Assessment of the impact of accurate green area index, water regime and harvest index on site-specific wheat yield estimation

Abstract

In recent decades, extensive research has focused on estimating winter wheat yields and developing methods for collecting the necessary field data. However, it may be advantageous to first evaluate which data types and levels of model complexity are truly essential. This study examined the explanatory power of water regime modeling, the green area index (GAI), site-specific soil texture, and harvest index (HI) data in estimating winter wheat yields throughout the growing season. Data collected over 13 years from measurements of GAI and soil moisture in winter wheat plot trials in northern Germany were integrated into a plant growth and soil water budget model (HUME). The soil moisture data were used to estimate site-specific soil textures. Monitoring GAI was identified as the key factor for explaining yield. The increased modeling effort of integrating GAI into HUME was found to be justified. The modelled transpiration provided a more accurate explanation of the yield at the end of the season (R² = 0.86) compared to radiation uptake (R² = 0.73). Additionally, predictors based on transpiration were less dependent on GAI senescence and HI data. By the time of the third N fertilization, the most effective predictor tested − transpiration standardized daily by the saturation deficit of the air – allowed for the prediction of a substantial portion of site- and year-specific grain yield variation (R² = 0.64). This suggests that it could serve as a valuable starting point for developing N management strategies. Site-specific soil textures only marginally improved yield estimation, with an increase in R² of less than 0.05. The findings indicate that interpreting GAI through soil water modeling can uncover water limitations that affect winter wheat yield, even in temperate climates. This underscored the importance of ongoing research to generate comprehensive, site-specific GAI data throughout the growing season. Alongside the results and methodological approach discussed here, such data could potentially enable nitrogen fertilization management driven by yield predictions in the future, thereby improving N efficiency in wheat cultivation.