Speaker
Description
1. Introduction
The increase in both the frequency and severity of drought occurrences poses a significant threat to the stability of crop yields and the quality of grains. Accurate yield prediction is crucial to mitigate the impact of climate change and optimize yield production. Early yield prediction during the growing season is critical for accurately informed decision-making regarding resource allocation, risk management, and policy-making (Bailey-Serres et al., 2019).
One promising approach is integrating multispectral data with predictive models, as it can provide valuable insights into environment x management x genotype interactions and aid in developing crop adaptation strategies (Maes & Steppe, 2019). This study aims to develop predictive yield models for five spring barley genotypes, i.e., Morex, Golden Promise, BCC1589, HOR7985, and RGT Planet, under different irrigation treatments and sowing patterns. Using vegetation indices derived from multispectral data, we seek to capture the optimal phenological stage to predict yield and identify phenotypical traits beneficial for drought tolerance.
2. Material and Methods
The field experiment was conducted in 2021 and 2022 at the experimental fields of the Julius Kühn Institute in Berlin-Dahlem. Three irrigation treatments were applied: rainfed, supplementary irrigation (applied when plant available water capacity (PAWC) < 30% and refilled to 70%), and non-limiting irrigation (applied when PAWC < 50% and refilled to 100%). Two sowing patterns were implemented with a sowing density of 290 seeds/m²: equidistant sowing in a triangular pattern with a 6 cm distance between single plants and conventional row drilling at an 11 cm row distance. Throughout the growing season, multispectral data was collected using the Micasense Altum camera mounted on a DJI M300 copter. The copter was operated at a flight height of 20 m with a speed of 2.1 m/s. The side overlap ratio of the images was 70%, and the front overlap ratio was set at 80%. From the multispectral data, we retrieved relevant vegetation indices, including the normalized difference vegetation index (NDVI), optimized soil-adjusted vegetation index (OSAVI), and crop water stress index (CWSI) using an image processing pipeline including Pix4D mapper, QGIS, and Rstudio. We used linear mixed models, including the different vegetation indices at different phenological stages as covariates under the combination of the different experimental treatments. The selection of the best-performing model for each genotype under separate treatments was based on the Akaike Information Criterion (AIC).
3. Results and discussion
An early yield prediction at the end of tillering was particularly possible for the Golden Promise genotype using SAVI and OSAVI, indicative of plant vigor, and for RGT Planet using NDRE (representative of chlorophyll content) and CWSI. There was no significant difference between predictive models for the two sowing patterns. Comparing predictive models under different irrigation treatments, we have concluded that the best predictive model under the rainfed treatment used NDRE and OSAVI as covariates at the end of anthesis. The selected models were cross-validated using observed data, which provided satisfying results. Through plotting partial regression of the different covariates contained in the selected predictive models of Golden Promise, BCC 1589, and Morex, early drought stress at the end of tillering was found to be associated with reduced drought stress response at flowering, which was reflected in increased yields after early drought stress in these three genotypes. Our findings highlight the potential efficacy of utilizing UAV-derived multispectral data and vegetation indices to predict yield in barley genotypes under varying drought situations.
4. References
Bailey-Serres, J., Parker, J. E., Ainsworth, E. A., Oldroyd, G. E. D., & Schroeder, J. I. (2019). Genetic strategies for improving crop yields. Nature, 575(7781), 109–118. https://doi.org/10.1038/s41586-019-1679-0
Maes, W. H., & Steppe, K. (2019). Perspectives for Remote Sensing with Unmanned Aerial Vehicles in Precision Agriculture. Trends in Plant Science, 24(2), 152–164. https://doi.org/10.1016/J.TPLANTS.2018.11.007
| Keywords | Spring barley; drought stress; sowing pattern; climate change; remote sensing |
|---|
