Modeling and predicting spring land surface phenology of the deciduous broadleaf forest in northern China

Using normalized difference vegetation index data, we fitted unified forcing and chilling phenology models to start date of the growing season (SOS) during 1982 to 2006 and selected optimum models for each pixel. Then, we validated performances of optimum models in extrapolating SOS dates during 2007 to 2011. Moreover, we reconstructed SOS time series over 1950 to 2005 and predicted decadal mean anomalies of SOS dates over 2006 to 2100 under scenarios of the representative concentration pathway (RCP) 4.5 and 8.5. Finally, we compared structures and parameters of the satellite data-based phenology model with those of the ground-based leaf unfolding model. Results show that the unified forcing model has higher simulation parsimony and efficiency than the unified chilling model at 96.4% of pixels. The external validation confirmed feasibility of the unified forcing model in predicting the SOS date. Across the entire region, predicted SOS dates indicate significant advancements by 0.65-1.79 days per decade. Taking the mean SOS date during 1961 to 1990 as the reference value, the absolute values of decadal mean anomalies of the SOS date represent an overall increase with the time. Consequently, the average SOS date during 2091 to 2100 would be 11.8-20.0 days earlier than the average SOS date during 1961 to 1990. At pixel scales, the absolute values of decadal mean anomalies of the SOS date under the RCP 8.5 scenario display mostly a latitudinal gradient from south to north. Similar distribution ranges of parameter d and e of the ground-based and satellite data-based phenology models demonstrate that the satellite-derived SOS date has consistent response to air temperature with the leaf unfolding date of dominant deciduous trees, which implies that the predicted start of the growing season may to some extent reflect changing tendency of the ground leafing phenology in the 21st century. (C) 2014 Elsevier B.V. All rights reserved.AgronomyForestryMeteorology & Atmospheric SciencesSCI(E)0ARTICLExiangzhong.luo@mail.utoronto.ca; cxq@pku.edu.cn; wanglingxiao@pku.edu.cn; linxu5048@163.com; tianyouhua@foxmail.com33-4119