A model for estimating transpiration from remotely sensed solar-induced chlorophyll fluorescence

Terrestrial evapotranspiration (ET) is an important flux that links global cycles of carbon, water and energy and is largely driven by transpiration (T) through leaf stomata in vegetated areas during the growing season. ET, however, remains one of the most uncertain hydrological variables at the global scale. In this study, we proposed a semi-mechanistic model for estimating terrestrial T by deriving an analytical solution between solar-induced chlorophyll fluorescence (SIF) and stomatal conductance (g(c)) as well as vapor pressure deficit (VPD), combining theories on the photosynthetic pathway and optimal stomatal behavior. The relationships of SIF-ETR and ETR-gc center dot VPD0.5 was calibrated by the Soil-Canopy Observation of Photosynthesis and Energy (SCOPE) model. This model was validated by hourly canopy SIF and concurrent eddy covariance flux observations at both forest and cropland ecosystems. Results showed that the SIF combined with VPD can better predict gc than using SIF alone with a more consistent seasonal trends found in both SIF and gc center dot VPD0.5. The correlation between gc center dot VPD0.5 and SIF was stronger than those between gc and SIF and between gc and VIs. Canopy T was accurately predicted from SIF at both hourly (R-2 > 0.65) and daily (R-2 > 0.76) scales and was also successfully estimated using SIF observations from the TROPOspheric Monitoring Instrument (TROPOMI) at cropland ecosystems. In comparison with empirical relationships of directly linking gc with SIF or VIs, the proposed model produced latent heat flux (lambda E) estimation in best agreement with measured values at all three sites. Our model could be a step forward in understanding the coupling of carbon and water cycles and may be used in ecosystem models for improving ET estimation over large areas.

Automated summary

This study proposed a semi-mechanistic model for estimating terrestrial transpiration by combining solar-induced chlorophyll fluorescence (SIF) with stomatal conductance (gc) and vapor pressure deficit (VPD), showing improved prediction accuracy and consistency. The model demonstrated that the combination of SIF and VPD can better predict gc and transpiration, with stronger correlations found between gc·VPD0.5 and SIF than other variables. The model's success in estimating transpiration and latent heat flux suggests its potential for improving ecosystem models and enhancing global evapotranspiration estimation.