Autumn phenology consistently delays in subtropical forests in China based on a new process-based model integrating temperature, photoperiod and precipitation

Abstract

Climate warming has delayed vegetation autumn phenology, which in turn influences terrestrial carbon and water cycles and their feedback to the climate. However, the performance of autumn phenology models, especially for subtropical forests, remains poor. In this study, we extracted the end-of-photosynthetic-growing-season (EOPS) dates in subtropical China over the period 2001–2018 based on high-resolution solar-induced chlorophyll fluorescence (SIF) dataset using three fitting methods. We developed a new autumn phenology model (DMP model) that integrates precipitation and photoperiod into the classic cold-degree days model, and the new model outperformed the classic model reducing the RMSE by approximately 2 d. We found that the EOPS date was delayed by an average of 4.1 d per decade in the subtropical forests of China from 2001 to 2018, and the precipitation (partial correlation coefficient, r = 0.45), rather temperature ( r = 0.29) determined the EOPS processes. We further studied future EOPS using the DMP model and found that the EOPS will delay by 1.0 d per decade under the shared socioeconomic pathway (SSP2-4.5) scenario and 2.7 d per decade under the SSP5-8.5 scenario from 2030 to 2100. Our study highlighted the role of precipitation in regulating EOPS in the subtropical forests of China and provided valuable insight for integrating multiple climatic determinants into autumn phenology models.

Keywords

• autumn phenology
• climate warming
• process-based autumn phenology model
• subtropical forests
• precipitation