Parameterization of Four Models to Estimate Crop Evapotranspiration in a Solar Greenhouse

TL;DR

This study compares four models for estimating tomato evapotranspiration in a solar greenhouse and finds that the parameterized Priestley-Taylor model performs best.

Contribution

The study introduces parameterized versions of four ET models and evaluates their performance in a solar greenhouse setting.

Findings

The PA-PT and PA-CC models performed better at all four growth stages of tomatoes.

The PA-PT model had the lowest error and highest agreement with measured ET over the entire growing season.

PA-PM and PA-SW models showed significant deviations at initial and late growth stages.

Abstract

Working to simplify mechanistic models on the basis of reliability for estimating crop evapotranspiration (ET) in a greenhouse is still worthwhile for horticulturists. In this study, four ET models (Penman–Monteith, Priestley–Taylor, and Shuttleworth–Wallace models, and the Crop coefficient method) were parameterized after taking the restriction effect of resistance parameters in these models on ET into account, named as PA-PM, PA-PT, PA-CC, and PA-SW, respectively. The performance of these four parameterized models was compared at different growth stages, as well as the entire growing season. Tomatoes that were ET-grown in a solar greenhouse without a heating device were measured using weighting lysimeters during 2016–2017 and 2019–2021, in which data from 2016 were used to adjust the model parameters, and data from the other four study years were used to examine the model performance. The results indicated that the PA-PT and PA-CC models have a better performance in estimating tomato ET at four growth stages, while the PA-PM and PA-SW performed well only at the development and middle stages. Compared to the ET that was measured with the weighting lysimeters, the ET that was predicted using the PA-PM model was 27.0% lower at the initial stage, and 8.7% higher at the late stage; the ET that was computed using the PA-SW model was 19.5% and 13.6% higher at the initial and late stages, respectively. The PA-PT model yielded the lowest root mean square error and the highest index of agreement against the other models over the entire growing season, indicating that the PA-PT model is the best recommended model for estimating tomato ET in a solar greenhouse.

Figures