Towards the Optimal Modis-based Photochemical Reflectance Index for Arid Areas

The diagnostic modeling of gross primary productivity is complicated by a variety of regulatory mechanisms acting at different time scales, e.g. variations in leaf area, chlorophyll content, rubisco activity, and stomatal conductance. A model formulation that can–theoretically–comprise all these variations was first described by [1, 2]: gross primary productivity (GPP) is computed as the product of the amount of absorbed photosynthetically active radiation (aPAR) and a light use efficiency term. Present-day diagnostic models have the light use efficiency (LUE) term implemented either as a constant (sometimes stratified according to plant functional type) or as a (biome-specific) maximum LUE reduced by scalars that represent environmental stress [3]. This look-up table approach is unable to represent the complete range of productivity dynamics, especially at shorter time scales [4, 5], primarily due to inaccurate maximum LUE estimates [6]. Global scale estimations of environmental drivers reducing maximum LUE cause more uncertainty [7]. Besides, current remote-sensing based models have difficulties to detect drought stress unless soil water content is accounted for [8], which is difficult on a global scale. So, we need an alternative approach for cases in which LUE and leaf area/ chlorophyll content do not vary synchronously. Another option to estimate LUE employs a narrow wavelength range centred around 531 nm [9]. Reflection there decreases as plants protect their photosynthetic system from excess sunlight trough xanthophyll pigment interconversion. The photochemical reflectance index (PRI) combines reflectance at this wavelength (ρ531) with a reference wavelength insensitive to short-term changes in light energy conversion efficiency (usually 570 nm, ρ570) and normalises it [10]: