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.
10 A Midwestern Perspective on Targeting Conservation Programs to Protect Soil Productivity C. Ford Runge, Wi!Ziam E. Larson, and Glaucio Roloff Soil erosion has been identified as an important potential threat to long-term agricultural productivity in the grain-growing regions of the Midwest. To date, however, much of the evidence supporting this view has been fragmentary or impressionistic. This study uses recent data developed as part of the National Resources Inventory (NRI) by the Soil Conservation Service (SCS) of the U.S. Department of Agriculture (USDA) to assess the potential onsite long-term productivity losses due to soil erosion in six Major Land Resource Areas (MLRAs) of the region. The essential purpose of this study is to demonstrate how the NRI can help implement policies that increase the efficiency of soil and water conservation by targeting those sections of the land mass most susceptible to damage from erosion. The NRI data, when combined with productivity measures developed by Larson et al. (1983), allow policymakers to go beyond simple measures of potential soil loss, such as topography, to investigate specific soil types that are highly susceptible to productivity declines. By carefully specifying the differential impact of water erosion on these soils, a clearer picture of potential productivity losses can be developed. This study, one of the first of its type, represents an initial effort in what is hoped will be an increasingly refined study of erosion impacts. It must be emphasized that this is preliminary and should not be interpreted as a sufficient basis for policy prescription. 273
274 METHODOLOGY The NRI is a USDA nationwide survey of private agri- cultural lands that contains data on approximately 22 parameters affecting potential agricultural productivity. Included are both physical characteristics of the land and water resource base and the impact of different agronomic practices on soil erosion. The 1982 NRI expands a 1977 data base, updated to encompass a variety of measures that would allow estimates of erosion potential on different land classes. In addition to the NRI data, this paper utilized the Soils-5 data base established by the SCS (USDA, 1983), which contains soil descriptions, ranges of soil and chemical properties, crop yields, and land capabilities and limitations for U.S. soils. Together, these data allowed the development of three scenarios that simulate the impact of three stylized programs of soil conservation. In this paper, attention is restricted to the onsite effects of water erosion upon soil productivity for land in row crops (corn and soy- beans) in six NLRAS of the Midwest (see Figure 1). Three of these--MLRAs 105, 109, and 113--are highly susceptible to erosion and have soils that may suffer large produc- tivity declines if erosion occurs. The other three-- MLRAs 103, 108, and 115--are comparatively less sus- ceptible to erosion or are less likely to suffer large productivity declines if erosion occurs (Pierce et al., 1984). The purpose of this exercise is to compare the impact of alternative soil conservation targeting policies on row-crop production in these areas, using concepts recently developed by Pierce, Larson, and others (Larson et al., 1983; Pierce et al., 1983). Soil and water conservation programs may be targeted according to myriad criteria, each of which may carry implications for the mix of crops grown and the future productivity of the targeted and nontargeted areas. It is therefore essential to define both the criteria employed and the measures used to estimate their effects In this paper, three basic scenarios and two measures are used to simulate alternative policies. In each scenario, the two measures reported are: the acres planted to corn and soybeans in each MLRA, and the 100-year impact of this pattern on soil productivity. The first scenario is a baseline estimate of the long-term effects of current erosion rates and the long-term productivity of soils in the six MLRAs if there are no changes in soil and water conservation programs or practices recorded by the 1982 NRI. .
275 100 _ 80 Al - c~ .- 60 ._ o 40 20, ) 1 1 1 An. \ ~ Group II ~` 1 1 0 1 2 an, an, 1 1 3 4 5 Reduction in Pl (%) Over 100 Years FIGURE 1 Major Land Resource Areas studied . The second scenario estimates the acreage in each MLRA that must be removed from row crops if a particular tolerance to soil erosion (T value) based on the Universal Soil Loss Equation (USLE) were chosen as a basis for policy. In this case, all land in row crops in the MLRAs under study with an erosion rate greater than the local soil loss tolerance limit (T) for the particular soil series would be put into forage. Given this, the T-value criterion leads to reductions in acreage planted to row crops. The magnitude of these reductions, together with the soil-productivity impact of the shift into forage, is estimated using a 100-year horizon. It is assumed, as in the baseline scenario, that conservation practices con tinue at current levels. In this scenario, however, these practices encompass those used on the row crops that remain in production, and those used on the land given over to forage (assumed to be the practices considered proper for an established stand with about 80 percent ground cover). This simple scenario can, of course, be modified to include improved conservation practices or other factors such as the relative impact of planting to other crops rather than shifts to forage. Here, however, the estimate is simply of the impacts that might result from such shifts into forage, assuming
276 T values are the targeting criterion. Clearly, more complex patterns of land use would actually occur. The third scenario applied to targeting is more refined and involves use of the Soils-5 data base in connection with the 1982 NRI. In this case, it is assumed that some of the land taken out of row crops and But into forage under the second scenario can be returned _ . . . · . ~ to row-crop production because the soil type is not highly vulnerable to losses in productivity. The potential productivity losses of soils in each MLRA, as calculated from the vulnerability (V value) of various soil types to erosion, is used as the basis for this targeting criterion. The total acres remaining in row crops are then reported. Because this vulnerability is a measure of potential loss in soil productivity due to erosion, results are reported for four levels of pro- ductivity over the 100-year horizon to test the sensi- tivity of the analysis to these levels. The four cases tested are based on estimates in which productivity declines at a rate 5.0, 2.5, 1.0, or 0.01 percent of its present level over a century. For ease of reference, the three scenarios employed in this study may be thought of as a no-change baseline, targeted programs based on T values alone, and programs based on new information concerning the vulnerability or V values of various soil types. The three simulations thus demonstrate the way in which alternative targeting policies can be tested using the NRI and Soils-5 data bases. (For more precise definitions and details of the methodology, see the Appendix.) . . RESULT S Baseline Scenario The first group (hereafter, Group I) of MLRAs (105, 109, and 113) represents areas highly susceptible to erosion and soils that may suffer large productivity declines if erosion occurs. The second group (hereafter, Group II) of MLRAS (103, 108, and 115) is comparatively (though not uniformly) less vulnerable. These regional characteristics are presented in Table 1, the f irst column of which shows the comparative erosion potential of the six NLRAs in tons/acre/year. This erosion potential, which varies from 11.2 in MLRA 103 to 71.6 in MLRA 10 5, does not necessarily correspond to levels of
277 Us or or U] a) Q o En o ·,1 a) o ~5 ~ Cal S of · - _' - o U] · - US U] .,. S C) O m 00 ~ 0 EN ~ .~t ·,~ ~ :> .-, .~. :S X 0 54 P4 0) Q U2 ~ a O ~ E~ O _ E~ ~: O ·,' U] O ~ ~: P4 ~ . - O · - Ul O ~ `4 0 O3 ~: cn ~ ~ tD ~ · * ·· . ~ o o oo o o ~ as ~0 a~ r~ ~CO a, co · . ·· . ~ o o oo o o a~ a, kD · . · · . ~ ~r o · . ~ ~ U) oo · . ~ ~ co ° ~ u:~ r~oo ~ u~ · . ·· . ~ o o oo o o M0 ~ ~ a~ 0 · . ·· . ~ ~ ~ u)~ oD H a, ~ O O O V H H CO ~ O O O ~n o ·,' _ a, O P. ~ E~ U] - . - ·,1 O ~n O ~ . - O O C~) U] 05 Q - 0O a, - U) H ~: z a' ·- C) o CO
278 conservation practice (CP value) in the MLRAs inves- tigated. In MLRA 105, however, the highest level of erosion potential (71.6) is combined with the lowest relative CP value (0.19), which signals the most intensive level of conservation practice. In MLRA 103, the lowest potential erosion level (11.2) is matched by the highest relative CP value (0.38), indicating the least intensive conservation. Still, these averages tend to obscure cases of potentially poor conservation practices on highly erosive soils within a given MLRA. This result is suggested by data in Table 1 showing actual and tolerable erosion rates. Tolerable as used here is defined as the soil loss tolerance limit (T value). Intensive levels of conservation practice on MLRAs with high erosion potential, such as 105, still resulted in actual erosion rates of 11.4 tons/acre/year in 1982, more than twice the tolerable level. In MLRA 109, on the other hand, a CP value of 0.36, indicating relatively nonintensive conservation practices, was associated with very high erosion rates of 15.2 tons/acre/year, about four times the tolerable level of 3.9. This suggests that within the NLRAs, as well as among them, more accurate targeting of conservation practices is required. In only one--MLRA 103--was the actual erosion rate in 1982 less than that considered tolerable. These findings also suggest the need for a sharper analytical tool than is provided by tolerance levels alone. Ideally, such a tool should be able to dis- tinguish, both between MLRAs and within them, which soils are most susceptible to productivity losses due to erosion. This is the purpose of the productivity index (PI) and vulnerability values reported in Table 1. These values indicate, based on specific depths and types of soils in the Soils-5 data base, which areas are a suitable environment for continued crop productivity. The PI of these NLRAs considers the sufficiency of available water capacity for each soil, the sufficiency of soil bulk density, the soil's acidity, and the depth of soil horizons in the zone of plant rooting. The V value is simply the slope of the productvity index/soil removal curve, which plots the loss in soil productivity resulting from incremental reductions in soil depth. (For purposes of this analysis, a linear approximation to this relationship is used. In general, this is quite accurate, although some soils manifest nonlinear pro- ductivity losses, leading V to change as successively
279 more soil is removed. V values are reported in this paper as absolute numbers.) Table 1 shows that, in general, low average levels of erosion potential are generally associated with high average levels of soil productivity, and vice versa. However, the relationship between soil productivity and erosion potential is far from straightforward. Productive soils may or may not be on highly erosive lands, sug- gesting the need for additional information if policy is to be correctly formulated. On shallower soils, for example, damage provides its own form of conservation incentive, while on deep soils greater rates of erosion may be economically rational (Walker, 1982). In cases where highly productive soils are found on erosive lands, it may nonetheless be appropriate from an economic perspective to continue farming them, and to focus the lion's share of conservation practices there. This requires a measure of soil vulnerability to losses in productivity (the last column of Table 1). Consider the situation in MLRA 105, in which the highest average level of erosion potential is paired with the most intensive average conservation practices. The productivity index is 0.84, greater than in both MLRA 109 and MLRA 113, where conservation practices are nearly half as intensive on average and where the average productivity indices are lower. The relative potential productivity of the soils in BRA 105 suggests strong reasons why it should continue to be targeted for improved conservation practices. This argument is reinforced by the overall vulnerability of its soils (0.23), which is greater than in either MLRA 109 (0.17) or MLRA 113 (0.21). A second example is MLRA 103, which pairs the lowest average level of erosion potential with the least inten- sive average levels of conservation practices. This relationship appears to indicate no excessive soil erosion losses. However, the high productivity index of the MYRA (0.88) combined with its relatively vulnerable soils (0.27) suggests that the observed levels of conservation practices may not be responsive to the vulnerability to productivity losses due to erosion. The efficiency losses resulting from a failure to target the soils most vulnerable to productivity declines have been estimated by Ervin et al. (1984). They con- cluded that "T values may not be appropriate compliance criteria across different soils," and that attempts to target conservation incentives based on T values alone may "give greatest incentives to control erosion on lands
280 for which the long-run social benefits are negative or smaller than for more erosive lands" (Ervin et al., 1984, pp. 277-278). The primary requirement for policy is, therefore, a more accurate targeting criterion based on potential productivity losses. An important point emerging from this analysis concerns the loss of information resulting from MLRA averages. Although such averages are the basis of this paper, the HRI and Soils-5 data bases allow the development of much more disaggregated simulations analogous to those presented here, which can then be used as a basis for more localized targeting policies. The next phase of this analysis concerns the compara- tive impacts of different types of targeting criteria on total acreage planted to row crops (see Table 2). A baseline scenario is given first, in which the acreage in corn and soybeans in each MLRA in 1982 is reported. Based on the assumed continuance of the level of con- servation practices reported in Table 1 above, the estimated impacts of these practices on soil productivity over 100 years are then calculated. The highest levels of soil productivity losses are observed in Group I, while the lowest levels would occur in Group II, (hence the rationale for the groupings). T-Value Scenario Scenario 2 estimates the impact of using T values as a basis for shifting lands out of row crops and into forage. Wherever the actual erosion rate exceeds T in a given sampling location, these acres are assumed to be taken out of row crops and shifted into forage. The acreage thus removed is subtracted from the baseline acreage. The acreage remaining in row crops is reported, together with the percentage reduction in this acreage and the change in soil productivity resulting from this shift in land use, again assuming no change in conservation practices. The result, as shown in Table 2, would be considerable reductions in 1982 row-crop acreage, ranging from a minimum of 22.0 percent in MLRA 103 to a maximum of 65.6 percent in MLRA 109. As expected, Group II MLRAs would show lower percentage reductions than Group I, although even these are hardly modest. In both groups, radical changes in land use are implied by the T-value targeting criteria.
281 U] ~ ' o 0 a, o ~ ' o To o .,, v ' o Us o s U] ~; o ~ U. ' ~ o a, ·-1 C) 1 U] a .,, L. a, U] a) Ll o U] o H - X · - ~ H a, 4) .,4 . - C) 1:5 o D3 ~ ~n 2 ~ ~ d o ~. ll o U] - o ~4 c dP U) 11 P4 d~ 11 E~ o ~ o c a, V) - ~s o C ·,4 · - C U) U] _ V V · P4 _ U] _ o ~ o U ~ ~:- c · ~ _ eq _ a, 0 Ll O U ~ 94 - - JJ C U · ~ U ~ <! _ U)- a, o Ll O U-I .e- - C U · h4 O I4 P4 <_ I,0- a, 0 ~ O U ~ #~ - JJ C U H ~ 04 ~ _ 4J C a' U · C) ~ P~ d- t.q- /v o `, o U ~ ,< _ - a) U 4 p4 <_ ~_ a, o L' o u ~ ,<:- ~: a~ ~ ~ kD CO · . · · . ~ ~ un O ~ tD ~ co co ~a~ a~ o: a~ ~o co u~ a'0 ~ 0 0 ~r ~0 oi ~ ~ ~ ~ ~ o co 0 u, ~r ~ ~ · . · · · ~ ut r~ ~o as ~ ~ ~u, ~ ~r o o CO U, ~ o N a~ ~1- ~ -I O4 ~ ~ ~D tD 0 0 ~ c~ ~r ~ ~ ~u~ oo a, u ~ c~ tD · . · · . ~ oo ~ ~r a' un `0 CO r~ un ~ 0\ ~o ~ ~ ~ a, ~ ~a, ~ o' ~ ~ ~., ~ r~ ~ ~ ~0 o~ co ~ · . ~ CD 1- 0 ~ 0 c~ ~ ~ a' ~ · . ~ co o, ~r 0 a' ~ u, ~ a~ 0o ~r ~ ~ u~ O oo ~ u~ r~ · . ·· · ~ ~ O OO oO r~ ~, ~O co \° · . ·· . ~ n ~ kD `0~ ~ ~ 0 0~ u~u~ 00 0 ~ ~ ~ 0 o, 0\ co ~er ~ a' L~ O O ~ ~ ~° r~ \0 ~ ~ra · . · · . . u, \0 co 0 co ~ u CD O ~ ~O u~ O ~ ~ ~D 0\ ~ `34 ~ C~ U~ ~ O o~ ~ O 0 - O ~ - - O ~ - - ~ ~4 C' ~ o .,. L' ~a C U U] ·_' · :a Ll ~ /L} h :>' O O O -~ U O U C O O .,' .,1 JJ U U ~ :' 'O ~ - CO a, - a U) ·` D: z c . - ·. o U]
282 Unsurprisingly, the consequence of these shifts in land use is to reduce substantially the loss in pro- ductivity shown in the baseline. In MLRA 109, for example, an estimated 100-year loss in productivity of 6.9 percent would be reduced to 0.8 percent. Comparable reductions occur in the other MLRAS, with the effects most pronounced in Group I. The opportunity cost of these reductions, in terms of acres of row crops foregone, appears to be very large, however, and would lead to major shifts in agricultural production away from these crops. V-Value Scenario Scenario 3 uses the information contained in the NRI and Soils-5 data files to target more accurately those soils highly vulnerable to productivity losses due to erosion. The V values listed in Table 1 were the basic criterion used to determine whether lands should be shifted out of row-crop production. This use of V values requires an explicit determination of the rate of . . . . . . . . . . . . reduction In the productivity index over the relevant time horizon. Four levels of reduction are used, both to test the sensitivity of the criterion to judgments concerning productivity declines and to indicate the importance of making such judgments explicit. These judgments must reflect both private and social values concerning the appropriate rate of depletion of soil resources--judgments that are ultimately normative. For some perspective on the rates chosen (5.0, 2.5, 1.0, and 0.01 percent over 100 years), they may be compared with either Scenario 1 or Scenario 2. A rate of 5.0 percent over 100 years, for example, is comparable to the Group I average of 5.6 percent in Scenario 1, resulting from a policy of no changes in row-crop production or conservation practices for these MLRAs. A rate of 2.5 percent is half this level of depletion, and it is comparable to the Group II average of 2.6 percent in Scenario 1, again implying a policy and set of conservation practices (for this group of MLRAs) essentially the same as at present. Scenario 2, as noted above, implies much lower rates of soil productivity depletion in return for major land use shifts. Even in Scenario 2, however, the lowest rate of decline is 0.5 (in MLRA 108). The choice of 1.0 as a rate of productivity decline would correspond to MLRA 105
283 under Scenario 2. The choice of 0.01 percent as an acceptable change in productivity index reflects the notion that almost no depletion is acceptable (a value of zero could not be used for computation). The resulting estimates in Scenario 3 are instructive. Where 5.0 percent is used as an acceptable 100-year rate of soil productivity loss, a major share of the acreage taken out of row-crop production in Scenario 2 would be returned in Scenario 3. In Group I, MLRA row-crop acreage reduction would average 24.1 percent of 1982 acreage. This compares quite favorably with the average reduction in Scenario 2 of 60.1 percent--roughly 2.5 times as much acreage shifted out of row crops. In Group II, the reduction would average only 11.0 percent of 1982 acreage. In contrast, use of the T-value criterion in Scenario 2 led to an average reduction of 38.8 percent, slightly more than 3.5 times as much acreage. When account is taken of the fact that 1982 was a high-production, nearly record-setting year for these row crops, the actual acreage reductions necessary to achieve 5 percent losses in soil productivity over 100 years if a V-value criterion is used would appear to be even less than suggested by these estimates. If a stricter soil-productivity-loss criterion of 2.5 percent is applied, the acreage that could be returned to row-crop production in Scenario 3 compared with Scenario 2 drops. In Group I, the reductions implied would average 41.4 percent. Although substantial, this is still considerably below the Grout I average reduction of .C:~=n=~;^ ~ ton 1 - Ear - ~\ ~ to-- _ ram ~ . In Group II, the acreage reduction would average 22.5 percent, again substantially less than the drop that would occur under Scenario 2 (38.8 percent). Results for a 1.0 percent rate of soil depletion led in general to reductions of the same order of magnitude as the T-value criterion used in Scenario 2. The only exception is in MLRA 103, where the relative vulnerability of soils would lead to greater reductions in acreage. This suggests the greater precision of the vulnerability measure. Overall, the implicit rate of depletion resulting from use of T values is approximately 1.0 percent over 100 years. The strictest assumption--of only a 0.01 percent reduction in soil productivity over 100 years--gives some indication of the magnitude of land use shifts that would be necessary to pursue essentially "steady-state" policies with respect to soil loss. In the case of Group I, 79.4
284 1~-~ r--- IdW_~ is' ~ MO -~,~ ! IS FIGURE 2 Average reduction in (corn and soybean) acreages for a given reduction in productivity index over 100 years, for MLRAs 105, 109, 113 (Group I) and 103, 108, 115 (Group II). percent of all row-crop acres in 1982 would have to be shifted to forage on average. In Group II, an average of 92.4 percent of all 1982 row-crop acres would be pulled from production and put into forage. In short, pursuit of a "steady-state" level of soil productivity implies the elimination from row-crop production of the vast majority of the acres in those MLRAs. The overall relationship between reductions in acreage planted and reductions in the percentage of productivity lost in Scenario 3 is shown in Figure 2. As less reduction in productivity is allowed over 100 years, proportionately larger shares of the acreage planted to row crops is removed from Groups I and II. As the figure shows, the distribution of vulnerability differs, and the acreage taken out of production rises at an increasing rate as the requirements for maintained productivity converge to the "steady state. n In all, these results suggest that substantially fewer acres could be shifted from row crops to forage if targeting policies for soil and water conservation were based on soil vulnerability to productivity losses rather
285 o x cat 20 15 in Q o 3 10 o cat lo 5 At: cat o o Do% 13.1%* 4 1.9% * 26.1%* S2.5%* 91.3%* Baseline *Reduction Compared to Baseline 5% 1% .01% T-Value s Reduction in Pi V-Values FIGURE 3 Total acres in row crops for all MLRAs under each scenario. than the customary T value. Although use of V values does not eliminate the need for shifts in land use, when 100-year productivity losses are set at 5.0 and 2.5 percent these shifts are far fewer than implied by T values alone (see Figure 3). Not only are fewer acres likely to be targeted for land use changes, but the particular acres chosen are more likely to exhibit specific soil characteristics damaging to long-term productivity. To simplify the analysis, a complete shift from row crops to forage has been assumed. Less extreme changes in rotation can and should be encouraged, based on local economic and soil characteristics. When the conservative assumptions used in this study are modified, substantial improvements in soil productivity may result without major disruptions in land use, provided policies are properly targeted. Finally, it must be reiterated that corn and soybean production in 1982 nearly broke records
286 for those crops, with many marginal acres in production. Use of 1982 as a baseline therefore may overstate the needed reductions in this acreage in other years. IMPLICATIONS FOR POLICY Use of more accurate targeting criteria for soil and water conservation policy can reduce onsite productivity losses and minimize the acreage affected by more restric- tive land use practices. Acres taken from production can and should be targeted, and those that are most vulnerable to erosion can increasingly be isolated. This study provides preliminary evidence that a targeting criterion can be developed, based on the recent NRI and Soils-5 data bases. Clearly, many difficulties and questions remain, although these appear to be less technical than institutional in nature. Two of the institutional issues are especially worthy of note. The first concerns the choice of an appropriate rate of depletion of soil resources. The results of Scenario 3 in this paper clearly demonstrate the impor- tance of this judgment and the magnitude of its effect on land use policy. Analysts are likely to differ over this rate, and no simple solution to the issue is possible (see Lind et al., 1982; Page, 1977). Nonetheless, current actions reveal an implicit rate that may well reflect existing preferences, as expressed by the 1982 baseline data reported above. Any policy applied with respect to targeting will have productivity implications over time and will reveal a similar implicit rate of depletion. It would be best, however, to make these judgments explicit. The vulnerability criterion developed in Scenario 3 does this. The choice of a "steady-state" rate, for example, appears to have major implications for future row-crop production in the Midwest, as is clearly revealed by this analysis. The second institutional issue worth noting concerns the impact of targeting on existing Soil and Water Conservation Districts and the wide range of other institutions developed since the 1930s to deal with related land use issues. Some have argued that targeting would make these institutions less important; one consequence of this has been the arousal of opposition to the targeting concept. Yet, targeting of soil and water conservation policy does not diminish the important role of these institutions. Rather, it changes their role to
287 one in which programs are more accurately directed and specifically fashioned to suit local needs. This implies, if anything, a broadened set of responsibilities for existing institutions, with ever greater emphasis on local autonomy over land use decisions based on improved technical information. Finally, the preliminary nature of these findings must again be emphasized, along with the need for continued improvements in technical methods to identify onsite productivity losses due to soil erosion. These losses are, of course, only one aspect of a larger problem that includes important offsite damages (see Christensen, this It seems, however, that important beginnings can be made by estimating onsite damages, with further and more difficult estimates of offsite damages to follow. As technical capabilities increase, a similar commitment to institutional innova- tions can result, leading to reduction in productivity losses arising from poor management of America's great inherited wealth: her soil resources. volume; Crosson and Stout, 1983). REFERENCES Crosson, P. R., with A. T. Stout. 1983. Productivity Effects of Cropland Erosion in the United States. Baltimore, Md.: Johns Hopkins University Press for Resources for the Future. Ervin, D. E., W. D. Heffernan, and G. P. Green. 1984. Cross compliance for erosion control: Anticipating efficiency and distributive impacts. Am. J. Ag. Econ. 66:273-278. Larson, W. E., F. J. Pierce, and R. H. Dowdy. 1983. The threat of soil erosion to long-term crop production. Science 219:458-465. Lind, R. C., K. J. Arrow, G. R. Corey, P. Dasgupta, A. K. Sen, T. Stauffer, J. E. Stiglitz, J. A. Stockfisch, and R. Wilson. 1982. Discounting for Time and Risk in Energy Policy. Baltimore, Md.: Johns Hopkins University Press for Resources for the Future. Page T. 1977. Conservation and Economic Efficiency: An Approach to Materials Policy. Baltimore, Md.: Johns Hopkins University Press for Resources for the Future. Pierce, F. J., W. E. Larson, R. H. Dowdy, and W. A. P. Graham. 1983. Productivity of soils: Assessing long-term changes due to erosion. J. Soil Water Conserv. 38:39-44.
288 Pierce, F. J., R. H. Dowdy, W. E. Larson, and W. A. P. Graham. 1984. Soil productivity in the Corn Belt: An assessment of erosion's long-term effects. J. Soil Water Conserv. 39:131-136. USDA (U.S. Department of Agriculture). 1983. Soils-5 Data Base. Ames, Iowa: Soil Conservation Service. Walker, D. J. 1982. A damage function to evaluate erosion control economics. Am. J. Ag. Econ. 64:690-698.
289 APPENDIX Definitions Universal Soil Loss Equation (USLE) The USLE is described in detail by Wischmeier and Smith (1978). The equation was developed from more than 10,000 plot years of basic runoff and soil erosion data measured at 49 research locations in the United States as well as data obtained from rainfall simulator studies. The equation takes the form: A = RKLSCP, (1) where A is the erosion rate in tons/acre/year. R is the rainfall and runoff factor and is the number of rainfall erosion index units, plus a factor for runoff from snow- melt or applied water where such runoff is significant. K is a soil erodibility factor, which expresses the rate of soil removed per erosion index unit for a slope of specified geometry. L is the slope-length factor, representing the ratio of soil erosion from the field slope length to that from a standard length (72 feet) under identical conditions. S is the slope-steepness factor and is the ratio of soil erosion from the field slope gradient to that from a 9 percent slope under identical conditions. C is the cover and management factor and is the ratio of soil erosion from an area with a specified cover and management factor to that from an identical area in tilled, continuous fallow. P is the support practice factor and is the ratio of soil erosion with a support practice like contouring, strip- cropping, or terracing to that with straight-row farming up and down the slope. Erosion Potential (EP) The right side of the USLE can be separated into two parts: the factors that are controlled by nature or are affected by humans only at great costs (RKLS) and the factors governed essentially by management (CP). RKLS is therefore the inherent potential for erosion at a given location. The actual erosion rate (A) depends on the values taken by C and P.
290 Soil Loss Tolerance Limit (T) This term is defined by Wischmeier and Smith (1978) as the maximum level of soil erosion that will permit a high level of crop productivity to be sustained eco- nomically and indefinitely. When substituted for A in the USLE, it allows an estimate of the maximum CP value necessary to keep erosion rates below the tolerance level, once RKLS is considered as a constant for a given location. T values range from 2.2 to 5.0 tons/acre/year and are based on experience and observations established through six regional workshops in 1961 and 1962 (Wischmeier and Smith, 1978). Productivity Index (PI) The PI used in this study is a modification by Pierce et al. (1983) of a model developed by Kiniry et al. (1983). The model indexes the soil according to its suitability as an environment for root growth. It was modified to include some additional concepts and to use data available in the Soils-5 data base (USDA, 1983). The modified model is: r PI = ~ (Al Ci Di WF), 1 = 1 where Al is sufficiency of available water capacity, Ci is sufficiency of bulk density (adjusted for permeability), Di is sufficiency of pa, WF is a weighting factor and r is the number of horizons in the depth of rooting. The model assumes that nutrients are nonlimiting to plant growth and that other factors are constant. Soil Vulnerability (V) The relative vulnerability of a soil to long-term erosion losses can be assessed by the slope of a PI-soil removal curve (Pierce et al., 1984), estimated as: V = PI/d, (2) (3)
291 where API is the percentage variation (I) in PI and Ad is the change in depth (cm) due to soil erosion. Pierce et al. (1983) used a constant arbitrary Ad of -50 cm, which was also adopted in this paper. Although a few values were equal to or greater than zero, the present study reports V as an absolute number. Simulation Methodology Scenario 1: Baseline The 1982 NRI furnished the erosion rate (Al), the C and P values, and the coded soil unit corresponding to each sampling location within a MLRA. The coded soil unit was matched with the proper unit stored in Soils-5, which allowed the calculation of PI by Equation 2 using A = Al. The variation in PI over time (API, %) was estimated by: API = (PIo ~ PIT) 100/PIo, where PIo is the productivity index at time zero (here, 1982) and PI1 is the productivity index after the removal of soil corresponding to 100 years of erosion at the present rate (Al). The C and P values were multiplied together and reported as CP values. All results were weight-averaged by acreage for each MLRA. Scenario 2: Use of Soil Loss Tolerance Limit Value Using the same basic data as in Scenario 1, each sampling location had its erosion potential (EP) calculated as EP = A1/CP and its Al value compared to its T value furnished by the 1982 NRI. All locations in which A exceeded T were assumed to be taken out of row crops and the acreage was summed and then deducted from the total row-crop acreage in Scenario 1. The acreage taken out of row crops was then assigned a CP value of 0.01, corresponding to an established, well- managed forage field with about 80 percent ground cover (Wischmeier and Smith, 1978). For these locations, a new erosion rate (A2) was calculated as A2 = EP ~ CP, using CP = 0.01. The PI values for acreage remaining in row crops as well as that shifted to forage was then
292 calculated as in Scenario 1, and all results were weight- averaged by acreage for each MLRA, to yield productivity losses over a period of 100 years. Scenario 3: Use of Soil Vulnerability Values Vulnerability values were calculated according to methods developed by Pierce et al. (1984) for each soil sampling location planted to corn and soybeans in each MLRA. These ~ values and the four different degrees of productivity loss were used to calculate acreage that would be taken out of row crops and put into forage. Changes in PI (API) of 5.0, 2.5, 1.0, or 0.01 percent over 100 years were used to calculate Ad using Equation 3. This allowed the calculation of the erosion rate (At) for a tolerable reduction in PI by: At = (Ad PI W)/(V t), where W is the weight of a soil layer 1 acre in area and 1-inch thick, determined using the bulk density value for the local soil series reported in Soils-5, and where t is time (here, 100 years). Equation 5 is a modification of Equation 3 in Pierce et al. (1984). The At values were then compared with local Al values. Where Al exceeded At, the area was taken out of row crops and assigned a C value of 0.01, as in Scenario 2. For these locations a new erosion rate (A3) was determined as AS = Ep CP, with EP calculated as in Scenario 2 and CP equal to 0.01. The acreage taken out of row crops was again deducted from the total row-crop acreage in Scenario 1. All results were weight-averaged by acreage using the new values where appropriate. REFERENCES Kiniry, L. N., C. L. Scrivner, and M. E. Keener. 1983. A soil productivity index based upon predicted water depletion and root growth. Res. Bull. 1051. Columbia: Missouri Agricultural Experimental Station. Pierce, F. J., W. E. Larson, R. H. Dowdy, and W. A. P. Graham. 1983. Productivity of soils: Assessing long-term changes due to erosion. J. Soil Water Conserv. 38:39-44. (5)
293 Pierce, F. J., R. H. Dowdy, W. E. Larson, and W. A. P. Graham. 1984. Soil productivity in the Corn Belt: An assessment of erosion's long-term effects. J. Soil Water Conserv. 39:131-136. USDA (U.S. Department of Agriculture). 1983. Soils-5 Data Base. Ames, Iowa: Soil Conservation Service. Wischmeier, W. H., and D. D. Smith. 1978. Predicting rainfall erosion losses: A guide to conservation planning. Agriculture Handbook 537. Washington, D.C.: U.S. Department of Agriculture. Discussion John A. Miranowski Runge, Larson, and Roloff have presented a framework for policy analysis that provides an excellent beginning in a new area of conservation policy research--using physical measures of soil productivity loss as the basis for soil conservation policy decisions. Simply targeting erosion control to the most erodible acres may not be the most efficient approach to the soil erosion problem. Gross soil loss may not be an accurate reflection of the potential productivity foregone. Using a measure of productivity loss, such as produced by the PI (produc- tivity index) model (employed in this analysis) or the EPIC (the Erosion-Productivity Impact Calculator) model, provides a more logical basis for targeting erosion control. The vulnerability index or measure discussed by Runge, Larson, and Roloff does allow for social judgment regard- ing the level of productivity loss that the public is willing to tolerate on croplands. But to some extent the vulnerability index criterion ignores the so-called "hard-core" economic information that should enter into this judgment. The vulnerability index does not provide a complete accounting of the added benefits and costs that may be involved. First, the decision to retire cropland acres from row-crop production should be based on social benefit-cost calculus. If the added benefits outweigh the added costs of retirement, then the policy can be justified in an economic sense. If not, society's welfare is reduced. Without a more explicit accounting of the costs incurred, it is difficult to ascertain how society will determine the allowable rate of productivity
294 decline, or as Runge and coauthors state, "it is our view that these judgments must reflect both private and social values concerning the appropriate rate of depletion of soil resources." Second, retiring acres that are eroding at rates that lead to productivity losses greater than 1.0, 2.5, or 5.0 percent will affect a significant portion of crop- land, as their Table 2 shows. Such large cropland retirements will have significant price effects, which in turn will have impacts on the mix of crops and tillage practices used on less erodible cropland. These adjustments may create related erosion problems but of lesser magnitude on the remaining cropland acres. In cases where highly productive soils are found on highly erodible lands, it may nevertheless be appropriate from an economic perspective to continue farming these soils but to employ more intensive conservation prac- tices. Retirement is not necessarily the most efficient alternative. Some recent work in Iowa (Miranowski and Hammes, 1984) considered land purchasers' willingness to pay for topsoil depth and erodibility (measured by the RKLS). Holding topsoil depth constant, the value of farmland decreased as erodibility increased. Landowners make investment decisions with respect to land purchases and conservation investments that reflect these tradeoffs between productivity and erosion control cost. It may prove costly to retire highly vulnerable cropland that is also highly productive. Additionally, some soils, precisely because they are highly productive and too costly to save, may be mined during periods of high commodity prices. These factors should be considered in any social decision to protect productivity of specific cropland through retirement programs. The vulnerability measure proposed by Runge, Larson, and Roloff also appears to ignore technological change and potential soil genesis. These omissions would tend to overstate vulnerability and the need for policy inter- vention. The vulnerability measure may also ignore nonlinearities, because it is looking at the marginal increment at a particular point in time. Unless tar- geting is continually readjusted over time, program managers using the vulnerability measure may initially fail to retire those croplands with a nonlinear productivity decline relationship, thus underestimating the need for policy intervention. It is also important to remember that there are two soil conservation goals: maintaining productivity,
295 which is very important, and avoiding the offside impacts of erosion, which may be of even greater benefit to society (Clark et al., 1985). Measures of these offsite impacts should be integrated into the analysis as well. Finally, the economics profession needs to be chal- lenged to focus greater attention on the economics of the soil erosion problem. Economists have made a major contribution to measuring productivity losses because they come from a tradition that tends to look for common factors to explain systematic behavior, that has been willing to aggregate individual decisions and draw broader generalizations, and that has emphasized model development and simulation. Although modeling and interpreting productivity impacts is an important first step, sight cannot be lost of the economic issues and the need to measure the dollar costs and benefits that are involved. In other words, we cannot afford to become too enamored with the indices and physical productivity measures that are being calculated or developed. Rather, creativity is needed in translating physical measures into economic measures, i.e., the net economic benefits of erosion control programs and policies. It is crucial that these productivity measures are now taken through the final step of the analysis to determine which soils should be targeted, retired, or saved because the social benefits of the policy action exceed the social costs. REFERENCES Clark, E. H., J. A. Haverkamp, and W. Chapman. Eroding Soils: The Off-Farm Impacts. Washington, D.C.: The Conservation Foundation. Miranowski, J. A., and B. D. Hammes. 1984. Implicit prices for soil characteristics for farmland in Iowa. Amer. J. Agr. Econ. 66:745-749.
