Optimizing the Canopy Resistance Models to Calculate Evapotranspiration from Summer Maize Fields

Abstract

【Objective】 Canopy resistance (rc) is an important parameter to calculate evapotranspiration (ET) from natural and planted fields, and its value varies between plants. Different models have been proposed to calculate rc and the objective of this paper is to propose a method to optimize these models. 【Method】 We took summer maize as the model plant, and the optimized canopy resistance model was used to estimate the evapotranspiration from the maize field using both meteorological data and eddy covariance data measured from a weather station at Huailai in China. Data and results measured and calculated in 2013 were used to calibrate the coefficients in the Jarvis (JA) model and the coupled surface resistance model (CO), using the ant colony optimization (ACO) and least square method (LSM) respectively; the backpropagation neural network model was used to analyze the sensitivity of rc to different factors. The ET calculated by the optimized model was tested against the ET measured from fields using an eddy covariance system in 2014. 【Result】 The meteorological factors to which rc was sensitive were ranked in the order radiation >leave area index>humidity >temperature > wind speed. The CO model optimized by the ACO gave the best calibrated rc model with R2=0.89, RMSE=410.90 s/m and d=0.88, and the ET estimated by it was also most accurate with R2=0.72, RMSE=1.07 mm and d=0.75. 【Conclusion】 Radiation and leave area index are the factors affecting rc the most. The CO model optimized by ACO was most accurate for calibrating the rc model and calculating ET of the maize field.

Keywords

• evapotranspiration
• canopy resistance model
• maize field
• sensitivity
• parameter optimization