aeolian transport of organic matter may be of greater importance than transport of soil. Soil losses resulting from disturbance are reduced over time as stabilization by vegetation or as lag gravels, and stones form a desert pavement. However, dead plant materials are susceptible to redistribution by wind and water. The redistribution of organic materials contributes to the development of the well-known “islands of fertility”



79

4.1  Wind Erosion

around desert shrubs (DeSoyza et al., 1997). Organic matter accumulations under shrubs are related to the morphologies of the shrubs, shrub densities, and productivity of the surrounding areas. Because of the small size and mass, much of the organic debris, the threshold velocity to dislodge and entrain leaves, stems, and fruits is quite low. Applying aerodynamic theory to the shape and mass of creosotebush, Larrea tridentata leaves, stems, and fruits, DeSoyza et al. (1997) calculated threshold velocities of 1.1 m s−1 (3.96 km h−1) for leaves, 1.72 m s−1 (6.2 km h−1) for fruits, and for a 25 mm stem segment 1.1 m s−1 (3.96 km h−1). Once entrained in a wind stream, leaves, small stems, and so on, are carried until turbulence and eddies reduce the wind velocity and the material is deposited. Turbulence develops as a result of resistance to flow by objects in the air stream path. In sparse shrubland, plants of different morphologies differ greatly in their effectiveness as litter “precipitators” (Fig. 4.2). Shrubs with a hemispherical shape accumulate more wind-borne litter than shrubs shaped like inverted cones. As shown in Fig. 4.2, based on aerodynamics theory, the streamline flow over and around a hemispherical shrub will cause a small local turbulence under the canopy, which will result in fragment deposition but not entrainment of that material. When that air flow pattern is compared to a conical-shaped shrub, the downflow will accelerate the flow of air around the base of the cone entraining litter in that flow. Although this analysis was based

Hemispherical shrub.

Upper flow

Turbulence at leeward side of shrub enhances deposition of litter

Inverted-cone shrub

Upper flow Side flow

Down flow with increase in velocity results in litter removal Little or no turbulence on leeward side of shrub

FIG. 4.2  Flow paths of wind encountering shrubs with cone-shaped canopies. Upper panel, example of a shrub with a hemispherical crown. Turbulence on the leeward side of hemispherical crowns causes deposition of litter. Lower panel, high wind velocity due to downflow at the base of conical shrubs causes turbulence and litter to be entrained in the wind stream. From DeSoyza, A.G., Whitford, W.G., Martinez-Meza, E., Van Zee, J.W., 1997. Variation in creosotebush (Larrea tridentata) morphology in relation to habitat, soil fertility and associated annual plant communities. Am. Midl. Nat. 137, 13–26. Reprinted with permission by The American Midland Naturalist, University of Notre Dame Press.

80

4.  Wind and Water Processes

on idealized leaf, stem, and fruit shapes and shrubs reduced to simple geometrical shapes that were perfectly symmetrical, the results provide an explanation for the observed patterns of litter accumulation and soil mounding under L. tridentata shrubs in the Chihuahuan Desert of New Mexico and are probably generally applicable to other shrub-dominated ecosystems.

4.2  WATER—REDISTRIBUTION OF RAINFALL Although wind is an important force in structuring of patches in the landscape and in the concentration of organic matter in some patch locations, water may be an equally important or more important force affecting patch structure and dynamics. Because water is essential for biological activity, those variables that affect redistribution of rainfall across a landscape are critical variables. The fate of rain falling on vegetation is very different from that falling on bare ground or rock. Rainfall can be partitioned into different units depending upon the pathway taken by the rain on its way to the soil or back to the atmosphere (Fig. 4.3). Gross rainfall is the amount of water measured in the open or above the canopy of the vegetation. Interception water is the amount of rainfall per event retained by the canopy or by litter and which is evaporated without adding moisture to the soil. Throughfall is that portion of the rainfall directly reaching the soil surface or litter through spaces in the canopy and as drip from leaves, twigs, and stems. Stemflow is water intercepted by the canopy and directed down

Ste

mf lo

w

Interception

Throughfall

Soil wetting front Root channelization

FIG. 4.3  Partitioning of rainfall and the fate of rainfall on patches in a desert ecosystem.



4.2  Water—Redistribution of Rainfall

81

the stems to the mineral soil or to the root crown. Infiltration is the movement or passage of water into the soil. Percolation is the movement of water through a soil column. Run-off is that portion of intercepted water that does not infiltrate into the soil. A simplified water balance for a landscape can be expressed by: Precipitation = Evapotranspiration + Run-off + Storage. In many landscapes there may be additional inflows and outflows but the principle of balance still applies. The hydrological landscape can be envisioned as a series of layers where different types of hydrologic flows and storages occur (Ferguson, 1991-92). The atmospheric layer includes fluxes of water vapor across, and to and from the land surface. The surface layer includes overland flows and surface water channels and bodies. The vadose layer includes soil moisture stored in soil voids or in transit toward groundwater, plant roots, or back to the surface. Vadose water is characterized by capillary tension (soil water potential). Vadose water supports plant growth and provides the hydraulic head for maintaining stream base flow. The phreatic layer includes water in any subsurface saturated zone either being stored or in transit toward deep aquifers (Ferguson, 1991-92). The hydrologic structure of arid and semiarid landscapes varies considerably from place to place. Water infiltrating landscape units are formed of permeable materials with or without a significant soil mantle and are formed of sandstone, carbonates, or unconsolidated material. Water spreading landscapes are formed of essentially impermeable materials including impermeable bedrock with little overlying soil. Lithology includes shale, slate, and granites. Drainage is mostly by surface run-off. Infiltration, subsurface storage, and subsurface flows are small. Many landforms are intermediate with impermeable bedrock overlain by a deep layer of permeable soil. These are the hydrologic units that characterize the basin and range topography of many desert regions. The growth form of the vegetation of a given landscape unit affects the distribution patterns of water and nutrients within that unit as well as the nature and quantity of nutrients and water exported to adjacent units. Factors affected by the growth form of the vegetation include (1) canopy interception of water, stemflow, leaf drip, and evaporation from leaf surfaces; (2) infiltration and run-off; (3) organic matter transport; (4) water movement in the soil; and (5) soil and organic matter accretion or loss.

4.2.1  Interception and Stemflow Many arid zone plants are morphologically well suited to promote stemflow (Fig. 4.4). It has been reported that up to 40% of the incident rainfall on an area in Australia covered by mulga (Acacia spp.) canopy was channeled down the main stems and entered the soil with minimal outwash at the bole (Pressland, 1976). Mallee (Eucalyptus spp.) has the inverted cone morphology and redirects 30% of the water in a rain event down the stems. That stemflow water is redirected along live root channels to depths as great as 28 m (Nulsen et al., 1986). Martinez-Meza and Whitford (1996) documented the importance of stem angle (with respect to the horizontal) and stem length as variables affecting stemflow in desert shrubs. In the three shrubs studied, stemflow yields were higher in shrubs with upward oriented branches than in shrubs with horizontally oriented branches. In mesquite (Prosopis glandulosa) maximum stemflow amounts were recorded on stems with stem angles of 70–75°. In tarbush (Flourensia cernua) there was some stemflow recorded from branches with stem angles less than 40° but maximum stemflow volumes were recorded from branches with stem angles

82

4.  Wind and Water Processes

FIG. 4.4  Morphologies of some desert shrubs and trees that facilitate stemflow and redistribution of stemflow to deep storage via live root channelization. (A) Mallee eucalyptus, Eucalyptus spp.; (B) mulga, Acacia aneura; (C) creosotebush, Larrea tridentata; and (D) Acacia spp.

greater than 45° and stem lengths greater than 70 cm. In tarbush, stem angle and stem length accounted for 50% of the variability of the collected stemflow. In creosotebush, L. tridentata, stem angle and stem length accounted for 41% of the variability in stemflow. Stemflow volume was maximum at stem angles equal to or greater than 65° and there was little stemflow from branches with stem angles <50°. This study reemphasizes the importance of morphology of the dominant vegetation as a factor affecting water redistribution. The importance of shrub morphology and stemflow water was reinforced by DeSoyza et al. (1997). They found that L. tridentata populations from the driest part of its range (Death Valley, California) were composed of plants with exterior stem angles >45°. Also, a population of young creosotebushes was almost entirely plants with exterior stem angles >45°. Populations of creosotebushes in the Chihuahuan Desert were composed of plants varying from inverted cone-shaped plants to hemispherical-shaped plants. Water that moves down the stems of shrubs and trees may follow live root channels and be stored at depths in the soil far below the rooting zone of ephemeral plants or perennial herbs and grasses that grow beneath the canopies of the shrubs. Nulsen et al. (1986) examined wetting and drying profiles and movement of dye. They found that stemflow water followed live root pathways and there was no evidence of that water moving along voided root channels, cracks in the soil, or channels occupied by dead roots. Although the soils in the area of their study had low hydraulic conductivity, the soil at 4.5 m under mallee gained water rapidly after a 50.5 mm rainfall. The hydraulic conductivity of the soils was not high enough to account for the apparent rate of moisture transmission even if the soil water flow was assumed to be saturated. The study by Nulsen et al. (1986) provides reasonable evidence that live roots enhance the movement of water into the deep soil profiles. Martinez-Meza and Whitford (1996) used a fluorescent dye on the stems of the shrubs to trace the fate of stemflow water resulting from simulated precipitation. Stemflow water moved to 40 cm when the soil wetting front was only 6 cm (Fig. 4.5). In L. tridentata, there is also a dense network of short-fine roots, originating from the root crown, in the upper 10 cm



4.2  Water—Redistribution of Rainfall

83

Soil depth (m)

0.1 1

3 0.6

7

2

6

5 4 1.00 (m)

Dye stained root Dye stained soil

FIG. 4.5  Rhodamine dyed roots and soil under creosotebush (Larrea tridentata) from dyed stems on shrubs subjected to simulated rainfall of 40 mm. From Martinez-Meza, E., Whitford, W.G., 1996. Stemflow, throughfall and channelization of stemflow by roots in three Chihuahuan Desert shrubs. J. Arid Environ. 32, 271–287.

of soil. This root distribution tends to maximize the water availability to this set of shallow roots at the stem base of the plant thus providing a relatively large quantity of water concentrated in the small area around the bole of the plant. The fluorescent dye technique did not provide information on the quantities of stemflow retained in the shallow soil compared with the quantity transferred to deeper soil by root channelization. In trees such as mallee, mulga, other acacias and shrubs such as creosotebush, water redirected to the root crown by stemflow may follow the channels provided by the living roots to depths in the soil greater than possible by infiltration. Relatively large amounts of creosotebush roots have been recovered from below indurated calcium carbonate layers. In excavations that cut through indurated layers, roots were found growing through the layers. Small channels that may be the result of termites excavating access tunnels also transmit water through indurated layers of calcium carbonate. A study of stemflow of Caragana korshinskii in the Shapotou desert in northwestern China found that 2.2 mm of precipitation was needed to initiate stemflow and that stemflow amount averaged 8% of the gross precipitation (Wang et al., 2011). They concluded that stemflow and root channelization were responsible for positive moisture balance in the root zone and for the replenishment of soil moisture in deeper soil layers. Stemflow accounted for 8.8 and 2.8% of the gross rainfall in two species of shrubs important in revegetation of degraded desert areas of China (Zhang et al., 2013). They reported that stemflow exhibited a linear increase with increasing rainfall depth. Stemflow at low rainfall intensity (<2 mm h−1) decreased at intensities >2 mm h−1. Ion concentrations in stemflow were higher than in throughfall and

84

4.  Wind and Water Processes

considerably higher than in bulk precipitation. The concentrations (nutrient enrichment ratios) of nitrogen, phosphorus, sodium, potassium, calcium, magnesium, chloride, and sulfate in stemflow were enriched between 122.8 and 1677.0 times for C. korshinskii and 12.6–1306.0 times for Artemisia ordosica in comparison to concentrations in bulk precipitation (Zhang et al., 2013). They concluded that stemflow has an important role in soil water and nutrient enrichment at deeper soil profiles of the root zone in revegetated arid ecosystems. Stemflow amount varies with canopy structure, intensity and duration of rainfall, and rainfall amount. It has been documented that stemflow increases with rainfall amount that exceeds the threshold value of stemflow generation. In that study, stemflow amount varied with canopy volume and stem diameter for one shrub, C. korshinskii but stemflow was not related to any of the morphological variables in A. ordosica. That study demonstrates the inherent variability of stemflow in shrubs with different canopy and stem morphologies and the need for more detailed studies of variables affecting stemflow and the effects of shrubs on infiltration and run-off. The fraction of bulk precipitation intercepted by shrub canopies that is funneled down the stems to the root crown varies as a function of stem density, stem angles (from the vertical), canopy area, leaf and stem morphology, and rainfall intensity. Concentrations of chemical nutrients in stemflow reflect species differences in retention of atmospheric dust deposition on leaf surfaces. The amount of stemflow at the root collar or base of stems that is transferred down root channels to deeper soil layers varies with rooting patterns and soil physical characteristics.

4.2.2 Throughfall The fraction of rainfall reaching the canopy of a plant that falls through the foliage or spaces in the canopy varies with storm depth. The percentage of the gross precipitation at the canopy surface that was measured as throughfall varied from 1% to 50%–70% at storm depths between 1 and 5 mm (Martinez-Meza and Whitford, 1996). At the lowest storm depths, the smaller fraction of gross rainfall that appears as throughfall is due to water retained on the leaf and stem surfaces and which is lost by evaporation. The throughfall asymptote was at 5 mm storm depth. At the asymptote, the throughfall percentage varied between 30% and 70% in creosotebush, 60% and 80% in mesquite, and 45% and 70% in tarbush. Throughfall drop energies are lower than the kinetic energies of raindrops that fall on surfaces not below vegetation cover (Brandt, 1989). Since raindrop energy is a function of drop size, raindrops intercepted by canopy leaves or stems are broken up and fall through the canopy as finer droplets. The drop size of throughfall water is independent of canopy characteristics and of rainfall intensity (Brandt, 1989). The reduced energy of throughfall water enhances the infiltration of the throughfall fraction. Small spheroid shrubs with dense foliage like many chenopods and tussock grasses are structurally capable of breaking up the large, high energy drops of intense, short duration rain storms, characteristic of summer convectional storm events of many hot desert areas. The density of foliage and/or fine branches of such plants close to the soil surface insures that fragments of leaves, and so on, will be trapped under the canopy. Rainfall intercepted by such canopies will enter the soil as a result of either leaf drip or stemflow with little or no transmission loss (Van Elewijck, 1989).



4.3 Splash Erosion—Kinetic Energy of Raindrops

85

4.3  SPLASH EROSION—KINETIC ENERGY OF RAINDROPS A raindrop impacts a bare soil surface something like a metal ball striking the surface. The energy of the raindrop impact is transferred to the soil particles directly under the drop. If the kinetic energy of the drop is sufficient, soil particles are detached from the surface and fly out in all directions. This is raindrop splash. Raindrop splash is the first step in erosion of soil. The volume of soil material detached by raindrop splash is determined by drop diameter, median grain size of the soil, and surface slope (Poesen, 1985). For a constant kinetic energy, the volume of splashed soil is not influenced significantly by diameter of raindrops nor of fall height of the raindrops except for low energy rainfall. Therefore kinetic energy is used to express the erosivity of rainfall in splash erosion equations. The kinetic energy of rainfall is translated into detachment of soil particles. Detachment of soil particles varies with the kinetic energy of raindrops and resistance of the soil to detachment. Empirical studies have shown that bulk density (mass per unit volume) of soil is the most significant parameter of resistance to detachment. Poesen (1985) provides an equation that expresses the total volume of soil detached by rainfall as: Vtot = R−1 ( B.D.)

−1

( K.E.)

where R is the resistance of the material to drop detachment (J kg−1), B.D. is the bulk density of the soil (kg m−3), and K.E. is the kinetic energy of the rainfall (J m−2 yr−1). Other factors are variables that affect the detachability of loose soil materials (Poesen, 1985). Slope may or may not affect detachability. The occurrence of positive relationships between slope and detachability probably depends on other factors such as texture and structure of the surface soil. The relationship between silty and clayey soils and volume of soil detached is greater than for sandy soils. Another variable that affects detachability of sloping soil is the effect of wind and the oblique trajectory of the raindrops that impact the soil surface. The variables other than those in the general model for total volume of soil detached are minor contributors to the variability in soil detachment. Once soil particles are detached by raindrop splash, their contribution to sediment transport is determined by the particle size of the detached grains of soil. Variation in transport as a function of time is explained by changing surface conditions such as surface water content, liquefaction, and water-film development (Poesen and Savat, 1981). The reduction in kinetic energy of rainfall that is intercepted or deflected by plant foliage is substantial. In rainfall simulation experiments, Wainwright et al. (1999a, b) found that there was a 10% reduction in the intensity of subcanopy rainfall. They reported that the kinetic energy of the subcanopy rainfall was reduced to 70% of that hitting unprotected soil. This reduction in kinetic energy contributes to the development of mounds under the canopies of trees and shrubs. Raindrop impacts cause both sand and litter particles to be displaced (Dunkerley and Geddes, 1999). They reported that the mass of sand displaced by water drops increased with the height from which the drop originated but the mass of litter particles moved did not vary with fall-height of water drops. Rates of sand moved by water droplets were greatly reduced by surface litter. The effect of undershrub litter is to limit soil splash which may contribute to the persistence of shrub mound microtopography that is common in desert shrublands (Dunkerley and Geddes, 1999). The effect of gravity drops from shrub canopies

86

4.  Wind and Water Processes

was ­confirmed by Hoffman et al., 2013, who reported that most sediment displacement resulted from raindrop splash erosion which was reduced by the shrub canopy and by annual plant cover. They concluded that shrubs and herbaceous annual plants play a significant role in shrub-mound growth and maintenance. The role of splash erosion in the formation and maintenance of shrub mound microtopography has been described as a stochastic advection–dispersion process (Furbish et  al., 2009). They suggest that gradients in raindrop intensity and thus grain activity (the volume of grains in motion per unit area) can be as important as gradients in grain concentration and surface slope as factors affecting transport. Growth of mounds below shrub canopies occurs where differential rain splash causes more sand grains to be splashed inward beneath the protective canopy than outward. This involves “harvesting” of nearby soil material. Because inward sand grain flux associated with differential rain splash is sustained over the lifetime of a shrub, the mound material is kept from erosional processes that would move this material downslope (Furbish et al., 2009).

4.4  INFILTRATION AND RUN-OFF Infiltration, which is defined as the rate at which water enters a soil column, is an important determinant of water availability. There are a large number of variables that affect infiltration rates: vegetation cover, soil surface cover, soil texture, soil bulk density, abundance and distribution of macropores, and antecedent moisture. In arid and semiarid regions factors such as extent of bedrock outcrop, areal extent of soil cover, slope area, and differences in bulk density between channels of rills and small drainages and interfluve areas all contribute to variability in infiltration and run-off (Berndtsson and Larson, 1987). Antecedent rainfall has a greater effect on infiltration and run-off than rain event magnitude, average intensity, maximum intensity, and duration (Istok and Boersma, 1986). In deserts where low intensity rainfall and frequent small magnitude rainfalls are characteristic of wet seasons, antecedent rainfall is an important consideration. Morphology of plant canopies and foliage densities affect infiltration rates by changing drop size and by the quantity of litter trapped below the canopy. Infiltration rates are higher under shrub or tree canopies than in intershrub or intertree spaces (Lyford and Qashu, 1969; Elkins et al., 1986) (Fig. 4.6). Infiltration rates are highest near stems of plants even on coppice dune sites. Coppice dunes have exceedingly high infiltration capacities (sometimes three to four times greater than interdune spaces) (Berndtsson and Larson, 1987). If there is a visible litter layer under the canopy, this can contribute to higher rates of infiltration of throughfall when compared to similar-sized shrubs that have little or no litter layer. However, woody shrubs and small trees probably have the greatest effect on infiltration by breaking up large drops of water that produce splash erosion and by stemflow as discussed in the previous section. The soils under shrub or tree canopies generally have higher organic matter content and lower bulk densities than soils in intercanopy spaces. These characteristics favor infiltration of the throughfall water. They also affect the water-holding capacity of the soil because of the larger total capillary space resulting from the higher soil organic matter. Stemflow and the higher below canopy organic matter therefore combine to produce the higher infiltration rates measured on plots with shrub cover.

87

4.4 Infiltration and Run-Off

Infiltration rate (mm h–1)



125

125

100

100

75

75

50

50

25

Termites present (0.06)

Termites present (0.12)

Termites absent (0.06)

Termites absent (0.29)

25

shrub 5

open

10 15 20 25 30 35 40 45

5

10 15 20 25 30 35 40 45

Time (min)

FIG. 4.6  Comparison of infiltration rates from plots centered on shrubs and plots without shrubs. Elkins, N.Z., Sabol, G.V., Ward, T.J., Whitford, W.G., 1986. The influence of subterranean termites on the hydrological characteristics of a Chihuahuan Desert ecosystem. Oceologia 68, 521–528. With permission from the Ecological Society of America.

4.4.1  Vegetation Effects Small grasses occurring at low density appear to have little if any effect on infiltration and run-off. However, larger tussocks have marked effects on infiltration. Large grass tussocks with standing dead foliage break up large drops and channel water along the leaf surface to the base of the clump (Van Elewijck, 1989). Tussock grasses have a dense network of relatively fine roots immediately below the vegetative clump. The presence of this root mass, the higher organic matter content of soils below grasses, and the reduction in drop size and energy by canopy interception combine to increase infiltration in comparison to soil with no vegetation or soil with a sparse cover of small grasses and herbs. Tree and shrub canopies affect infiltration indirectly by their effect on certain soil properties, that is, bulk density and organic matter content. Lyford and Qashu (1969) measured infiltration rates at several distances from the plant centers of paloverdi trees (Cercidium microphyllum) and creosotebush (L. tridentata). They found infiltration rates nearly three times greater under the plants than in the intercanopy spaces. Since soil characteristics are important determinants of the type of vegetation that will develop in a given climate, it is difficult to assess the importance of soil characteristics per se on infiltration. Branson et al. (1981) summarize infiltration data for a variety of soils with predominantly grass cover. On sand or sandy soils with 65% vegetative cover, infiltration rates ranged between 42.4 and 79.5 mm h−1 compared with 36.8 mm h−1 on clayey soils. On dense clay, infiltration rate was only 12.7 mm h−1 but vegetative cover was only 40%. The only generalization that can be made is that sand or coarse soils have higher infiltration rates than fine-textured clayey or silty soils. Another soil parameter that affects infiltration is soil depth. Shallow soils, that is, soils overlying a relatively impervious argillic (clay) layer or indurated calcium carbonate hardpan will have reduced infiltration because the capillary space and pore space in the overlying soil quickly fills and percolation into and through the relatively impervious layer is slow. Thus shallow

88

4.  Wind and Water Processes

sand overlying caliche or clay will produce ponding very quickly in an intense storm whereas some storm intensity on that same land in a dune 1–2 m deep will not produce ponding. Rainfall on shrub-dominated landscapes reaches the ground either as stemflow or throughfall when intercepted by a shrub canopy. An experimental study using simulated rainfall found that once equilibrium was achieved 6.7% of the rain falling on the shrub canopy was funneled to the base as stemflow (Abrahams et al., 2003). Surface run-off rate is the proportion of the rate at which rain water arrives at a point. Analysis of data by multiple regression revealed that 75% of the variance in run-off was explained by three variables: the rainfall rate, ratio of canopy area to collar (shrub base) area, and presence or absence of subcanopy vegetation. Subcanopy vegetation accounted for 66.4% of the run-off. Therefore the density and extent of subcanopy vegetation is the single most important control of partitioning rainfall between run-off and infiltration under shrubs (Abrahams et al., 2003). A study of the intensity of subcanopy rainfall found that the kinetic energy as measured by leaf drip contributes only 10% of the subcanopy kinetic energy. Leaf drip makes up 28.9% of the subcanopy rainfall. Comparison of the effective kinetic energy (effective kinetic energy is energy of raindrops with sufficient energy to detach sediment) below the canopy with the kinetic energy of raindrops outside the canopy shows that the below canopy energy is 55% of the energy of raindrops outside the canopy (Wainwright et  al., 1999a, b). The differential splash contributes to the buildup of mounds under desert shrubs and contributes to the understanding of the development of resource islands in desert ecosystems.

4.4.2 Run-Off Numerous methods have been used to measure infiltration rates and all have limitations with respect to the interpretation and use made of the data (Branson et al., 1981). These methods include sprinkling infiltrometers, ponding infiltrometers, lysimeters, and run-off plots, which provide information about the infiltration characteristics of a small part of a landscape unit. The other end of the scale are the run-off data collected from instrumented weirs on drainages from whole watersheds that often include several landscape units. Anyone who has studied in semiarid or arid regions will quickly point out that there appears to be little relationship between plot infiltration rates and water loss from a watershed. Comparison of run-off from grassland and shrubland habitats showed that run-off was initiated at lower thresholds of rainfall in shrublands than in grasslands (Schlesinger et al., 2000). The run-off in shrublands averaged 18.6% of the rainfall over a seven-year period. Run-off from grassland plots varied between 5.0% and 6.3% over a five and one-half year period. Nutrient losses were higher from shrubland plots (0.33 vs. 0.15 kg ha−1 yr−1) because of higher run-off volumes. However, in these Chihuahuan Desert sites, there was a net accumulation of most nutrients due to inputs from atmospheric deposition (Schlesinger et al., 2000). Comparison of run-off and sediment yield from two desert grassland types on the same soil type and geomorphic surface found that water run-off was three times greater from burro-grass areas (a low-­growing rhizomatous grass) than from tobosa grass areas (a bunch grass with larger bare patches between plants) (Devine et al., 1998). Run-off generated by simulated rainfall on grassland, and shrubland communities produced hydrographs that showed that run-off begins earlier on vegetated plots than on unvegetated plots. Hydrographs from degraded grassland and intershrub plots rose continuously



4.4 Infiltration and Run-Off

89

throughout the 30 min simulated rainfall indicating that these plots do not achieve equilibrium run-off. The continuously rising hydrograph was attributed to development of raindrop produced surface seals (Neave and Abrahams, 2001). The hydrographs of most grassland and shrubland plots level out after the initial rise indicating that equilibrium run-off is achieved. Some shrub plots exhibited a decline in discharge which was attributed to the delayed infiltration of water into macropores beneath shrubs. This was attributed to vegetation in the understory of shrubs (Neave and Abrahams, 2001). Most undershrub macropores are probably the result of termite foraging tunnels in the litter accumulations under shrubs (Nash et al., 1999). Sediment concentration was not correlated with any surface property in the grassland or shrub environments (Neave and Abrahams, 2001).

4.4.3  Biological Soil Crusts In Australian chenopod shrublands, vegetative cover by shrubs and herbaceous plants is around 50% and the remaining soil surface has a cover of lichen crusts. Final steady-state infiltration rates with intact lichen crusts were < 10 mm h−1 and with the crust removed averaged 46 mm h−1 (Graetz and Tongway, 1986). This relationship held even at low rates of applied rainfall (28 mm h−1) when the infiltration rate with intact lichen crust was 7 mm h−1 and with crust removed was 14 mm h−1. However, the relationship between biological soil crust cover and infiltration is not simple and varies with the species composition of the soil crust. Some studies have shown that the presence of crytogamic crusts improves infiltration (Fletcher and Martin, 1948) or has no effect (Loope and Gifford, 1972; Eldridge et al., 1997). Removal of soil crusts only affected infiltration rates on plots where soil had been disturbed by livestock (Eldridge et al., 1997). Cyanobacterial sheaths and fungal hyphae may reduce infiltration in poorly developed soil crusts if they exclude water by occupying capillary pores (Greene and Tongway, 1989). The hydrophobic nature of microphytic crusts has been reported from sand dune soils in Australia and the Negev Desert of Israel and removal of these crusts greatly enhanced infiltration and reduced or eliminated run-off (Eldridge et al., 1997). Removal of cyanobacterial crusts on a loess floodplain had no effect on ponded infiltration (Eldridge et al., 2000). The absence of cyanobacterial crusts effects on infiltration on the loess floodplain was attributed to exposure of surface silts to water and subsequent clogging of matrix pores and sealing of the surface. These studies indicate that variation in the effects of biological soil crusts depends upon soil texture, species composition of the crust, and the impacts of domestic livestock. “Biological soil crusts are the dominant living cover in many drylands of the world” (Belnap, 2006). Biological soil crusts or microbiotic crusts influence many aspects of the hydrology of arid systems: soil porosity, absorptivity, roughness, aggregate stability, texture, pore formation, and water retention. The influence of soil crusts on these factors depends upon their internal and external structure (function of species composition) which varies with climate, soil, and disturbance history. Most studies that examine the effects of biological soil crusts on hydrology are done by comparing undisturbed sites with sites that were recently disturbed by researchers which complicates the interpretation of results. Applied disturbances have an effect on many soil features such as texture, roughness, aggregate stability, porosity, bulk density, and physical crusting. Such studies present conflicting results on how biological soil crusts affect water infiltration or run-off. In studies of biological soil crust

90

4.  Wind and Water Processes

types that use naturally occurring differences among the types, indicate that biological crusts in hyperarid regions reduce infiltration and increase run-off and have mixed effects in arid regions but increase infiltration and reduce run-off in semiarid cool and cold deserts (Belnap, 2006). Almost all studies show that biological soil crusts reduce sediment yields regardless of crust type or desert type. Microbiotic crusts play an important role in the hydrology of arid ecosystems by changing run-off and sediment yield. When studies of the effects of microbiotic crusts were compared on gypsum sand dunes, quartz sand dunes, and silty alluvial soils, plus crust removal at all sites, it was found that run-off was related to the combined effects of (a) parent material and (b) crust properties: species composition, microrelief, and exopolysaccharide content. No runoff was generated on the gypsum dunes on either the crusted or crust removal sites. Effective rainfall and crust microrelief was related to the amount of run-off on quartz sand and silty alluvial soils sites. It was concluded that soil parent material, species composition of the crust, and microrelief interact to affect run-off from microbiotic crusts in arid ecosystems (Kidron et al., 2012). Since biological soil crusts are prone to disturbance by humans and animals in many arid ecosystems, a simple classification of the stages of soil crust development has been constructed based on the external features of the crusts, which reflects their level of development (Belnap et al., 2013). The classification system has six levels of development classes from low (1) to high (6). There were large differences in infiltration, run-off, and erosion among crusts with different levels of development. Under dry antecedent conditions approximately 50% of the water applied ran off the lowest level of development plots whereas less than 10% ran off the plots of the two highest levels of development. Sediment loss was 400 g m−2 from the lowest level of development plots and almost zero from the higher level of development plots. Modeling indicated that erosion increases as slope length and slope gradient increase, especially beyond the threshold value of 10 m for slope length and 10% for slope gradient (Belnap et al., 2013). A study on soils of the Loess Plateau of China on artificially cultured biological soil crusts on different slopes showed that moss-dominated biological soil crusts will almost completely cover the soil surface in about 10 months. The moss crusts increase infiltration in proportion to the extent that the crusts cover the soil which decreases overland flow proportionately. Moss crusts are particularly important for protecting steep slopes (Xiao et al., 2011). Exopolysaccharides secreted by microbes in biological soil crusts affect the hydraulic conductivity of the crusts and their water retention. Exopolysaccharides in biological soil crusts affect their moisture capturing capacity and the uptake of moisture from the atmosphere was positively correlated with the amount of low molecular weight carbohydrates in the biological soil crust. Exopolysaccharides produced by microbes in biological soil crusts trap and retain moisture in sandy soils increasing water content of the crusts, reduce infiltration, and protect the soil from erosion (Colica et al., 2014).

4.4.4  Physical Crusts Another factor that affects infiltration in soils devoid of vegetation is the development of surface crusts (Tarchitzky et al., 1984; Moore and Sanger, 1990). These crusts are impervious in comparison to soils without crusts. They apparently form as a result of raindrop impact disrupting soil aggregates. This results in fine particles aligning to form a “skin,” 0.1 mm



4.4 Infiltration and Run-Off

91

thick and a 2–3 mm layer immediately below the “skin” with higher bulk density because soil aggregates have been destroyed (Tarchitzky et al., 1984). This hypothesis of crust formation was documented in sandy soils where the kinetic energy of rainfall was the major crusting factor for sandy soil (Valentin, 1991). Waterlogging and subsequent slaking of aggregates led to structural and depositional crusts in a clay-loam soil which resulted in reduced infiltration. The development of physical crusts results from breakdown of soil aggregates either by the kinetic energy of raindrops or by slaking of soil aggregates by standing water on fine- textured soils.

4.4.5 Animals Considering factors that affect water infiltration and influence the redistribution of water across a unit of landscape, the importance of the activity of animals must not be overlooked. The influence of large grazing herbivores either native or domestic has large effects on infiltration and run-off plus sediment losses and has been the subject of considerable research effort in arid regions (Graetz and Tongway, 1986). Obvious effects of grazing herbivores on infiltration are reduction in vegetative cover or changes in vegetative composition that affect infiltration and run-off as previously described. Somewhat less obvious are the effects that trampling by hoofed animals have on destruction of surface crusts and/or destruction of cryptogamic crusts. Disruption of the “sealed” crust resulting from raindrop impact can result in greater infiltration and less run-off. Disruption of cryptogamic crusts may increase or decrease infiltration and thereby affect run-off. Probably the most important factor affecting water infiltration is the abundance of soil macropores (1–5 mm holes (Anderson et  al., 1990)). It is only in the past 20–25  years that soil scientists have rediscovered the importance of soil macropores in the spatial distribution of water in soil and the effects of these pores on water infiltration (Phillips et al., 1989). Continuous pores or tubes transport flowing water to the deeper parts of the soil profile faster than predicted by theory. Macropores are widespread in many if not most ecosystems and are extremely important for infiltration in desert soils. Since the numbers of macropores per unit area need not be large to affect water infiltration, burrow/tunnel-constructing invertebrates are the most important contributors to macroporosity of desert soils (Fig. 4.7). The best-known sources of macropores are the burrows of earthworms (Lee and Foster, 1991). In arid and semiarid ecosystems, termites probably replace earthworms as the most important producers of macropores. Other invertebrates that produce pores in the soil include larvae of insects such as beetles and cicadas, burrowing spiders, and ants. In studies in the Sahel comparing plots from which termites were excluded with plots with termites present, Mando and Miedema (1997) reported an average of 88 ± 25 large voids (0.8–1.2 cm diameter) per square meter. No voids were detected in the termite excluded plots. They reported that termites accounted for more than 60% of the macropores in this region. Mando (1998) in a study of water infiltration on plots with and without termites found significant increases in infiltration on plots with termites (Fig. 4.8). In his study an average of 51.3% of the precipitation infiltrated on plots with termites but only 36.3% of the annual precipitation infiltrated on plots from which termites were excluded. Elkins et  al. (1986) reported on studies that utilized plots from which termites had been eliminated by a long-term, selective insecticide, chlordane. Infiltration rates on plots with scattered small grasses were 51.3 ± 6.8 mm h−1 on

92

4.  Wind and Water Processes

FIG. 4.7  A run-off plot centered on a creosotebush, Larrea tridentata, shrub in the Chihuahuan Desert, New Mexico. The perimeter of the plot is a steel frame that is carefully driven into the soil. Water running off the plot passes over a sheet metal tray and flows into a slotted pipe. The pipe empties into a series of calibrated containers.

plots with no termites and 88.4 ± 5.6 on plots with termites removed. Termite removal affected soil bulk density: 1.99 g cm−3 compared to 1.70 g cm−3 and porosity 24.9% vs. 35.8% on the termite free plots in comparison to plots with termites present. Additional evidence for the role of macropores produced by the feeding galleries of subterranean termites on water infiltration was provided by Eldridge et al. (1997). They found that these macropores were responsible for the deep levels to which infiltrating water percolated. When they compared saturated to unsaturated infiltration they found high values which has been reported to be a useful index of the macropore status of a soil. High values (1,6 or greater) suggest that most of the flow is through macropores. Ratios of 1:5 or less indicate that macropores are scarce and that capillary pores are the most important avenue for water movement into the soil. Data from other semidesert systems in more tropical areas suggest that activities of termites reduce infiltration at least with respect to the nest structure itself (Tongway and Noble, personal communication). In degraded grassland and shrub-dominated communities, sediment concentration in overland flow was related to the average diameter of small mammal disturbances. Small mammal diggings produce ejected soil piles that are friable and easily transported by overland flow (Whitford and Kay, 1999a, b). In addition to soil piles from small mammals, many species of ants inhabiting such plant communities produce mounds of erodible soil around the nest entrances (Whitford and Eldridge, 2013). However, in arid south-western Australia, holes produced by woylies (Bettongia penicillata) that forage on the fruiting bodies of hypogeous fungi become filled in with subsurface soil from the digging and with organic litter with the result that the soils are water repellant. Buried organic material was found in many decayed diggings that were severely water repellant and where the organic matter supported



93

4.4 Infiltration and Run-Off

Cumulative infiltration (mm)

500

Termite plots Nontermite plots

400 300 200 100 0 1993

(A)

1994 Year

1995

Infiltration rate (mm h–1)

100

(B)

80 60 40 20 0

Termites present

Termites excluded

FIG. 4.8  Effects of excluding termites on rainfall infiltration in Sahelian soils. From data in Mando, A., 1998. Soildwelling termites and mulches improve nutrient release and crop performance on Sabelian crusted soil. Arid Soil Res. Rehabil. 12, 153–164. Elkins, N.Z., Sabol, G.V., Ward, T.J., Whitford, W.G., 1986. The influence of subterranean termites on the hydrological characteristics of a Chihuahuan Desert ecosystem. Oceologia 68, 521–528.

fungal hyphae, the decayed diggings were very severely water repellant (Garkaklis et  al., 2000). These studies demonstrate the importance of examining generalizations about the effects of biota processes like water infiltration in many types of deserts because those processes may vary with different soil types, digging behaviors of animals, and the processes examined. In a study that measured water infiltration and soil physical and chemical properties on and off nests of two ant species (Pogonomyrmex rugosus and Aphaenogaster cockerelli) at different locations on a Chihuahuan Desert watershed, infiltration was slowest on areas with the highest silt and clay contents and fastest on sandy soil sites (James et al., 2008). Infiltration was generally higher in ant nest-modified soils than on adjacent soils without ant nests and this was attributed to macropores associated with the ant nests. There were differences in infiltration among sites on the watershed related to soil differences and also between the ant species at some locations. Despite considerable variation in soil chemical properties on the watershed, soils associated with ant nests of either species exhibited higher levels of carbon,

94

4.  Wind and Water Processes

nitrogen, sulfur, phosphorus, and electrical conductivity than was measured on adjacent soils not associated with ant nests. While nests of long-lived persistent ant colonies, especially seed-harvesting species, have generalized effects on nutrient levels and infiltration, differences in soil type, nest density, and ant species are likely to moderate these effects. In a sand dune area in China where biological soil crusts were used to stabilize the dunes, the presence of ant nests enhanced infiltration. There were more ant nests in late succession biological soil crusts where the lichen-moss-dominated crusts were thicker and the surface more stable than the early cyanobacteria-dominated crusts (Li et al., 2014).

4.4.6  Rocks and Stones Many desert regions have large areas of surface covered with rock fragments or stones. Rock cover decreases the erosion potential compared to that of unvegetated soil with little or no stone cover. Run-off volume is controlled by stone cover and stone size (Abrahams and Parsons, 1991). Empirical studies have shown that surface cover of rock and vegetation provided better protection (less sediment yield) than either 100% vegetation cover or 100% rock cover (Benkobi et al., 1993). Stony soils with large surface area stones provide microhabitat for soil animals. Termites build galleries on the undersurfaces of rocks that are embedded in the soil and some species of ants build nests under rocks. In stony soils with scattered large rocks, run-off from the rock surfaces may not contribute to run-off because that water infiltrates rapidly at the buried margins of the rock because of the macropores produced by invertebrates. Poesen and Ingelmo-Sanchez (1992) examined the effect of rock and macropores on interill run-off and sediment yield. They found that when macropores at the soil surface close to water infiltration is greatly reduced the effect of rock cover switches from reduction of runoff to an increase in run-off. Rock-covered soils in deserts may increase run-off and sediment transport under certain conditions. The size distribution of rocks on the soil surface is important for the functional relationship of rock cover, infiltration, run-off, and sediment yield. Empirical studies in a desert shrubland showed that infiltration rate was negatively related to gravel cover. Sediment production was also negatively related to gravel cover. This apparent dichotomy results from the obstruction value of the gravel fragments that slow overland flow and the resistance to infiltration by the gravel

4.5  EXCHANGES AMONG LANDSCAPE UNITS Exchange of materials among units of a landscape is the result of transport of dissolved and suspended materials in water from upslope to lower positions on a catena or watershed. Fluvial transport is the most important agent that produces soil gradients along catenas. The number of soil units that characterize a catena is at least partially determined by the length of the catena, the slope angle and/or variety of slope angles along the catena, the nature of the parent material at the top of the catena, and the infiltration characteristics of the soil series of the catena. The soil units on a catena support different vegetation assemblages. The growth form, cover, and productivity of the vegetation assemblages vary with soil type and catenary position and influence run-off, infiltration and accumulation of sediment and transported materials. Infiltration and run-off at the watershed scale are very different from



4.5  Exchanges Among Landscape Units

95

the rates and volumes estimated by studies of small plots on parts of the watershed. Flows in ephemeral stream beds monitored by flumes are considerably lower than the estimates based on summing the run-off for square meter units on the watershed. Although transmission losses to the sediments of the ephemeral channels account for some of the reduction in measured run-off, the within drainage soil and vegetation heterogeneity are more important variables (Murthy and Chandrasekharan, 1985). Within arid and semiarid regions, sparse vegetation and the microtopography related to that vegetation are important variables affecting run-off at fine spatial scales (Bergkamp, 1998). Increased infiltration near vegetation reduces overland flow. Higher infiltration under plants relative to bare soil is attributed to lower soil bulk density, greater soil aggregate stability, and greater density of macropores. Microtopography affects run-off by concentrating or diverging run-off flows. These spatial characteristics must be considered together with discontinuity in run-off that is related to storm depth and duration.

4.5.1  Ephemeral Streams (Arroyos, Dry Washes, Wadis) In order to understand run-off production, it is necessary to understand the discontinuity in run-off related to both the short duration and small size of storms and the heterogeneity in surface properties (DeBoer and Campbell, 1989). Microtopographic effects were documented in a study of a landscape made up of small terraces centered on grasses and small shrubs (terracettes), and hummocks that included the deep roots of the small trees that formed the center of the hummocks (Bergkamp, 1998). He found that there was flow from the terracette to the surrounding bare space even during short duration and relatively low intensity storms. However, there was no flow connection between terracettes. The plants on the hummocks redistributed water to deep soil layers, which controlled water movement on parts of the slope. At the partslope scale (1000–2000 m2) and the slope scale (1 ha) even with longer duration, higher intensity rainfall, the scattered distribution of saturated overland flow patches and the sinks (hummocks) inhibits the continuity of flow at this scale. Bergkamp concluded that run-off from the catchment cannot be directly related to run-off generation on the slope. It is suggested that flow in channels normally results from run-off in valley bottoms near stream channels. The spatial distribution of vegetation and bare patches on the slopes is the important landscape structural element that causes discontinuity in flow between fine scales and broader scales (Bergkamp, 1998). Noy Meir (1981) lists four examples of arid ecosystems in which water redistribution by run-off at small distances is important: (1) plains and alluvial fans with a dense network of small runnels and channels; (2) plains and gentle slopes with regular (or irregular) slight depressions, for example, gilgais or mulga grove-intergrove system; (3) rocky slopes interspersed with soil pockets; and (4) areas with strong microvariation in surface soil characteristics, in particular infiltrability. This list is certainly not exhaustive and the examples were included to demonstrate the need to incorporate spatial redistribution into models for plant production. If we are to develop models that incorporate spatial patterns of water, organic matter, and nutrient redistribution as effectors of primary production, we need also to consider how these spatial patterns vary with rainfalls of different quantity and intensity and how the return time of particular “threshold” event affects the ecosystem. Overland flow forms the major component of basin run-off that rapidly reaches the channel network. The typical flood in a major drainage channel is initiated between 4 and 16

96

4.  Wind and Water Processes

min depending on storm intensity which is responsible for the idea of “flash flood.” Flash floods are characterized by high turbulence and large amounts of sediments from hillslopes (Reid and Frostick, 1987). The importance of rainfall intensity and duration was confirmed by running the Hydrological River Basin Environment Assessment Model to provide virtual storms with different intensities and duration. Virtual storms were compared with flash flood events in Wadi Quena in the eastern desert of Egypt. The model incorporated thirtyeight geomorphometric parameters of topography, scale, shape, and drainage characteristics of the ephemeral drainages. Scale and topography were the most important parameters as found by the model while shape and drainage network were less important parameters. The most significant factor was total rainfall amount and duration affecting the hydrologic response of the wadi. The importance of wadi geomorphology increased with increasing rainfall intensity (Abdel-Fattah et al., 2017). That modeling study demonstrated that rainfall intensity and duration are the important factors resulting in “flash floods” and that features of the landscape such as slope, vegetation, drainage network are only factors when rainfall intensity and duration pass a threshold for overland flow and initiation of a “flash flood.” Flood water that infiltrates ephemeral channels (transmission losses) recharges local and regional aquifers and is the main water source in hyperarid regions (Morin et al., 2009). Their study verified the importance of large and medium floods to sustain the vegetation lining ephemeral channels. When the evidence from the study of the Kuiseb River in Namibia was generalized to other ephemeral streams with diverse channel and watershed characteristics the importance of high magnitude floods to recharge channel soils is increased. Increased channel lengths and greater active channel widths result in higher transmission losses. In these conditions, medium and small floods often fail to reach the downstream reaches of arid ephemeral streams or rivers (Morin et al., 2009). That study emphasizes the need for measurement of transmission losses (stream bed infiltration) for ephemeral streams with different lengths, widths, channel depths, stream bed sediments in order to interpret and predict the effects of rainfall intensity and duration plus the distance of subsequent flooding downstream. Desert mountains are subject to flash floods that produce debris flows. Debris flow occurs when water-filled masses of soil and fragmented rock are carried down the sides of mountains; funnel into stream channels; carry vegetation, soil, and animal feces and form thick muddy accumulations on the valley floor. It is thought that rainfall intensities >25 mm h−1 or amounts of rainfall totals of 50 mm in a single event are needed to initiate debris flows. Such events may result from convective storms or very slow moving frontal storms. Because flash floods have caused human deaths and debris flows have caused many deaths and extensive property damage, there is a need to predict debris flows in arid regions (Stolle et al., 2015). Smaller rain events were found to initiate debris flow in the Dead Sea region (David-Novak et  al., 2004). A regional storm with maximum rainfall intensities of 19–27 mm h−1 with a duration of 45 min initiated three debris flows. An earlier storm with a depth of 30 mm and intensity of 30 mm h−1 initiated debris flows in the same area.

4.6  EXCHANGES WITHIN LANDSCAPE UNITS Canopy interception and stemflow frequently eliminate run-off from the soils under the canopies of the established shrubs. As a consequence, leaf litter and plant parts accumulate under the canopies of such shrubs and trees. If slope angle and soil infiltration characteristics



4.6  Exchanges Within Landscape Units

97

produce run-off from most rain events, the shrubland develops insular topography. In such habitats, transport of organic matter by run-off is limited to that material in the intercanopy space. The transport of organic matter resulting in its redistribution on a landscape is a function of variables discussed in the previous section. Shrubs and tree canopies break up raindrops reducing the energy of the drops such that throughfall does not disrupt the litter layer. The litter layer and higher organic matter content of subcanopy soils insures that virtually all of the throughfall infiltrates. Thus there is no surface movement of water to remove litter from under canopies. Because there is little or no erosion from subcanopy soils, shrubs are often growing on slight mounds with an organic matter layer surrounded by the bare or nearly bare intershrub soil. Organic matter that is moved by sheet flow is limited to the leaves, stems, fecal pellets, and so on, located in the intercanopy space. The distance that this material is moved downslope depends on sheet flow velocity and spacing of impeding vegetation downslope. The net result from most rainfall events is an internal redistribution within a unit or accumulation in a downslope unit. Run-off in arid and semiarid regions frequently results in the formation of litter dams and microterraces. These structures are commonly observed especially immediately after large rain events. Litter dams form quickly when there is general sheet flow across a landscape unit. Sheet flow causes leaf litter, animal feces, stem segments, and so on, to be transported as “litter rafts” (Mitchell and Humphreys, 1987). Obstructions on the surface such as protruding rocks, tussocks of grass, stems of shrubs or large logs (such as decaying Yucca logs) capture the rafting litter and produce a dam. When a dam is formed, sediment accumulates in the ponded water behind the dam. Occasionally breaches form in a dam wall, which will temporarily drain the pond and cause rilling in the sediment infill. Typically, these breaches are closed in subsequent run-off events. Microterraces form from the coarser mineral sediments that accumulate behind the litter dams. In most deserts the litter dams are rapidly colonized by foraging termites and other arthropods. These organisms reduce the structural integrity of the litter dams and may reduce the whole structure to a microterrace of sediment and fine particulates. Litter dams and microterraces are transient features of arid landscapes that develop as a result of within-landscape unit transport. The factors important in litter dam and microterrace evolution were listed by Mitchell and Humphreys (1987). These include relatively bare slopes, slope angles less than about 10°, supply of organic debris, surface soil conducive to run-off and sediment yield, availability of obstructions, seed sources, sources of soil algae, and presence of litter breakdown mechanisms and bioturbating organisms. Despite the lack of quantitative data on litter dams and microterraces in arid and semiarid regions, it is obvious that these are important features of these environments and contribute significantly to the patch dynamics of these landscapes.

4.6.1 Rills In areas with hillslope topography with a vegetative cover that is primarily shrubs, rills are a frequent feature. Rills are shallow channels of no more than a few tens of centimeters deep. Rills are cut into the soil by flowing water. On relatively steep slopes or slopes where the contributing area is increasing, rills are generally continuous extending to a drainage channel or its immediate floodplain (Wainwright et al., 2002). Rills exhibit different transmission losses depending upon bed characteristics. Transmission losses in sand-bedded rills are more than

98

4.  Wind and Water Processes

60% higher than those in gravel-bedded rills (Parsons et al., 1999). Sand-bedded rills exhibit increasing transmission loss with increasing depth of flow but gravel-bedded rills do not. Transmission losses in rills are about 10X greater than infiltration losses in adjacent interrill areas (Parsons et al., 1999). Discontinuous rills can occur where hillslope surfaces cause the catchment area to decrease relative to the area required to maintain sediment transport or where local slope causes deposition. In areas of deposition, transported sediments generally splay (spread out around the channel) and water flow often continues again below the sediment deposition area. Net accumulation of the depositional surfaces seems to occur during times of lower than average rainfall but the depositional surfaces are dynamic and exhibit net erosion in wetter years. Analysis of the vegetation of the splay area revealed higher total vegetation cover and proportionately more grasses than surrounding areas on the watershed (Wainwright et al., 2002). Discontinuous rills result in vegetation patches with higher productivity and larger plants than is typical of most hillslope areas. Some areas that are degraded by livestock overgrazing may be subject to rapid gulley formation and development of badlands. Badlands form where sediments are loosely consolidated. Badlands develop in semiarid zones and are characterized by erosion features, slope failures, deeply incised dry stream beds, and highly dissected slopes. In the South African Karoo Desert, badlands and gullies developed since European settlement (Boardman et al., 2003). Many footslopes developed in shales, clays, and colluviums have closely spaced gullies of up to 1.5 m deep. In valley bottoms and low slopes, gullies up to 8 m deep have developed usually to bedrock. Simulated rainfall experiments demonstrated that badland areas become active erosion areas under high-frequency, low magnitude rainfall events (Boardman et al., 2003).

4.6.2  Steep Hillslopes A different kind of within-unit exchange occurs on steep hillslopes. Here run-off and sediment transported from the upper slopes have no chance of reaching the slope base during most rain events. Steep hillslopes with sedimentary rock lithography tend to be structured with narrow flat or concave benches between the steep-sided exposed rock surfaces. In the Negev, water moves from the top of the limestone hillslopes and ponds in the bench areas. Run-off and sediment yield measurements on these hillslopes clearly show that the bulk of the run-off and sediment transport is between upper and lower units of the hillslope (Yair et al., 1980). On volcanic and granitic slopes, boulders and ridge structures probably tend to keep the run-off and sediment transport local.

4.7  MODELING OVERLAND FLOW AND EPHEMERAL CHANNEL FLOW Modeling soil erosion involves an equation for predicting sediment transport capacity by interrill overland flow on rough surfaces (Abrahams et al., 1998). In an examination of surface roughness and other variables on sediment transport, regression analysis revealed that almost 90% of the variance in transport capacity was accounted for by excess flow power and flow depth (Abrahams et al., 1998). Another examination of a model for overland flow



4.8  Episodic Events

99

found that the model gave good results for low sediment concentrations but flows with higher concentrations of sediment yielded higher sediment rates than predicted by the model (Abrahams et  al., 2001). Models that focus on the effects of cover by functional groups of vegetation provide some insights into overland flow and sediment transport (Howes and Abrahams, 2003). That modeling effort indicated that run-off experiments used to calibrate the model need to be conducted after the rainy season has resulted in better developed soil seals. Other hydrological models of overland flow incorporate a number of measured parameters of roughness plus constants (Hu and Abrahams, 2006). A model for the development of pavements in desert landscapes by raindrop erosion processes demonstrated the importance of these processes. When the sensitivity of the model was tested on small plots by repeated 5 min rainfall events it was found that desert pavements can reform within 10 events which is similar to the recovery of surfaces under natural rainfall annually. However, the details of particle size fractions produced were less well simulated by the model (Wainwright et  al., 1999a, b). It was concluded that the problem is not related to soil characteristics but is an indication of poor understanding of all of the feedbacks in raindrop erosion processes. It is probable that subsoil transported to the surface by the activity of small animals like termites and ants may have contributed to the differences in particle fractions in the model simulation. Wainwright et  al. (2008a) developed the model MAHLERAN (Model for Assessing Hillslope-Landscape Erosion, Run-off, and Nutrients) based on particle travel distance that overcame limitations of earlier models by incorporating parameterizations of different detachment and transport mechanisms that occur in water erosion on hillslopes and small catchments. When the model MAHLERAN was developed and subjected to sensitivity analysis, the model was able to reproduce sediment yield, spatial and temporal patterns of particle size distributions of eroded sediments after erosion parameters were optimized for specific soil conditions (Wainwright et al., 2008b). This model was also tested by monitoring plots of different sizes. Although the model produced the correct ranking of the magnitude of erosion events, it did less well in reproducing the values and particle size distributions of eroded sediment. It was concluded that a more holistic evaluation is required and that appropriate parameterization data are needed to support the use of the model in a wide range of environmental conditions (Wainwright et al., 2008c). Their conclusion demonstrates that there are many variables that have an effect on overland flow and sediment transport in arid ecosystems and not all of those variables are based on topography, soil types, and vegetation.

4.8  EPISODIC EVENTS If there is one generalization that can be made about arid and semiarid environments, it is that episodic events have a greater impact on the ecosystems and landscapes than the cumulative effect of 25–30 years of events within the “average range.” Episodic events may be a wind storm with wind velocities more than 2 standard deviations above the long-term mean wind velocities for a season and location or an extremely intense, large storm depth rainfall that produces flooding that reaches or exceeds the capacity of the large drainage channels. Flood frequency analysis is based on dimensionless curves that relate return period in years to floods of different magnitudes (Farquharson et al., 1992). The problem of predicting return period for floods of a given magnitude is that there are short records for arid areas. Return

100

4.  Wind and Water Processes

­ eriods are based on calculation of mean annual flood for a given geographic location. In p their synthesis of flood frequencies, Farquharson et al. (1992) examined the relationship between catchment area and mean annual flood for arid regions. Regression of all the available data gave an r2 of 0.55 and regression of regions, that is, North Africa, Southern Africa, Iran, and Queensland and the Arabian Peninsula, yielded r2 values between 0.43 and 0.92. Obviously catchment area is an important variable affecting flood size (measured in m3 s−1). However, the probability of a low frequency, large volume, intense storm that produces extreme flooding is unpredictable for arid and semiarid regions. Studies of paleofloods in the drainage of a 1400 km2 catchment in the hyperarid area of the Negev Desert found 28 floods that ranged from 200 to 1500 m3 s−1 over the last 2000 years (Greenbaum et al., 2000). This study demonstrates the low frequency of episodic floods that affect the geomorphology and structure of desert landscapes. The results of a study of erosion and deposition plus effects on vegetation in a wadi in the Negev Desert of Israel document the importance of episodic events in restructuring ecosystems. The watershed of Wadi Zin experienced a 60 mm rainfall within a period of 1 h (Naomiish-Shalom-Gordon and Gutterman, 1991). The resulting flood was rated at a 100-year recurrence interval. The flood not only damaged vegetation, but it also lowered the surface of the wadi, uncovered bedrock in several areas, and deposited large quantities of sediment as the flood waters waned. In this flood, vegetation was washed away from a portion of the wadi confined between vertical rock walls. These authors describe a two-phase flood regime. The high-energy first phase produced large loss of soil from the channel and benches. The second phase resulted from low-energy flow that produced soil deposition especially in the wider parts of the wadi. The result of the erosion and deposition from a single storm was a complete restructuring of the wadi that will affect vegetation development for years into the future. Storms with different rainfall characteristics produce very different types of flow events in ephemeral channels. White (1995) characterized storm intensities by the kinds of flow events that occur in drainage channels. High-intensity storms produce debris flows (transport of organic debris, small cobbles, and soil sediments) whereas low intensity storms produce fluvial flows (flows with low quantities of suspended sediments). He characterized channels based on channel morphology and characteristics of the uplands that drain into the channels as debris-type flow channels or as fluvial-type flow channels. The flood bore in debris-type flow frequently “scours” the watershed sufficiently that a subsequent rain event of similar intensity and duration will not produce a debris flow in a channel that is a debris-type flow channel (White, 1995). Scouring events are episodic in most deserts. In more than 50 years of research on the Jornada Experimental Range in New Mexico, I have experienced only one very large intense storm that “scoured” the watershed that we were studying at the time. Although we had no gauging stations on the ephemeral drainages of the watershed, the bank cutting, bank overflooding, and large debris deposits at the terminus of identifiable drainage channels provided evidence that the volume of water moving across the uplands of the watershed reached or exceeded the channel capacities of the drainage channels. The scouring action of the large overland flow removed all debris dams, dead wood from intershrub spaces, leaf litter (except under mounded shrubs), and dung. Large quantities of this material were deposited at the periphery of the ephemeral lake at the base of the watershed. The ephemeral lake basin overflowed to several meters beyond the finetextured lacustrine soils of the basin. This event was a small, episodic flood in comparison



4.9  Ephemeral Ponds and Lakes

101

to a flood in the local area at the turn of the century that was reported to have produced all of the head cutting of the present ephemeral stream channels. These anecdotal observations reinforce the need for “long-term” measurements in desert ecosystems. Even with a long-term ecological research program in place, we were not prepared to make all of the measurements needed to document all of the changes in the ecosystems of the watershed that resulted from this single storm. Except for studies conducted in the ephemeral lake, all of the changes in distribution patterns of organic materials, and so on, were qualitative. Episodic events “reset” spatial patterns of organic matter, “reset” soil processes, and may “reset” many other ecosystem processes.

4.9  EPHEMERAL PONDS AND LAKES Ephemeral lakes and ponds are basins that remain flooded for short periods of time during a year but may not hold water for several years if the rainfall regime is not suitable to produce flooding. Filling of ephemeral lakes is episodic. Water entering lake basins may originate as run-off from a relatively small area surrounding the basin or may represent water in excess of the transmission loss capacity of the ephemeral streams that drain a large watershed. Motts (1969) proposed a flooding ratio, Rn, the ratio of the number of days a basin contains water to the number of days it is dry. He proposed that a value of 0.33 or less is indicative of an ephemeral lake. Ephemeral lakes vary in their origin, size, flood timing, flood frequency, areal extent of flooding, salinity, relation to groundwater, vegetation, soils, ratio of basin size to watershed size, surface morphology, mineralogy, and stratigraphy (Van Vactor, 1989). These physical variables affect the biota that inhabit ephemeral lakes and also affect the ecosystem processes of a lake when flooded. In the desert rangelands of North America, the filling of ephemeral lakes does not necessarily require that the ephemeral stream channels reach the edge of the lake. Indeed, the more typical geomorphology is for the ephemeral stream channels to terminate at some distance from the dry lake basin (Wondzell et al., 1996). In the mulga woodlands of northwestern New South Wales, Australia, some ephemeral lake basins received water directly from ephemeral drainage channels whereas other dry lake basins did not (personal observation). The importance of discharge from ephemeral stream channels has been questioned because of the probability of greatly diminished discharge because of transmission losses in the ephemeral stream channels. Measured infiltration rates in ephemeral stream channels in the Chihuahuan Desert varied from 57.3 to 22.4 cm h−1 with an average rate of 39.5 ± 14.8 cm h−1. Infiltration rates of 90–100 cm h−1 were measured in another Chihuahuan Desert ephemeral stream bed using a different technique (Van Vactor, 1989). These high infiltration rates result in large transmission losses into the stream bed and suggest that only the most intense storms produce sufficient run-off to produce discharge and overland flow of that discharge from the terminus of the ephemeral stream channel. Basically then, ephemeral lake flooding occurs when rainfall inputs to the lake watershed exceed losses by infiltration thereby producing run-off that is transported by a combination of overland flow and channel flow to the dry lake basin (Van Vactor, 1989). When flooding occurs, the water in the lake basin contains suspended organic materials and suspended soil particles. The velocity and depth of the run-off that enters the basin

102

4.  Wind and Water Processes

­ etermines the quantity and characteristics of the suspended matter. After a particularly ind tense storm, the suspended matter in a dry lake that we studied included leaf fragments, small sticks, and rabbit dung in addition to the unidentifiable suspended organics and inorganics. Ephemeral lakes thus function as the primary sink for organic matter in desert landscapes. The soil organic matter (soil carbon) of an ephemeral lake bed at the base of an intensively studied Chihuahuan Desert watershed was 3.4% compared with values ranging between 0.4% and 1.2% for the soils of the upland portions of the watershed (Nash and Whitford, 1995). The higher organic matter content of ephemeral lake soils may also result in higher rates of nitrogen mineralization because of the larger quantities of organic nitrogen in comparison to the soils on the remainder of the watershed (Nash and Whitford, 1995). When an ephemeral lake is flooded, the allochthonous (origin outside the water body) and the autochthonous (origin in situ) materials are rapidly decomposed by a variety of organisms. These materials form the basis for the complex food webs that develop in desert ephemeral lakes. If the ephemeral lake basin supports terrestrial vegetation, the grazing management determines the residual vegetation remaining on the basin when a flooding event occurs. Most ephemeral lake basins are devoid of terrestrial vegetation or nearly devoid of such vegetation. The characteristics of the lake basin with respect to presence or absence of terrestrial vegetation is the primary factor that determines the structure of the developing aquatic food web during a flooding cycle. With terrestrial vegetation present, the aquatic food web starts as a heterotrophic system, which gradually shifts to an autotrophic system. Physical and chemical characteristics of ephemeral ponds are extremely variable both temporally and spatially. In most ephemeral waters, there are large diurnal fluctuations in temperature and dissolved oxygen (Podrabsky et al., 1997). Dissolved oxygen content, water temperature, pH, and conductivity values differ depending upon the characteristics of the watersheds supplying the body of water. Most ephemeral ponds have a high amount of suspended solids which results in high turbidity. Dissolved oxygen varies greatly among ponds in arid and semiarid regions which is probably the effect of differences in biological activity within each pond (Clark, 1988).

SUMMARY 1. Wind erosion is one feature of dryland ecosystems not shared with more mesic ecosystems. Dust whirlwinds produced by thermals suck up soil, plant fragments, and other organic materials that are transported for varying distances across the landscape. 2. The erosive power of wind is related to the threshold shear velocity: the velocity below which saltation (flux of sand particles) does not occur. Saltation is affected by soil moisture, relative humidity, aggregate stability of soil, cover of biological soil crusts, density and cover of vegetation, vegetation height, and porosity of vegetation cover. 3. Grasses are more effective at keeping soil from erosion that are shrubs, trees, and other vegetation. 4. Plants redistribute rainfall. Precipitation intercepted by plant canopies that is retained and evaporated does not contribute to soil moisture. Throughfall is that portion of rainfall that reaches the soil or litter by falling through spaces in the canopy or by leaf drip. Stemflow of trees and shrubs follows root channels into the deeper soil profiles than throughfall.

REFERENCES 103

5. Soil with no vegetative cover is subject to splash erosion which results from the energy of raindrops that generate small craters in the soil, pushing fine particles into the surrounding space. 6. Water moves into the soil by infiltration. Variables that affect infiltration are vegetation cover, soil surface cover, soil texture, soil bulk density, abundance and distribution of macropores, and antecedent rainfall. 7. Rainfall that exceeds infiltration rate for a soil produces run-off. Run-off is affected by vegetation cover, topography (slope, rock cover, gravel surfaces), and surface roughness. 8. Run-off that reaches rills or ephemeral streams is transmitted various distances downslope but the amount of water moving in a channel is reduced by transmission losses of water moving into the stream bed sediments.

References Abdel-Fattah, M., Saber, M., Kantoush, S.A., Khali, M.F., Sumi, T., Sefelnasr, A.M., 2017. A hydrological and geomorphometric approach to understanding the generation of wadi flash floods. Water 9, 553. https://doi.org/10.3390/ w9070553. Abrahams, A.D., Li, G., Krishman, C., Atkinson, J.F., 1998. Predicting sediment transport by interrill overland flow on rough surfaces. Earth Surf. Process. Landf. 23, 481–492. Abrahams, A.D., Li, G., Krishman, C., Atkinson, J.F., 2001. A sediment transport equation for interrill overland flow on rough surfaces. Earth Surf. Process. Landf. 26, 1443–1459. Abrahams, A.D., Parsons, A.J., 1991. Relation between sediment yield and gradient on debris­covered hillslopes, Walnut Gulch, Arizona. Geol. Soc. Am. Bull. 103, 1109–1113. Anderson, S.H., Peyton, R.L., Gantzer, C.J., 1990. Evaluation of constructed and natural soil macropores using x-ray computed tomography. Geoderma 13, 13–29. Abrahams, A.D., Parsons, A.J., Wainwright, J., 2003. Disposition of rainwater under creosotebush. Hydrol. Process. 17, 2555–2566. Alvarez, L.J., Epstein, H.E., Li, J., Okin, G.S., 2012. Aeolian process effects on vegetation communities in an arid grassland ecosystem. Ecol. Evol. 2, 809–821. Belnap, J., Gillette, D.A., 1998. Vulnerability of desert biological soil crusts to wind erosion: the influences of crust development, soil texture, and disturbance. J. Arid Environ. 39, 133–142. Benkobi, L., Trlica, M.J., Smith, J.L., 1993. Soil loss as affected by different combinations of surface litter and rock. J. Environ. Qual. 22, 657–661. Bergkamp, G., 1998. A hierarchical view of the interactions of runoff and infiltration with vegetation and microtopography in semiarid shrublands. Catena 33, 201–220. Belnap, J., 2006. The potential roles of biological soil crusts in dryland hydrologic cycles. Hydrol. Process. 20, 3158–3178. Belnap, J., Wilcox, B.P., Van Scoyoc, M.W., Phillips, S.L., 2013. Successional stage of biological crusts: an accurate indicator of ecohydrological condition. Ecohydrology 6, 333–505. Berndtsson, R., Larson, M., 1987. Spatial variability of infiltration in a semi-arid environment. J. Hydrol. 90, 117–133. Boardman, J., Parsons, A.J., Holland, R., Holmes, P.J., Washington, R., 2003. Development of badlands and gullies in the Sneeberg, Great Karoo, South Africa. Catena 50, 165–184. Brandt, C.J., 1989. The size distribution of throughfall drops under vegetation canopies. Catena 16, 507–524. Branson, F.A., Gifford, F.F., Renard, K.G., Hadley, R.F., 1981. Rangeland Hydrology, second ed. Range Science Series No. 1. Society for Range Management, Kendall/Hunt Publishing, Dubuque, Iowa. Chepil, W.S., 1951. Properties of soil which influence wind erosion: IV state of dry aggregate structure. Soil Sci. 72, 387–401. Chepil, W.S., 1953. Factors that influence clod structure and erodibility of soil by wind I. Soil texture. Soil Sci. 75, 473–483. Clark, W.J., 1988. Pond limnology, with applications to arid-area ponds. In: Thames, J.L., Ziebell (Eds.), Small Water Impoundments in Semi-Arid Regions. University of New Mexico Press. 162 pp.

104

4.  Wind and Water Processes

Colica, G., Li, H., Rossi, F., Li, D., Liu, Y., De Philippis, R., 2014. Microbial secreted exopolysaccharides affect the hydrological behavior of induced biological soil crusts in desert sandy soils. Soil Biol. Biochem. 68, 62–70. David-Novak, H.B., Morin, E., Enzel, Y., 2004. Modern extreme storms and the rainfall thresholds for initiating debris flows on the hyperarid western escarpment of the Dead Sea, Israel. Bull. Geol. Soc. Am. 116, 718–728. DeBoer, D.H., Campbell, I.A., 1989. Spatial scale dependence of sediment dynamics in a semiarid badland drainage basin. Catena 16, 277–290. DeSoyza, A.G., Whitford, W.G., Martinez-Meza, E., Van Zee, J.W., 1997. Variation in creosotebush (Larrea tridentata) morphology in relation to habitat, soil fertility and associated annual plant communities. Am. Midl. Nat. 137, 13–26. Devine, D.L., Wood, M.K., Donart, G.B., 1998. Runoff and erosion from a mosaic tobosagrass and burrograss community in the northern Chihuahuan Desert grassland. J. Arid Environ. 39, 11–19. Dunkerley, D., Geddes, N., 1999. The influence of organic litter on the erosive effects of raindrops and of gravity drops released from desert shrubs. Catena 36, 303–313. Eldridge, D.J., Leys, J.F., 2003. Storage of carbon in soil and vegetation in paired roadside site in the box woodlands of eastern Australia. J. Arid Environ. 53, 457–466. Eldridge, D.J., Tozer, M.E., Slangen, S., 1997. Soil hydrology is independent of microphytic crust cover: further evidence from a wooded semiarid Australian rangeland. Arid Soil Res. Rehabil. 11, 113–126. Eldridge, D.J., Zaady, E., Shachak, M., 2000. Infiltration through three contrasting biological soil crusts in patterned landscapes in the Negev, Israel. Catena 40, 323–336. Elkins, N.Z., Sabol, G.V., Ward, T.J., Whitford, W.G., 1986. The influence of subterranean termites on the hydrological characteristics of a Chihuahuan Desert ecosystem. Oceologia 68, 521–528. Farquharson, F.A.K., Meigh, J.R., Sutcliffe, J.V., 1992. Regional flood frequency analysis in arid and semi-arid area. J. Hydrol. 138, 487–501. Furbish, D.J., Childs, E.M., Haff, P.K., Schmeeckle, M.W., 2009. Rain splash soil grains as a stochastic advectiondispersion process, with implications for desert plant-soil interactions and land surface evolution. J. Geophys. Res. 114, F0A03. https://doi.org/10.102/2009JF001265. Ferguson, B.K., 1991-92. Landscape hydrology, a component of landscape ecology. J. Environ. Syst. 21, 193–205. Field, J.P., Breshears, D.D., Neff, J.C., Whicker, J.J., Painter, T.H., Ravi, S., Reheis, M.C., Reynolds, R.L., 2010. The ecology of dust. Front. Ecol. Environ. 8, 423–430. Fletcher, J.E., Martin, W.P., 1948. Some effects of algae and moulds in the rain crust of desert soils. Ecology 29, 95–100. Gillette, D.A., Chen, W., 2001. Particle production and aeolian transport from a "supply-limited" source area in the Chihuahuan desert, New Mexico, United States. J. Geophys. Res. 106 (D6), 5267–5278. Graetz, R.E., Tongway, D.J., 1986. Influence of grazing management on vegetation, soil structure and nutrient distribution and the infiltration of applied rainfall in a semiarid chenopod shrubland. Aust. J. Ecol. 11, 347–360. Greenbaum, N., Schick, A.P., Baker, V.R., 2000. The palaeoflood record of a hyperarid catchment, Nahal Zin, Negev Desert, Israel. Earth Surf. Process. Landf. 25, 951–971. Greene, R.S.B., Tongway, D.J., 1989. The significance of (surface) physical and chemical properties in determining soil surface condition of red-earths in rangelands. Aust. J. Soil Res. 30, 213–225. Garkaklis, M.J., Bradley, J.S., Wooller, R.D., 2000. Digging by vertebrates as an activity promoting the development of water-repellent patches in sub-surface soils. J. Arid Environ. 45, 26–42. Hoffman, O., Yizhaq, H., Boeken, R., 2013. Small-scale effects of annual and woody vegetation on sediment displacement under field conditions. Catena 109, 157–163. Howes, D.A., Abrahams, A.D., 2003. Modeling runoff and runon in a desert shrubland ecosystem, Jornada Basin, New Mexico. Geomorphology 53, 45–73. Hu, S., Abrahams, A.D., 2006. Partitioning resistance to overland flow on rough mobile beds. Earth Surf. Process. Landf. 31, 1280–1291. Hupy, J.P., 2004. Influences of vegetation cover and crust type on wind-blown sediment in a semi-arid climate. J. Arid Environ. 58, 167–179. Istok, J.D., Boersma, L., 1986. Effect of antecedent rainfall on runoff during low-intensity rainfall. J. Hydrol. 88, 329–342. James, A.I., Eldridge, D.J., Koen, T.B., Whitford, W.G., 2008. Landscape position moderates how ant nests affect hydrology and soil chemistry across a Chihuahuan Desert watershed. Landsc. Ecol. 23, 961–975. Kidron, G.J., Monger, H.C., Vonshak, A., 2012. Contrasting effects of microbiotic crusts on runoff in desert surfaces. Geomorphology 139–140, 484–494. Lee, K.E., Foster, R.C., 1991. Soil fauna and soil structure. Aust. J. Soil Res. 29, 745–775.

REFERENCES 105

Leys, J.F., Raupach, M.R., 1991. Soil flux measurements using a portable wind erosion tunnel. Aust. J. Soil Res. 29, 533–552. Li, X.R., Gao, Y.H., Su, J.Q., Jia, R.L., 2014. Ants mediate soil water in arid desert ecosystems: mitigating rainfall interception induced by biological soil crusts? Appl. Soil Ecol. 78, 57–64. Li, J., Okin, G.S., Epstein, H., 2009. Effects of enhanced wind erosion on surface soil texture and characteristics of windblown sediments. J. Geophysical Res. Biogeosci. 114, G02003. 8 pp. Li, J., Okin, G.S., Herrick, J., Belnap, J., Miller, M.E., Vest, K., Draut, A.E., 2013. Evaluation of a new model of aeolian transport in the presence of vegetation. J. Geophysical Res. Earth Surf. 118, 288–306. Loope, W.L., Gifford, G.F., 1972. Influence of a soil microfloral crust on selected properties of soil under PinyonJuniper in southeastern Utah. J. Soil Water Conserv. 27, 164–167. Luo, W., Zhao, W., 2015. Effects of wind erosion and sand burial on growth and reproduction of a clonal shrub. Flora 217, 164–169. Lyford, F., Qashu, H., 1969. Infiltration rates as affected by desert vegetation. Water Resour. Res. 5, 1373–1376. Mando, A., 1998. Soil-dwelling termites and mulches improve nutrient release and crop performance on Sabelian crusted soil. Arid Soil Res. Rehabil. 12, 153–164. Mando, A., Miedema, R., 1997. Termite-induced change in soil structure after mulching degraded (crusted) soil in the Sahel. Appl. Soil Ecol. 6, 26–263. Martinez-Meza, E., Whitford, W.G., 1996. Stemflow, throughfall and channelization of stemflow by roots in three Chihuahuan Desert shrubs. J. Arid Environ. 32, 271–287. Mayaud, J.R., Wiggs, G.F.S., Bailey, R.M., 2016. Dynamics of skimming flow in the wake of a vegetation patch. Aeolian Res. 22.141–151. Mitchell, P.B., Humphreys, G.S., 1987. Litter darns and microterraces formed on hillslopes subject to rainwash in the Sydney Basin, Australia. Geoderma 39, 331–357. Morin, E., Grodek, T., Dahan, O., Benito, G., Kulls, C., Jacoby, Y., Van Langenhove, G., Seely, M., Enzel, Y., 2009. Flood routing and alluvial aquifer recharge along the ephemeral arid Kuiseb River, Namibia. J. Hydrol. 368, 262–275. Moore, D.C., Sanger, M.J., 1990. Crust formation effects on soil erosion processes. Soil Sci. Soc. Am. J. 54, 1117–1123. Motts, W.S., 1969. Geology and Hydrology of Selected Playas in Western United States. Air Force Cambridge Research Laboratories Report No. AFCRL-69-0214, Bedford, MA. Murthy, K.N., Chandrasekharan, H., 1985. Spatial variability of infiltration in an arid watershed in India. J. Arid Environ. 8, 189–198. Musick, H.B., Trujillo, S.M., Truman, C.R., 1996. Wind-tunnel modeling of the influence of vegetation structure on saltation threshold. Earth Surf. Process. Landf. 21, 589–605. Naomiish-Shalom-Gordon, Gutterman, Y., 1991. Soil disturbance by a violent flood in Wadi Zin in the Negev Desert Highlands of Israel. Arid Soil Res. Rehabil. 5, 253–263. Nash, M.H., Whitford, W.G., 1995. Subterranean termites: regulators of soil organic matter in the Chihuahuan Desert. Biol. Fertil. Soils 19, 15–18. Nash, M.S., Anderson, J.P., Whitford, E., 1999. Spatial and temporal variability in relative abundance and foraging behavior of subterranean termites in desertified and relatively intact Chihuahuan Desert ecosystems. Appl. Soil Ecol. 12, 149–157. Neave, M., Abrahams, A.D., 2001. Impact of small mammal disturbances on sediment yield from grassland and shrubland ecosystems in the Chihuahuan Desert. Catena 44, 285–303. Noy Meir, I., 1981. Spatial effects in modelling arid ecosystems. In: Goodall, D.W., Perry, R.A. (Eds.), Arid Land Ecosystems: Structure, Functioning and Management. vol. 2. Cambridge University Press, Cambridge, pp. 411–432. Nulsen, R.A., Bligh, K.J., Baxter, I.N., Solin, E.J., Imrie, D.H., 1986. The fate of rainfall on a mallee and heath vegetated catchment in southern western Australia. Aust. J. Ecol. 11, 361–371. Okin, G.S., Gillette, D.A., Herrick, J.E., 2006. Multi-scale controls on and consequences of aeolian processes in landscape change in arid and semi-arid environments. J. Arid Environ. 65, 253–273. Parsons, A.J., Wainwright, J., Stone, P.M., Abrahams, A.D., 1999. Transmission losses in rills on dryland hillslopes. Hydrol. Process. 13, 2897–2905. Phillips, R.E., Quisenberry, J.M., Zeleznik, J.M., Dunn, G.H., 1989. Mechanisms of water entry into simulated macropores. Soil Sci. Soc. Am. J. 53, 1629–1635. Pickup, G., 1985. The erosion cell- a geomorphic approach to landscape classification in range assessment. Aust. Rangeland J. 7, 114–121.

106

4.  Wind and Water Processes

Poesen, J., 1985. An improved splash transport model. Z. Geomorphol. 29, 193–211. Poesen, J., Savat, J., 1981. Detachment and transportation of loose sediments by raindrop splash. Part II. Detachability and transportability measurements. Catena 8, 19–41. Poesen, J., Ingelmo-Sanchez, F., 1992. Runoff and sediment yield from topsoils with different porosity as affected by rock fragment cover and position. Catena 19, 451–474. Pressland, A.J., 1976. Soil moisture redistribution as affected by throughfall and stemflow in an arid zone shrub community. Aust. J. Bot. 24, 641–649. Podrabsky, J.E., Hrbek, T., Hand, S.C., 1997. Physical and chemical characteristics of ephemeral pond habitats in the Maracaibo basin and Llanos region of Venezuela. Hydrobiologia 382, 67–77. Ravi, S., D’Odorico, P., Over, T.M., Zobeck, T.M., 2004. On the effect of air humidity on soil susceptibility to wind erosion: the case of air-dry soils. Geophys. Res. Lett. 31, L09501. Reid, I., Frostick, L.E., 1987. Flow dynamics and suspended sediment properties in arid zone flash floods. Hydrol. Process. 1, 239–253. Schlesinger, W.H., Ward, T.J., Anderson, J., 2000. Nutrient losses in runoff from grassland and shrubland habitats in southern New Mexico. II. Field plots. Biogeochemistry 49, 69–86. Stolle, A., Langer, M., Blothe, J.H., 2015. On predicting debris flows in arid mountain belts. Glob. Planet. Chang. 126, 1–13. Sweeney, M.R., McDonald, E.V., Elyemezian, V., 2011. Quantifying dust emissions from desert landforms, eastern Mojave Desert, USA. Geomorphology 135, 21–34. Tarchitzky, J., Banin, A., Morin, J., Chen, Y., 1984. Nature, formation and effects of soil crusts formed by water drop impact. Geoderma 33, 135–155. Tisdall, J.M., Oades, J.M., 1982. Organic matter and water stable aggregates in soils. J. Soil Sci. 33, 141–163. Valentin, C., 1991. Surface crusting in two alluvial soils of northern Niger. Geoderma 48 (20), 1–222. Van de Ven, T.A.M., Fryrear, D.W., Spaan, W.P., 1989. Vegetation characteristics and soil loss by wind. J. Soil Water Conserv. 44, 347–349. Van Elewijck, L., 1989. Influence of leaf and branch slope on stemflow amount. Catena 16, 525–533. Van Vactor, S., 1989. Hydrologic and Edaphic Patterns and Processes Leading to Playa Flooding. MS. Thesis, New Mexico State University, Las Cruces, NM. Wainwright, J., Parsons, A.J., Abrahams, A.D., 1999a. Field and computer simulation experiments on the formation of desert pavement. Earth Surf. Process. Landf. 24, 1025–1037. Wainwright, J., Parsons, A.J., Abrahams, A.D., 1999b. Rainfall energy under creosotebush. J. Arid Environ. 43, 111–120. Wainwright, J., Parsons, A.J., Schlesinger, W.H., Abrahams, A.D., 2002. Hydrology-vvegetation interactions in areas of discontinuous flow on a semi-arid bajada, Southern New Mexico. J. Arid Environ. 51, 319–338. Wainwright, J., Parsons, A.J., Muller, E.N., Brazier, R.E., Powell, D.M., Fenti, B.A., 2008a. A tansport-distance approach to scaling erosion rates: 1 Background and model development. Earth Surf. Process. Landf. 33, 813–826. Wainwright, J., Parsons, A.J., Muller, E.N., Brazier, R.E., Powell, D.M., Fenti, B.A., 2008b. A transport-distance approach to scaling erosion rates. 2. Sensitivity and valuation of MAHLERAN. Earth Surf Process. Landf. 33, 962–984. Wainwright, J., Parsons, A.J., Miller, E.N., Brazier, R.E., Powell, D.M., Fenti, B.A., 2008c. A transport-distance approach to scaling erosion rates. 3. Evaluating scaling characteristics of MAHLERAN. Earth Surf. Process. Landf. 33, 1113–1128. Wang, X.-P., Wang, Z.-N., Berndtsson, R., Zhang, Y.-F., Pan, Y.-X., 2011. Desert shrub stemflow and its significance in soil moisture replenishment. Hydrol. Earth Syst. Sci. 15, 561–567. White, K., 1995. Field techniques for estimating downstream change in discharge of grave-bedded ephemeral streams: case study in southern Tunisia. J. Arid Environ. 30, 283–294. Whitford, W.G., Kay, F.R., 1999a. Biopedturbation by mammals in deserts: a review. J. Arid Environ. 41, 203–230. Whitford, W.G., 2000a. Keystone arthropods as webmasters in desert ecosystems. In: Coleman, D.C., Hendrix, P.F. (Eds.), Invertebrates as Webmasters in Ecosystems. CABI Publishing, Wallingford, UK, pp. 25–41. Whitford, W.G., 2000b. Effects of habitat characteristics on the abundance and activity of subterranean termites in arid southeastern New Mexico (Isoptera). Sociobiology 34, 493–504. Whitford, W.G., Kay, F.R., 1999b. Biopedturbation by mammals in deserts: a review. J. Arid Environ. 41, 203–220. Whitford, W.G., Eldridge, D.J., 2013. Effects of ants and termites on soil and geomorphological processes (Editor in Chief). In: Shroder, J., Butler, D.R., Hupp, C.R. (Eds.), Treatise on Geomorphology. vol. 12. Academic Press, San Diego, CA, pp. 281–292. Ecogeomorphology.

FURTHER READING

107

Wiggs, G.F.S., Livingstone, I., Thomas, D.S.G., Bullard, J.E., 1994. Effect of vegetation removal on airflow patterns and dune dynamics in the southwest Kalahari Desert. Land Degrad. Rehabil. 5, 13–24. Wondzell, S.M., Cunningham, G.L., Bachelet, D., 1996. Relationship between landforms, geomorphic processes and plant communities on a watershed in the northern Chihuahuan Desert. Landsc. Ecol. 11, 351–362. Xiao, B., Wang, Q., Zhao, Y., Shao, M., 2011. Artificial culture of biological soil crusts and its effects on overland flow and infiltration under simulated rainfall. Appl. Soil Ecol. 46, 11–17. Yair, A., Sharon, D., Lavee, H., 1980. Trends in runoff and erosion processes over an arid limestone hillside, northern Negev, Israel. Hydrol. Sci. 25, 243–255. Zhang, Y.-F., Wang, X.-P., Hu, R., Pan, Y.-X., Zhang, H., 2013. Stemflow in two xerophytic shrubs and its significance to soil water and nutrient enrichment. Ecol. Res. 28, 567–579.

Further Reading Neave, M., Abrahams, A.D., 2002. Vegetation influences on water yields from grassland and shrubland ecosystems in the Chihuahuan Desert. Earth Surf. Process. Landf. 27, 1011–1020. Navar, J., Bryan, R., 1990. Interception loss and rainfall redistribution by three semi-arid shrubs growing in northeastern Mexico. J. Hydrol. 115, 51–63.