Objective The purpose of this study was to explore the applicability of sunlight-induced chlorophyll fluorescence (SIF) in phenology of different forest types, and to provide a new method and scientific basis for more accurately revealing the dynamics of vegetation photosynthesis and monitoring vegetation phenology.

Method Taking three forest stations of different typical forest types in eastern China (Changbai Mountain, Qianyanzhou and Dinghu Mountain) as the research object, using the SIF remote sensing dataset from 2003 to 2010, we estimated the phenological period (SOS at the beginning of phenology and EOS at the end of phenology) of each forest type through a dual-logic growth model. Then using EVI remote sensing dataset and station dataset based on ChinaFlux (gross primary productivity, GPP; ecosystem net carbon exchange, NEE), we compared and analyzed the estimation results using SIF, furtherly assessing the potential of SIF in estimating phenological periods of different forest types. Using correlation analysis, the effects of environmental factors on phenological periods in spring and autumn were explored.

Result (1) The variation characteristics of time series curves of SIF (GOSIF, CSIF) and EVI, GPP, NEE in each forest type were basically the same. (2) Compared with GOSIF, CSIF had a stronger correlation with GPP and NEE. Compared with NEE, the correlation between three remote sensing indices and GPP was stronger. (3) In the three forest types (Changbai Mountain, Qianyanzhou and Dinghu Mountain), the CSIF estimation results of SOS and EOS were both close to NEE estimation results; the beginning of growing season estimated by GPP was earlier than NEE estimate, and the end of growing season estimated by GPP was lagging behind NEE estimation. (4) In Changbai Mountain station, SOS was mainly affected by air temperature and soil temperature, and EOS was mainly affected by soil temperature and saturated water pressure. In the Qianyanzhou site, both SOS and EOS were mainly affected by solar radiation. In the Dinghu Mountain site, SOS was mainly affected by solar radiation, and EOS was mainly affected by solar radiation and air humidity.

Conclusion CSIF with higher spatiotemporal resolution can more accurately capture the subtle changes of vegetation at different growth stages, and monitor vegetation phenology dynamics more finely.