Loading Events

« All Events

  • This event has passed.

DSSS: The carbon balance of forests from machine learning using age and climate

July 11, 2018 @ 10:30 am - 12:00 pm

Terrestrial gross primary production and energy fluxes have been successfully scaled from ecosystem eddy covariance (EC) observations to the global scale. However, this has not been possible for net ecosystem productivity (NEP) because forest age is an important determinant of NEP and up to now there was no globally gridded forest age map available. Moreover, the age distribution of EC sites is not representative of global forests. We combined new maps of forest age with NEP-age relationships and environmental predictors using two different machine-learning algorithms to produce gridded estimates of forest NEP. Globally, forest NEP when accounting for age is a sink of CO2 with a large range (5.8 to 8.4 Pg C a-1) depending on the algorithm used. After removing from NEP the losses of carbon that are not observed at flux tower locations: fire emissions, harvested biomass, emissions from carbon leached to rivers, and non-CO2 compounds released to the atmosphere, our best estimate of the net carbon balance (NBP) is 3.5 ± 1.2 PgC a-1 matching independent estimates from forest inventories and atmospheric inversions. Temperate forest NBP is a carbon sink of 1.5 ± 0.3 PgC a-1, larger than in inventories, but boreal forest NBP is a small net source of 0.3 ± 0.2 PgC a-1. Tropical forest NBP is estimated to be 2.1 ± 0.7 PgC a-1, a larger sink than inventories. Forests and woodlands in the dry tropics are the most uncertain component of NBP. On the one hand, there is almost no inventory data for this biome.


About the Speaker: Dr. Philippe Ciais

Philippe Ciais Researcher at Laboratoire of Sciences for Climate and Environment (LSCE) near Paris. He studied Physics at Ecole Normale Supérieure, received a doctoral degree in 1991, and was a post-doctoral fellow at NOAA (National Oceanic and Atmospheric Administration), US. At LSCE he deve,loped research on inverse modelling of GHG fluxes and helped to establish the French GHG monitoring network. Over the last two decades Philippe Ciais worked on terrestrial greenhouse gas fluxes including natural and cultivated ecosystems. He is the author of more than 650 peer- reviewed publications in the field of carbon cycle and climate change. Laureate of the Copernicus Medal of the European Geophysical Union in 2016 and of the CNRS Medal of Silver in 2017. His main current project is to investigateinteractions between carbon, nitrogen and phosphorus in the Earth System, as part of a European ERC-synergy grant.


Hosts: Jennifer Holm and Rose Abramoff


1 Cyclotron Road
Berkeley, CA
+ Google Map


Distinguished Scientist Seminar Series Committee