Detecting leaf-water content in Mediterranean trees using high-resolution spectrometry

•We aimed to identify an index for remote leaf water content (EWT) estimation.•Field spectral measurements were correlated to leaf water content.•Right slope 970 nm index is best suited to estimate EWT.•Simulation of spectra illustrated sensitivity of the index to leaf area index.•Sensitivity to other environmental variables is small.

Water content of the vegetation canopy or individual leaves is an important variable in physiological plant processes. In Mediterranean regions where water availability is an important production limiting factor, it is a strong indicator of vegetation stress. Spectroscopic earth-observation techniques in the solar part of the electromagnetic spectrum provide opportunities to determine leaf and canopy-water content due to the presence of water-absorption bands around 970 and 1200 nm. We investigated the possibilities to predict leaf-water content of three dominant tree species in a study area in Mediterranean France using spectral indices. During a field campaign leaf-water content (EWT) was determined and high-resolution spectra were measured of the same leaves. The spectra were measured in two ways: using an optical cable with a field of view of 25° and using a leaf clip with its own artificial illumination source. The spectra were analyzed and related to leaf-water content as original reflectance spectra and as continuum-removed spectra using eight spectral leaf-water indices. Next, reflectance spectra were simulated to explore their sensitivity to environmental conditions like leaf area index and illumination angle using a radiative transfer model. Results show that a good correlation (0.70) exists between leaf-water content and spectral indices using the right slope of the 970 nm water-absorption band. Continuum-removal correction of the spectra improved the relations. The model sensitivity analysis illustrated that from a set of five environmental variables leaf area index has, as may be expected, an important impact on leaf-water estimates. This field and model study illustrates that it is feasible to determine foliar water content on the basis of spectral indices located around the minor water-absorption bands with a limited effect of environmental conditions.