Tracking seasonal drought effects on ecosystem light use efficiency with satellite-based PRI in a Mediterranean forest

Gross primary productivity (GPP) changes occur at different time-scales and due to various mechanisms such as variations in leaf area, chlorophyll content, rubisco activity, and stomatal conductance. Diagnostic estimates of primary productivity are obviously error prone when these changes are not accounted for. Additional complications arise when factors inuencing a biome-specific maximum light use efficiency (LUE) must be estimated over a large area. In these cases a direct estimation of ecosystem LUE could reduce uncertainty of GPP estimates. Here, we analyse whether a MODIS-based photochemical reectance index (PRI) is a useful proxy for the light use efficiency of a Mediterranean Quercus ilex forest. As the originally proposed reference band for PRI is not available on MODIS, we tested the reference bands 1 (620–670 nm), 4 (545–565 nm), 12 (546–556 nm), 13 (662–672 nm), and 14 (673–683 nm) using different atmospheric correction algorithms. We repeated the analysis with different temporal resolutions of LUE (half-hourly to daily). The strongest correlation between LUE and PRI was found when considering only a narrow range of viewing angles at a time (especially 0–10° and 30–40°). We found that the MODIS-based PRI was able to track ecosystem LUE even during severe summer time water limitation. For this Mediterranean-type ecosystem we could show that a GPP estimation based on PRI is a huge improvement compared to the MODIS GPP algorithm. In this study, MODIS spectral band 1 turned out to be the most suitable reference band for PRI, followed by the narrow red bands 13 and 14. As to date no universally applicable reference band was identified in MODIS-based PRI studies, we advocate thorough testing for the optimal band combination in future studies.