Vegetation photosynthetic phenology metrics in northern terrestrial ecosystems: a dataset derived from a gross primary productivity product based on solar-induced chlorophyll fluorescence

Abstract. Vegetation phenology can profoundly modulate the climate-biosphere interactions and thus plays a key role in regulating the terrestrial carbon cycle and the climate. However, most previous phenology studies are based on the traditional vegetation indices, which are inadequate to characterize the seasonal activity of photosynthesis. Here, we generated an annual vegetation photosynthetic phenology dataset with a spatial resolution of 0.05 degree from 2001 to 2020, using the latest gross primary productivity product based on solar-induced chlorophyll fluorescence (GOSIF-GPP). We combined smoothing splines with multiple change-point detection to retrieve the phenology metrics: start of the growing season (SOS), end of the growing season (EOS), and length of growing season (LOS) for terrestrial ecosystems in the Northern Hemisphere. We found that the derived phenology metrics agreed better with in situ observations from the flux tower sites than vegetation indices and MODIS-GPP. Our phenology metrics captured the spatial-temporal patterns of the single and double growing season in the Northern Hemisphere. The double season was mainly from the cropland rotation and ecosystems having two different phenological cycles. In addition, we observed a trend toward advanced SOS in about 62.98 % of the land area, with a mean rate of 0.14 ± 0.01 days year-1, a trend toward delayed EOS in about 61.87 % of the area, with a mean rate of 0.19 ± 0.16 days year-1, and a trend toward extended LOS in about 70.52 % of the area, with a mean rate of 0.33 ± 0.17 days year-1. Our phenology product can be used for validating and developing phenology models or carbon cycle models, for evaluating satellite remote sensing phenology, and for monitoring climate change impacts on terrestrial ecosystems. The data are available at https://doi.org/10.6084/m9.figshare.17195009.v2 (Fang et al. 2021).