Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
3 The Measures of Soil Erosion It is possible today to locate the most highly erodible croplands in most regions of the country. Likewise, land not subject to serious ero- sion under any management system also can be readily identified. The effect of alternative conservation practices on erosion losses can be accurately estimated for a diversity of field conditions. Although the effects of the concentrated flow of water on erosion and sediment yield are not well established, many of the basic physical factors contributing to erosion and accounting for erosion control are well known. These analytical capabilities made possible by the development of physi- cally based erosion equations have revolutionized conservation plan- ning and program administration. The Universal Soil Loss Equation (USLE) is used to estimate the long- term average amount of soil displaced by the forces of rainfall and water runoff along a specified slope. Similarly, an equation has been devel- oped that represents soil loss caused by wind; however, the Wind Erosion Equation (WEE) is far less accurate than the USLE. Currently, there is no widely accepted, practical method for estimating another form of erosion known as ephemeral gully erosion. The topography of most fields causes runoff to collect and concentrate in a few major natural waterways or swales before leaving the fields (Foster, 1982; Thorne, 1984~. These features are often ephemeral, and the erosion that occurs in them can be called ephemeral gully, concentrated flow, or megarill erosion. Ephemeral gully areas within fields are plowed in and tilled across annually, in contrast to the permanency of classical gullies 34
THE MEASURES OF SOIL EROSION 35 (Foster, 1986). (Ephemeral gully erosion is discussed in detail in the final section of this chapter.) NRI Estimates of Sheet and Rill Erosion: The USLE Developed in the late 1450s, the USLE is designed to predict long- term average soil losses through sheet and rill erosion from specific land areas under specified cropping and management systems. In a sense, soil loss is a misnomer; movement or displacement are better terms. Eroding soil is never lost in the sense of disappearing. Often it is merely moved from one part of a field to another, to be deposited in low-lying parts of the landscape. In other cases, soil is moved from the land surface and transported in streams and rivers. (Soil eroding off and down a sloping field, however, is "lost" from its point of origin.) USLE estimates, commonly expressed in terms of tons/acre year, do not accurately represent the amount of soil that leaves a field, enters a body of water, or otherwise contributes to offsite erosion damage. Heavy sheet and rill erosion on sloping cropland, probably exceeding 30 tons/ acre year (Pottawattamie County, Iowa). Severe sheet and rill erosion is limited to a relatively small proportion of cropland in the United States. For example, approximately 7 percent of the land in row and close-grown crops accounts for 41 percent of the total tonnage of sheet and rill erosion on land in that use. Credit: U.S. Department of Agriculture, Soil Conservation Service.
36 SOIL CONSERVATION (Models designed for this purpose are in use, as noted in Chapter 2.) Nor do erosion rates estimated by the USLE necessarily correspond directly to the severity of onsite damages to land productivity, since such factors as soil depth and subsoil quality are also important deter- minants of a soil's vulnerability to erosion. Rather, the USLE estimates the long-term average amount of soil displaced by the forces of rainfall and water runoff along a specified slope. The segment represented may be on or off the field, at the bottom of a hill, in a terrace channel, or in a natural depression along a slope. The form of the equation is: A = RKLSCP where A is the computed soil loss per unit area over a specified time; it is usually expressed as tons/acre year. The factors R. K, and S reflect characteristics of climate and land that generally cannot be modified by human activity designed to influence erosion rates: amount and inten- sity of rainfall (R), soil erodibility (K), and steepness of field slope (S). The factor reflecting length of slope (L) can be reduced by installing terraces, which effectively break the naturally occurring slope length into smaller segments. (The effective slope length can also be shortened by stripcropping and grassed waterways, but this is reflected in the P factor.) As noted in Chapter 1, however, terraces are in use on limited acreage nationwide. The remaining two factors reflect the effects of human activities on erosion rates: soil cover and management practices (C) and supporting conservation practices (P). R is a numerical indicator of the erosive forces of rainfall and runoff, developed from rainfall data averaged over several decades. The R factor is the single most influential factor in the equation. Its value varies from 550 along parts of the high rainfall Gulf Coast to 20 in the arid West. The values for R are generally less reliable for western regions in any given year. This occurs because data from the West were insufficient to develop the equation, and rainfall there is intermittent and often torrential, causing average annual sheet and rill erosion esti- mates to vary greatly in those areas. For these and other reasons, con- cern has been expressed that the equation is not suited to evaluate potential erosion from rangelands (Society for Range Management, 1985~. The committee has not dealt specifically with rangeland in this report. Other sources of runoff are also important in parts of the West. For the 1982 NRI, modifications of the it-factor values were made for sam- ple points on frozen soils in the Pacific Northwest. K reflects the inherent susceptibility of a soil to erode, if it is barren of
THE MEASURES OF SOIL EROSION 37 crop cover or residue and exposed to rainfall and runoff. The values for K (ranging from 0.7 for highly erodible soils to 0.2 for soils resistant to erosion) are a function of texture (tine percentages of sand, silt, andclay- sized particles), organic matter content, physical structure, and perme- ability to water. L and S factors represent the effects of slope length and steepness, respectively. The erosive force of water runoff increases as slope steep- ness and length increase; the values of these factors increase accord- ingly. In field applications, including the 1982 NRI, these factors are combined into a single LS factor according to procedures in SCS techni- cal guides or Agriculture Handbook No. 537 (Wischmeier and Smith, 1978~. Determination of a value for LS in the field can be an imprecise exercise. If a prevailing slope can be identified, the surveyor can usually estimate slope steepness fairly accurately with a clinometer, an instru- ment designed to measure angles of elevation or inclination. Slope length is a parameter that is difficult to evaluate. Once the vector is selected, however, slope length can be determined by measuring with a tape or by pacing. If slopes vary greatly in a field, it is difficult to determine which is the prevailing slope. An average slope steepness is often assigned to a field after clinometer readings have been made for several slopes. Similar judgments must be made to estimate slope length. The product of the four factors, RKLS, yields an estimate of the average annual sheet and till erosion (in tons) expected if an area of land were tilled continuously up and down any prevailing slope and kept barren of vegetation. These conditions correspond to a C value of 1 and a P value of 1. The value for the RKLS product thus represents a soil's inherent potential for sheet and rill erosion. Any erosion control prac- tices reduce soil loss on a particular field by lowering either or both of the C or P values below 1. C is the vegetative cover and management factor. Values for the C factor are multiplied by the product RKLS to represent the reduction from inherent erodibility brought about by cropping sequences, tillage practices, and plant residues on the soil surface. For example, a C-factor value of 0.30 the average value for cropland in the 1982 NRI means that vegetative cover and management reduced sheet and rill erosion on cropland to 30 percent of its inherent potential rate, estimated by the product RKLS. C-factor values represent the ratio of soil loss from land cropped under specified conditions to the corresponding loss from clean-tilled, continuous fallow. Values for the C factor range from as low as 0.003 for a dense vegetative cover such as permanent, high-
38 SOIL CONSERVATION quality pasture to 0.7 for crops that produce very small amounts of residue cotton, for example and fields that are extensively tilled (see the boxed article Key Role of C Factors in Controlling Erosion). The factor P reflects the erosion control effects of such supporting conservation practices as contouring, stripcropping, and terracing. These practices break up the lengths of downslope segments traveled by runoff water into shorter segments, thus limiting the volume and velocity of moving water. As a result, less soil is displaced and trans- ported. The P factor is the ratio of soil loss with a specific support practice or practices to the corresponding loss with up-and-down slope cultivation. Values for the P factor are multiplied in the equation in the same manner as C values. A value of 0.91 the average value for P on cropland in the 1982 NRI means that the estimated sheet and rill ero- sion would be reduced 9 percent below the rate that would occur with- out the supporting practice. NR! Findings On Average USLE-Factor Values National average 1982 NRI values for the RKLS product (inherent potential for sheet and rill erosion) and for C and P factors on major cropland uses are shown in Table 3-1. In each case, individual values for each factor were extracted from the NRI sample point files in respective cropland uses. The acreage associated with each factor was determined by applying the relevant NRI acreage expansion factor to each sample point. In Table 3-1, the column labeled RKLS indicates the sheet and rill erosion rates that would be expected if land in each use were contin- ually tilled up and down the slope and left barren of vegetation, a worst-case scenario. On all cropland, average soil displacement under these conditions would tee 21.8 tons/acre~year. Cropland used for close- grown crops such as wheat would erode at the lower rate of 15.4 tons/ acre~year. This rate reflects the fact that a substantial portion of the close-grown crop acreage is in the Great Plains, where rainfall and R- factor values tend to be lower. Cropland used for hay generally has a high inherent potential for sheet and rill erosion (35.2 tons/acre year). Use of that land in hay crops, however, alleviates most of the problem, because of the dense soil cover maintained throughout the year, charac- terized by the low C-factor values associated with sod-based crops. The USLE was developed for a single crop per year; therefore, its applica- tion to double-cropping is limited. The distribution of acreage by the RKLS product presents a pattern similar to the distribution of acreage according to sheet and rill erosion rates. The practical significance of
a, - · - Q o 5- o o cn o 5- - a, a' En o CD ._ CD ._ Cal ~ 5- o O ·CD ~ so LO o u En . ¢ Ed o Lou ~ o ~ . _ U ~ 5 . - o u ~ ~ o .-o o I YOU u - o In QK' ^' 5-'1 U o o CD AL ho 39 ~ ~ ~ Do oo oo oo oo . . . . . o o o o o oo O ~ ~ ~D CN ~ o ~ ~ o o o o o ~D ~ ~ ~ ~ . . . . . ~ ~ oo oo o oo oo oo ~ ~ ~ . . . . . d4 00 ·n CN ~ ~ C~ . . . . . ~ ~ 0 ~ ~ d~ ~ Lr, oo r' ~ ~ o~ =3 ~ ° e,. U , ~ ~ o ~: U X O ¢ ~; U o CD b4 ~ ~ ._ · C ._ ~ oc ,o=, ~a o ,S, .= O 9 O a, O U O ~ O U ~ .= ~ ~ ci, :: 00 C~ . - 5 - a; cn o o a' . ~ z 00 . . u o U)
40 SOIL CONSERVATION this distribution of potential erosion problems and related policy impli- cations are discussed in Chapter 5. Across the various land uses, including cropland, pastureland, and forestland, the highest average value for the C factor, 0.28 (see Table 3-1), is for land used for row crops. This value reflects the fact that row- crop farming usually involves an annual disturbance of the soil surface in the course of preparing a seed bed. Disturbed land is often exposed for at least a few months to the erosive effects of rainfall. The C-factor value of 0.04 reported for hayland is about one-eighth the average value reported for row crops. The difference explains why average sheet and rill erosion rates on hayland are much lower than on land used for row crops, even though land planted to hay is estimated to be about 40 percent more erodible on average than cultivated cropland. Average values for the P factor tend to vary only slightly among cropland uses. Average P-factor values are concentrated at the high end of the theoretical range of values for this factor, averaging 0.91 for all cropland. This distribution of P values indicates that a fairly small per- centage of cropland acreage has been treated with supporting conser- vation practices such as contour farming and stripcropping. Table 3-1 provides an indication of the impact on potential erosion rates of cropping, management, and conservation practices repre- sented by the C and P factors. The column labeled reduction factor is a ratio of the inherent potential erosion (RKLS-product value) and the USLE estimate (RKLS x CP) for each major cropland use. For all crop- land, the inherent potential for sheet and rill erosion was 21.8 tons/ acre~year, but the effect of the C- and P-factor values was to reduce the estimated USLE rate to 4.3 tonslacre~year. The resulting reduction- factor values in Table 3-1 show that potential erosion rates were reduced most significantly on cropland used for hay (from 35.2 to 0.6 tons/acre~year, a 58.7 reduction factor and a 98 percent decrease in erosion), followed by other land (used mostly for vegetables, orchards, and other crops) and by land used for close-grown crops. Potential erosion was reduced considerably on land used for row crops by a factor of 3.7 (or a 73 percent reduction) from an average annual rate of 22.3 to 6.1 tons/acre. The reduction of potential erosion on row- cropped land was less than that for other cropland uses. The USLE is routinely used by USDA conservation agencies for on- farm planning, program evaluation, and analysis. The committee believes, however, that USDA state- and national-level program man- agement and analytic activities could be strengthened by similarly encouraging routine use of the concept of inherent potential for sheet and rill erosion.
THE MEASURES OF SOIL EROSION 41 Recently, the SCS has begun to use the concept of inherent erosion potential, as estimated by the RKLS product, in evaluating alternative land classification systems (see Chapter 5~. The committee believes that this is a positive step. Promotion of proper use of this concept would advance understanding about erosion problems among soil scientists, conservationists, analysts, and farmers. The USDA should prepare and encourage adherence to a special publication that presents common concepts, terminology, and defini- tions of key land use and conservation measures and indicators. Such a publication would thoroughly explain the uses of the concept of inher- ent erosion potential for sheet and rill erosion. Similar guidelines could be issued for wind erosion, when appropriate. The publication could be used by relevant agencies, university researchers, extension service personnel, and other analysts. The SCS is the appropriate lead agency to draft these guidelines with the assistance of soil scientists and engineers working for the Agricultural Research Service (ARS), state experiment stations, and other scientific centers. The publication should be distributed to rele- vant personnel in the ARS, the Economic Research Service, the Forest Service, the Agricultural Stabilization and Conservation Service, and other federal and state agencies. Key Role of C Factors in Controlling Erosion Because val ues for the C factor are contained in the N Rl data fi les for most sample points, the inventory is useful for analyzing the geo- graphic distribution of cropping and management practices and their potential effectiveness in control I ing sheet and ri 11 erosion. Analysis of N Rl data may also help define further research that is needed on the C factor and help refine factor values. Recorded N Rl val ues for the C factor reflect probable sheet and ri 11 erosion conditions nationwide and within such smaller geographic regions as Major Land Resource Areas (MLRAs). Ideally, as the poten- tial for erosion (indicated by the RKLS product) increases on cropland, C values should decline, indicating a greater effort at erosion control through cropping and management practices. However, Figure 3-1 demonstrates that average nationwide C values for land with low erosion potential do notdiffer markedly from values for land with high erosion potential. It also shows the relationship between C- and P-
42 SOIL CONSERVATION NATIONAL SU M MARY 40 35 >' 30 - z o 20 oh O 15 10 25 5 o _~ P-FACTOR ~ 1 ~ EROSION - 0.2 ~ . . . . o.o 180 C-FACTOR 0 30 60 90 120 150 AVERAGE RKLS IN CLASS t/ac/yr 0.8 0.6 O lo 0.4 ~ FIGURE 3-1 Plots of the 1982 NRI weighted average erosion rate (tons/ acre year), C and P factor versus the potential for erosion (tons/acre~year) as expressed by the RKLS product of the USLE. Data were summarized nation- al Iy from the 1 982 N Rl. Each data point is plotted at the midpoint of a range or class of RKLS values. Source: Pierce et al., 1986. factor values and erosion rates as a function of progressively rising inherenterosion potential. The relationship between C-factor values and inherent erodibility differs across the country (see Figure 3-2~. In MLRA 105 (the northern Mississippi Valley loess hills of Wisconsin, Iowa, Minnesota, and Illinois), for example, the average C-factor value is tower than that for the nation as a whole. C-factor values decrease significantly as the potential forerosion increases. ThefactthatCvalues change littlewith inherent erosion potential in MLRAs 103 (the central till prairies of Minnesota and lowa), 134 (the southern Mississippi Valley silty uplancis of Mississippi, Tennessee, and Kentucky), and 1 36 (the south- ern Piecimont of Virginia, North Carolina, South Carolina, Georgia, and Alabama), however, ir~clicates that conservation management techniques are not widely used on erosion-prone soils, nor are they concentrated on the most erodible soils.
THE MEASURES OF SOIL EROSION Similar analyses can be performed for virtually any grouping of cropland. In Figure 3-3 the national average values for the C factor are arrayed by potential erosion rate for class I and subclasses lie, Ille, and IVe of the USDA Lancl Capability Class System (LCCS). (The letter e followingthe class denotes a subclass of land that has suffered erosion damage in the past or is vulnerable to it.) Under this system, the very best land is designated class 1. It has few natural limitations for inten- sive cultivated crop uses. Successive classes have progressively greater limitations for intensive crop production. Classes IV and above are cleemed unsuitable for intensive cropping. Forthenation'sclassiland,C-factorvaluesgenerallydeclineasthe erosion potential increases. However, on subclasses lie, Ille, and IVe, which comprise the majority of erodible U.S. cropland, the C-factor values remain fairly constant as erosion potential increases. This fincl- M LRA ~ 03 30 P- FACTO R ~ - / EROSION _ _ ~ C-FACTOR _af 25 20 5 0 o . o 20 - - it o C`' 30 _ O 2s _ ~ 20 _ 1 5 1 o _ 1 n _ _ 0.8 0.6 nd . . _ 0.2 o.o 40 60 80 100 M L RA 1 34 P-FACTOR _ EROSION _ ~ - C-FACTOR ~ 1.0 0.8 0.6 _ 0.4 _ 0.2 o ~ ~ . . o.o o 20 40 60 80 1 Do MLRA 105 20 0 s o 30 _ _ 1 .o 25 ACTOR _ 0.8 15 EROSION 0.6 ~C- FACTO R _ ~~ o . 0.0 80 1 00 0.4 _ 0.2 o 20 40 60 MLRA 136 _ 20 1 5 . _ 1 o P-FACTOR EROSION 6,' C-FACTOR _ 1.0 ~ 08 ot5 . 0.6 0.4 0.2 o l l l l o.o o 20 40 60 80 1 00 AVE.RKLSIN CLASS t/ac/yr FIGURE 3-2 Plots of the 1982 NRI weighted average erosion rate (tons/ acre~year), C and P factor versus the potential for erosion (tons/acre~year) as expressed by the RKLS product of the USLE for MLRAs 103 (central Iowa and Minnesota till prairies), 105 (northern Mississippi Valley loess hills), 134 (southern Mississippi Valley silty uplands), and 136 (southern Piedmont). Source: Pierce et al., 1986. 43
44 SOIL CONSERVATION NATIONAL SU M MARY 0.4 0.3 o c' 0.2 11 0.1 o.o . ~- _ Fir &, class I class lle · · class Ille 0~0 class IVe 0 30 60 90 120 AVE.RKLSIN CLASS t/ac/yr 150 180 FIGURE 3-3 Plots of the weighted average C factor versus potential for erosion (tons/acre year) as expressed by the RKLS factor of the USLE for land capability subclasses 1, lie, me, and IVe. Data were summarized nationally from the 1982 NRI. Source: Pierce et al., 1986. ing suggests that management and cropping practices beneficial in reducing soil erosion have not been concentrated on those lands that are in greatest need of erosion control. Uncertainties Associated With C-Factor Values A sizable scientific literature exists on the USLE and on the character- istics and proper values of each of its individual factors. As noted later in this chapter, further research is needed on concentrated flow and on the relationship between erosion and the transport of sediment from farm fields to watercourses. The equation is continually being refined to reflect specific climatic, soil, or vegetative conditions and improvements in the understanding of the dynamics of water erosion processes. Such work will continue to
THE MEASURES OF SOIL EROSION 45 be necessary, because land use and tillage practices will change over time. To evaluate the validity and usefulness of the USLE estimates included in the 1982 NRI, the committee thought it advisable to exam- ine some of the conceptual and empirical issues associated with con- temporary C-factor values. Values assigned to C factors are critical determinants of sheet and rill erosion estimates in the USLE. Those values thus have important scientific and policy implications. Although a considerable amount of research has been directed toward refining C-factor values and testing C factors in field experiments, the committee believes that they can be improved through further revision and refinement. The research community, including scientists within the ARS and state agricultural experiment stations, must take a leader- ship role in carrying out this work. Theoretically, the range of possible variation in C-factor values, from 0.001 for undisturbed forestland to 1.0 for clean-tilled, fallow land, makes it the most important variable controlled by human activity in determining estimated sheet and rill erosion rates. In application, how- ever, the full range of C-factor values is primarily most useful when erosion rates are compared among land uses- for example, when land is converted from row crops (average C value of 0.28) to permanent pasture (C value of 0.01 or lower). Within a given land use, C-factor values for most cropland are within a narrower range. For cropland used for row crops, for example, nearly 57 percent of the acreage has C- factor values between 0.25 and 0.45 (Rosenberry and English, 1986~. The C-factor values for a particular subcategory of land use cropland used to grow corn, for example would tend to be even more tightly clustered around the mean value. Despite the narrower range for C-factor values for a given crop or land use, it is important to emphasize that of all the factors in the USLE, the C factor is the one most subject to change as a result of changes in farming practices. Continuous corn production with conventional till- age might yield a C-factor value of 0.37, for example; theoretically, corn produced on comparable land in the same area under no-till planting in sod would be assigned a C value 37 times lower (C = 0.01~. If sheet and rill erosion were 50 tons/acre~year under the conventional system, it would be reduced to less than 2 tons/acre~year through use of the no-till method. Mulch-Factor Value The C-factor values encountered in the NRI data bases are subject to several types of errors. One source of bias arises from the relationship between the percentage of residue cover on the
46 ~ a: I_ SOIL CONSERVATION Corn growing in soybean residue in a no-till system, Jackson County, Iowa. No-till C-factor values are greatly reduced compared with those for continuous corn production with conventional tillage. Credit: U.S. Department of Agricul- ture, Soil Conservation Service. soil surface and its corresponding mulch, or residue, value. The mulch- factor value accounts for the erosion reduction produced by crop resi- dues buried just below the surface. As contained in the USLE, the relationship tends to skew estimated rates of erosion upward by over- estimating C-factor values under some conditions. The mulch value is used to adjust C-factor values in the equation to reflect differences in tillage systems. Recent research (Pierce et al., 1986) indicates that, for a given percent- age of residue cover on the soil surface, the mulch factor can be quite variable (ranging from 0.5 to 0.01), depending on the roughness of a soil surface and the type of residue and extent of its incorporation into the soil. Others have shown less variation. The more extreme findings indicate that values for the mulch factor might be three to four times too
THE MEASURES OF SOIL EROSION 47 high in the USLE used in the NRIs. It implies that the values of the C factor and, hence, estimated erosion rates for cropland reported in the NRI are also too high. The bias is most pronounced when the percentage of surface cover is relatively low (from about 15 to 30 per- cent). Thus, where fairly high C-factor values are encountered in the NRIs, sheet and rill erosion rates might be overstated. Field Application of the USLE Other sources of error in the values for the C factor arise as a result of the application of the equation in the field. Ascribing the correct value for C for a particular location is largely subjective. A standard method is to measure the amount of ground cover along a transect. However, a recent study of variability in residue cover revealed considerable variation among transects for the same field and among observers along the same transect (Richards et al., 1984~. A study of the application of the L, S. and K factors would probably reveal similar variations on fields that are difficult to charac- terize because of variable soils and topography. In the 1982 NRI, as in most field applications of the USLE, SCS personnel generally did not attempt to measure the amount of ground cover. Rather, they determined the type of tillage practice used at a specific sample point and then recorded the appropriate C-factor val- ues from standard generalized tables. Field studies have shown, how- ever, that these tables may be misleading. The use of specific tillage implements or practices can result in a wide range of crop residue levels on the soil surface. A review (Colvin et al., 1981) of three tillage studies found that spring tillage with a disk harrow resulted in crop residue cover ranging from 42 to 73 percent of the soil surface. Such widely ranging crop residue levels have very different implications for soil erosion conditions and raise questions about the actual erosion control afforded by specific practices. Conservation tillage is a case in point (see Chapter 1 for definitions of tillage systems). Estimates of acreage treated with conservation tillage in the 1982 NRI (and with minimum tillage in the 1977 NRI) are based primarily on assignment of C-factor values for that practice taken from standard tables. Had actual residue levels been measured for most sample points, very different C-factor values and resultant erosion rates might have been recorded. A 1980 survey by Peter Nowak (University of Wisconsin, Madison, personal communication) indicates that erosion control benefits might be overstated when field interpretations of this kind are used (see Pierce et al., 1986~. Nowak interviewed 200 farmers from three Iowa watersheds; 78 percent professed to use conservation tillage systems.
48 SOIL CONSERVATION Upon checking residue levels on their farms, however, Nowak found that only 7 percent of the corn acreage and 26 percent of the soybean acreage met SCS technical residue specifications for conservation till- age. If in fact what is: reported as conservation tillage is not achieving expected levels of erosion control implied by the term, such tillage might be providing considerably less erosion control on less acreage across the United States than is suggested by the rapid increase in the use of conservation tillage equipment (chisel plows, for example, and implements other than moldboard plows). Errors of Omission Finally, C-factor values are subject to what might be called errors of omission. In some parts of the United States, existing C-factor values do not adequately represent vegetation-erosion rela- tionships. In certain rangeland areas, for example, a pavement effect or flat rock fragments on exposed soils actually protect those soils from the erosive forces of raindrops and runoff, though nearby soils can be eroded as runoff concentrates along natural channels and drainage- ways. Some researchers believe that this phenomenon should be reflected in K rather than C factors. Others propose that a subfactor should be applied to the USLE in rangeland watersheds prone to soil pavement and concentrated flow effects. The subfactor approach also has been proposed to improve the USLE procedure and C-factor values for forest conditions in the southeastern United States. Many uses can be made of this information for scientific and policy analysis and program design, administration, and evaluation. Rates and total tons of soil displaced through sheet and till erosion can be tabulated for any land use, crop, or land capability class and can be aggregated for the entire country, states, or MLRAs. Acreage in a given land use can be tabulated according to its erosion rate or its erosion rate in relation to its T value. Erosion rates under alternative cropping and conservation conditions can be simulated for any land use (see the boxed article Erosion Under Alternative Cropping, Management, and Conservation Practices). Recommendations on the USLE The data in the 1977 and 1982 NRIs can tee put to many uses, including those illustrated in this report. The sheet and rill erosion information contained in the 1982 NRI is reliable and sufficiently accurate for many analytical applications at the national, regional, and state levels. Those uses include conservation program planning and analysis, refinement
THE MEASURES OF SOIL EROSION 49 Erosion Under Alternative Cropping, Management, and Conservation Practices The sheet and ri 11 erosion estimates in the NRls reflect values for the C factor recordecl at each sample point. By altering those C-factor values in the computerized NRI data base, it is possible to simulate erosion rates that would be expected from alternative cropping and management conditions or the addition of supporting conservation practices. In effect, the USLE files for any geographic aggregation in -the NRls can be used to ask questions about the alternative effects of varying combinations of these conclitions. For comparative purposes other factors such as slope, length, and sol ~ materials are constant. A sample analysis is presented in Figure 3-4. The C-factor values for all NRI sample points falling on land in row and clos~grown crops were altered in the NRI data files to reflect a range of C-factor values. At one end of the continuum, with uniformly assigned C-factor values of 0.30 (the national average value for row and close-grown crops), about 73 percent of this acreage would have sheet and fill erosion NATIONAL SU M MARY 1 00 o to 80 o 111 >, 60 - u~ 40 V CO 20 o _ ROW & CLOSE GROWN HIGH POTENTIAL MEDIUM POTENTIAL l · I 1 1 0.1 0.2 0.3 C FACTOR (ASSIG N ED) FIGURE 3 - Percentage of acres nationally with USLE erosion rates greater than 5 tons/acre year at assumed levels of C factors for land in row and close- grown crops in 1982. Note: High- and medium-potential curves illustrate land with a high or medium potential for conversion to cropland. Source: Pierce et al., 1986.
50 SOIL CONSERVATION rates of less than 5 tons/acre~year. This corresponds closely to the estimated 75 percent of the row and close-grown crop acreage reported as eroding below 5 tons/acre~year in the 1982 NRI. Under a more optimistic assumption, all land in row and close- grown crops is assigned a C-factor value of 0.1, about the average value expected if all land were farmed by no-till methocis the ulti- mate form of conservation tilIage and if heavy crop residue levels were maintained throughout the cropping year. Under these circum- stances about 93 percent of the 323 million acres would experience sheet and rill erosion rates well below 5 tons/acre~year. At least 7 percent of the land now in principal crops over 20 million acres- woulcl still erode at rates above 5 tons/acre~year, even if they were farmed with the most effective, commonly available technology. Some aclditional structural practices inclucling stripcropping, terrac- ing, or permanent vegetative cover would be required to bring ero- sion rates to within the 5-ton maximum T-value level on these lancis. Similar analyses at the MLRA level reveal striking differences in the erosion rates under alternative farming conditions. In such heavily cropped parts of the country as MLRA 103 (the central Iowa and Minnesota till prairies) a much higher proportion of land would be adequately protected from erosion with an average level of conserva- tion management (see Figure 3-5~. By contrast, in MLRA 105 (the northern Mississippi Valley loess hills of Wisconsin, Iowa, Minnesota, and Illinois) and MLRA 136 (the southern Piecimont areas of Virginia, North Carolina, and South Carolina, Georgia, and Alabama), even a high degree of conservation management would leave 20 to 40 per- cent of the acreage vulnerable to erosion rates in excess of 5 tons/ acre~year. Similar analyses can be performed to simulate the effect of convert- ing land from use for row and close-grown crops to less intensive uses. As part of the 1 977 and 1 982 N Rls, officials of the local SCS, the conservation district, and the extension service assessed the potential for future conversion of pastureland, forestland, and rangeland to cropland use, based on the physical characteristics of the land and recent patterns of lancl use in each county or conservation clistrict. Because USLE data were recorded for this potential croplancl, it is possible to estimate what sheet and rill erosion conditions would be like if the land were converted from its present use to more intensive crop uses in the future. Results of such an analysis for land with high and medium potential for conversion are shown in Figures 3~4 and 3-5. Even the high-poten- tial cropland presents, on average, a more significant erosion control
THE MEASURES OF SOIL EROSION M LRA 1 03 100 80 60 40 20 At o En o CE LL - loo V On 111 is: 80 60 40 20 o o.o ROW ~ CLOSE GROWN ~ HIGH POTENTIAL MEDIUM POTENTIAL O , , I . o.o o. 1 0.2 0.3 MLRA 134 0.1 0.2 0.3 MLRA 105 loor 80 40 20 .~ o _. l I , , o.o 0.1 0.2 0.3 80 60 40 C FACTOR (ASSIGN ED) 100 _ MLRA 136 _ l _ 20 _ o- o.o 0.1 \~ 0.2 0.3 FIGURE 3-5 Percentage ofacres in MLRAs 103 (central lowa end Minnesota till prairies), 105 (northern Mississippi Valley loess hills), 134 (southern Mis- sissippi Valley silty uplands), and 136 (southern Piedmont) with USLE erosion rates greater than 5 tons/acre year at assumed levels of C factors for land in row and clos~grown crops in 1982. Note: High- and medium-potential curves illustrate land with high or medium potential for conversion to crow land. Source: Pierce et al., 1986. challenge than land used for row and close-grown crops in 1982. With a C-factor value of 0.30 the average conditions for land in row and close-grown crops in 1982 only 60 percent of the country's high-potential cropland and 57 percentofthe medium-potential crop- land would have sheet and rill erosion rates less than 5 tons/acre year. Even with much lower C-factor values of 0.1, representative of no-till farming systems, 1 1 percent of the high-potential cropland and 18 percent of the medium potential cropland would be expected to have erosion rates greater than 5 tons/acre~year. Among regions, consider- able variation is found in the vulnerability of high- and medium- potential cropland to average annual erosion rates greater than 5 tons/acre. 51
52 SOIL CONSERVATION These analyses of alternative C-factor values indicate the diversity of sheetancl rill erosion problems acrossthe country. No single cropping and management approach, including no-till farming, will solve the erosion problem on all croplancl. These analyses also suggest the value of the 1 977 and 1 982 N Rls for assessing sol ~ erosion conditions and alternative conservation practices. Similar analyses can be per- formecl to assess the erosion conditions that wou Id be expected under alternative conservation supporting practical conclitions, such as stripcropping, by altering P-factor values in the NRI data files. of procedures for estimating erosion rates, and analysis of basic erosion processes and effects. The USLE data can also be used to test new techniques that might be incorporated in future resource inventories to estimate erosion or assess other phenomena, such as nonpoint pollution. While erosion estimates based on USLE data in the 1982 NRI are site specific and indicate soil displacement rather than direct erosion damages, the esti- mates are sufficiently accurate to serve as primary guides for state, regional, and national analyses of conservation needs and opportuni- ties until better analytical tools and data become available. Nevertheless, the committee recommends that efforts continue to improve the USLE, for conservation planning and future resource inventories. A special need exists to improve the accuracy of the equa- tion, or to develop a separate equation, for areas where water sources other than rainfall, including snowmelt and irrigation, are important contributors to total soil displacement. Further work much of it under way is needed to improve the methods of measurement and accuracy of individual factors in the equation. In addition, as noted earlier, the relationship between sheet and rill erosion, and concentrated flow . ~ remains uncertain. Some of the limitations of USLE data addressed in this report stem from efforts over the last decade to apply the USLE in ways for which it was not originally designed. The equation was not designed to directly measure damages from erosion, either onsite or offsite. The committee notes that several mathematical models of erosion-productivity inter-
THE MEASURES OF SOIL EROSION 53 actions (the Productivity Index model and the Erosion Productivity Impact Calculator) have been designed to make use of NRI data. Simi- larly, models such as the Agricultural Runoff Model (ARM); the Hydro- logic Simulation Program in Fortran (HSPF); the Nonpoint Source Model (NPS); and later, Chemicals, Runoff, and Erosion from Agricul- tural Management Systems (CREAMS) have been developed, which incorporate USLE data for evaluation of offsite erosion effects using NRI data. Other models need to be assessed; new models should be developed, building on and extending USLE data for specific applica- tions. - The SCS and other agencies at the USDA should continue to rely on the USLE as the major analytical device for conservation program planning, man- agement, and analysis. SCS should take the lead in continuously upgrading field-level expertise in USLE applications. The agency should also introduce new techniques as soon as they are available that more directly relate erosion rates observed in thef~eld to actual onsite and offsite damages. The committee believes that the erosion equations and NRIs have not yet been used to full advantage. If continued public investments in these tools are to be justified, the committee believes that the USDA has a continuing obligation to demonstrate that important, practical bene- fits are accruing from the use of these tools in ongoing efforts to control serious erosion problems. Sheet and rill erosion problems on much of the land currently used for row and close-grown crops often can be adequately controlled through changes in cropping and management practices. Conserva- tion tillage no-till farming systems, in particular demonstrates exceptional promise for erosion control, provided that adequate levels of crop residues are maintained. On many soils, additional soil conservation benefits can be achieved by using traditional supporting conservation practices such as contour farming systems, vegetated waterways, terracing, and stripcropping in conjunction with conservation tillage systems. However, rainfall, type of soil, and slope characteristics beyond the control of farmers make it extremely difficult to economically control erosion on the approxi- mately 30 million acres of erosion-prone soils currently in conventional cultivation. The committee stresses that basic land use patterns are of great significance in conserving highly erodible lands. Where the potential for sheet and rill erosion is great, economically feasible control of erosion levels can be achieved only through shifts away from con- ventionally cultivated crops and into such uses as pasture, hay, range, some tree crops, or forest.
54 Wind Erosion Estimates SOIL CONSERVATION The 1977 NRI provided estimates of wind erosion for the 10 Great Plains states, using the WEE described by Skidmore and Woodruff (1968~. In the 1982 NRI, wind erosion estimates were attempted for all nonfederal lands in the United States on all land uses (see the boxed article The Wind Erosion Equation). The Wind Erosion Equation The version of the WEE used in the 1982 NRI is a variation of equations developed in the early 1960s. It has the following func- tional form: E = ffl,C, K, L,V) Unlike the USLE, the factors in the WEE are not Elirectly multiplied together to determine a value for E, which is the potential annual wind erosion rate in tons/acre year. Rather, this potential is a function of the fol lowing: I The soil eroclibility value, or 1 factor, reflects the size of soil particles or aggregates. The value for the I factor is the single most important source of variance in the equation. The l-factor value is expressed as the average annual soil loss expected to occur from an isolated, level, smooth, unsheltered fielcl, barren of vegetation and with a noncrusted surface. Originally, the I factorwasclerivecl from the total annual erosion (in tons/acre year) for a number of farm fields located nearGarden City, Kansas. In applications such es the N RI, the I values are assigned to soils based on the percentage of soil mass occurring in aggregates smaller than 0.84 mm in diameter. This value is obtained from SCS technical guides, in which values of I correspond to wind erosion groups (WEGs). The WEGs are in turn based on predominant soil textural classes (percentages of sancl, silt, and clay), wh ich were cleterm i ned by reference to sol I su rveys or other tech n ical guides. C C is an adjustment value to correct for areas having wind speed and rainfall patterns different from those of the reference area used to construct the original equation. The correction is based on mean quantities of wind speed and rainfall evaporation. The value for C is obtained from SCS technical guicles.
THE MEASURES OF SOIL EROSION K The soil ricige roughness value, K, is assigned to a sample point based on field inspection. The lower the value, the greater the vulner- ability of a field to erosion, if other conditions are equal. Deeply furrowed fields tend to trap wincl-blown soil in their furrows. Only two K-factor values could be assigned in the NRI: K = 0.05, for a smooth fielcl, and K = 1 .0, for a ridged field. L The value representing the unsheltered distance across a field along the prevai ti ng wi nd di rection is L. The unsheltered area of a field begins leeward of a protected area or from a barrier at a distance of 1 0 ti mes the barrier's effective height and perpendicu lar to the prevai l i ng wind clirection. AccorclingtotheWEE, afield wincibreakcomposed of trees 40 feet in height will theoretically protect a field on the leeward side of the trees for a distance of 400 feet. This value is based on field inspection. V The vegetative cover value V combines residue quantity, type, and orientation (flat or standing). The correct value is selected from SCS technical guides based upon field inspection. 55 ~ ~1 1 1 l : : :: ~L: Sand moves across the highway in 50 mph winds near Pacific City, Oregon. The new beach grass planting (left) will offer some protection against wind erosion. Credit: U.S. Department of Agriculture, Soil Conservation Service.
56 SOIL CONSERVATION The development of the WEE marked a major conceptual advance in soil erosion science. Use of the equation in the 1977 and 1982 NRIs was an important step toward identifying lands subject to serious damage from wind erosion. The committee believes that while the wind erosion data contained in the NRIs are useful for a number of important scien- tific and policy applications, the quantitative estimates are not suffi- ciently reliable for many uses. For example, the NRI estimates probably provide reliable indicators of the relative hazard from wind erosion in parts of the United States, especially for areas in the Great Plains states where wind erosion is chronically severe. It is likely, however, that the absolute values are too high (Gillette, 1986~. They do not appear to provide accurate estimates in more humid regions, on heavier soils, or in some arid regions where conditions of the soil surface in relation to wind erosion are not yet well defined. Limitations of the WEE Compared with the USLE data, the 1977 and 1982 NRI data for wind erosion are less accurate and less reliable. Moreover, because of limita- tions of time and personnel, as well as added computer costs, values for the individual factors of the equation are not currently part of the com- puterized NRI data bases. For this reason, it is not yet possible to use NRI data directly to refine the factors or the overall equation, or to undertake analyses of alternative options described previously in refer- ence to the USLE. There are several reasons for the present variation in estimates of sheet and rill erosion rates in contrast to those for wind. First, scientists have identified a number of conceptual shortcomings of the equation used to estimate wind erosion, some of which are related to the rela- tively narrow empirical and experimental basis of the WEE (Gillette, 1986) (see the boxed article Improving Estimates of Wind Erosion). Second, wind erosion estimation techniques used in the field have received less attention from scientific researchers than techniques for sheet and rill erosion, despite great interest in wind erosion at the inception of the modern soil conservation movement in the 1930s. Until uncertainties are resolved, it will be difficult for analysts to interpret estimates of wind erosion in the 1982 NRI. Improving Wind Erosion Predation An expanded research effort is needed to develop and validate basic components useful in improving the WEE. (Some avenues of research
THE MEASURES OF SOIL EROSION 57 Improving Estimates of Wind Erosion To identify some of the problems and research needs in the area of wind erosion prediction, the committee commissioned a paper for the project on the subject of the WEE and its use in the 1 982 NRI (Gillette, 1 986~. A summary of findings follows. Several sources of possible errors in the WEE, and thus in the NRI data sets for wind erosion, were identified. Some new work has been completed on the threshold wind velocities required to initiate ero- sion of soil particles of various sizes. This research indicates that the 1, or erodibi lity, values used in the WEE might lead to systematic overes- timates of erosion when 20 to 65 percent of the surface soil mass is composed of particles smaller than 0.84 mm in diameter. If this find- ing is correct, the NRI wind erosion estimates would generally be too high for certain soils, particularly those with textures other than sand and sandy loam. In practical terms, the estimated average annual erosion rates reported for humid areas in the 1982 NRI are generally low. Yet those estimates are still probably too high. A single value for I is used for an entire year when the WEE is used. However, soil aggregates can alter in one season. During drought, for example, soil aggregates of clay-textured soils from western Texas disintegrated. As the average size of aggregates decreased, the thresh- oicl wind velocity decreased, too. The C factor in the WEE is used to adjust the I factor to reflect climatic circumstances of wind speed and rainfall evaporation for areas other than those found in the reference area (Garden City, Kan- sas). Recent research on wind threshold velocities indicates that the C factor leads to overestimates of wind erosion wherever mean wind speeds are lower than those for Garden City. Wind speeds are in fact lower for most regions of the United States. An attempt was made to compare wind erosion estimates for a humid area (Minneapolis) using the WEE with estimates derived from a provisional WEE that reflects new research on wind threshold veloc- ities. The results showed moderately good agreement for soils in that area. The supplemented WEE predicts that the soils generally suscep- tible to high rates of wind erosion are very fine-to-medium sands or sandy loams in WEGs 1 and 2. Such soils do not predominate in this midwestern area. The WEE appears to overstate wind erosion by a factorofaboutfiveforthemorecommonmoderatelyerodiblesoilsin WEGs 3, 4, and 4L (textures ranging from very fine, sandy loams to silty clay loams). For soils in WEGs 5 and 6, the WEE appears to overestimate wind erosion rates by a factor of about 10.
58 SOIL CONSERVATION have been suggested by Gillette [1986~.) In the near term, however, efforts should be made to compare application of the WEE in the NRI with other methods. Comparisons might help identify the need or the opportunity to adjust NRI wind erosion estimates to make them more accurate and usable. Values for individual factors of the WEE should be coded and made available on a special supplemental NRI computer tape, if at all possible. They should also be available on the raw data tapes dunug any future inventories, at least for some regions. As a first step, wind erosion factor data should be entered for selected MLRAs representative of a diverse range of wind erosion conditions. Such data would facilitate basic research on wind erosion prediction and lend greater authority to short-term policy and program applica- tions in arid regions. These computer tapes should be made available to interested analysts inside and outside of the USDA. The USDA should design and fund a comprehensive plan of research to improve the prediction of wind erosion over the long term and to compare NRI wind erosion estimates with those denved by other methods during the nextiew years. A similar program should be developed to improve water erosion predic- tion. The research plan should specify the roles and responsibilities of the SCS, the ARS, experiment stations, university scientists, and scientists from other agencies. A realistic mechanism for coordination, along with solid USDA commitment and leadership, will be essential to ~ in, , , ensure that the research program is effective, efficient, and of high quality. The overall research plan and strategy for evaluating and improving wind erosion estimation techniques ideally should be devel- oped or at least reviewed by a panel of experts working outside of USDA. Field verification should tee performed by independent contrac- tors, using criteria developed by an independent panel of experts. The committee recommends that this research program include · An assessment of an alternative formulation of a WEE; · Verification of WEE estimates, basic parameters, and concepts through field-level measurements; · Study of the deposition patterns of soil eroded by wind, with a special focus on air and water pollution consequences; · Reconsideration of the need for nationwide wind erosion data. It might be possible to reliably apply certain variables of the WEE, perhaps with modifications, as indicators of relative wind erosion haz- ard. Such an approach has been proposed by the Resource Conserva-
THE MEASURES OF SOIL EROSION 59 tion Act (RCA) Fragile Soils Work Group as a means of identifying highly erodible lands (McCormack and Heimlich, 1985~. This group of USDA soil scientists and analysts evaluated six major options for classi- fying erodible soils. The recommended option uses the soil loss toler- ance (T) factor and soil and site factors of the USLE and the WEE as the basis for defining erodible soils, RKLS/T and CUT, for example. Some method for classifying land according to its relative wind ero- sion hazard will be needed if government agricultural policies are revised in the future to take soil erosion problems into account. The NRI could provide a useful data base for evaluating some of the effects of sodbuster and conservation reserve initiatives, for example, in areas where wind erosion is a problem. First, however, some scientific uncer- tainties about the WEE and NRI wind erosion estimates must be resolved. Evaluation of the usefulness of the WEE and its individual factors as criteria for identifying lands that are highly erodible by wind should be a priorly of the USDA. The evaluation should build upon the work of USDA's RCA Fragile Soils Work Group. Sufficient personnel and resources should be allotted to this activity to enable the USDA to propose in 1986 scientifically defensible and administratively feasible criteria to identify highly credible land that is cur- rently in cultivation or subject to cultivation in and or semiarid regions. The rapid development of a representative computer data base of WEE-factor values collected in the 1982 NRI would expedite and support this activity. Erosion by Concentrated Flow: Ephemeral Gullies Soil scientists, conservationists, and farmers have recognized for many years that serious erosion problems sometimes occur within nat- ural drainageways, usually on cropland where water runoff concen- trates and washes soil from wide, shallow channels. Unlike smaller rills, the sues of these concentrated flow channels are relatively perma- nent topographic features. Like rills, however, tillage operations usu- ally obliterate evidence of concentrated flow erosion; the term ephemeral gully erosion reflects this condition. Ephemeral gully areas tend to erode, sometimes at very high rates, unless appropriate steps are taken to divert water runoff or to plant the channel area in a sod-type crop. Establishing grass-based sod in these areas is a highly effective erosion control practice, and they are referred to as grassed waterways. There is no commonly accepted, practical method for estimating ephemeral gully erosion in the field; USLE estimates of sheet and rill erosion do not independently predict or quantify ephemeral gully ero-
60 SOIL CONSERVATION sign. Neither the 1977 nor the 1982 NRI includes estimates of ephemeral gully erosion. The extent and magnitude of this type of erosion are not well established, because methods for doing so are lacking. In some areas, however, this form of erosion is thought to equal or surpass others as a soil conservation problem. Evidence from field observations, measurements, and water quality models suggests that ephemeral gully erosion can result in sediment removal rates compara- ble to those for sheet and rill erosion (Foster, 1986~. (Sheet and rill erosion can be soil displacement not necessarily soil loss from a given field.) Concentrated flow may accentuate sheet and rill erosion in adjacent areas and add to crop loss within the Sullied area. The flow itself may transport sediment and chemicals directly from fields to water courses. In some cases, conservation tillage is adequate to control ephemeral gully erosion. More serious problems may require grassed waterways to directly protect the eroding channel or the construction of terraces, diversions, and outlets to slow and divert runoff as it moves down- slope. Maintenance of grassed waterways, however, can be a problem in farming systems that use herbicides that are toxic to grasses. Special structures in effect, small dams and spillways are sometimes installed to prevent some ephemeral gully areas from evolving into deep, permanent gullies. SCS personnel in about 30 states are making field estimates of ephemeral gully erosion. Several estimation methods are being used (Foster, 1986~. Plans are being developed to assemble all data for analy- sis within the coming year. In the committee's judgment, there is reason to believe that ephem- eral gully erosion may pose significant problems on cultivated cropland of sloping, uneven topography. Ephemeral gully erosion may prove to be the predominant and, perhaps, most damaging form of erosion in some regions, in terms of soil productivity and offsite problems. A comprehensive and coordinated research program involving the ARS, the SCS, experiment stations, the Extension Service, other fed- eral agencies, and universities should be developed to meet the follow- ing objectives: · Adoption of a consistent terminology to describe ephemeral gully erosion that will help distinguish it from other forms of erosion. · Evaluation of field methods for estimating ephemeral gully ero- sion. Initially, evaluations should emphasize the means for identifying areas affected to varying degrees by ephemeral gully erosion rather than procedures for estimating rates of erosion with great precision. In
THE MEASURES OF SOIL EROSION 61 many cases, identification of specific problems, combined with general (as opposed to site-specific) information about the damages associated with those problems, will aid in developing recommendations for land users regarding the adoption of cost-effective conservation measures. · Acceleration of research and educational programs conducted by the ARS, the SCS, experiment stations, universities, the extension service, and other USDA agencies on ephemeral gully erosion in regions where this problem is widespread. · Assessment of deposition patterns associated with ephemeral gully erosion to ascertain the extent of any changes in the productivity of landscapes where soil sediment is deposited. · Development of an interagency work group should be convened within USDA to study the feasibility of including methods for estimat- ing the extent and severity of ephemeral gully erosion in future NRIs. It may be desirable to focus future inventory activities on areas where preliminary assessments identify serious ephemeral gully erosion problems. One criterion for severity should include the potential impact of ephemeral gully erosion on water quality. To the extent that areas and specific sites can be identified where ephemeral gully erosion constitutes a significant problem: · Greater emphasis should be placed on targeting cost-sharing pro- grams and technical assistance to conservation systems that include grassed waterways and other proven, affordable methods for control- ling ephemeral gully erosion.
