Evapotranspiration (ET) is the link between energy, water and carbon cycle. The ET remote sensing model is the main approach to obtain ET spatial distribution. However, these models are often subject to great uncertainty in their estimation due to their difficulty to distinguish the contribution of vegetation canopy and soil to ET and too many PFT-specific empirical parameters of vegetation. To address the above problems, Yang Kun’s Research Group of the Department of Earth System Science (DESS), Tsinghua University has, by integrating the eco-evolutionary optimality principle and the land surface model scheme, developed a universal vegetation transpiration module and a module that can separate remote sensing thermal information and built a dual source eco-evolutionary optimality-based ET (EEOET) model with stronger mechanism and fewer calibration parameters.

Existing ET models based on remote sensing observations can be generally divided into two categories from the perspective of core variables: temperature-based and stomatal resistance-based models. Although the temperature-based models avoid the complex calculation of transpiration and stomatal resistance, yet the errors resulting from inputs and parameterization are all attributed to ET; and such models generally depend on multi-angle remote sensing observations to separate thermal information, leading to limited applicability. The stomatal resistance-based models express the impact of meteorological conditions, vegetation characteristics and soil water forcing, yet too much dependence on the parameters of vegetation types leads to increasing uncertainties in resistance calculation.

Fig. 1. Flowchart of the eco-evolutionary optimality-based ET (EEOET) model.

Different from past models, the eco-evolutionary optimality-based ET (EEOET) model uses land surface temperature of remote sensing as inputs and calculate the resistance parameters in ET (Fig. 1). In this study, the estimation scheme of physiological and ecological parameters based on the Eco-Evolutionary Optimality (EEO) principle developed by Professor Wang Han of Tsinghua University is introduced to estimate the maximum rate of carboxylation, which avoids the use of parameters specified according to vegetation functional types and thus greatly reduces the parameters for stomatal resistance calculation. In addition, the energy balance and water-carbon coupling scheme of the Noah-MP land surface process model can calculate stomatal resistance and distribute energy between canopy and soil at the same time.

Fig. 2. Comparison of the validation from the PT-JPL (light blue), SiTH (blue) and EEOET models (units:

. EEOET models provide the Ball-Berry method and Fick’s law to calculate stomatal resistance, with the results of validation represented by EEOET_B (dark blue) and EEOET_F (white) respectively. ALL denotes the average of the corresponding statistics at all sites.

The estimated ET was validated using measurements from 74 FLUXNET sites around the world, and comparison was made with the results of two typical ET models (Fig. 2). The results show that the EEOET model provides reasonable ET estimations for different PFTs, and without any calibration, the model shows comparable performance to other models with more tuned parameters or more empirical processes.

The study further discusses the effectiveness of the EEO principle (Fig. 3). The results show that reasonable ET estimation can be provided by calculating vegetation ET with look-up tables and having the energy balance and carbon-water coupling schemes of the land surface model as the framework. The effectiveness of ET estimation is further improved by optimizing the stomatal resistance schemes with the EEO principle. Therefore, the introduction of the EEO principle not only reduces the number of parameters related to vegetation functional types but also improves the estimation accuracy of the model.

Fig. 3. Comparison between the validation results for the ET estimated before (blue) and after (red) introducing the EEO principle. The results shown in the right column represent the statistics for all sites. Table: using the look-up tables and original schemes in Noah-MP to calculate GPP. EEO: using the EEO principle to calculate GPP.

The main advantages of the model are as follows: (1) The energy distribution scheme of land surface model is adopted, and the canopy temperature and soil surface temperature can be obtained from a single surface temperature; meanwhile, compared with the method of treating ET as the residual, the EEOET model solves multiple energy components simultaneously through iteration, making the estimated results less sensitive to input data errors. (2) The EEO principle is used to reduce the number of PFT-specific parameters to no more than two, thus reducing the uncertainty caused by parameter calibration. In addition, EEOET model considers the adaptation of vegetation to climate change, making it suitable to construct ET data products reflecting the impact of climate change.

The research results have been published in Journal of Hydrology as an article titled “Integrating eco-evolutionary optimality principle and land processes for evapotranspiration estimation”. Zou Mijun, a postdoctoral fellow of the Department of Earth System Science (DESS), Tsinghua University, is the first author of the article, and his co-supervisor Professor Yang Kun is the corresponding author. Co-authors are from the DESS, Tsinghua University and the School of Geographical Sciences, Southwest University, China. The research is supported by the National Natural Science Foundation of China (42105122) and the Center for the Earth System Basic Science on the Tibetan Plateau (41988101).

Article information: Zou, M., Yang, K., Lu, H., Ren, Y., Sun, J., Wang, H., Tan, S., and Zhao, L, 2022: Integrating eco-evolutionary optimality principle and land processes for evapotranspiration estimation. Journal of Hydrology, 128855.

Full-text link: https://www.sciencedirect.com/science/article/pii/S0022169422014251

Written by Zou Mijun

Edited by Wang Jiayin

Reviewed by Zhang Qiang