SOYBEAN IRRIGATION MANAGEMENT AND YIELD ESTIMATION BASED ON UAV IN THE MID-SOUTH

Optimizing irrigation management is critical for maximizing soybean yield, particularly in regions with variable rainfall and diverse soil types. This study aimed to (1) quantify the effects of irrigation timing (R1, R3, and R5 growth stages), rates (0.5, 1, and 1.5 in/wk), and tools (water balance method and soil moisture sensors) on soybean yield in silt loam soil; (2) evaluate how these effects vary across different soil types; (3) assess the utility of one-year rainfed yield data for field zoning; and (4) develop a predictive model using drone-derived vegetative index by comparing in-season and post-season models. Field experiments were conducted in Jackson, TN, with various irrigation treatments. Drone-based multispectral data, soil electrical conductivity (EC), ground-penetrating radar (GPR), and weather data were collected to capture dynamic crop responses. Results showed that (1) initiating irrigation at the R3 stage with a rate of 1.5 in/wk maximized yield whith less water in silt loam soil in 4 out of 5 experimental years; (2) soil type significantly influenced irrigation effectiveness, with sandy soil requiring earlier irrigation initiation at R1 with 1.5 in/wk to achieve maximum yield; (3) one-year rainfed yield data proved effective for field zoning, comparable to EC and GPR, in identifying areas with high yield variability; (4) the incorporation of soil features consistently enhanced yield prediction model performance across all experimental configurations. In conclusion, this study highlights the importance of tailoring irrigation strategies to soil type and growth stage, demonstrates the effectiveness of one-year rainfed yield data for zoning, and underscores the potential of drone-based tools for optimizing soybean production under varying environmental conditions.