In recent years, forest decline is affecting large parts of Spain and Europe. Detection of forest stress levels has been the target of numerous works based on remote sensing based primarily on the relationship between biophysical parameters and vegetation indices sensitive to changes in structure as the NDVI or vegetation indices sensitive to chlorophyll content. However, when the chlorophyll or the structure of the canopy is affected by stress, plant damage has already occurred, and the status of it is compromised.  The detection of stress in its early phase, usually defined as pre-visual and is a critical information needed for early assessment of stress.

The Photochemical Reflectance Index (PRI) is a physiological index sensitive to the epoxidation state of xanthophyll cycle pigments and photosynthetic efficiency. The PRI was proposed as a normalized difference of the band 530 nm, related to the absorption of xanthophyll pigments, and one reference strip located at 570 nm. The assessment of water stress at canopy scale from PRI is a complex problem because it is an index highly affected by light conditions and structure

.

Figure 1. Photochemical Reflectance Index calculated form AHS images on three study sites with different levels of water stress.

Figure 2 (left). High-resolution thermal imaging of a conifer forest affected by water stress acquired with airborne data from the sensor AHS. Figure 3. High-resolution thermal imaging of a conifer forest affected by water stress acquired with a fixed-wing unmanned aerial vehicle (UAV).

Water stress can also be analyzed through thermal information. Under conditions of plant water deficit stomatal closure is induced, reducing the transpiration rate and consequently, increasing the leaf temperature. However, very limited references have successfully demonstrated the relationship between canopy temperature and water stress in Mediterranean forests.

Treesatlab has shown progress in the application of hyperspectral images with high spatial resolution for the early detection of stress levels in Mediterranean conifers subjected to forest decline. The results demonstrate the potential application of thermal data and vegetation index PRI obtained from high spatial resolution image data to detect water stress on forest environment.

The detection of water stress in conifers from thermal imaging and narrow bands vegetation indices has important implications for early detection of forest decline. The investigations have been made from high spatial resolution sensors acquired from airborne sensors (AHS-Airborne Hyperspectral System) and unmanned UAV acquired (Untripulated air vehicle). Also carried out an extensive comparative analysis of broad-band multispectral images (QuickBird, and Landsat ETM +) versus narrow-band hyperspectral imagery (Hyperion, ChrisProba).