Introduction
 

Agriculture ranks among the most important economic activities of the Great Lakes region, accounting for more than $15B in annual cash receipts (USDA/NASS, 1997a). Livestock, including dairy, is the number one agricultural commodity group, comprising over half of the total. Dairy production alone produces almost $5B in receipts. Other major commodity groups include; grains/oilseeds, vegetables, ornamentals, and fruit. Crop diversity is an important characteristic of agriculture in the region due at least partially to the moderating influence of the Great Lakes on regional climate (Changnon and Jones, 1972). Over 120 commodities are grown or raised commercially in the region (USDA/NASS, 1997a).
 

Geographically, agriculture in the region follows a south to north gradient, with intensive rowcrop monoculture in southern sections gradually giving way to forests and other natural vegetation across the north. Southern sections of the region form the northern boundary of the U.S. Corn Belt region, with corn, soybeans, hogs, and cattle as major commodities. Dairy and associated alfalfa production are common in the driftless area of southeastern Minnesota and southwestern Wisconsin and scattered across central and northern sections of the region. Vegetable production is centered in the Central Sands area of central Wisconsin, across Lower Michigan, and in the Red River Valley of Minnesota, which also forms the eastern boundary of the Great Plains small grain production area. Fruit and ornamental crops are grown intensively along the eastern shore of Lake Michigan.
 

The amount of water and the frequency of its availability are primary climatological constraints for the production of most annual crops (Thompson, 1986; Andresen, 1989). Approximately 2-3% of farmland in the region is commercially irrigated, largely for potato, sugarbeet, and seed corn production (USDA/NASS, 1997b). During drought periods, such as 1988, large area yields for corn and soybeans were less than 50 % of average trend yields (USDA/NASS, 1997a). Conversely, farmers lost entire crops during the extensive flooding years of 1993 and 1997, particularly in Minnesota and Wisconsin. Early season cool, wet spells can lead to delays in planting which require a shift to a shorter season cultivar or to a different crop such as soybeans (Schwartz, 1999) and may lead to substantial crop yield and quality reductions due to plant diseases (McMullen, 1997).
 

Regional Agricultural Stresses
 

The major stresses on agriculture in the Upper Great Lakes region can be generally categorized as economic, environmental, social, and regulatory. Most important among these categories currently are economic stresses, which include commodity prices near or below the costs of production, price volatility, difficulty in obtaining finance and capital, high farm labor costs and low labor availability, and loss of industry infrastructure (Olson, 1999; Lotterman, 1998). Environmental factors may also create direct stress on regional agriculture. Major environmental factors include climate and its inherent variability, long-term degradation of soil resources, geographical concentration of livestock production and the associated management of large amounts of livestock waste, and the contamination of surface waters and groundwater by agricultural chemicals (Lakshminarayan,et al., 1996). Societal issues influencing the agricultural industry include population shifts to and from rural areas, loss of farmland to urban sprawl, and increasing land values and property taxes. Finally, one category of stresses that integrates many of the above factors is governmental regulation, which may drastically change standards or alter the economics of the production system. One current example is the gradual implementation of the 1996 Food Quality Protection Act, which may ultimately result in the loss of many pesticides used commercially in agriculture (especially in fruit and vegetable production) and for which few, if any, substitutes now exist (DiFonzo and Olsen, 1997).
 

Potential Impacts from Climate Change: Past Studies
 

The potential impacts of climate change on regional agriculture will depend greatly on the magnitude, timing, and the variability associated with the change. Of these elements, variability is generally considered the most difficult with which to cope and adapt, especially should it increase in the future (Parry and Carter, 1985). The majority of research on climate change in the Great Lakes region suggests a warmer and wetter climate in the future (Wigley, 1999), with relatively more warming occurring in the winter and spring than in other seasons (Mortsch and Quinn, 1996). Agriculturally, this would most likely lead to a longer growing season and greater potential productivity, but also to greater potential rates of evapotranspiration. An additional critical factor in determining potential productivity is CO₂ enrichment, which has been associated with increases in total plant dry matter accumulation and improved crop water use efficiency through decreases in transpiration rates. While some research studies have shown that yield increases due to higher CO₂ levels may decrease when other resources are limiting and that the enrichment effect may decrease over time for some plant species, most scientific literature suggests significant long term benefits to agriculture as ambient atmospheric concentrations of CO₂ increase in the future (Rosenzweig and Hillel, 1998; Idso and Idso, 1994).
 

There are relatively few previous research studies pertaining to potential agricultural impacts of climate change within the Great Lakes region itself. Early research on crops in this region suggested future increases in the yields of several major grain crops, largely the result of a longer growing season and shift to crop cultivars and varieties with greater potential productivity (e.g. Rosenzweig et al., 1994; Ritchie et al., 1989). Subsequent studies have yielded mixed results, however, with yield increases associated with CO₂ enrichment at least partially offset by a hastening of crop phenological cycles and increasing levels of water stress (Brown and Rosenberg, 1999). Impact studies which have included estimates of the costs and benefits associated with changes in crop productivity at locations within or close to the Great Lakes region have also tended to suggest a net cancellation of the potential positive and negative effects of climate change, resulting in overall minor economic impacts on agriculture (Lewandowski and Schimmelpfennig, 1999; Doering et al., 1997; Adams et al., 1995; Kaiser et al., 1993).
 

Adaptations and Coping Mechanisms
 

If the magnitude of regional climatic changes in the future reaches values suggested by general circulation studies, farmers will be forced to adapt to the changes or become uncompetitive and unprofitable. Future adaptations could include longer season or different crop varieties, double cropping systems, use of irrigation, and other unforseen technological improvements. There is evidence based on past performance that agriculture could at least partially adapt to a changing climate and that the costs of such adaptations would be small compared to costs associated with an expansion of or changes to major production areas (Easterling, 1996; Doering et al., 1997; Mendelsohn, et al., 1996). Ultimately, however, the ability to adapt will likely depend upon the nature of the climatic change, as increases in variability could make future adaptations difficult (Mearns et al., 1997).
 

Current Regional Study
 

Given relatively few past studies concerning climate and agricultural production in the Great Lakes region, the major objective of this study was the deterministic simulation of crop behavior at a local level as a function of weather and climate alone under both historical and projected future climates. Particular attention was given to areas within the region where agricultural activities have historically been limited by climatological and soil constraints, but which could become more favorable for agriculture in the future given a warmer climate. Three crops commonly grown in the region were chosen for the study: alfalfa, a forage used extensively for dairy production; maize, a coarse grain; and soybean, an oilseed. Simulation models based on the physiological processes which govern growth and development of the crops were used: DAFOSYM (Rotz et al., 1989), CERES-Maize (Jones and Kiniry, 1986), and SOYGRO (Jones et al., 1988) for alfalfa, maize, and soybean crops, respectively. These simulation models have been successfully used in a wide number of past studies and applications (e.g. Easterling, 1998; Adams et al., 1995; Pickering et al., 1995). The models require large amounts of input data including daily values of precipitation, maximum and minimum temperatures, and solar radiation. Thirteen locations across the region (5 in Minnesota, 3 in Wisconsin, and 5 in Michigan) were chosen for the study on the basis of climatological series record length and homogeneity, as well as geographical coverage across the region and its major land use zones (see Table 1). Three of the study locations were specifically chosen from northern, historically non-agricultural areas of the region to investigate the potential for agriculture under a warmer climate. The study was divided into two major categories; historical and future. Historical scenarios were based on observed daily weather data at each of the locations. The individual lengths of the data series ranged from 67-102 years and all but 4 of the series began in the 1895-1900 period. Potential future weather scenarios were derived from two of the general circulation models chosen for use in the U.S. National Assessment: 1) the United Kingdom Meteorological Office/ Hadley Centre for Climate Prediction and Research HADCM2 model (Johns et al, 1997); and 2) the Canadian Climate Centre First generation Coupled General Circulation Model (CGCM1) (Flato et al., 1998). Both models were run with assumed future transient increases of 1% equivalent CO2 increase per year plus the effects of atmospheric aerosols through year 2100 as outlined in the Intergovernmental Panel on Climate Change (IPCC) scenario IS92a (IPCC, 1994). Future daily weather series used in the study were taken from the gridded VEMAP2 0.5X 0.5 resolution data set for the United States (UCAR, 1999) developed by Kittel et al. (1997). Daily series for each grid were generated synthetically on the basis of monthly differences between GCM projections and historical (observed) data. For all 13 station locations in the study, data were taken from the nearest available VEMAP2 grid cell.
 

In order to isolate the effects of weather and climate on crop performance, fertility levels in all models were assumed to be non-limiting for crop growth and development. Soils data in the simulations were based on profile data typical of agricultural soils in the vicinity of each location in the study and were taken from laboratory pedon data available at the National Soil Survey Center (USDA/NRCS/NSSC, 1999). Other agronomic input data necessary for the crop simulations (e.g. crop cultivar characteristics, plant populations) were chosen to reflect typical current (i.e. late 1990's) technology and growing conditions. Planting dates of both maize and soybean crops each season were determined automatically by the models based on user-specified weather and soil conditions. For all alfalfa simulations, a first-year crop stand was assumed. Maize and soybean were harvested each season at physiological maturity while the automatic harvest option was used for alfalfa, with four seasonal cuts of alfalfa taken at the 7 southernmost stations and three seasonal cuts taken at the remaining 6 northern stations depending on weather conditions.
 

Verification of the crop simulation models used in the study consisted of a comparison of simulated yields vs. observed county-levels yields on an independent 5-station set of data for the period 1961-1990. Because of significant upward trends in the observed yields over the verification period due to improvements in technology and other factors, the simulated and observed series were not directly comparable as the simulated series assumes constant levels of technology. As a result, the observed data series were first statistically detrended with linear or nonlinear regressions. Series of residuals defined as the difference of observed (simulated) and trend (1961-1990 simulated average) yields were then computed. General agreement was found in comparisons of the observed and simulated residuals, with mean absolute differences between simulated vs. observed residuals ranging from 0.41-0.72, 0.48-0.65, and 0.16-0.26 ton/acre for alfalfa, maize, and soybean crops, respectively. In general, the models tended to overestimate interannual variability relative to the observed yields, which is expected since the observed series are spatial averages of smaller scale yields within a county and the model simulations, by definition, are single, plot-level estimates. Most importantly, however, the models appeared to account for the majority of within-season shortages and surpluses of plant available moisture and the resulting negative and positive deviations from mean yields, and were, therefore, judged to be acceptable for use in this study.
 

To investigate the effects of possible future CO₂ enrichment, the ambient atmospheric concentration of CO₂ for future crop model simulations was based on the IPCC IS92a scenario following Joos et al.(1996). CO₂ concentrations in this scenario are expected to increase to just above 700ppm by the year 2100. For all historical simulations, ambient atmospheric CO₂ concentrations were held constant at the model default level of 330 ppm. CO₂-dependent changes in rates of biomass production for each crop were taken from Curry et al. (1990) while changes in stomatal resistance were obtained from Rogers et al. (1983).
 

Historical and future model simulations for all three crops were run in chronological order for each location and scenario, with soil moisture in the models initialized each season at the drained upper limit (DUL) on the 1^st of March for DAFOSYM and on the 1^st of April for CERES-Maize and SOYGRO. Model output including daily estimates of crop moisture usage, soil water content, and water stress intensity was saved and archived, with growing season statistics calculated in a secondary processing step. For statistical summation and averaging purposes, the growing season was defined as the date of planting through physiological maturity for maize and soybean, and from 1 March through 31 October for alfalfa. In order to determine the nature of changes in the model output variables studied, estimates of both trend magnitude and significance were calculated.
 

Model Simulation Results
 

Historical Trends
 

When averaged across the historical record of the study, simulated seasonal crop yields at the 13 regional study locations are comparable to observed yields with regional ranges of 3.77-4.93 ton/acre (8.46-11.06 ton/ha), 17.88-116.52 bu/ac (1.20-7.82 ton/ha), and 15.58-41.34 bu/ac (0.98-2.60 ton/ha) for alfalfa, maize, and soybean, respectively. Highest simulated yields for all crops occurred at southern locations with decreasing yields to the north. Greatest interannual yield variability for maize and soybean was found at northern stations, where crop failures due to extreme low temperatures and lack of seasonal growing degree day accumulations (i.e. insufficient heat unit accumulation) were common. Yield variability of alfalfa was also greatest at northern locations, but less pronounced than for maize and soybean. The impact of the soil profile chosen for the simulations could also be identified at some of the locations, with relatively higher yields on deeper soil profiles (greater than 60 in. (150 cm) in depth) and lesser yields on shallow (1m or less) profiles.
 

Several trends were identified in the historical climatological data and from agroclimatological output variables derived from the model simulations. Across the study period, increases in growing season precipitation were found at 10 or more of the 13 locations for all three crops. These increases have generally taken place from the 1920's and 1930's through the present, are in agreement with larger regional scale trends (Karl et al., 1994), and are at least partially due to greater frequency of wet days and wet days which follow wet days (Andresen, 1999). Increases in simulated soil moisture available to the plant at mid-season, a key variable in determining ultimate yield potential, were also found for maize (1l of 13 locations with increases) and soybean (12 of 13 locations with increases). In contrast, simulated potential evapotranspiration, the potential loss of water due to soil evaporation and plant transpiration, was found to decrease at 11, 10, and 9 of the 13 locations for alfalfa, maize, and soybean crops, respectively. As a result of the trends towards wetter, less stressful conditions, there were increases in both maize (positive trends at 11 out of 13 locations) and soybean (positive trends at all 13 locations) yields occurred across much of the region. Alfalfa yield trends were mixed, with decreases at 8 locations and increases at 5 locations. Overall, greatest increases in simulated yields for all crops over time were found at western and northern study locations.
 

Projected Future Results
 

Projected data for the period 2000-2099 from the United Kingdom Meteorological Office Hadley Centre HADCM2 and Canadian Climate Centre CCGCM1 general circulation model simulations (see Appendix B for additional information) suggest an overall warmer and wetter climate by the year 2099 across the region. The CGCM1 model is the warmer of the two models, with a 7.2F (4C) or greater increase in mean annual temperatures at the study locations by 2099 relative to historical averages (versus a 4.5F (2.5C) increase for HADCM2). Average annual precipitation totals across the region generally increase from approximately 31.5 inches (800mm) at 2000 to 39.4 inches (1000mm) at 2099 for both GCMs. However, the rate of precipitation increase for the HADCM2 GCM is much more consistent over the 100 year period than for the CGCM1 GCM, in which much of the overall 100-year increase occurs during the last 20 years of the period.
 

A comparison of historical and potential future simulated yields for the three crops averaged from 2000-2099 and across all 13 study locations for both GCM and CO2 enrichment scenarios is given in  Table 2. In general, the warmer and wetter climate suggested by both GCMs leads to increases in average simulated non-CO ₂ enriched crop yields relative to historical yields, ranging from 6% for alfalfa for both GCMs to 26% for maize in the CGCM1 model. When the impacts of CO ₂ enrichment are also considered, the yield differential relative to the historical period increases to a range of 16% for alfalfa with the CGCM1 model to 81% for soybean with the CGCM1 model. Largest percentage increases in yield across the 2000-2099 study period were at northern locations. The ratios of the future scenarios with and without CO ₂ enrichment suggest that the majority of yield increases during this period are due to CO ₂ enrichment.
 

It is important to note that the ratios given in  Table 2 represent averages over the entire 100-year future period. The majority of the simulated yield series actually exhibited consistent increases through the period, especially with the HADCM2 model data. Other yield series tended to decrease across the period, or increase during the initial decades of the period, followed by decreases later in the period. The latter pattern was especially true for crop simulations at southern and western study locations with the CGCM1 model. Decadal box plots illustrating these trends at Waseca, MN and Madison, WI relative to historical yields are given in 1a and  1b. In  Figure 1a with the HADCM2 model and CO₂ enrichment, simulated soybean yields increase more than 100% relative to historical yields across the period. With maize and the CGCM1 model at Madison, WI (Fig. 1b), however, CO₂-enriched yields gradually decrease during future decades. Also, with a few exceptions, interannual yield variability (represented by the width of the box/whiskers) is generally equal or less than for the historical series, especially for the CGCM1 model simulation, and appears to decrease towards later decades of the period. These decreases are likely associated with corresponding decreases in the interannual variability of GCM-derived annual mean temperature and precipitation series.
 

Fig. 1a

Fig. 1b

The water balance for a crop at a given location can be described as a balance between the water source, precipitation, and the sum of evapotranspiration, runoff, and drainage (the sinks). Assuming no other water is added or lost from the soil profile, the seasonal change in water from soil storage can then be obtained as the residual of the sources and sinks. A water balance averaged over time for historical and future maize simulations at Madison, WI is given in  Table 3. On a historical basis, the simulations indicate that growing season precipitation is insufficient to meet the evapotranspiration demands of the crop, and that 5.2 inches (132mm) of water (or 29% of the evapotranspiration) must be supplied from soil storage, usually from accumulated non-growing season precipitation. In the two selected future decades, the average fraction of water supplied to the crop by storage decreases to 16%(22%) of the historical values for HADCM2 (CGCM1) models in the 2025-2034 decade and to 7%(24%) of historical values in the 2090-2099 decade. For the HADCM2 model, the reduction in the storage term is a result of an increase in growing season precipitation, runoff, and drainage, and a reduction of evapotranspiration. For the CGCM1 model, the reduction is primarily a result of a reduction in evapotranspiration alone. The reduction in evapotranspiration from the historical values through 2090-2099, ranging from 2.32 inches (59mm) for the HADCM2 model to 3.82 inches (97mm) for CGCM1, is primarily a result of decreases in the transpiration rate of the crop due to CO₂ enrichment.
 

The projected GCM data suggest a warmer and wetter climate for the Great Lakes region by the end of the next century, possibly leading to a northward shift of some current crop production areas (Brklacich and Smit, 1992; Smit et al., 1990). Even with less suitable soils agronomically, the model simulations suggest an improving yield potential (relative to climatology) at three of the northernmost study locations currently outside of major agricultural production areas: Chatham, MI, East Jordan, MI, and Grand Rapids, MN. Simulated CO₂ enriched crop yields averaged over these locations for historical and two future decades are given in Table 4. The average yields for maize and soybean increase dramatically by the 2090-2099 decade relative to historical yields, ranging from 276%(263%) for soybean with HADCM2 (CGCM1) models to 343%(373%) for maize. The increases for alfalfa were smaller, ranging from 26% with the CGCM1 model to 29% with the HADCM2 model.
 

Should climate change occur across the region at some point in the future, it is highly likely that the producers of the crops would adapt to the changes as long as the adaptations were economically feasible (Easterling, 1996). As a simple example of an adaptative strategy, agronomic input data in the CERES-Maize crop model were modified to better suit the warmer future climate suggested by the GCMs at Coldwater, MI, a location typical of the northern Corn Belt region. In particular, the crop was planted 15 days earlier each season (on or after 1 May, depending on weather and soil conditions) and the total number of base 46.4F growing degree days required for the crop to advance from silking to physiological maturity, was increased from 1224 units to 1440 units. Cumulative frequency distributions of simulated yields, derived by ranking the yields from both 'adapted' and 'non-adapted' cultivars from 2000-2099, are given in  Figure 2. Due to total crop failures in 4 of the 100 years of simulation (due to early freezes in the beginning decades of the future scenarios), the adapted crop exhibited a probability of zero yields for a small portion of the distribution. At a probability of 0.11 or greater, however, the adapted yields exceed non-adapted yields and continue as such for the remainder of the distribution, with the magnitude of the differences generally ranging from 14.9-26.1 bu/ac (1.0-1.75 t/ha).
 

Fig. 2

Conclusions
 

Overall, this study attempts to identify changes in the potential productivity of three crops in the region as a function of weather and climate alone. Increases in simulated crop yields during the past several decades were likely associated with corresponding trends toward wetter, less stressful growing season weather. The model simulation results from the two GCMs differ somewhat, but suggest that crop yields in the future may be substantially greater than those observed during the past century due to the effects of CO₂ enrichment and because of more favorable growing season weather, especially in northern sections of the region. The simulations also suggest that the amount of water from soil storage necessary to meet crop demands will decrease, making in-season water shortages (and moisture stress) less likely than in the past. Finally, the majority of projected future yield series exhibit decreasing interannual variability with time, which was associated with decreases of growing season temperature and precipitation variability.
 

Limitations and Further Study
 

There are many limitations to this study. Perhaps most importantly (and due to its primary objective of isolating the effects of weather and climate), it did not consider the impacts of major limiting factors in agriculture such as inadequate fertility or pressure from weeds, diseases, and insect pests. In addition, the projected future weather scenarios are simplistic synthetically-derived series from the coarse-scale, monthly grid output values of the GCMs and represent the output of only two GCM simulations. Future studies based on more representative regional- or local-scale climate simulations which include these and other limiting factors as well as resulting economic impacts are needed for future risk assessment and for the development of new technologies necessary for commercial adaptative strategies if and when climate change does occur.
 
 

1a) Simulated historical and future soybean yields by decade, HADCM2 model data with CO₂ enrichment, Waseca, MN. Lower, middle, and upper horizontal bars in the box represent 25^th, 50^th, and 75^th percentiles, respectively while bottom and top whiskers represent 10^th and 90^th percentiles.
 

1b) Simulated historical and future maize yields by decade, CGCM1 model data with CO₂ enrichment, Madison, WI. Lower, middle, and upper horizontal bars in the box represent 25^th, 50^th, and 75^th percentiles, respectively while bottom and top whiskers represent 10^th and 90^th percentiles.
 

2) Cumulative simulated frequency distributions of adapted vs. non-adapted crop cultivars, 2000-2099, with HADCM2 data, Coldwater, MI. The Y-axis (p(x)) value indicates the fraction of all yields less than or equal to the corresponding number on the X-axis. The adapted cultivar required 18% more growing degree days between silk and maturity than non-adapted cultivar and was planted 15 days earlier.
 
 
 

References
 

Adams, R. M., Fleming, R. A., Chang, C.C., McCarl, B., and C. Rozenzweig, 1995: A reassessment of the economic effects of global climate change on U.S. agriculture. Climatic Change, 30, 147-167.
 

Andresen, J.A., 1999: Climatological Trends and Potential Agricultural Impacts in Michigan: 1930-1995. Submitted to Michigan Agricultural Experiment Station Research Report Series, Michigan State University, East Lansing, MI.
 

Andresen, J.A., R.F. Dale, J.J. Fletcher, and P.V. Preckel, 1989: Prediction of County-Level corn Yields Using and Energy-Crop Growth Index. J. Clim. 2:48-56.
 

Brklacich, M. and B. Smit, 1992. Implications of changes in climatic averages and variability on food production opportunities in Ontario, Canada. Climatic Change 20, 1-21.
 

Brown, R. A. and N. J. Rosenberg, 1999: Climate change impacts on the potential productivity of corn and winter wheat in their primary United States growing regions. Climatic Change 41, 73-107.
 

Changnon, S.A. and D.M.A. Jones, 1972: Review of the influences of the Great Lakes on Weather. Water Resour. Res. 8, 360-371.
 

Curry, R.B., R.M. Peart, J.W. Jones, K.J. Boote, and L.H. Allen. 1990. Simulation as a tool for analyzing crop response to climate change. Trans. ASAE 33:981-990.
 

DiFonzo, C. and L.G. Olsen, 1997: All About the Food Quality Protection Act. The IPM Report Vol. 6, No. 1. Michigan State University Extension, East Lansing, MI.
 

Doering, O., M. Habeck, J. Lowenberg-DeBoer, J. C. Randoph, J. J. Johnston, B. Z. Littlefield II, M. A. Mazzocco, M. Kinwa, and R. Pfeiefer, 1997: Mitigation Strategies and Unforseen consequences: A systematic assessment of the adaptation of upper midwest agriculture to future climate change. World Resource Review, 9 (4), 447-459.
 

Easterling W.E., A. Weiss, C.J. Hays, and L.O. Mearns, 1998: Spatial scales of climate information for simulating wheat and maize productivity: the case of the US Great Plains. Agric. For. Met. 90:51-63
 

Easterling, W. E., 1996: Adapting North American agriculture to climate change in review. Agric. For. Met., 80, 1-53.
 

Flato, G.M.; Boer, G.J.; Lee, W.G.; McFarlane, N.A.; Ramsden, D.; Reader, M.C.; Weaver, A.J., 1998: "The Canadian Centre for Climate Modelling and Analysis Global Coupled Model and its Climate". Submitted to Climate Dynamics.
 

Idso, K. E., and S. B. Idso, 1994: Plant responses to atmospheric CO2 enrichment in the face of environmental contraints: a review of the past 10 years' research. Agric. For. Meteorol. 69:153-203.
 

IPCC, 1994: Climate Change 1994: Radiative Forcing of Climate Change and an Evaluation of the IPCC IS92 Emission Scenarios, J.T. Houghton, L.G. Miera Filho, J. Bruce, Hoesung Lee, B.A. Callander, E.F. Haites, N. Harris, and K. Maskell (eds.). Cambridge Univ. Press, Cambridge, UK.
 

Johns T.C., Carnell, R.E., Crossley J.F., Gregory J.M., Mitchell J.F.B., Senior C.A., Tett S.F.B. and Wood R.A., 1997: The Second Hadley Centre coupled ocean-atmosphere GCM: Model description, spinup and validation. Climate Dynamics 13 103-134.

Jones, C.A. and J. Kiniry, 1986: CERES-Maize: A Simulation Model of Maize Growth and Development. Texas A&M University Press, College Station, TX, 194pp.
 

Jones, J., K. Boote, S. Jagtap, G. Hoogenboom, and G. Wilkerson, 1988: SOYGRO v5.41 Soybean Crop Growth Simulation Model. User's Guide. Florida Agricultural Experiment Station Journal No. 8304, IFAS, University of Florida, Gainesville, FL.
 

Joos, F., M. Bruno, R. Fink, U. Siegenthaler, T. Stocker, T. Le Queue, and J.L. Sarmiento, 1996: An efficient and accurate representation of complex oceanic and biospheric models of anthropogenic carbon uptake. Tellus 48B:397-417.
 

Karl, T.R., D.R. Easterling, R.W. Knight, and P.Y. Hughes, 1994: U.S. national and regional temperature anomalies. pp. 686-736. In Boden, T.A., D.P. Kaiser, R.J. Sepanski and F.W. Stoss (eds.), Trends '93: A Compendium of Data on Global Change. ORNL/CDIAC-65. Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory, Oak Ridge, TN, USA.
 

Kaiser, H. M., Riha, S. J., Wilks, D. S., Rosister, D. G., and R. Sampath, 1993:

A farm-level analysis of economic and agronomic impacts of gradual warming. American Journal of Agricultural Economics, 75(2), 387-398.
 

Kittel, T.G.F., J.A. Royle, C. Daly, N.A. Rosenbloom, W.P. Gibson, H.H. Fisher, D.S. Schimel, L.M. Berliner, and VEMAP2 Participants. 1997. A gridded historical (1895-1993) bioclimate dataset for the conterminous United States. Pages 219-222, in: Proceedings of the 10th Conference on Applied Climatology, 20-24 October 1997, Reno, NV. American Meteorological

Society, Boston.
 

Lakshminarayan, P. G., Bouzaher, A., and J. F. Shogren, 1996: Atrazine and Water Quality: An Evaluation of Alternative Policy Options. J. Env. Management, 48(2), 111-126.
 

Lewandrowski, J. and D. Schimmelpfennig, 1999: Economic Implications of Climate Change for U. S. Agriculture: Assessing Recent Evidence. Land Economics, 75 (1), 39-57.
 

Lotterman, E., 1999: Is Agriculture in Crisis? Working research paper in fedgazette vol. 11, no1. Woodrow Federal Reserve Bank of Minneapolis, Minneapolis, MN.
 

McMullen, M., Jones, R., and D. Gallenberg, 1997: Scab of wheat and barley: a re-emerging disease of devastating impact. Plant Disease,81(12), 1340-1348.
 

Mearns, L. O., C. Rosenzweig, and R. Goldberg, 1997: Mean and variance change in climate scenarios: Methods, agricultural applications, and measures of uncertainty. Climate Change, 35, 367-396.
 

Mendelsohn, R., Nordhaus, W. D., and D. Shaw, 1996: Climate impacts on aggregate farm value: accounting for adaptation. Agric. For. Meteorol., 80,55-66.
 

Mortsch, L. D., and F. H. Quinn, 1996: Climate change scenarios for Great Lakes Basin ecosystem studies. Limnol. Oceanogr., 41(5), 903-911.
 

Olson, K., 1999: Mixed News from 1998 Farm Records. Agricultural Economist Newsletter, No. 696, University of Minnesota Extension Service, St. Paul, MN.
 

Parry, M.L. and T.R. Carter, 1985: The Effect of Climatic Variations on Agricultural Risk. Climate Change, 7, 95-110.
 

Pickering, N.B., J.W. Jones, and K.J. Boote, 1995. Adapting SOYGRO v5.42 for Predicting under Climate Change Conditions. In Climate Change and Agriculture: Analysis of Potential International Impacts. ASA Special pub. 59, Am. Soc. Agron., Madison, WI.
 

Ritchie, J. T., Baer, B. D., and T. Y. Choo, 1989: Effect of global change on agriculture in the Great Lakes region. In Potential Effects of Global Change on the United States, J. B. Smith and D. A. Tirpak (eds.), USEPA, Washington, DC.
 

Rogers, H.H., G.E. Bingham, J.D. Cure, J.M. smith and K.A. Surano. 1983. Responses of selected plant species to elevated carbon dioxide in the field. J. Environ. Qual. 12:569-574.
 

Rosenzweig, C. and D. Hillel, 1998: Climate Change and the Global Harvest: Potential Impacts of the Greenhouse Effect on Agriculture. Oxford University Press, New York, 324 pp.
 

Rosenzweig, C., B Curry, J. T. Ritchie, J. W. Jones, T.-Y. Chou, R. Goldberg, and A. Inglesias, 1994: The effects of potential climate change on simulated grain crops in the United States. In

Implications of Climatic Change for International Agriculture: Crop Modeling Study. U.S. Environmental Protection Agency Report EPA 230-B-94-003. 24pp.
 

Rotz, C.A., J.R. Black, D.R. Mertens and D.R. Buckmaster. 1989. DAFOSYM: A

model of the dairy forage system. J. Production Agric. 2(1):83 91.
 

Schwartz, D, 1999: Management Strategies for delayed planting. Minnesota Crop News, 5(7), 45.
 

Smit, B., M. Brklacich, R.B. Stewart, R. McBride, M. Brown, and D. Bond, 1990: Sensitivity of Crop Yields and Land Resource Potential to Climatic Change in Ontario, Canada. Ag. For. Met. 2: 153-173.
 

Thompson, L.M., 1986: Climatic change, weather variability, and corn production. Agron. J. 78:649-653.
 

UCAR, 1999: VEMAP2 Product Description, available online at: http://www.cgd.ucar.edu/naco/summary.html. University Center for Atmospheric Research, Boulder, CO.
 

USDA/NRCS/NSSC, 1999: National Soil Survey Laboratory Online Database Access, http://vmhost.cdp.state.ne.us/~nslsoil/soil.html. United States Dept. of Agriculture, Natural Resources Conservation Center, National Soil Survey Center, Lincoln, NE.
 

USDA/NASS, 1997a: USDA Economics and Statistics System data sets: Crops, available online at: http://usda.mannlib.cornell.edu/usda/usda.html. U.S. Dept. of Agriculture, National Agricultural Statistics Service, Washington, DC.
 

USDA/NASS, 1997b. 1997 Census of Agriculture. Available online at: http://www.usda.gov/nass/. U.S. Dept. of Agriculture, National Agricultural Statistics Service, Washington, DC.
 

Wigley, T.M.L., 1999: The Science of Climate Change: Global and U.S. Perspectives. Pew Center on Global Climate Change Special Report. The Pew Trust, New York, NY. 48p.