Evaluation of Federal Crop Insurance Corporation methods to estimate soft winter wheat grain yield in the Eastern United States

Abstract

Soft winter wheat (SWW) (Triticum aestivum L.) is vulnerable to environmental stressors throughout winter and early spring. To assess yield potential of SWW, crop insurance adjustors estimate grain yield by multiplying the number of stems ft⁻² by a yield factor of 0.50. However, crop insurance adjustors believe the yield factor of 0.50 is too low. A 3-year experiment was conducted in Michigan, Ohio, and Kentucky to compare predicted SWW yield to harvested yield. The existing yield factor underestimated SWW yield in 243 out of 246 comparisons. Average predicted yield was 40 bu acre⁻¹ (range of 6 to 122 bu acre⁻¹) while actual yield averaged 93 bu acre⁻¹ (range of 51 to 124 bu acre⁻¹). Due to the discrepancy in predicted and actual yield, data from a planting date and seeding rate experiment conducted at four site-years in Ohio was used to establish a new yield factor based on the number of stems ft⁻² and fractional green canopy cover (FGCC) measured with the Canopeo app at Feekes 5 growth stage. The new methods were applied to the original multi-state dataset. Using a logarithmic function based on the number of stems ft⁻², 50% of the predicted yield values were within −8 to 18 bu acre⁻¹ of the actual yield values. A logarithmic function based on FGCC resulted in 50% of the predicted yield values within 3 to 18 bu acre⁻¹ of the actual yield values. Overall, our results showed that new models performed better than the current method used by crop insurance adjustors.