Standards Vary in Studies Using Rainfall Simulators to Evaluate Erosion

UC Agriculture & Natural Resources California Agriculture

Title Standards vary in studies using rainfall simulators to evaluate erosion

Permalink https://escholarship.org/uc/item/1nk6k310

Journal California Agriculture, 66(3)

ISSN 0008-0845

Author Grismer, Mark E

Publication Date 2012

Peer reviewed

eScholarship.org Powered by the California Digital Library University of California ReVIEW Article ▼ Standards vary in studies using rainfall simulators to evaluate erosion

by Mark Grismer

Rainfall simulators are often employed Drake Kevin to measure erosion rates, in order to estimate stream loading of sediment and nutrients in California foothill watersheds. The rainfall simulator en- ables the precise application of artificial rain with controlled drop sizes, intensity and duration. In addition to rain factors such as drop energy and intensity, several soil- and cover-related factors affect erosion rates. While computational models have evolved to quantify erosion based on field measurements taken by rainfall simulators, there has not been a consensus on the methodology to be deployed, especially in forested and remote landscapes. In addition, it is challenging Rainfall simulators are used by researchers to measure erosion and the loading of sediments into streams. UC Davis hydrology professor Mark Grismer (right) and GIS specialist Lee Perlow collect to apply study results from small plots data using a simulator installed on a bare soil slope above Kings Beach, Lake Tahoe. to entire watersheds. To guide future use needle tanks or nozzle sprayers to ap- transport. Runoff, as overland flow, car- fieldwork on sediment loading to water ply water at desired rates and durations. ries with it the most erodible silt and very bodies, we review key concerns related to Since 2000, extensive studies across the fine sand particles from the soil surface rainfall simulator studies. Lake Tahoe Basin have used drop-former as the water flows downhill. When rills type rainfall simulators (Battany and (small streams or rivulets) form they initi- Grismer 2000) to help determine the im- ate small channels, eventually forming he ability to estimate how land-use pacts of road and forest management on gullies, which can result in massive soil Tpractices affect soil erosion has be- sediment loading to the lake (Grismer and losses. Rainfall simulators are typically come critically important. The U.S. En- Hogan 2004, 2005a, 2005b), and research- used to determine inter-rill erosion rates vironmental Protection Agency, under ers overseas have used rainfall simulators and their dependence on rainfall and soil section 303(d) of the Clean Water Act, has in similar studies. parameters. listed dozens of streams in California and A major handicap in this area of criti- European researchers have tried to the Western states as impaired or threat- cal research, however, is that there is no develop standards for the use of rainfall ened due to excessive sediment concentra- standardized methodology for measur- simulators. Parsons and Lascelles (2006) tions, which adversely affect fish habitat. ing erosion rates. This article reviews the detailed efforts to catalog the rainfall If erosion rates specific to site and land- literature on rainfall simulator techniques simulators in use and their specifications use practices can be measured adequately, and their applicability to forest, rangeland and performance characteristics, and to estimates of stream sediment loading can and ski-run areas in the Sierra Nevada of develop a standard evaluation and test be developed. California (the complete report is avail- methodology so that data from various The unpredictable and infrequent na- able at http://ucanr.org/u.cfm?id=48). studies could be compared. Facing ero- ture of rain makes it difficult to study its sion and stream sedimentation problems Rainfall simulator approaches eroding effects on soils while it is raining. from vineyards and rangelands similar To overcome these difficulties, rainfall In general, rainfall runoff and ero- to those in California, researchers in simulators can be used to apply precisely sion are initiated by the impact of rain Spain, including Cerdà (1997), have been defined “storms” over frames designed drops on bare or nearly bare soils, which interested in California rainfall simula- to capture and enable the measurement detaches and splashes soil particles and tor research. Agassi and Bradford (1999) of runoff and erosion rates. A variety of subsequently transports them downslope rainfall simulators have been developed as part of overland flow. Net erosion rates Online: http://californiaagriculture.ucanr.edu/ over the past two decades and deployed (sediment mass/unit area) are a function landingpage.cfm?article=ca.v066n03p102&fulltext=yes in the field. Rainfall simulators typically of both rain splash and overland flow DOI: 10.3733/ca.v066n03p102

102 CALIFORNIA AGRICULTURE • VOLUME 66, NUMBER 3 reviewed inter-rill erosion measure- Soil cover. The effect of plant canopy value of erosion models lies primarily in ment studies using rainfall simulators cover on reducing runoff and erosion in conservation planning, as tools to predict and found inadequate characterization rangeland is attributed primarily to in- soil loss. Increasingly, though, they are of (1) the type of rainfall simulator and creased litter cover, soil macro-porosity used to develop regulatory guidelines deployed rainfall intensities, mean drop and soil structure, rather than the di- and evaluate compliance when monitor- size, drop size distribution and water rect interception of rainfall. Similarly, ing information is lacking. quality, (2) the soil plot’s physical and rock cover tends to reduce erosion rates Physically or process-based models chemical properties and (3) the type proportionally to the area of coverage. employ mathematical representations of of results obtained and how they were Overland flow on a specific site is difficult flows of mass, momentum and various presented. Later, Kinnell (2005, 2006) re- to measure, and little is known about the forms of energy to describe soil-water viewed several raindrop-affected erosion mechanics of soil loss by this process. processes. They consist of a number of processes in the laboratory and noted that Slope changes. All other factors being linked equations with parameters that conceptual models and measurements equal, it has been established that erosion have direct physical significance and can failed to adequately characterize observed rates increase as slope angles increase; each be evaluated by independent field erosion processes from bare soils. Due to presumably as overland flow velocities measurements. In principle, physically difficulties in comparing rainfall simula- increase, so does the erosive power and based processes only require representa- tor studies across rangelands and forested transport capacity of runoff to carry sus- tive physical characteristics of the soil- areas of the Tahoe Basin (Foltz et al. 2012; pended sediments. Slope angle is also water system in the model for the results Grismer and Hogan 2004), members of the important to how raindrop splashes affect to be realistic. Tahoe Science Consortium have recently erosion; as steepness increases, more soil Universal Soil Loss Equation. The raised concerns about the variety of rain- is splashed downhill. However, the runoff Universal Soil Loss Equation (USLE) fall simulator methods and the lack of rate is most sensitive to slope change; be- was codified in 1965 in the U.S. standardization in measuring infiltration yond a soil- and cover-dependent thresh- Department of Agriculture (USDA) and erosion rates. old, it is the dominant erosive process. Agriculture Handbook No. 282 and re- Interrelated processes. Erosion from vised by Wischmeier and Smith (1978) in The erosion process soil surfaces involves interrelated pro- Agriculture Handbook No. 537. The USLE Raindrop energy. A raindrop’s kinetic cesses that combine in complex spatial was derived from statistical analyses energy (KE) is one-half of the product of and temporal variations. These processes of natural runoff and erosion data and its size (mass) and velocity squared. Lal include particle (aggregate) breakdown equivalent rainfall simulator–derived plot (1988) maintained that kinetic energy is a and detachment; rain splash effects fol- data largely gathered in the central United major factor in the soil detachment pro- lowed by particle suspension and trans- States. The authors emphasized that the cess, and therefore that the total energy of port as part of overland flow or wind USLE was an erosion model designed a storm is proportional to its “erosivity.” transport; particle filtration by covers to predict the long-term average annual It has been shown in statistical analyses or mulch layers; and particle movement soil losses from rill and inter-rill erosion that kinetic energy is insufficient to de- into the soil profile. These processes are that might be expected from specific field scribe erosivity; the terms “erosivity,” or affected by basic hydrologic phenomena “erodibility,” in fact stem from qualitative such as precipitation form and rates, soil descriptions and lack quantitative defini- infiltration rates and capacity, and soil tions based on physical processes. surface conditions. Infiltration and erosion.The impact In contrast, most water erosion re- Rachel McCullough of raindrops on bare soil compacts the search assumes the simplest conditions: surface and may detach soil particles; the bare soils (no cover or mulch) of known soil surface may become sealed, reducing texture and bulk density, on mild slopes the infiltration rate. For mild bare slopes, (< 10%) with no infiltration-limiting layer. detachment and rain splash are the domi- With the exception of areas that have nant factors causing erosion. As the slope roads or are disturbed, such conditions angle increases, runoff becomes the domi- are rarely found in rangelands or forests nant factor. When the rainfall intensity of the California foothills. exceeds the infiltration rate, surface water Erosion loss models accumulates on the soil, and when surface depressions are filled, runoff can occur. Following the Dust Bowl era and the Increased surface roughness due to soil consequent dramatic losses of soils due to textural variations, tillage, residues on cultivation and grazing, research efforts the surface or the presence of living plant were directed at determining the primary stems reduces the velocity of overland factors contributing to soil losses from flow. Soil surface cover, in the form of liv- agriculture. These efforts included the ing vegetation or residues, reduces the im- development of simple-to-use equations This simulator uses a frame with nozzles (top) to pact (kinetic energy) of the raindrops and and models for estimating erosion rates apply water at various rates and a frame on the prevents them from striking bare soil. under various agricultural practices. The ground to catch sediments.

http://californiaagriculture.ucanr.edu • July–September 2012 103 areas under various cropping and management systems. The USLE identified six major erosion factors, the product of which represents average annual soil loss: (1) A = R × K × L × S × C × P where: A = estimated soil loss (tons per acre per year), R = rainfall runoff, K = soil erodibility, L = slope length, S = slope steepness, C = cover and management and P = supporting practice. Rainfall runoff (R) is a key factor of the USLE model and is determined by local climate conditions. The erodibility factor (K) is determined from the soil type, and C the management and practices factors ( Drake Kevin and P) are estimated from tables of values The models for analyzing rainfall simulator data are based on rainfall and soil loss rates, slope length associated with management and practice and steepness, percentage of vegetative cover and other factors. Monitoring specialist Mike Ukraine descriptions. The USLE equation was de- filters samples at the Integrated Environmental Restoration Services lab in Tahoe City. rived from soil loss data measured from erosion plots after about 1 year of runoff, the WEPP model and related equations (the product of runoff velocity and land and as such the equation predicts annual should be replaced by more careful defini- slope). A rise in stream power likely in- accumulated soil losses rather than indi- tions of the forces acting on hypothetical creases possible aggregate disintegration, vidual rain event losses. soil particles or aggregates. Presumably and there may be a practical threshold WEPP model. Later, the Water Erosion from there the forces or energy needed for of stream power effects to consider in and Prediction Project (WEPP) model was aggregate breakdown could be applied detachment modeling (Fristensky and developed (Nearing et al. 1990) with the to determine the extent of finer particle Grismer 2009). Thus, either the physical concept of developing a physically based liberation and subsequent transport process description given by equations mathematical description of erosion pro- (Fristensky and Grismer 2009). Owoputi such as those in the WEPP model is inad- cesses, but it also uses the equivalent of and Stolte (1995) underscored the need equate, or erodibility needs greater clarifi- the K, C and P factors of the USLE equa- to account for the moisture dependence cation and evaluation. tion. Both the USLE equation and WEPP of soil strength and seepage, though in a Natural and simulated rainfall model need estimates of inter-rill erod- rainfall- or runoff-induced erosion event it ibility (K), which can be obtained using is likely that the surface soil layers are at The role of raindrop velocity, or energy, rainfall simulators. or near saturation, their weakest state. in the splash detachment of soil particles Similarly, in a thorough review of has been a concern for decades (Bisal Post-WEPP developments erosion induced by raindrops on mildly 1960; Ellison 1947). Debate centers on While the WEPP model and its re- sloping bare soils, Kinnell (2005) claimed whether raindrop size, velocity, momen- lated equations represent accumulated that current models “do not represent all tum, kinetic energy or some combination research of the past several decades, they of the erosion processes well.” None of of these is the key parameter in the design originated from Ellison’s (1947) paradigm the models deal with temporal changes of rainfall simulators used for erosion that “erosion is a process of detachment in surface properties, and all simplify the studies. In addition, a threshold concept and transport of soil materials by erosive process descriptions to a planar surface must account for the limited erosion rates agents.” Such a view has come under lacking the variations in microtopography encountered during low-intensity storms criticism, because erosion processes are or surface roughness found in even rela- (for which the use of kinetic energy alone sufficiently complex that many questions tively smooth field soils. Grismer (2007) tends to overestimate erosion rates). remain unresolved, including laminar noted that the research briefly summa- Nonetheless, in contrast to earlier studies, versus turbulent flows in the field; the rized here, and similar studies, by neces- recent work includes determinations of fundamental applicability of the turbu- sity were conducted on bare soils and as rainfall kinetic energy as a measure of to- lent flow–based shear stress equations to a result may not apply to duff-covered or tal energy available for aggregate disinte- slopes greater than 10%; the discrepancy litter- and mulch-matted range and for- gration, detachment and transport. These between measured and modeled soil est soils in which the dominant sediment estimated kinetic energies depend in part shear strength; and raindrop effects on detachment and transport processes are on drop sizes and their distribution. steeper, relatively undisturbed forest perhaps better characterized as filtration. The median drop size of natural rain- soils. As a result, the precise definition of According to Zhang et al. (2003), soil fall varies with intensity. Several studies erodibility remains elusive (Agassi and erodibility would ideally be quantitatively suggest that drop sizes of around 2.5 mil- Bradford 1999). defined as a detachment or transport limeters may be appropriate for simulated Owoputi and Stolte (1995) suggested coefficient relating soil detachment rates rainfall at the intensities often employed that the semi-empiricism implicit in to an appropriate form of stream power in the field. When drop size distributions

104 CALIFORNIA AGRICULTURE • VOLUME 66, NUMBER 3 are expressed as a fraction of the rain sizes, their distributions and kinetic ener- been used in erosion-related research, as event’s volume and intensity, relatively gies. It is not clear if the variability of nat- reported in more than a dozen journals, low-intensity events are dominated by ural rainfall duration, intensity and drop of which around 80% were the nozzle drop sizes of less than 1 millimeter, while size is critical in terms of soil detachment type and the remainder variations on rainfall intensities between 40 and 120 and erosion, if the mean or maximum ki- the drop-former type. (See full report at millimeters (1.6 to 4 .8 inches) per hour netic energies are known or estimated. http://ucanr.org/u.cfm?id=48 for a sum- are associated with a median drop size mary of rainfall simulator characteristics.) Rainfall simulator designs of around 2 millimeters. Few direct mea- Two rainfall simulators used in a variety surements of kinetic energy for simulated Rainfall simulators must be designed of field environments across a range of and natural rainfall exist; rather, kinetic to meet competing demands: replication slopes for roughly 1-square-meter plots energies are estimated from drop sizes, of natural rainfall, ease of portability have emerged as de facto standards: the assumed distributions and fall heights or across remote and steep terrain, reason- oscillating veejet nozzle system (Paige terminal and nozzle velocities. able costs of construction, and uniformity et al. 2003) and the needle drop-former Van Dijk et al. (2002) reviewed studies across the test plots in terms of rainfall (Battany and Grismer 2000). Assuming of the relationship between rainfall drop intensity, drop size and kinetic energy. cost and portability of the two are rel- sizes, intensity and kinetic energies from Duplicating the range of drop sizes and atively equivalent, the differences are around the world and found that in good kinetic energies of natural rainfall has related to their simulated rainfall charac- quality data, kinetic energy ranged from proven difficult. teristics. Simple drop-former designs are 11 to 36 joules per square meter per mil- Two types of rainfall simulators have commonly used where access is more dif- limeter depth (J/m2-mm) with maximum emerged in field research, broadly catego- ficult or water availability is limited. values averaging around 29 J/m2-mm rized as the spray/sprinkler nozzle and Method evaluations. Rainfall simula- and minimum values of about 12 J/m2- the drop-former, which simulate rain in- tors have been widely used to assess ero- mm. Particular kinetic energy values tensities of 10 to 200 millimeters per hour sion control or treatment technologies. depended on location, type of storm and and drop sizes of 0.1 to 6 millimeters. In Sutherland (1998a, 1998b) noted that the storm pattern. They found that high- terms of size, rainfall simulators range “formative years” prior to around 1990 intensity storms typical of rainfall simu- from a simple, small, portable infiltrom- produced a mass of information that lator studies (> 40 millimeters per hour) eter with a 6-inch-diameter rainfall area lacked scientifically credible, standard- result in average kinetic energies of 23 to (Bhardwaj and Singh 1992) to the complex ized methods or data from actual applica- 24 J/m2-mm. Kentucky rainfall simulator, which covers tions. His arguments for standardized Overlooked by van Dijk et al. (2002) a plot 14.75 feet by 72 feet (4.5 meters by evaluation methods that have field ap- were earlier studies (Madden et al. 1998) 22 meters) (Moore et al. 1983). plicability, with greater emphasis on the that used piezoelectric crystals to directly Many original laboratory rainfall simu- study of surface or near-surface processes measure natural and simulated raindrop lators were of the nozzle type, presumably controlling erosion, remain valid more power (kinetic energy per unit of time). due to ease of construction. Laboratory than a decade later. Simulated rains at intensities of 23 to 48 drop-former simulators emerged later in Relatively portable rainfall simulators millimeters per hour developed powers response to uncertainties associated with have been more commonly deployed in of 200 to 1,320 joules per square meter nozzle-generated drop sizes, distribu- the past two or three decades with cor- per hour (J/m2-hr), while natural rainfall tions and intensities. In the past decade, responding plots of 1 or 2 square yards powers for 85 events ranged from around about 40 different rainfall simulators have that are well suited to a wide range of 200 to 3,000 J/m2-hr at intensities of 1 and field studies, particularly where access 42 millimeters per hour, but reached as is difficult or when multiple replications much as 6,000 J/m2-hr for short, high- are needed across a large area. They have intensity storm events. been used to study runoff and erosion What this range of kinetic energies at mechanisms in a wide range of environ- given intensities means with respect to Rachel McCullough ments; however, in practice these rainfall the evaluation of erodibilities remains simulators necessarily fail to accurately unclear. Van Dijk et al. (2002) commented, replicate natural rainfall characteristics, “In terms of process-based research, it due to their portability, cost design or appears that our knowledge of the distri- management limitations. While runoff bution of drop size and terminal velocity and erosion rates from rangeland and for- in natural rainfall is well ahead of our est soils are generally much lower than understanding of the way in which these that from bare and disturbed soils, these interact to detach and transport soil par- latter soils often comprise substantially ticles by splash.” larger areas within watersheds and as a In another review, Dunkerley (2008) result may contribute significant loading lamented that most rainfall simulator– to streams. based studies employ extreme rainfall Standardization of erosion studies using rainfall However, there have been few di- simulators would allow more effective com- intensities for the application region or parisons and data analysis. Above, a collection rect field measurements of runoff and duration, with an overemphasis on drop frame after a simulation near Lake Tahoe. erosion rates, or modeling approaches

http://californiaagriculture.ucanr.edu • July–September 2012 105 capable of predicting these rates, from issues relative to the larger landscape. Scalability issues less-disturbed forest and rangeland soils Methodological variations and sources In most forested catchments, the main (Grismer 2012). Meyer (1988) contended of uncertainty regarding the comparison sources of stream sediment are erosion that simulated rainfall results only give of results include water supply (water associated with disturbances such as dirt relative, rather than absolute, erosion data, chemistry and soil interaction); simu- access roads (for logging and fire control), and that to correlate the simulation results lated rainfall characteristics (e.g., drop and log skid trails and channel inci- to that of natural events, data from similar size, intensity and kinetic energy); plot sions linked to increased overland flows plots subject to long-term natural rainfall runoff frame size and installation; runoff following disturbances. Nonetheless, events must be available for comparison sampling size, frequency and duration; rainfall simulator erosion evaluations are (Hamed et al. 2002). Nonetheless, rainfall the identification of plot cover, slope and conducted in the field to guide general simulators in the field continue to be de- surface soil conditions; the measurement assessments of hill slope and catchment veloped and used as few replacements are of inter-rill erosion, rill erosion or com- runoff and the erosion rates associated available for generating physical process– binations; plot replication or the degree with different soils and land uses. Scaling based erosion information. to which plots represent hill slope condi- up to the hill slope or catchment involves Field methodologies. The area of tions; and the interpretation of runoff at least three issues beyond the scope simulated rainfall coverage is inherently sediment sampling relative to local soil, of small-plot rainfall simulator studies: limited by the rainfall simulator, slope, cover and climate conditions. At a mini- (1) the natural heterogeneity of soil condi- available water and the possibility of rep- mum, each of these should be addressed tions (e.g., infiltration and erosion rates) lication, so small field-plot erosion studies in research deploying rainfall simulators across the hill slope, or plot-to-plot vari- are necessarily compromised by sampling to facilitate comparisons between studies. ability; (2) the interconnectivity between measured and nonmeasured areas, or between eroding and depositional areas; TABLE 1. Issues to address in the standardization of field rainfall simulator erosion studies and (3) soil plot disturbance effects as a result of the rainfall simulator measure- Issue/question Comments ments themselves. How do the local natural rain characteristics Include drop size, drop distribution and kinetic energies. Le Bissonnais et al. (1998) noted the compare to those of the rainfall simulator? need to consider the spatial structure of Which rainfall characteristics are expected to be Depends on cover conditions. the catchment, while García-Ruiz et al. important for determining local erosion rates or (2010) and others have highlighted that erodibility? connectivity with fluvial channels is the Are there soil-related thresholds that are critical Aggregate strength is a dynamic soil property that is to determining erodibility? If so, how can they be largely unknown. most important factor linking plot to determined or measured? catchment studies. Both studies under- How do we quantify the soil hydrophobicity effects Hydrophobicity is a dynamic property that increases scored the importance of considering common in range and forest soils of the California runoff rates in late summer or after fire (Rice and Grismer various spatial and temporal scales, since foothills? 2010). it is well known that geomorphic and hy- What is erodibility in the context of forested The definition of erodibility depends on the conceptual landscapes or deeply mulch- or duff-covered soils? equation applied. Information is required about drological processes are scale dependent. How can it best be defined or measured in this infiltration rates, soil compaction, antecedent moisture Some of the issues associated with field situation? and depth to the less-permeable layer. variability, including that introduced by How many replications in studies of runoff or Plot variability effects increase with decreasing measured experimental design of the erosion plot erosion rates are sufficient to characterize the sediment yields; the assumption of evenly distributed (Zöbisch et al. 1996), were recognized sample area of interest? erosion rates may not be valid. With the considerable plot-to-plot variability in measured erosion rates from more than a decade ago (Bagarello and seemingly homogeneous areas, standard replication and Ferro 1998). Unexplained variability statistical analyses should be promoted. between erosion test-plot results, even While erosion rates conceptually increase with Plot variability may have a greater effect on measured in apparently homogeneous, cultivated increasing slope and the associated increased erosion rates than increased slope at less than about 20% fields (Rüttimann et al. 1995), remains runoff rate for given rainfall intensities, are there for bare soils and 50% for forest soils. thresholds below which slope effects are negligible perplexing and limits the development and above which they are significant? of more generalized conclusions about In addition to rainfall and runoff rates, are there Maybe; see above. runoff and erosion rates (Gómez et al. slope-related thresholds, especially on steep slopes, 2001). Variability of 30% to 75% between that are critical to determining their erodibility? plots located on a seemingly homoge- At what combinations of bare-soil slope length, Open question; appears to depend on soil type. surface runoff rate, slope angle and surface neous landscape is common (Foltz et al. condition does rill erosion become dominant 2012; Grismer and Hogan 2004; Nearing compared to inter-rill erosion? et al. 1999). At the same time, knowledge What are the effects of frame installation No studies quantify the effects of frame installation on is needed about soil erosion processes methods, and do frames capture surface erosion measured erosion rates. occurring in field plots across a range of appropriately? sizes, the threshold limits at which dif- What are the most appropriate methods for There may be a threshold-based effect that needs further quantitatively characterizing the plant, mulch and/ clarification or definition (Grismer et al. 2009). ferent processes are significant and the or duff covers? factors that determine natural variability (Bagarello and Ferro 2004).

106 CALIFORNIA AGRICULTURE • VOLUME 66, NUMBER 3 Boix-Fayos et al. (2006) sought to re- to relate sediment yields to soil type, rainfall simulators will continue to be view these issues within the following condition and slope, in order to model used for decades to come. A standard- framework: “(i) temporal and spatial daily sediment loads from paired 630- ized methodology for rainfall simulator scales, (ii) representation of natural con- and 1,300-acre watersheds on the Tahoe design, runoff frame installation and the ditions, (iii) the disturbance of natural Basin west shore. Unfortunately, actual analysis of results needs to be developed conditions and (iv) accounting for the field data on infiltration and erosion rates and applied to all studies as they attempt complexity of ecosystem interactions.” at different spatial scales from one to tens to address key issues (table 1). Ultimately, the uncertainties associated of meters are difficult to obtain, and little with these issues are set aside so that ero- can be found in the literature (Bagarello sion predictions can be made as part of and Ferro 2004, 2010), since most field watershed process modeling, to evaluate measurements have concentrated on wa- the effects of changing landscape condi- ter erosion at the runoff plot scale (Poesen tions on watershed health and the quality and Hooke 1997). M. Grismer is Professor of Hydrology, Depart- of discharge water. ments of Land, Air and Water Resources, and Bio- Key questions logical and Agricultural Engineering, UC Davis. The research related to scaling erosion Co-workers at Integrated Environmental Res- estimates from plot-based measurements The lack of quantitative, physical- toration Services of Tahoe City supported the rain- to the entire hill slope or catchment is process-based information about infil- fall simulator fieldwork, and Marta Ruiz-Coleme, conflicting. Grismer (2012) used 1-square- tration, runoff and erosion rates from Ph.D. Student, University of Madrid, Spain, pro- meter rainfall simulator erosion test plots forest and range conditions suggests that vided valuable library research.

References Gómez JA, Nearing MA, Giráldez JV, Alberts EE. 2001. Moore ID, Hirschi MC, Barfield BJ. 1983. Kentucky rainfall Analysis of sources of variability of runoff volume in a simulator. Trans ASAE 26:1085–9. Agassi M, Bradford JM. 1999. Methodologies for inter-rill 40-plot experiment using a numerical model. J Hydrol soil erosion studies. Soil Till Res 49:277–87. Nearing MA, Deer-Ascough L, Laflen JM. 1990. Sensitiv- 248:183–97. ity analysis of the WEPP hill-slope profile erosion model. Bagarello V, Ferro V. 1998. Calibrating storage tanks for Grismer ME. 2007. Soil restoration and erosion control: T ASAE 33(3):839–49. soil erosion measurement from plots. Earth Surf Proc Quantitative assessment in rangeland and forested areas. Land 23:1151–70. Nearing MA, Govers G, Darrell Norton L. 1999. Variability T ASABE 50(5):1619–26. in soil erosion data from replicated plots. Soil Sci Soc Bagarello V, Ferro V. 2004. Plot-scale measurement of soil Grismer ME. 2012. Erosion modeling for land manage- Am J 63:1829–35. erosion at the experimental area of Sparacia (southern ment in the Tahoe Basin, USA: Scaling from plots to small Italy). Hydrol Process 18:141–57. Owoputi LO, Stolte WJ. 1995. Soil detachment in the forest catchments. Hydrol Sci J 57(5):1–20. physically based soil erosion process: A review. T ASAE Bagarello V, Ferro V. 2010. Analysis of soil loss data from Grismer ME, Hogan MP. 2004. Simulated rainfall evalua- 38:1099–110. plots of differing length for the Sparacia experimental tion of revegetation/mulch erosion control in the Lake area, Sicily, Italy. Biosyst Eng 105:411–22. Paige GB, Stone JJ, Smith JR, Kennedy JR. 2003. The Tahoe Basin: 1. Method assessment. Land Degrad Dev Walnut Gulch rainfall simulator: A computer-controlled Battany MC, Grismer ME. 2000. Development of a por- 15(6):573–88. variable intensity rainfall simulator. Appl Eng Agric table field rainfall simulator for use in hillside vineyard Grismer ME, Hogan MP. 2005a. Simulated rainfall evalu- 20(1):25–31. runoff and erosion studies. Hydrol Process 14(6):1119–29. ation of revegetation/mulch erosion control in the Lake Parsons AJ, Lascelles B. 2006. Rainfall simulation in geo- Bhardwaj A, Singh R. 1992. Development of a portable Tahoe Basin: 2. Bare soil assessment. Land Degrad Dev morphology. Earth Surf Proc Land 25:679. rainfall simulator infiltrometer for infiltration runoff and 16(4):397–404. erosion studies. Ag Water Manage 22(3):235–48. Poesen JW, Hooke JM. 1997. Erosion, flooding and chan- Grismer ME, Hogan MP. 2005b. Simulated rainfall evalu- nel management in Mediterranean environments of Bisal F. 1960. The effect of raindrop size and impact veloc- ation of revegetation/mulch erosion control in the Lake southern Europe. Prog Phys Geog 21(2):157–99. ity on sand splash. Can J Soil Sci 40:242–5. Tahoe Basin: 3. Soil treatment effects. Land Degrad Dev 16(5):489–501. Rice EC, Grismer ME. 2010. Dry-season soil-water re- Boix-Fayos C, Martinez-Mena M, Calvo-Cases A, et al. pellency affects Tahoe Basin infiltration rates. Cal Ag 2006. Measuring soil erosion by field plots: Understand- Grismer ME, Shnurrenberger C, Arst R, Hogan MP. 2009. 64(3):141–8. ing the sources of variation. Earth-Science Rev 78:267–85. Integrated monitoring and assessment of soil restoration Rüttimann M, Schaub D, Prasuhn V, Rüegg W. 1995. Cerdà A. 1997. The response of abandoned terraces treatments in the Lake Tahoe Basin. Env Monit Assess 150:365–83. Measurement of runoff and soil erosion on regularly to simulated rain. In: Rickson RJ (ed.). Conserving Soil cultivated fields in Switzerland — some critical consider- Resources: European Perspectives. Oxford, UK: CAB Int. Hamed Y, Albergel J, Pépin YJ. 2002. Comparison be- ations. Catena 25:127–39. p 44–55. tween rainfall simulator erosion and observed reservoir Sutherland RA. 1998a. Rolled erosion control systems for Dunkerley D. 2008. Rain event properties in nature and sedimentation in an erosion-sensitive semi-arid catchment. Catena 50:1–16. hillslope surface protection: A critical review, synthesis in rainfall simulation experiments: A comparative review and analysis of available data. I. Background and forma- with recommendations for increasingly systematic study Kinnell PIA. 2005. Raindrop-impact-induced erosion pro- tive years. Land Degrad Dev 9:465–86. and reporting. Hydrol Process 22:4415–35. cess and prediction: A review. Hydrol Process 19:2815–44. Sutherland RA. 1998b. Rolled erosion control systems for Ellison WD. 1947. Soil erosion studies, Part I. Agr Eng Kinnell PIA. 2006. Simulations demonstrating interac- hillslope surface protection: A critical review, synthesis 28:145–6. tion between coarse and fine sediment loads in rain- and analysis of available data. II. The post-1990 period. Foltz RB, Elliot WJ, Wagenbrenner NS. 2012. Soil erosion impacted flow. Earth Surf Proc Land 31:355–67. Land Degrad Dev 9:487–511. model predictions using parent material/soil texture- Lal R. 1988. Erodibility and erosivity. In: Lal R (ed.). Soil Van Dijk AJM, Bruijnzeel LA, Rosewell CJ. 2002. Rainfall based parameters compared to using site-specific pa- Erosion Research Methods. Soil and Water Conservation intensity — kinetic energy relationships: A critical litera- rameters. T ASABE 54(4):1347–56. Society, Ankeny, IA. p 141–60. ture appraisal. J Hydrol 261:1–23. Fristensky A, Grismer ME. 2009. Evaluation of ultrasonic Le Bissonnais Y, Benkhadra H, Chaplot V, et al. 1998. Crust- Wischmeier WH, Smith DD. 1978. Predicting Rainfall aggregate stability and rainfall erosion resistance of dis- ing, runoff and sheet erosion on silty loamy soils at vari- 2 Erosion Losses: A Guide to Conservation Planning. Agri- turbed and amended soils in the Lake Tahoe Basin, USA. ous scales and upscaling from m to small catchments. culture Handbook, No. 537. USDA/Science and Education Catena 79:93–102. Soil Till Res 46:69–80. Administration, Washington, DC. 58 p. García-Ruiz JM, Lana-Renault N, Beguería S, et al. 2010. Madden LV, Wilson L, Ntahimpera N. 1998. Calibration Zhang GH, Liu BY, Liu GB, et al. 2003. Detachment of un- From plot to regional scales: Interactions of slope and and evaluation of an electronic sensor for rainfall kinetic disturbed soil by shallow flow. Soil Sci Soc Am J 67:713–9. catchment hydrological and geomorphic processes in energy. Phytopathology 88(9):950–9. the Spanish Pyrenees. Geomorphology 120:248–57. Zöbisch MA, Klingspor P, Oduor AR. 1996. The accuracy Meyer LD. 1988. Rainfall simulators for soil conservation of manual runoff and sediment sampling from erosion research. In: Lal R (ed.). Soil Erosion Research Methods. Soil plots. J Soil Water Conserv 51(3):231–3. and Water Conservation Society, Ankeny, IA. p 75–96.

http://californiaagriculture.ucanr.edu • July–September 2012 107