Development of a new UAV-thermal imaging based model for estimating pecan evapotranspiration

Pecans are a specialty crop in New Mexico’s Lower Rio Grande Valley (LRGV), a region that produces around 30% of pecans in USA. Pecans are also a major water consumer, requiring 1200–1300 mm depth for maximum yield in this region. The combination of prolonged drought and increasing competition for water among various water consumers has created an urgency for the efficient use of scarce water resources in the LRGV. More efficient water management through the real-time irrigation scheduling is one method to promote reduced water application in agriculture. This study was conducted to calibrate and validate a new modified model for estimating the pecan actual evapotranspiration (ETa) based on canopy temperature using thermal images taken from an Unmanned Aerial Vehicle (UAV) during three growing seasons in a drip irrigated pecan orchard. A capacity to estimate the relation between ETa and canopy temperature provides an important information to guide water management choices. The Simplified Surface Energy Balance (SSEBop) model was modified and used for calibration and validation. Applied irrigation water based on ETa was used to calibrate and validate the proposed modified model. The scaling factor of K in the SSEBop model was calculated as 0.75 through the calibration process. Findings showed a good agreement between estimated pecan ETa using modified SSEBop model and applied water based on ETa during calibration (R2 = 0.72, RMSE = 0.6 mm/d, MAE = 0.48 mm/d) and validation period (R2 = 0.90, RMSE = 0.24 mm/d, MAE = 0.22 mm/d). Also, findings confirmed the utility of modified model for estimating monthly pecan ETa (RMSE = 8.87 mm/month, MAE = 6.55 mm/month). The proposed modified model provides pecan farmers with a simple real-time irrigation scheduling tool where they can better practice precision irrigation. Although the modified model was calibrated and validated for irrigation scheduling in the LRGV, it has potential to see application for other locations with different crops using similar calibration approach.