A comparison of plot-based satellite and Earth system model estimates of tropical forest net primary production

Net primary production (NPP) by plants represents the largest annual flux of carbon dioxide (CO2) from the atmosphere to the terrestrial biosphere, playing a critical role in the global carbon (C) cycle and the Earth's climate. Rates of NPP in tropical forests are thought to be among the highest on Earth, but debates about the magnitude, patterns, and controls of NPP in the tropics highlight uncertainty in our understanding of how tropical forests may respond to environmental change. Here, we compared tropical NPP estimates generated using three common approaches: (1) field-based methods scaled from plot-level measurements of plant biomass, (2) radiation-based methods that model NPP from satellite-derived radiation absorption by plants, (3) and biogeochemical model-based methods. For undisturbed tropical forests as a whole, the three methods produced similar NPP estimates (i.e., similar to 10 Pg C yr(-1)). However, the three different approaches produced vastly different patterns of NPP both in space and through time, suggesting that our understanding of tropical NPP is poor and that our ability to predict the response of NPP in the tropics to environmental change is limited. To address this shortcoming, we suggest the development of an expanded, high-density, permanent network of sites where NPP is continuously evaluated using multiple approaches. Well-designed NPP megatransects that include a high-density plot network would significantly increase the accuracy and certainty in the observed rates and patterns of tropical NPP and improve the reliability of Earth system models used to predict NPP-carbon cycle-climate interactions into the future.