Multiple satellite remote sensing datasets open new opportunities to monitor global terrestrial radiation, energy, and CO2 balances at high spatial resolutions. Although these components are tightly coupled to each other, there have been few efforts to produce global high resolution maps in a coupled manner. Here, we report Breathing Earth System Simulator (BESS v2) which combines multiple satellite data streams to run a coupled biophysical model that computes atmosphere and canopy radiative transfer, photosynthesis, transpiration, energy balance, and ecosystem respiration. It produces global maps of solar radiation, net radiation, latent and sensible heat fluxes, photosynthesis and ecosystem respiration at daily, 1 km resolution between 2001 and 2016. At site level, we test BESS v2 against in-situ FLUXNET2015 datasets, which show reasonable performance (R2>0.6, relative bias<2%). At a global scale, BESS v2 show comparable spatial and temporal patterns with data-driven MPI-BGC products. At a decadal scale, BESS v2 show increasing strengths of land carbon sink and water use efficiency contributed by CO2 fertilization effects. Compared to other datasets, BESS v2 offers unprecedented high spatiotemporal resolution land surface flux maps across the global land, which will contribute to carbon and water management from local to the global scales. Finally, we discuss challenges in extending global land surface flux records from 1999 back to 1982.