Impact of spring phenology variation on GPP and its lag feedback for winter wheat over the North China Plain

Spring green-up date (GUD) is a sensitive indicator of climate change, and of great significance to winter wheat production. However, our knowledge of the chain relationships among them is relatively weak. In this study, based on 8-day Enhanced Vegetation Index (EVI) data from Moderate Resolution Imaging Spectroradiometer (MODIS) from 2001 to 2015, we first assessed the performance of four algorithms for extracting winter wheat GUD in the North China Plain (NCP). A multiple linear regression model was then established to quantitatively determine the contributions of the time lag effects of hydrothermal variation on GUD. We further investigated the interactions between GUD and gross primary production (GPP) comprehensively. Our results showed that the rate of change in curvature algorithm (RCCmax) had better performance in capturing the spatiotemporal variation of winter wheat GUD relative to the other three methods (Kmax, CRmax, and cumCRmax). Regarding the non-identical lag time effects of hydrothermal factors, hydrothermal variations could explain winter wheat GUD variations for 82.05% of all pixels, 36.78% higher than that without considering the time lag effects. Variation in GUD negatively correlated with winter wheat GPP after green up in most parts of the NCP, significantly in 35.75% of all pixels with a mean rate of 1.89 g C m^-2 yr^-1 day^-1. Meanwhile, winter wheat GPP exerted a strongly positive feedback on GUD in >82.42% of all pixels (significant in 28.01% of all pixels), characterized by a humped-shape pattern along the long-term average plant productivity. This finding highlights the complex interaction between spring phenology and plant productivity, and also suggests the importance of preseason climate factors on spring phenology.