Submitted to: Field Crops Research
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/16/2007
Publication Date: 8/3/2007
Citation: Kemanian, A.R., Stockle, C.O., Huggins, D.R., Viega, L.M. 2007. A simple method to estimate harvest index in grain crops. Field Crops Research 103:208-216. Interpretive Summary: Background: Developing crop simulation models is important for predicting crop yields and for understanding the factors that influence yield. Using relatively simple methods and relationships in crop models is preferable if adequate prediction can be achieved. In this research we present a relatively simple method to estimate the harvest index of grain crops. The harvest index is the proportion of total above-ground crop growth that is actually harvested as grain and represents a fundamentally important factor in crop simulation models. Description: We proposed that there is a simple relationship between the harvest index and the amount of crop growth that occurs after the crops flower. We developed and tested this simple relationship for wheat, barley and sorghum and found that the simple models were able to reasonably represent the harvest index. A major advantage of defining these simple harvest index relationships is that they can be easily determined from crop growth and yield measurements. Impact: This research contributes to the development of crop models that are relatively simple and require less information to achieve reasonable prediction of crop yields. The estimates of harvest index will be incorporated into crop models such as CropSyst to enhance simulation of yield.
Technical Abstract: Several methods have been proposed to simulate yield in crop simulation models. In this work we present a simple method to estimate harvest index (HI) of grain crops based on fractional post-anthesis growth (fG = fraction of growth that occurred post-anthesis). We propose that there is a linear or curvilinear relationship between HI and fG. The linear model has two parameters, the intercept (HIo) and the slope (s). The curvilinear model was assumed to be monotonic: HI = HIx – (HIx – HIo) • exp(–k • fG); where HIx is the asymptote, HIo is the intercept and k is a crop or cultivar dependent constant. Both the linear and curvilinear models were fitted to actual data for barley and wheat (Pullman, WA) and sorghum (Australia). For barley, the linear model appropriately represented the response of HI to fG, while for wheat the curvilinear was better than the linear model. For sorghum, both linear and linear-plateau models fitted data reasonably well. It is shown that the models work reasonably well in crops source-limited or source-sink co-limited, but in sink-limited conditions the magnitude of the limitation needs to be characterized to compute HI. Cases of sink limitation are discussed for the three crops. A major advantage of these models is that the parameters are readily calibrated from yield data and biomass measurements at anthesis and harvest.