Improving Jarvis Canopy Resistance Model by Modifying the Effective Leaf Area Index in a Venlo-Type Greenhouse

Abstract

Canopy resistance (r_c) is a vital parameter for further estimating crop evapotranspiration (ET_c), and Jarvis model is one of the most widely used models to estimate r_c with parameterizations based on the environmental factors and leaf area index (LAI). However, previous researches on the Jarvis model mainly focused on optimising environmental parameters, which could not explain the great fitting disparity for the r_c estimations of distinct crop species and the uncertain model performances in different growing stages. Therefore, we modified the effective leaf area index (LAI_e) in Jarvis model using a multi-layer method in which we measured stomatal resistance (r_s) and LAI at different layers (0–50, 50–100, 100–150 and 150–200 cm of plant height) of greenhouse cucumber to calculate LAI_e to describe the influences of leaf spatial distribution and photosynthesis efficiency. We compared the performances of the improved and the original Jarvis models on r_c estimation and found the coefficient of determination (R²) and root mean squared error (RMSE) of 0.68 and 163 s m⁻¹, 0.64 and 171 s m⁻¹, respectively, the corresponding errors for hourly ET_c calculation were 0.81 and 0.10 mm h⁻¹, 0.88 and 0.098 mm h⁻¹, respectively. The accuracy of both models gradually raised with the increasing in LAI, and an obvious improvement on ET_c estimation appeared when LAI ≤ 1 m² m⁻² with R² = 0.81 for the improved Jarvis model while R² = 0.77 for the original Jarvis model. Although the accuracy of the improved Jarvis model was not prominently increased for all growing stages, the improved Jarvis model is of great significance in considerations of crop physiological and growth diversity to enhance the model adaptability for distinct crops and reduce the uncertainty of model performance for different growing stages.