| 摘 要： |
Accurate phenological extraction is important for estimating carbon uptake in terrestrial ecosystems under climate change. The emergence of remotely sensed vegetation indices (VIs) and solar-induced chlorophyll fluorescence (SIF) provides multiple approaches for extracting land surface phenology. However, there is lacking studies to track phenological metrics via multiple VIs and SIF. Therefore, the advantage of combining VIs and SIF to estimate more accurate phenology requires exploration. In this study, we combined the advantages of the normalized difference, enhanced, green-red, near-infrared reflectance vegetation indices from MCD43A4 data set, and SIF from CSIF data set to estimate hybrid phenology at 20 eddy flux sites in North America. Results showed that the hybrid phenology derived from the best-performing start (SOS) and end (EOS) of the growing season among multiple VIs and SIF for each plant functional type and site were both more consistent with those derived from gross primary production (GPP). Specifically, the R-2 of hybrid phenology increased by 0.11-0.4 (0.04-0.4) for SOS, 0.01-0.24 (0.09-0.22) for EOS, 0.01-0.7 (0.05-0.34) for the length of the growing season (LOS) based on Gaussian (logistic) method. Moreover, hybrid phenology can improve the explanation of the seasonal and annual variations in GPP. The explanatory power of hybrid phenology for GPP variations increased by 0.05-0.15 (0.02-0.23) for SOS, 0-0.36 (0.11-0.27) for EOS, 0.01-0.51 (0.03-0.4) for LOS, 0.04-0.18 (0.04-0.16) fo....r LOS x seasonal GPP maximum based on Gaussian (logistic) method. These findings highlight the potential of combining high-spatiotemporal structural and coarse-spatiotemporal physiological vegetation indicators in tracking phenology and GPP. |
在线留言
联系我们
所长信箱