CIESIN Reproduced, with permission, from: Easterling III, W. E., P. R. Crosson, N. J. Rosenberg, M. S. McKenney, L. A. Katz, and K. M. Lemon. 1993. Agricultural impacts of and responses to climate change in the Missouri-Iowa-Nebraska-Kansas (MINK) region. In Towards an integrated impact assessment of climate change: The MINK study, ed. N. J. Rosenberg, 23-62. Dordrecht, The Netherlands: Kluwer Academic Publishers.



Present address

1 Institute of Agriculture and Natural Resources, Department of Agricultural Meteorology, University of Nebraska, Lincoln, Lincoln, NE 68583-0728, U.S.A.

2 Resources for the Future, Washington, D.C. 20036, U.S.A.

3 Battelle, Pacific Northwest Laboratories, 901 D Street, SW, Washington, D.C. 20024-2115, U.S.A.

4 c/o EA Engineering, Science, and Technology, 121 S. 13th Street, Lincoln, NE 68508, U.S.A.

5 Department of Ecology & Systematics, Cornell University, Ithaca, N.Y. 13853, U.S.A.

6 Nature Conservancy, 1818 N. Lynn Street, Arlington, VA 22209 U.S.A.

Abstract. The climate of the 1930s was used as an analog of the climate that might occur in Missouri, Iowa, Nebraska and Kansas (the MINK region) as a consequence of global warming. The analog climate was imposed on the agriculture of the region under technological and economic conditions prevailing in 1984/87 and again under a scenario of conditions that might prevail in 2030. The EPIC model of Williams et al. (1984), modified to allow consideration of the yield enhancing effects of CO2 enrichment, was used to evaluate the impacts of the analog climate on the productivity and water use of some 50 representative farm enterprises. Before farm level adjustments and adaptations to the changed climate, and absent CO2 enrichment (from 350 to 450 ppm), production of corn, sorghum and soybeans was depressed by the analog climate in about the same percent under both current and 2030 conditions. Production of dryland wheat was unaffected. Irrigated wheat production actually increased. Farm level adjustments using low-cost currently available technologies, combined with CO2 enrichment, eliminated about 80% of the negative impact of the analog climate on 1984/87 baseline crop production. The same farm level adjustments, plus new technologies developed in response to the analog climate, when combined with CO2 enrichment, converted the negative impact on 2030 crop production to a small increase. The analog climate would have little direct effect on animal production in MINK. The effect, if any, would be by way of the impact on production of feedgrains and soybeans. Since this impact would be small after on-farm adjustments and CO2 enrichment, animal production in MINK would be little affected by the analog climate.

1. Introduction

Agriculture's ability to adjust to a possible future climate change is an issue of major interest. Although a number of analysts have confronted the question of how climate change may affect agriculture (e.g., Bach, 1979; Newman, 1982; Rosenzweig, 1985; Parry and Carter, 1988; Smith and Tirpak, 1989; Adams et al., 1990), they have not dealt with the adjustment issue. These studies are limited in other ways a. well: (a) they impose a scenario of future climate change on the world of today; (b) they represent the direct effects of CO2 on crop photosynthetic and water use efficiencies in an overly simplistic way; and (c) they fail to examine linkages between agriculture and other resource sectors such as water and energy that are bound to be affected by climate change.

The research described in this paper moves beyond these several limitations. In particular, we place major emphasis on the strategies farmers and agricultural research institutions in MINK likely would follow in adjusting to the consequences of climate change in that region.

The general approach taken in the research was to simulate crop production for a set of representative farms across the MINK region and for the region as a whole under the 1951-80 normal, or what we call our control climate, and under an historic analog of climate change. We selected the notoriously hot and dry decade of the 1930s as our historical analog (see Paper 1 for more details on our choice of climate scenarios).

The representative farms are meant to typify farming systems within a set of Major Land Resource Areas.* We then scaled the simulations from the individual farms up to the regional level in order to assess the effect of climate change on MINKs crop output. The initial simulations were of farms using existing technologies and management practices. Subsequent simulations estimated the impact of the analog climate on crop output after allowance for adjustments in management practices and for adaptive technologies that farmers could plausibly be expected to use in response to the changed climate. We distinguish between cumulative, long-term responses to climate change, adaptations, and short-term responses, adjustments. For our purposes the easy, low-cost, currently available responses are called adjustments. Long-term research and changes in institutional policy are called adaptations.

2. Crop Modeling

EPIC (Erosion Productivity Impact Calculator), a simulation model, was used to determine the relationship between climate and crop growth. EPIC was developed by USDA-ARS, SCS, and ERS (Agricultural Research, Soil Conservation and Economic Research Services of the U.S. Department of Agriculture) at the Grassland Soil Water Research Laboratory in Temple, Texas (Williams et al., 1984) to simulate relationships between long-term soil erosion and soil productivity throughout the United States.

EPIC is a physical-process model that simulates the interaction of the soil-climate-management environment of submodels capable of simulating hydrology, climate, erosion-sedimentation, nutrient cycling, plant growth, tillage, soil temperature, management and a simple accounting of costs and returns.

EPIC operates on a daily time step. It converts photosynthetically-active radiation into biomass and the biomass is portioned into above- and below-ground components. The above-ground component is further divided into economic yield and nonharvestable biomass. Daily biomass accumulation can be reduced by environmental stress factors (temperature, moisture, nutrient). Crop phenology in EPIC is a function of the accumulation of growing degree-days. The only input factors that were considered optimal were the timing and application of irrigation water and, in some cases, nitrogen availability.

EPIC is composed of deterministic and empirical relationships. The CERES family of models used in the recent study by the U.S. Environmental Protection Agency (Smith and Tirpak, 1989) is relatively more deterministic and, hence, in principle more easily transported to different regions than is EPIC. However, EPIC is more flexible in modeling alternative farming systems than most simulation models. Such flexibility plus validation testing - discussed below - gave us confidence in using EPIC to simulate crop response to CO2-induced climate change. More details on the use of EPIC in the MINK study can be found in Easterling et al. (1992).

CO2 Direct Effects

Modifications were made in EPIC to permit mechanistic simulation of photosynthetic and evapotranspirational responses to increasing ambient CO2 concentration and changing climatic conditions (see Stockle et al., 1992a and b for full details of these modifications). EPIC uses the concept of light-use efficiency in calculating photosynthetic conversion to biomass. Light-use efficiency in EPIC was made sensitive to atmospheric CO2 concentration based on response curves in the literature assembled by Morison (1987). It is also made sensitive to vapor pressure deficit because it is known that dryness of the air influences the rate of photosynthesis.

Ambient CO2 concentration is also known to reduce the stomatal conductance of water vapor in leaves, thereby reducing transpiration and making plants more efficient water users. EPIC has been modified to track evapotranspiration with the Penman-Monteith model (Monteith, 1965) which considers canopy resistance. Penman-Monteith was made sensitive to changes in CO2 concentration and vapor pressure deficit through empirical relations between these factors and the canopy resistance term (Stockle et al., 1992a, b). Differences between C3 and C4 plants in response to higher atmospheric CO2 concentrations were accounted for.

Representative Farms in EPIC

EPIC is a generic model in that, with proper localized inputs, it can represent farming in a variety of situations and locations. The crop growth model in EPIC ran at the scale of hectare on enterprises that we call representative farms. A representative farm is a description of a cohesive, functional farm enterprise which typifies most of the farms in its particular region. Data which describe each representative farm were collected primarily from detailed interviews with experts from each of the land grant universities in the four MINK states and were supplemented with information in published reports.

Each representative farm was defined by a unique combination of crop rotation, soil type and weather station. A large number of attributes of each farm was recorded including, for example, tillage practices, farm size, input amounts and the like. In addition, information on costs of production was collected for each farm. This resulted in forty-eight farms being modeled for the MINK region, exclusive of some farms that were added later in the analysis with altered farming practices in order to represent adaptations to climate change (Table I).

Input data sets for each representative farm were constricted for the EPIC model. Factors considered as inputs to EPIC are listed in Table II. Note that irrigation water and, in some cases nitrogen were assumed, unrealistically, to be non-limiting and were the only input variables to be treated in this way.

Climate Data Inputs to EPIC

The crop growth model in EPIC requires daily data on maximum and minimum temperature, precipitation, solar radiation, relative humidity and windspeed. Long-term records of daily temperature and precipitation were assembled for 18 NOAA Cooperative Climatological stations distributed across the MINK region (Figure 1). NOAA daily data for the other climatic elements were also assembled for the 1951-1980 control climate. Incomplete records for the 1930s prevented us from relying solely on NOAA data and from using a larger number of stations to represent the region.

Daily records on relative humidity, solar radiation and windspeed were not readily available for the 1930s. Monthly values for relative humidity were calculated from dewpoint temperatures at the First Order stations in Table III. These data are recorded in back issues of Monthly Weather Review. Monthly values for solar radiation were estimated in a univariate regression using percent possible sunshine as the independent predictor of solar radiation (see Easterling et al., 1992 for details on the techniques used to reconstruct climate data for the 1930s). Daily values of relative humidity and solar radiation were then developed from the monthly data in a stochastic weather generator in EPIC.

Daily records on windspeed were not available for the 1930s. Though it was windier in the MINK region in the 1930s than at present, we did not alter wind values from their current normals for the EPIC simulations. This was justified on the basis of sensitivity analyses that suggested that EPIC simulated crops are relatively unresponsive to slight changes in windiness.

The ten-year record of the 1930s is too short to allow EPIC to achieve stable results, particularly in the case of multi-crop rotations. We extended the 1930s to 60 years by simply repeating the decade sequentially. Comparison with a 60 y sequence assembled by selecting years from the decade (with replacement) showed virtually identical results in terms of means and variances in crop yields and other outputs.

Validation of EPIC

Before EPIC, with our modifications, can pass muster as a tool for estimating the effects of climate change, it must realistically simulate current crop yields and evapotranspiration (ET). EPIC was validated for a select number of representative farms by comparing its simulated yields under the control (1951-80) climate with USDA (NASS) data, with estimates by the experts (EXPERT) who helped build the representative farms,** and with data from agronomic field experiments reported in the literature. NASS provided data on average yields in the period 1984-87, which we take to be the 'current' baseline against which climate change effects on crop production are compared in all subsequent analyses. Details of these verification procedures are given in Rosenberg et al. (1992).

In general the EPIC yield simulations of all crops modeled in the MINK region were within +/-20% of both the NASS and EXPERT data. with some outliers. Figures 2a and 2b show that despite some variation, particularly in dryland corn, EPIC yields cluster with NASS and Expert yields. The clustering is sufficiently strong (r[2] of 0.84 and 0.88 for the NASS and Expert yields, respectively) in our judgment to justify use of EPIC to estimate the impacts of the analog climate on crop yields. Better agreement of EPIC yields with the independent data would have been unlikely since EPIC was run with weather records from 1951-80 while the NASS and EXPERT data were from the period 1984-87. The technologies and management practices in EPIC, however, were those prevailing in 1984-87.

EPIC simulations of yield and et also agreed well with the results of selected agronomic experiments conducted either in the MINK states or in adjacent regions. A summary comparison of EPIC simulations with the experimental yields and ET is shown in Tables IV and V respectively. Simulated yields and ET fell well within the ranges of the experimental values. We conclude that the EPIC model reflects farming systems across the MINK region in the baseline period with sufficient accuracy for our purposes.

3. The Impact of Climate Change on Current Agricultural Production

In this section we ask how crop and livestock production as they are currently practiced in MINK would be affected by a permanent shift from the 'current' (i.e. 1951-1980) climate to that of the 1930s. Using EPIC, we consider a benchmark 'worstcase scenario in which the 1930s climate is imposed on representative farms with present ambient CO2 concentration and no attempts by farmers to adjust to the climate change. (We call this task B1). (See Paper 1 of this series for details of the task structure.) We then increase ambient CO2 by 100 ppm over present levels (task B2). Finally, we allow a set of easy, low cost adjustment strategies to be used on the farms to deal with the climate change, both at today's ambient CO2 concentration and at the higher level (task B3).

Crop Yield Effects without On-Farm Adjustments

Across the Region as a Whole. Table VI shows the effects of the 1930s climate on baseline, or control, yields with and without the CO2 enrichment effect and in the absence of on-farm adjustments. Without CO2 enrichment the effects ranged from a decline of 25% for corn and soybeans to an increase of 9% for irrigated wheat. The effect on yields was negative for all crops except irrigated wheat. The negative effects occur primarily because the higher temperatures under the analog climate force growing degree-days to accumulate rapidly which causes the modeled crops to mature before achieving adequate grainfill. Water stress also claimed some of the maximum potential yield of these crops.

The effect of the analog climate on wheat yields may be due, in part, to more moderate climate conditions during the cool seasons across large parts of the wheat production area.

Table VI indicates that the addition of 100 ppm of CO2 to the atmosphere would reduce the simulated yield loss for dryland corn, soybeans and sorghum yields. Dryland and irrigated wheat yields would be 10-11% higher than the control, and dryland alfalfa yields would be up 8%.

Geographic Variation in Crop Yields. Differences in yield effects across the region were closely related to the regional differences between the analog climate (1931-40) and the control climate (1951-80). Figures 3a and 3b show 10 individual dryland corn farms and 10 individual irrigated corn farms distributed across the MINK region. Farms can be located on Figure 1. Aside from large yield losses in southeastern Iowa where the analog climate was unusually severe, the greatest yield losses were in the more climatically marginal western MINK areas. Western MINK areas were more droughty in the 1930s than eastern MINK areas.

Interannual Variation in Cop Yields. The analog climate would cause an increase in the proportion of years in which the farms experienced poor versus bumper harvests. Figures 4a and b show the interannual distribution of yields for a Nebraska corn farm and a Missouri soybean farm with no CO2 enrichment and no adjustments. Bumper harvests would disappear for both farms under the analog climate and poor harvests would become much more frequent. In the Missouri soybean case, it is interesting to note that, despite the increase in 'normal' harvests ones clustered in the middle of the distribution - the increased frequency of poor harvests at the expense of bumper harvests would increase the riskiness of farming there.

Impact of the Analog Climate on Evapotranspiration. In the EPIC simulation lower crop yields in the analog period were associated with reduced evapotranspiration (ET) in the dryland crops. Among the irrigated crops ET increased since water supply, for the purpose of this phase of the simulations, was assumed non-limiting. The reduction in ET on dryland was primarily the result of the shortfall in precipitation and abbreviated growing season. The analog climate with the additional 100 ppm of CO2 decreased ET still more in the dryland grain crops an wheatgrass. The additional CO2 also decreased ET in irrigated crops as well.

For irrigated crop production EPIC estimates the amount of water the plants require to avoid water stress. Under the analog climate, and in the absence of CO2 enrichment, this requirement would increase by 11% for wheat relative to the 'control' requirement, by 21% for sorghum, and by 29% for corn. The estimates in Table VI of the effects of the analog climate on irrigated crop yields assume that these higher water requirements are met.

This provides useful technical information about the impacts of the analog climate, but economic considerations suggest the assumption is quite unrealistic. Frederick (1991) in the accompanying paper of water resources argues that groundwater supplies for irrigation in Kansas and Nebraska, where most of the region's irrigated production is located, are inadequate to long sustain even current rates of withdrawal. Were the analog climate to become the new normal, the increased withdrawals implied by EPIC likely would quickly become uneconomic and therefore unsustainable.

Although we have not developed a formal analysis of the economics of irrigated crop production in MINK under the analog climate, we have addressed some aspects of these economic issues. This discussion is presented below.

Crop Yield Effects with On-Farm Adjustments

Farmers would surely attempt to adjust their farming operations in response to climate change. In this research, we examined a variety of currently-available, low-cost adjustment strategies in search of ones that might help MINK farmers deal with the impacts of the analog climate (see Easterling et al., 1991, for a more detailed discussion of the methods used to represent and evaluate these adjustments). Sensitivity analyses on a small set of representative farms indicated that the only EPIC-simulated adjustment strategies that were effective were earlier planting in combination with longer season varieties in the annuals (except wheat) and simply shorter season varieties in the perennials (wheatgrass and alfalfa), and the use of furrow diking to conserve moisture in the dryland warm season crops. Based on sensitivity analyses, all warm season crops were planted 14 days earlier with cultivars that require an additional 200 heat units to reach maturity in order to simulate a longer-season cultivar. Early planting apart, wheat was also treated in this way. Crop substitutions such as, for example, drought-hardier sorghum or wheat in place of corn, were examined as well.

When these adjustments, crop substitutions apart, were applied to a larger subset of representative farms across the MINK region, yields of all irrigated crops, alfalfa and dryland wheat were marginally higher under the analog climate with current CO2 levels than under the control climate. Yields of dryland corn, sorghum and soybeans still were less than under the control, but by smaller amounts than before adjustments. (Compare Tables VI and VII).

For all crops except irrigated and dryland wheat analog climate yields with CO2 enrichment are higher with than without adjustments (Figure 5). Except for dryland corn and soybeans, yields with CO2 enrichment and adjustments are higher than control yields.

Yields after adjustment were higher than without adjustment for all crops except irrigated wheat (Table VII). For example, before adjustments, dryland corn yields under the analog climate were 4.9 t/ha, down 25% from control yields (Table VI). After adjustment dryland corn yields were 5.2 t/ha, 20% less than control yields but 6% more than without-adjustment yields (Table VII). Irrigated wheat yields before adjustments were 4.9 t/ha, 9% higher under the analog climate than under the control (Table VI). After adjustments irrigated wheat yields were 4.6 t/ha, only 3% above control after adjustments (Table VII). Consequently, the adjustments reduced irrigated wheat yields 5% relative to yields without adjustment (Table VII).

The adjustments were also evaluated in terms of economic profitability. obviously farmers would not adopt an alternative practice if it did not pay off economically. In treating the effects of adjustments on profitability we considered only the scenarios without CO2 enrichment.

To estimate the effects of the adjustments on farm profits (net revenues) we need crop prices to estimate the value of gross crop production, and estimates of production costs. We could have used actual crop prices in 1984-87, the baseline period, but decided against this because these prices were well below their long-term trend, reflecting the sharp decline in demand for U.S. crop exports in the mid-1980s. We judged that trend prices better represented long-term equilibrium levels than actual 1984-87 prices. Consequently, with U.S. Department of Agriculture data, we calculated the price trend for each crop in 1945-1984, and used the trend to extrapolate prices for 1984-87. The annual averages of these trend prices were used to value crop production in that period.

EPIC tracks fixed and variable costs of production based on data for each of the representative farms. Net revenue for each farm, therefore, was calculated as the difference between the farm's gross value of crop production and its production costs as estimated by EPIC.

We used these calculations only to judge whether the various adjustments would reduce the negative impacts of the analog climate on farm profitability. We did not address the question of whether the effect of the adjustments on profits would be sufficiently favorable to hold people in agriculture who otherwise would be forced out by the analog climate. This is a much larger question, and well beyond the scope of our study.

Our analysis showed that in wheat farming the adjustments would actually reinforce rather than offset the effects of the analog climate on net revenues. It can be assumed, therefore, that wheat farmers would not adopt any of the alternative practices we considered. For all other dryland crops the adjustments would reduce the negative effects of the analog climate on net revenues.

For irrigated sorghum the adjustments would accentuate rather than offset the loss of net revenue. As noted above, the analog climate would increase irrigation water requirements for sorghum by 21% relative to the control climate. The principal adjustment practice for sorghum, the use of longer season varieties, would increase the irrigation requirement by 28% relative to the control. Although this practice converts a 10% yield loss for sorghum (Table VI) to a 13% yield gain (Table VII), the increased cost of meeting the higher irrigation requirement more than offsets the increased benefits of the higher yield. Consequently, farmers engaged in production of irrigated sorghum would not likely find it economically attractive to switch to longer season varieties.

With irrigated corn the outcome is different. The longer season varieties turn a 7% loss of yield (Table VI) into a 1% gain (Table VII), more than enough to offset the negative net revenue effect of an increase in the irrigation water requirement from 29 to 32%. Consequently, farmers producing irrigated corn likely would find the longer season varieties more attractive than those currently in use.

Crop Rotation Substitutions

Crop rotation substitutions were examined to determine their efficacy in offsetting climate-induced productivity declines. Crop substitutions were also evaluated in terms of economic profitability. Under the current climate, for example, farmers may find a corn-soybean rotation more profitable than a sorghum-soybean rotation. However, sorghum is drought-hardier than corn and could be more profitable than corn under climate change. Table VIII shows revenues under the analog climate for corn and sorghum on 7 representative farms, where each crop grew under the same conditions (as if in adjacent fields). The comparisons are intended as an example of the calculations farmers might make in deciding on crop switching as a response to the analog climate. For this purpose we consider responses in the absence of the other adjustments farmers might make.

Even though sorghum yields would be less negatively affected by the analog climate than corn yields - with and without CO2 - corn, because of its higher price and its higher yields (Table VI), still would earn higher net revenues than sorghum on 3 of the five dryland farms and on all of the irrigated farms. As far as they go, these results suggest some shifting out of dryland corn into dryland sorghum in the eastern part of MINK but no such changes in the irrigated areas of the western part.

Scaling up EPIC Results

The final objective of the crop modeling was to derive regional changes in production of each crop under the analog climate by aggregating the individual farm simulations. The aggregation was done in two steps. First, the MLRA average percentage changes in yields were weighted by the percent of land in a given soil type and then the statewide changes were weighted by mean production totals for each MLRA (see Easterling et al., 1991, for details on the scaling up procedure).

Frederick's account (1991) of increasing water scarcity in the western part of MINK suggests that it probably would not be economical for farmers with irrigated land in that part of the region to respond to higher crop water requirements under the analog climate by pumping more, holding the total amount of irrigated land constant. Our analysis of the profitability of adjustments to the impacts of the analog climate in irrigated corn and sorghum production also indicated that the amounts of groundwater withdrawals probably would not be those indicated by EPIC.

We think a more plausible response by farmers would be to let presently marginal land go out of irrigated production and to increase per hectare applications of water on the better remaining land. To represent this response we assume that total water applied would remain at the 1980s baseline rate and that per hectare applications for each crop would rise by the percentage increase in water required, according to EPIC. For example, where the analog climate caused a 25% increase in per hectare irrigation requirements, the farmer would irrigate his better land at the required rate and put 20% of the less favored land into a hardier dryland crop like sorghum or wheat. This would hold irrigation water withdrawals at 1984-87 base-line levels.

Region-wide Impacts on Crop Production. The decline in yields under the analog climate would increase the costs of producing the various crops. The resources available to us did not permit estimation of the cost curves of the crops so we are unable to estimate the amounts of the cost increases and the subsequent declines in production. Instead, we assumed that costs of each crop in the scenarios with no on-farm adjustments would rise such that the decline in production would be proportional to the decline in yields. In the scenarios including on-farm adjustments the declines in production reflect both the decline in yields and some shifting of land out of irrigated corn, because of the increased scarcity of water, and into dryland production of wheat and sorghum.

Table IX shows the impacts of the analog climate on crop production in the region before on-farm adjustments, with and without the CO2 enrichment effect on crop yields. The impacts are calculated as changes in production from average amounts in 1984-87. Table X shows the impacts after on-farm adjustments, with and without the CO2 effect. The changes in value were found by multiplying the changes in production volume by average 1984/87 trend prices of the various crops expressed in 1982 dollars.(Trend prices rather than actual prices were used for reasons given above.)

When no account is taken of the CO2 enrichment effect, and no adjustments are permitted, the value of MINK production of the five crops under the analog climate declines 17.1% from the average 1984-87 value of protection. Corn alone accounts for 60% of the decline and soybeans for 29%. Because over half of MINK's production of corn and soybeans is in Iowa, 56% of the total production decline of $2,709 million is in that state. Because it specializes in wheat production. which is not much affected by the analog climate, Kansas would suffer the smallest decline in the total value of crop production.

Because CO2 enrichment ameliorates the effect of the analog climate on crop yields, the decline in the value of 1984-87 production with CO2 enrichment is 8.4% instead of 17.1% (still with no on-farm adjustments). In this scenario Iowa absorbs 72% of the total production decline, a substantially higher percentage than in the without-CO2 scenario. Iowa's total loss, however is less than in that scenario. In the with-CO2 scenario the value of production in Kansas actually increases slightly from the 1984-87 level primarily because wheat yields increase with more CO2 and wheat is the dominant crop in Kansas.

The on-farm adjustments to the analog climate further reduce the impact on regional production. Comparison of the without CO2 cases in Tables IX and X indicates that the simulated on-farm adjustments would reduce the loss of crop production from $2,709 million, or 29%. The combination of CO2 enrichment and on-farm adjustments reduces the loss further to %532 million (Table X) or by 80% compared with the no adjustment, no CO2 enrichment case. In these simulations, CO2 enrichment is more important than on-farm adjustments in moderating the effects of the analog climate on crop yields.

Region-Wide Impacts on Animal Production. In 1984-87 MINK accounted for 31% of national production of cattle, calves and hogs (Table XI), about the same as its share of national corn and soybean production. The similarity of these percentages suggests that corn and soybean production is spatially complementary with production of cattle and hogs. Other evidence supports this. Data for corn and hay (the other principal animal feed) indicate that in 1977-79 (the last years for which the data are available), 39% of the corn and over 85% of the hay produced in MINK were used to feed animals on the farms where the crops were produced (USDA, 1980). Soybeans must be processed before feeding to animals, so only 1-2% of soybean production in the region was consumed on the farm where it was produced. However, IMPLAN, the input-output model for MINK discussed in Paper 2 in this volume, shows that a little over half of soybean production in MINK is consumed in the region.

A study of the cattle industry in Nebraska also indicated a strong spatial complementary between feedgrain production and animal production. The study found that most of the cattle marketed in Nebraska spent some time in feedlots before sale and slaughter (Azzam et al., 1987). Of the 11,600 feedlots in the state in 1982, 97% were on farms and fed no more than 1,000 cattle per year. The 3% of the commercial (off-farm) feedlots handling more than 1,000 animals per year were responsible for two-thirds of the 4 million head of fed cattle marketed in 1982.

Byrkett et al. (1976) found that availability of grain was a major factor in explaining the location of feedlots in the United States. Studies done in the four MINK states suggest that a reason for the locational importance of grain is the high percentage of feed costs in total value added in animal production (Jacobs, 1988; Iowa State University, 1988; Agricultural Economics Extension Staff, 1988; Fausett and Barnahy, Jr., 1988). The percentages ranged from 40% to well over 50% in cattle and hog production, depending in part on the price of corn.

The study by Byrkett et al. (1976) indicates that regional differences in land values also affect the location of the livestock industry. Industrial and commercial activities typically generate higher returns to the land than animal production. So, in regions where these activities are concentrated, animal production cannot compete successfully for the land. Moreover, Byrkett et al. hypothesize, industrial regions typically have high human population densities, bringing conflict between people and animal production because of the odors, noise, and waste disposal problems animal production entails. In their statistical analysis, Byrkett et al. (1976) found that regional differences in land use intensity were highly significant in explaining the spatial distribution of animal feeding activities. MINK, of course, is one of the more lightly populated regions of the country, with comparably low densities of industrial and commercial activities.

The locational tie between feedgrain/soybean production and animal production in MINK suggests that the impacts of the analog climate on crop production would indirectly affect animal production in the region.*** In the short run - say up to five years - the effect likely would be small if animal producers in MINK could import feedgrains and soybeans from neighboring states at prices not much different than those prevailing before imposition of the analog climate. Over the long run. however, there might be some shift of animal production out of MINK...

* Major Land Resource Areas, as designated by the Soil Conservation Service (1981), are regions of relatively homogeneous climate, soils, vegetation and land use. Eleven such areas were chosen within the MINK region to represent farming there.

** Experts were asked to judge 'typical' yields on a given farm in a 'normal' year. The expert judgments for yield, though subjective, relate more specifically to the representative farms than do the other sources.

*** The analog climate might also affect animal production directly through effects on animal health, fertility, and productivity. However, discussion with scientists at the U.S. Department of Agriculture's Meat Animal Research Center in Clay Center, Nebraska indicated a strong consensus that small adjustments in animal management practices could ameliorate most, if not all, of any direct negative effects of the analog climate on animal production. The main effects, it was held, would be indirect by way of effects on feedgrain and soybean production.