Module 2

Climatic Change: Impacts on Grain Yields,
Price Variability, and Risk Management

S. Elwynn Taylor, Iowa State University
Jeffrey A. Andresen, Michigan State University

Topics

• Introduction
• Yield Trends
• Regional Climate Trends
• Variability and Weather Trends
• Anticipating Yearly Outcomes
• Table 1. Frequency (number of years) of Trend Corn Yields vs. Southern Oscillation Index Values by State, 1900-1994

• Use of Weather Forecasts
• Sources of International and Domestic Crop and Weather Information
• Future Climatic Changes
• References
• End of Module

Introduction

Weather remains one of the most important uncontrollable variables involved in agricultural production systems. A recent rash of extreme weather conditions across the Midwest region included abnormally cool-growing-season temperatures in 1992, unprecedented heavy growing- season rainfall and flooding in 1993, and record-breaking winter cold, snowfall, and subsequent spring flooding in 1997. These extremes and associated impacts on crop performance, whether positive or negative, are generally considered a characteristic of the climate of a particular area, defined as the average of day-to-day weather conditions over an extended period of time. However, while commonly thought of as a stable or static system, climate is in reality a changing, dynamic process that has shifted significantly in the past (due to factors such as changes in earth-sun geometry and solar irradiance, continental drift and uplift, and changes in atmospheric composition and aerosol content) and will likely change again in the future.

This segment of the Managing Risks and Profits series focuses on documented past and projected future changes of climate and crop performance in the Midwestern U.S., and possible implications for regional agriculture. We also include a review of the creation and use of weather forecasts and a list of pertinent crop/weather information available to farmers, much of it free across the Internet. Because there is no sure way to totally control the weather (other than the impractical solution of moving all agricultural activities indoors!), perhaps the best strategy for coping with the uncertainty of weather is a good understanding of weather/climate systems. This includes a good knowledge of the climatology of an area of interest, the potential short- and long-term agricultural impacts of weather on agriculture, and a list of potential managerial strategies. The management strategies can used to hedge risk involved with future unforeseen weather events.

Yield Trends (or go to Topics )

Between 1900 and 1940, corn yields in the U.S. Plains (Figure 1) showed little indication of trend. Year-to-year variability was observed to be related to temperature and precipitation conditions. Dramatic increases in yield were realized between 1940 and 1970, and a diminished rate of yield improvement has been observed since 1972. Carlson, Todey and Taylor (1996) analyzed the yield trend as a logistic, "S-shaped," trend. They considered a deviation of 10% above or 10% below the "trend line" to be significant. The rapid increase in yield during the 1960-1970 period is typically attributed to technology, favorable weather, and the development of hybrids that respond well to high levels of production management. It may be argued that the technological potential yield has increased linearly since 1930; a straight line almost perfectly connects the "best" years from 1930 to the present, and it is assumed that all yields below the line resulted from adverse environmental or sub-optimal conditions. The flat period from 1900-1930 may reflect a period of little technological improvement or a period of declining weather that balanced technological advancements.

Regional Climate Trends (or go to Topics )

Karl, Easterling, Knight, and Hughes (1994) showed diminishing annual precipitation from 1900 through 1930 in the Northern Plains and the Great Lakes regions of the United States (U.S.). Precipitation trends were mixed from 1930-1960. Increasing annual precipitation has been noted since 1960. Annual mean temperature in the Northern Plains region of the U.S. was increasing during the 1900-1940 period, decreased from 1940-1972 and shows a warming trend since 1972. Near the Great Lakes, the temperature trend also shows warming, but of a lesser magnitude, possibly because of the moderating effects of water.

Variability and Weather Trends (or go to Topics )

The 1940-1972 interlude of decreasing temperature was a period of relatively consistent weather in the central U.S. Year-to-year yield variability was minimal during the period. Yield variability since 1972 has been notable and is similar to the variability observed previous to 1940 when analyzed as a percentage variation. The "leveling" of the yield trend in recent years in much of the Midwest appears to be caused by the impact of unfavorable years. That is, "bad" years hurt more than favorable years "benefit." The leveling is more pronounced in the western Corn Belt.

Anticipating Yearly Outcomes (or go to Topics )

Carlson, Todey, and Taylor (1996) showed that the SOI (Southern Oscillation Index) has a significant correlation with crop yield in the major corn-producing states. The SOI is a meteorological index associated with periodic major shifts of normal weather patterns and surface ocean currents across the equatorial Pacific region. Because of the potential interaction of the Pacific region climatic anomalies with the jet streams that circle the middle and upper latitudes of both hemispheres, these shifts may affect both summer temperatures and precipitation in the U.S. Corn Belt. The apparent impact of the SOI varies in magnitude across the Corn Belt. If the SOI is strongly negative the probability of having an adverse year is reduced and the possibility of a favorable season enhanced. A positive SOI is associated with increased risk of adversity. The dependence of trend corn yields on SOI from 5 Midwestern states, 1900-1994, is illustrated in Table 1.

Table 1. Frequency (number of years) of Trend Corn Yields vs. Southern Oscillation Index Values by State, 1900-1994* (from Carlson et al., 1996). (or go to Topics )

Table footnotes:

* Due to missing data, the years 1906-1908, 1914, 1915, 1921, 1927, 1931, and 1932 were not included in the analysis.

¹ The yield codes are broken down into categories of 10% less than the state trend yield in a given year, 10% greater than trend yield, and all yields in between 10% above and 10% below the trend yield.

² The SOI code represents the standardized SOI value during the summer months.

The probability for yield to exceed the plus 10% of trend is enhanced when the SOI is strongly negative. Note that Illinois yield did so in 7 of 22 years.

Use of Weather Forecasts (or go to Topics )

Given the strong dependence of so many agricultural activities on weather, forecasts of future weather conditions have the potential to be a key factor in managerial decision-making. Weather forecasts are frequently categorized into one of three major time frames: short term, which includes forecasts up through 48-60 hours into the future; medium range, which includes outlooks from 3-14 days in advance; and long lead forecasts, which include all outlooks longer than 2 month into the future. The forecast categories differ both in how they are prepared and what variables are being forecast. For short-range and medium-range forecasts, meteorologists use output from highly quantitative computerized process models that describe the motions and physical processes of the atmosphere. The models are started at a given time with representative real data from the surface and upper levels of the atmosphere, then allowed to run into the future following the given initial conditions. Even though these models are run on powerful supercomputers and could theoretically be run out days, weeks, and months into the future, there is one very important limitation. Since the atmosphere is a chaotic system (i.e., movements are characterized by occasional sudden and unpredictable changes), any small error in the initial condition the model is given at the start time of the model run can lead to increasing errors as the model runs into the future, eventually leading to a total loss of predictability. Given this limitation, the theoretical forecast limit of a computer process model is thought to be on the order of 10-14 days into the future. Thus, in general, the direct use of computer process models is limited to short and medium-range forecasts.

Beginning with medium-range outlooks and continuing into the long-range time frame, forecasters generally use empirical, statistical tools based on the behavior of the atmosphere in the past to generate forecasts. Medium-range outlooks are usually a "hybrid" forecast made up of a jet stream/upper air orientation taken from a computer process model and a statistical forecast of mean temperature and precipitation based on that particular jet stream pattern. The most common public medium-range forecast, the 6-10 day outlook, is issued by NOAA’s Climate Prediction Center (CPC) every Monday, Wednesday, and Friday afternoon at approximately 3 p.m.EST/EDT. The jet stream/upper air orientation forecast portion of this outlook is heavily based on output from two computer models: the National Weather Service (NWS) Medium range forecast (MRF) model, and the European Centre for Medium range Weather Forecasting (ECMWF) model. A third model from the United Kingdom Meteorological Office (UKMO) is also sometimes available for inclusion into the outlook. These models are often mentioned in news from the commodities markets, as medium-range forecasts are frequently used by meteorologists and other analysts for detecting major changes or shifts of weather over a region such as the Corn Belt. There are many sources of short and medium range weather information available on the worldwide web. An excellent site for those unfamiliar with meteorology and forecasting is the NOAA National Weather Service Interactive Weather Information Network (IWIN) at; http://iwin.nws.noaa.gov. Among the sites specializing in providing high quality satellite imagery is the University of Wisconsin's Space Science Engineering Center at http://www.ssec.wisc.edu/data/. For those seeking more detailed, in-depth information, including individual computer forecast output, try sites at Purdue University (http://wxp.atms.purdue.edu/) and Ohio State University (http://twister.sbs.ohio-state.edu/). Both of these sites include medium range forecast output from MRF and ECMWF models.

Until the mid-1990s, NWS long-range outlooks covered only 1-month and 3-month forecast periods. NWS long-lead outlooks are now available in 3-month increments out to 12 months into the future. As mentioned above, long-range outlooks of 1 month or more are generally based on statistical methods, including forecasts based on land/ocean surface temperature and precipitation anomalies during the past 12 months and forecasts based on the evolution of past upper air flow patterns similar to those at the forecast time (analog method). An important exception is the National Center for Environmental Prediction (NCEP) Coupled Ocean/Atmosphere Model, which is a process-based computer simulation of Pacific Ocean sea surface temperatures, which in turn are related to the El Nino Southern Oscillation, ENSO for which the Southern Oscillation Index (SOI) monitors as an index. Long-lead outlooks and other international climatic information are available from the CPC across the Internet at http://nic.fb4.noaa.gov. Another site with excellent information on the status of ENSO, including a visual time lapse of see surface temperatures across the Pacific Ocean is the NOAA Climate Diagnostics Center (http://www.cdc.noaa.gov/). Other long-lead outlooks (global precipitation) are available on the Internet from the Queensland Dept. of Natural Resources/Dept. of Primary Industries (Australia) at http://www.dnr.qld.gov.au/longpdk/. Finally, long-range outlooks may also be obtained from many different private meteorological firms.

The utility of weather forecasts for decision-making on the farm generally boils down to two major factors: 1) the historical skill of the forecast (i.e., how consistently is it correct and by how much?); and 2) how much risk is a grower willing to take? Public NWS or private forecasts of conditions during the upcoming 1-2 days are frequently correct more than 90% of the time, with declining skill as the length of the forecast increases. Short-term forecasts can be made even more effective when combined with real-time weather information, such as weather radar data, significantly reducing the odds of a grower making a weather-related mistake (e.g., a grower who decides to cut hay on the basis of a forecast of less than 30% chance of measurable precipitation during the upcoming 72 hours holds off his decision at the last minute due to the development of an isolated thundershower upwind from his farm that he has detected on a TV weather radar display). In contrast, claims of a long-lead outlook, 3-12 months in the future, being "useful" for agricultural applications, depend greatly on whether a manager can afford to make a decision based on forecast odds being 55-60% correct. Before using forecasts operationally in your management scheme, we advise obtaining some type of record indicative of past forecast performance (including all forecasts issued in a given time period, not just the successful ones!). Be cautious if this type of information is not available.

Some generalizations of forecast skill and accuracy: 1) the expected skill of a forecast generally decreases as one goes out in time; 2) forecast skill depends on the variable being forecast and the season (e.g., precipitation frequency and amounts are generally more difficult to forecast than max./min. temperature); and 3) both public NWS and private meteorologists generally start with the same information. The difference in the resulting forecasts is in the interpretation of the data/information by the forecaster. Finally, due to improvements of the simulation of the climate system, including the interaction between the earth-atmosphere interface, and increases in computational power, the skill of forecasts is increasing. For example, a recent study of forecasting skill indicated that medium-range forecast models, which in the early 1980s had a useful length (i.e., the forecast is more accurate than the use of climatological statistics alone) of 5.5 days, now have a useful life of over 7 days, an improvement of more than 25% (Kerr, 1996).

Sources of International and Domestic Crop and Weather Information (or go to Topics )

In the fast-paced environment of major U.S. commodity markets, rarely a day goes by without some mention of crop conditions or weather problems in major international production areas. If you’ve ever wondered about agriculture in other parts of the world (and given the international nature of today’s markets, you should have), especially which crops are planted where and when, we offer a couple of pieces of information that you will find valuable and helpful in making sense of international trends.

The first is the Weekly Weather and Crop Bulletin, published weekly by the NOAA/USDA Joint Agricultural Weather Facility (JAWF) in Washington, D.C. This bulletin, which has been published since 1895, gives detailed crop and weather information for the U.S. and all major international crop areas, with occasional articles on major weather events such as heat waves, droughts, etc. The consistent format (e.g., weekly tables of average temperatures and total precipitation, U.S. crop conditions by state and crop) of the bulletin allows readers to keep track of growing conditions and anomalous weather trends year-round. The bulletin is available at JAWF’s Internet web site at ( http://www.usda.gov/oce/waob/jawf/index.html ). Otherwise, one major drawback of the bulletin is that it is available only via first class mail, resulting in information about 1 week old by the time it reaches your doorstep. For information about subscriptions, call (202) 720-7917 or FAX number (202) 720-1455. Specific domestic and international crop supply and demand figures for major production areas are also available over the Internet from the USDA World Outlook Board (one of the parent organizations of JAWF) at (http://www.usda.gov/oce/waob/waob.htm ).

A second publication recommended for reference is Major World Crop Areas and Climatic Profiles Agricultural Handbook no. 664, published by USDA/WAOB/JAWF in 1994. This reference provides a very current framework for assessing weather’s impact on crop production around the world. It includes geographical information on international crop areas and crop calendars, as well as graphical and tabular climatic profiles (for temperature and precipitation) at representative locations. Crops covered include coarse grains, wheat (spring and winter), rice, major oilseeds, sugar, and cotton. Besides general information on the topics above, the 279-page volume contains detailed articles on the El Nino Southern Oscillation phenomenon, the Indian monsoon (which generally governs the success or failure of agriculture on the Indian subcontinent), and the climate and agriculture of the former Soviet Union. A copy of this handbook is available via Internet at the address listed for JAWF above. Paper copies can be ordered over the phone from USDA/ERS/NASS at (800) 999-6779 or by writing (with a $20 check payable to ERS-NASS) to: ERS-NASS, 341 Victory Drive, Herndon, VA 22070.

Future Climatic Changes (or go to Topics )

Improvements in the physical understanding of the climate system during the past few decades have led to concern about levels of atmospheric carbon dioxide and other trace gases, which have steadily increased in concentration since the industrial revolution, and may ultimately result in significant increases in mean global surface temperatures of 2-6^oF by the end of the next century (IPCC, 1995).

These estimates represent the consensus of a large international group of scientists working in this discipline and are largely the result of research with General Circulation Models (GCM), which are complex, physically based models of the earth’s atmospheric/oceanic system (similar to models used for short- and medium-range forecasting, except on a larger scale). For the Midwestern U.S., the GCMs generally suggest increases in both mean temperatures and precipitation, which could lead to either crop yield increases or decreases, depending on location and crop. New research studies that also include the fertilization or enrichment effect of increasing carbon dioxide levels on crop growth suggest that any yield decreases due to unfavorable climatic shifts should be at least partially offset by CO₂ enrichment, especially for C-3 crops such as wheat and soybean (Curry et al., 1995).

There are several very important factors to consider regarding global climate change. First, while current theories strongly suggest some type of global surface temperature increases associated with increasing trace gases, there remains a great deal of uncertainty about the magnitude and timing of any changes. The uncertainty is due to a number of factors, including an incomplete knowledge of a few key feedback mechanisms within the climatic system (such as the amount of clouds in tropical areas of the world) and a lack of computer power needed to run GCMs for longer time periods on finer spatial and temporal scales. Secondly, for the bulk of agricultural activities, it is not long-term changes in the means of individual climatic variables, but changes in variability, such as the frequency of extreme temperatures or rainfall events, that likely pose the greatest potential threat to producers (Parry and Carter, 1985). Research of observed trends of climatic variability in the U.S. since 1895 indicated some increases in extreme events such as heavy 1-day rainfall events (Karl et al, 1996). However, recent studies of results derived from GCM output indicate mixed trends, including some variability decreases (Winkler et al, 1997; IPCC, 1992). Finally, based on past performance, there is a likelihood that new technological innovations in agriculture and related fields in future years will allow farmers to keep pace with most climatic changes, provided that they do not include rapid changes in variability (IPCC, 1995).

References (or go to Topics )

Curry, R.B., J.W. Jones, K.J. Boote, R.M. Heart, L.H. Allen, and N.B. Pickering, 1995. Response of soybean to predicted climate change in the USA. In Climate Change and Agriculture: Analysis of Potential International Impacts, C. Rosenzweig, L.H. Allen, L.A. Harper, S.E. Hollinger, and J.W. Jones, eds. ASA Special Publication No. 59, American Society of Agronomy, 677 South Segoe Rd., Madison, WI 53711-1086.

Intergovernmental Panel on Climate Change, 1995. Climate Change 1995: The Science of Climate Change Contribution of Working Group I to the Second Assessment Report of the Intergovernmental Panel on Climate Change Editors J.J. Houghton, L.G. Meiro Filho, B.A. Callander, N. Harris, A. Kattenberg and K. Maskell. 1996. 584 pages.

Karl, T.R, R.W. Knight, D.R. Easterling, and R.G. Quayle, 1996. Indices of Climatic Change in the United States. Bull. Am. Met. Soc. 77, No. 2: 279-292.

Kerr, R.A., 1996. Budgets stall but forecasts jump forward. Science 273: 1658-1659.

Carlson, R.E., D.P. Todey, and S.E.Taylor, 1996. Midwestern Corn Yield and Weather in Relation to Extremes of the Southern Oscillation. J. Prod. Ag. 9:347-352.

Curry, R.B., J.W. Jones, K.J. Boote, R.M. Heart, L.H. Allen, and N.B. Pickering, 1995. Response of soybean to predicted climate change in the USA. In: Climate Change and Agriculture: Analysis of Potential International Impacts, C. Rosenzweig, L.H. Allen, L.A. Harper, S.E. Hollinger, and J.W. Jones, eds. ASA Special Publication No. 59, American Society of Agronomy, 677 South Segoe Rd., Madison, WI 53711-1086.

Intergovernmental Panel on Climate Change, 1995. Climate Change 1995: The Science of Climate Change Contribution of Working Group I to the Second Assessment Report of the Intergovernmental Panel on Climate Change. Editors J.J. Houghton, L.G. Meiro Filho, B.A. Callander, N. Harris, A. Kattenberg and K. Maskell. 1996. 584 pages

Karl, T.R, R.W. Knight, D.R. Easterling, and R.G. Quayle, 1996. Indices of Climatic Change in the United States. Bull. Am. Met. Soc. 77, No. 2: 279-292.

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 T.A. Boden, 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.

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

Winkler, J.A., J.P. Palutikof, J.A. Andresen, and C.M. Goodess, 1997. The Simulation of Daily Temperature Series from GCM Output. Part II: Sensitivity Analysis of an Empirical Transfer Function Methodology. Accepted for publication in Journal of Climate.

End of Module (or go to Topics )

Go to Module 3 | Introduction | MRP Introduction
Universities and Agribusinesses |Table of Contents | or Go to Modules :

1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | Questionnaire
Supplementary Material