Comparing the value of Southern Oscillation Index-based climate forecast methods for Canadian and US wheat producers

Southern Oscillation Index (SOI) based forecasting methods are compared to determine which method is more valuable to Canadian and US wheat producers. Using decision theory approach to valuing information, the more commonly used three-phase method of El Niño, La Niña, and other is compared to a five-phase system. Because of differences in growing season and yearly SOI classification schemes, two different three-phase methods are used. The five-phase system is based on the level and rate of change of the SOI over a 2 month period. Phases are consistently negative, consistently positive, rapidly falling, rapidly rising, and near zero. As expected, results vary by the method used. Winter wheat producers in Illinois place no value on either of the SOI-based forecasting systems. Producers at seven of the 13 sites prefer the five-phase method over either of the three-phase method (spring wheat producers in Manitoba, Alberta, North Dakota and South Dakota, along with winter wheat producers in Oklahoma, Texas, and Washington). The value of the five-phase approach is up to 70 times more valuable than the three-phase approach. Producers growing spring wheat in Saskatchewan and Montana, along with winter wheat producers in Ohio and Kansas value the three-phase approach more than the five-phase. In this case, the value of the three-phase system is up to two times more valuable than the five-phase system. Depending on expected price and region, the values of the SOI-based forecasts range from 0 to 22% of the value of perfect forecasts. In both absolute and percentage of perfect forecasts, producers in Oklahoma, Texas, Manitoba, Saskatchewan, and South Dakota value either system more than producers in the remaining regions. Economic value and distributional aspects of the value of climate forecasts have implications for producers, policy makers, and meteorologists. Finally, the results clearly suggest all producers will not prefer one forecast type. Forecasts need to be tailored to specific regions. ©2000 Elsevier Science B.V. All rights reserved.