Abstract
The annual peak growth of vegetation is critical in characterizing the capacity of terrestrial ecosystem productivity and shaping the seasonality of atmospheric CO 2 concentrations. The recent greening of global lands suggests an increasing trend of terrestrial vegetation growth, but whether or not the peak growth has been globally enhanced still remains unclear. Here, we use two global datasets of gross primary productivity (GPP) and a satellite-derived Normalized Difference Vegetation Index (NDVI) to characterize recent changes in annual peak vegetation growth (that is, GPP max and NDVI max ). We demonstrate that the peak in the growth of global vegetation has been linearly increasing during the past three decades. About 65% of the NDVI max variation is evenly explained by expanding croplands (21%), rising CO 2 (22%) and intensifying nitrogen deposition (22%). The contribution of expanding croplands to the peak growth trend is substantiated by measurements from eddy-flux towers, sun-induced chlorophyll fluorescence and a global database of plant traits, all of which demonstrate that croplands have a higher photosynthetic capacity than other vegetation types. The large contribution of CO 2 is also supported by a meta-analysis of 466 manipulative experiments and 15 terrestrial biosphere models. Furthermore, we show that the contribution of GPP max to the change in annual GPP is less in the tropics than in other regions. These multiple lines of evidence reveal an increasing trend in the peak growth of global vegetation. The findings highlight the important roles of agricultural intensification and atmospheric changes in reshaping the seasonality of global vegetation growth.
| Original language | English |
|---|---|
| Pages (from-to) | 1897-1905 |
| Number of pages | 9 |
| Journal | Nature Ecology and Evolution |
| Volume | 2 |
| Issue number | 12 |
| DOIs | |
| State | Published - Dec 1 2018 |
Funding
This work was financially supported by the National Key R&D Program of China (2017YFA0604603), National Natural Science Foundation (31430015, 41601099 and 41630528) and National 1000 Young Talents Program of China. We thank all the people who worked to provide data for this study, particularly the MsTMIP modelling group. We are grateful for receiving MTE GPP products from MPI-BGC, biweekly NDVI data from the GIMMS team, MODIS EVI products from USGS, climate-forcing data from CRU and Princeton University, CO2 site data from NOAA, and GOME-2 SIF retrievals from Eumetsat. We further thank the TRY initiative for plant traits (http://www.try-db. org). The TRY initiative and database are hosted, developed and maintained by J. Kattge and G. Bönisch (MPI-BGC). The eddy-covariance data of FLUXNET used in this study were mainly acquired by the following networks: AmeriFlux, GHG-Europe, SOERE, FORE-T, the Fluxnet-Canada Research Network (supported by CFCAS, NSERC, BIOCAP, Environment Canada and NRCan), GreenGrass, KoFlux, LBA, NECC, OzFlux, TCOS-Siberia and USCCC. The vector map data were made using Natural Earth. J.B.F. contributed to this paper from the Jet Propulsion Laboratory, California Institute of Technology, under a contract with NASA, and support was provided by the IDS programme.
