Between Dec 19, 2024 and Jan 2, 2025, datasets can be submitted to DRUM but will not be processed until after the break. Staff will not be available to answer email during this period, and will not be able to provide DOIs until after Jan 2. If you are in need of a DOI during this period, consider Dryad or OpenICPSR. Submission responses to the UDC may also be delayed during this time.

Browsing by Author "da Silva, Maykon"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    Managing Soybean Iron Deficiency Chlorosis with Agronomics, Economics, and Remote Sensing
    (2023-03) da Silva, Maykon
    Iron deficiency chlorosis (IDC) is a major yield-limiting factor for soybean [Glycine max (L.)] grown on high pH calcareous soils. In soybean, IDC is caused by a lack of soluble (Fe2+) iron to the plants, with symptoms characterized by interveinal chlorosis of the leaves and stunting of the growth. In order to overcome the problem and ensure profitability, effective and economical solutions are needed. This study examined the effectiveness and interactive effects between three of the most often used management strategies for soybean IDC across a range of IDC stress levels: varietal tolerance [highly tolerant (HT) and moderately tolerant (MT)], iron chelate rates (0, 2.24 and 4.48 kg Fe- EDDHA ha-1) and seeding densities (309,000 and 433,000 seeds ha-1). Given that there is a trade-off in cost and yield relative to the adoption of each of these management strategies, profitability and economic risk analysis were performed to evaluate the impact of variety selection, seeding densities, and iron chelate rates on economic returns. Overall, our findings determined that planting a HT variety, applying Fe-EDDHA in-furrow at planting, and increasing the seeding rate were all effective at minimizing yield losses due to IDC. For every one-point increase in IDC severity measured by our environmental index (EI), the HT variety yielded 0.21 Mg ha-1 more, on average, than the MT variety. A similar trend was observed with iron chelate application. As IDC became very severe (EI of 4, for example), yield improvements averaging 1.4 and 1.7 Mg ha-1 were observed with soil applications of 2.24 and 4.48 kg Fe-EDDHA ha-1, respectively, compared to the untreated plots. Although a smaller effect was observed with increased seeding rates (150 kg ha-1 yield increase with 433,000 seeds ha-1 relative to 309,000 seeds ha-1), this effect was consistent across environments and treatment combinations. For profitability and economic risk analysis, environments were classified into low-moderate and severe IDC based on their EI. Our economic risk analysis showed that the best option in terms of risk to reward for low-moderate IDC conditions was the HT variety with 2.24 kg Fe-EDDHA ha-1 at 309,000 seeds ha-1. Conversely, when IDC was severe, the single best alternative for IDC management considering the amount of risk per unit of reward was the HT variety with 4.48 kg Fe-EDDHA ha-1 planted at 433,000 seeds ha-1, which provided a sharpe ratio only 0.02 points lower than the optimum portfolio. Because IDC most frequently occurs in complex and discontinuous patterns, creating low, moderate, and severe areas interspaced within a field, growers very often do not know the extent of these areas and amount of yield loss being caused by this abiotic stress. Thus, this study also investigated the utility of UAV-based vegetation indices for estimating grain yield of soybean grown under IDC stress conditions as a tool to aid growers, researchers, industry and policy makers with crop management, market planning, market research, and policy writing. Results from this study showed that in-field assessment of IDC symptoms using vegetation indices (VI’s) generated from UAV imagery is more precise, objective, and efficient than ground-based methods such as visual chlorosis scores and ground-based canopy sensing tools such as Crop CircleTM. In addition, we found that NDVI provides the highest predictive power for yield estimation at R1, while NDRE provided the most accurate yield estimations at R5.5. These two VI’s were then used as explanatory variables for yield prediction model development using linear regression. Performance analysis showed that NDRE at R5.5 was more accurate in predicting yield than NDVI at R1, but the overall performance of both models could have been better when validated with testing data. As such, an alternative approach was proposed to improve yield forecasting accuracy. A path analysis was performed to identify the cause-and-effect relationship between VI’s and grain yield, which indicated NDRE at R5.5, OSAVI at R5.5, and NDVI at R1 as most relevant for yield estimation. These VI’s were used as predictors in a regression tree algorithm, which was able to predict soybean yield with a relatively low RMSE (0.53 Mg ha-1) and MAE (0.45 Mg ha-1), while explaining more than 93% of the yield variability. Results from this study can help soybean growers increase productivity, improve economic returns while controlling economic risk, and provide an advantage when it comes to agricultural decision making.