Angular chlorophyll indices estimates derived from ground-based diurnal course data and multiangular CHRIS-PROBA data: Two case studies

Abstract

At leaf and plant level chlorophyll indices have shown strong correlations with chlorophyll content and photosynthesis-related processes. However, at canopy level additional abiotic and biotic factors confound the fidelity of these indices. For instance, the Photochemical Reflectance Index (PRI) is known to be sensitive to viewing angles and canopy structure. In this paper we present case studies of two natural canopies at different scales where the influence of sun-target-sensor geometry and canopy structure is inter-compared for a range of chlorophyll indices. In the first case study, surface reflectance was measured in a montane grassland ecosystem located at the Bily Kriz experimental study site (Czech republic) using a stationary mounted AISA (Airborne Imaging Spectrometer for Applications) spectrometer. The experimental set-up resulted in a ground pixel resolution of ~2mm. The effects of changing sun angles on the indices were assessed through a diurnal sampling between 9:00 and 15:00 hrs (local time). Classes of shaded and illuminated photosynthetic (PV) and non-photosynthetic vegetation (NPV) were distinguished per image using a pixel wise classification. The relative contributions of confounding factors as well as the influence of the diurnal variability on performance of the selected chlorophyll indices were evaluated. In the second case study, surface reflectance was measured over an Alpine coniferous ecosystem in the Swiss National Park (Switzerland) using multiangular hyperspectral CHRIS-PROBA (Compact High Resolution Imaging Spectrometer onboard the Project for On-board Autonomy) satellite system with a ground pixel resolution of 17 m. The angular signature of PRI and the structure invariant pigment index (SIPI) was assessed using CHRIS data. Besides, we evaluated the influence of varying tree crown composition and varying viewing angles to the chlorophyll indices with a radiative transfer model FLIGHT. In both cases, the PRI and the green NDVI (gNDVI) responded extremely sensitively to the considered confounding factors at canopy level. The Transformed Chlorophyll Absorption in Reflectance Index normalized by the Optimized Soil-Adjusted Vegetation Index (TCARI/OSAVI), designed to be insensitive to background and LAI variations, responded more sensitively than the conventional NDVI. No certain sensitivity was found for SIPI. The pronounced sensitivity of e.g. PRI and gNDVI, on one hand, and the inconsistency between the chlorophyll indices, on the other hand, erodes the fidelity to use these spectral indices as an effective non-destructive chlorophyll detector.