Combing both simulated and field-measured data to develop robust hyperspectral indices for tracing canopy transpiration in drought-tolerant plant.

Transpiration plays a key role in water and energy fluxes at various scales. While in recent remote sensing offers a fast and convenient method for tracing transpiration at multiple scales, the approach is mostly indirect and relies on energy balance. Although several hyperspectral indices have been reported to show potentials for tracing transpiration directly, both at leaf and canopy scales, they remain in pioneer stages and need extensive validations. In this study, we used the Soil, Canopy Observation, Photochemistry and Energy fluxes (SCOPE) model calibrated to arid ecosystems in Central Asia, to generate a simulated dataset for validation. Furthermore, new and robust indices have been developed by combining both simulated and in situ measured datasets. Results suggested that the SR(1525, 2150), ND(1425, 2145), and previously reported index of dSR(660,1040) have significant relationships with both simulated and in situ measured transpiration. Further analyses revealed that the ND(1425,2145) shows consistent performance, even with different methodologies of combining simulation and field-measured datasets. Statistically significant results were obtained in this study, even for a dominant drought-tolerant species in arid land, a place that typically has weak vegetation reflectance under strong background radiation. We foresee the approach being conducted in other regions where vegetation reflectance dominates. This may lead to robust hyperspectral indices being developed for directly tracing transpiration at various scales.