Estimating the peak of growing season (POS) of China’s terrestrial ecosystems

Abstract

Vegetation phenology is a sensitive indicator of climate change and has substantial impacts on ecosystem carbon uptake. Previous studies have revealed the importance of the start and end of growing season and investigated their dynamics with climate change. However, the interannual variability of the peak of growing season (POS), which also plays a crucial role in carbon cycle of terrestrial ecosystems, has not been well documented. Here, we used field observations from regional flux measurements to evaluate four methods in POS extraction based on time series of satellite observed normalized difference vegetation index (NDVI). Results showed that the combination of Savitzky-Golay filter and cubic spline (SG-cubic spline) method had overall higher accuracy (R2 = 0.77, RMSE = 8.3 days) than current approaches for POS detection. We then analyzed the spatial and temporal patterns of POS in China during 1982–2015 and found an overall advanced trend of POS in most regions, but variations were also observed among biomes. Preseason temperature was the most important factor controlling POS for about 25.6% of all pixels, followed by radiation (19.7%) and precipitation (14.6%), respectively. A higher temperature overall advanced POS, while influences of precipitation and radiation were highly biome-dependent. We further investigated the impacts of POS on plant productivity and found that an advanced POS would increase annual GPP, but this relationship was opposite in arid regions, highlighting the role of soil moisture in dry areas. Our study is meaningful for deepening the understanding between phenology and carbon cycling with respect to future climate change.