Heterogeneity of land cover data with discrete classes obscured remotely-sensed detection of sensitivity of forest photosynthesis to climate

•The climatic sensitivities of SIF were detected across the coarse forest pixels.•The heterogeneity was indicated by tree coverage (Tc) and landscape diversity (SHDI).•Overall, the climatic sensitivities of SIF were amplified with an increase in Tc.•The sensitivities also varied with SHDI, generally showing a negative correlation. Satellite-based land cover data with discrete classes provide a critical basis for studies on response of forest photosynthesis to climate change at the regional scale. However, this discrete classification oversimplifies reality due to widely observed mixed pixels, which may obscure the remotely-sensed detection at moderate/low spatial resolution. To this point, the purpose of this study was to explore whether and how heterogeneity of land cover data with discrete classes impacts remotely-sensed detection of climatic sensitivity of photosynthesis across coarse grid cells of forested area. First, coarse forest grid cells at moderate resolution (0.05°) with tree coverage > 50% were derived from a finer land cover dataset (300 m) in central-eastern China. Then, solar-induced fluorescence (SIF) was taken as a proxy of forest photosynthesis to detect the sensitivities to three climatic variables (i.e., temperature, precipitation and radiation) in each forest pixel of interest. After that, variabilities in the climatic sensitivities with grid cell-based tree coverage and landscape biodiversity were investigated across the study area by gradient analysis. The findings show an explicit gradient change in the climatic sensitivities of photosynthesis with forest coverage and landscape diversity. Overall, the climatic sensitivities are amplified with an increase in tree coverage in the grids. Similarly, the climatic sensitivities of forest photosynthesis vary with landscape diversity, generally showing a negative correlation in most of the area. These quantitative results built a strong evidence to answer the proposed questions. Hence, we suggest that it is necessary to consider heterogeneity of land cover data with discrete classes in parameterization of modeling and inversion. The achievement of this study is helpful to provide insight into reasons for ecological response to climate change.