|
|
Data products
|
AmeriFlux data (gap-filled and partitioned) - We gap-filled, partitioned flux data for nearly 300 AmeriFlux and NEON sites. The gaps of the meteorological data were filled with ERA5-Land reanalysis data, and the flux measurements were gap-filled and partitioned using standardized procedures. The data cover the period from the early 1990s to 2021. EC-MOD - The EC-MOD data set consists of gross primary productivity (GPP), net ecosystem exchange (NEE), ecosystem respiration (ER), and evapotranspiration (ET) at 1km spatial resolution and 8-day time step over North America for the period from March 2000 to December 2012 (Xiao et al. 2014). It was developed from eddy covariance (EC) flux data, MODIS data streams, micrometeorological reanlaysis data, stand age, and aboveground biomass data using a data-driven approach (Xiao et al. 2008, 2010, 2011). GIMMS NDVI3g Phenology - A spring and autumn phenology product for the northern hemisphere with 1/12-degree spatial resolution and annual time step over the period 1982-2014. This phenology product is derived from the the GIMMS NDVI3g dataset and consists of five different phenology estimates for both start of growing season (SOS) and end of growing season (EOS). GOSIF - The GOSIF product is a new global, OCO-2-based SIF (solar-induced chlorophyll fluorescenc) data set with high spatial and temporal resolutions (i.e., 0.05-degree, 8-day) over the period 2000-2022 based on discrete OCO-2 SIF soundings, remote sensing data from MODIS, and meteorological reanalysis data (Li and Xiao, 2019a). GOSIF GPP - The GOSIF GPP product is a new global, fine-resolution dataset of GPP (gross primary productivitiy). GOSIF GPP is derived from GOSIF and SIF-GPP relationships. This data set has high spatial and temporal resolutions (i.e., 0.05-degree, 8-day) over the period 2000-2022 (Li and Xiao, 2019b). MF-CW - The datasets of Machine learning and FLUXNET based Carbon and Water Fluxes (MF-CW), including monthly GPP and ET, and yearly WUEeco at 0.5-degree resolution over the global vegetated lands for 1982-2016, were generated using 24 machine learning methods (see list below), FLUXNET in-situ observations of CO2 and water vapor fluxes, satellite-derived observations and climate reanalysis data.
|
