Global distribution of leaf maximum carboxylation rate derived from the TROPOMI solar-induced chlorophyll fluorescence data

Abstract

Photosynthesis plays an important role in the terrestrial carbon cycle and is often studied using terrestrial biosphere models (TBMs). The maximum carboxylation rate at 25 °C (V_cmax25) is a key parameter in TBMs, and yet the information on the spatiotemporal distribution of this parameter is uncertain. In this study, we retrieved the global distribution of V_cmax25 at 0.25° resolution based on TROPOMI-observed solar-induced chlorophyll fluorescence (SIF) and meteorological forcing data using a parameter optimization technique. This study improves global mapping of V_cmax25 using TROPOMI's SIF and MODIS photochemical reflectance index (PRI) for accurate GPP estimation by sunlit leaves in the following aspects: the previous method relied on an empirical estimation of the ratio of SIF per unit sunlit leaf area to that per unit shaded leaf area (β), while β here was derived from a look-up table (LUT) constructed using the Soil-Canopy Observation of Photosynthesis and Energy (SCOPE) model. Validated at two flux tower sites, the LUT method explained most of the variation in β with R² = 0.71 and 0.67, RMSE=0.19 and 0.15 and Slope=0.84 and 0.70 for two ground validation sites. We calculated the global ratio of SIF from sunlit to that from shaded leaves (SIF_ratio), and found that the $SIF\_ratio$ had a strong spatio-temporal variability with a global average of approximately 4.6, and that the contribution of SIF from shaded leaves to the canopy total was <20 %. The optimized V_cmax25 from TROPOMI was validated against V_cmax25 derived from concurrent flux data at 27 sites distributed globally using an independent method (R² = 0.39 - 0.65, RMSE = 6.47 - 21.74 μmol m^-2 s^-1 and rRMSE =0.14–0.36). Based on the improved global V_cmax25 map, we found that, spatially, V_cmax25 varies significantly with latitude and between- and within-plant function types (PFTs), and temporally, it has strong seasonal variation in all PFTs except evergreen broadleaf forests. The new global V_cmax25 dataset would be useful for improving terrestrial GPP modelling from the current state of the art of using constant V_cmax25 values by plant functional type.