Solar-induced chlorophyll fluorescence detects photosynthesis variations and drought effects in tropical rubber plantation and natural deciduous forests

Past assessments of the global carbon balance have shown to have high uncertainty particularly in Mainland Southeast Asia (MSEA) owing to the increasing climate extremes and land use changes. Recently, remotely sensed solar-induced chlorophyll fluorescence (SIF) and near-infrared reflectance of vegetation (NIRv) have been found to have great potential to assess the dynamics of vegetation gross primary production (GPP). Using flux tower GPP from two rubber plantation and one natural forest sites, our study assessed the utility of SIF and NIRv to detect GPP and drought stress for these two major land uses in this tropical monsoon region, and the underlying mechanistic link among SIF, NIRv, and GPP based on path analysis. The results indicate that the native SIF from OCO-2 satellite and its two derived finer-resolution SIF products (i.e., GOSIF and ) outperformed the normalized difference vegetation index (NDVI) and the enhanced vegetation index (EVI) in detecting GPP and drought effects at the rubber plantation sites. Although the young natural forest exhibited much lower EVI and NIRv, it had comparable magnitude of GPP to the rubber plantations, which was captured by SIF. Path analysis shows that the SIF-GPP relationship was mainly controlled by NIRv, which represents the canopy structural effects including both the absorption and scattering traits. In line with site-scale analysis, the regional patterns indicated higher consistency between SIF and NIRv than betwen SIF and EVI for the tropical forests. The quantum yield of SIF () was sensitive to precipitation anomaly, while its negative effect on photosynthesis was offset by high PAR under drought. Both tropical deciduous forest and plantation forest were more vulnerable to drought than evergreen forest. We highlight the usage of NIRv and SIF for monitoring canopy structure and photosynthesis, respectively, and the needs of higher-resolution observations to reduce the uncertainties in quantifying SIF-GPP relationships.