The goal of this proposal is to improve the CarbonTracker flux estimates by optimizing the a priori fluxes from a simple biosphere (SiB) model using atmospheric measurements of OCS and CO2.The proposed work will be focused on the North American domain, where the atmospheric observations are densest. The background condition is a major part of the observed signals for both OCS and CO2. Empirical boundary curtains are built based on observations at the NOAA/ESRL marine boundary layer stations and from aircraft vertical profiles, and are utilized as the lateral boundary conditions for COS and CO2 for North America. These empirical boundary curtains will be evaluated against aircraft observations from the HIPPO campaign. The non-GPP related COS fluxes, i.e. soil uptake, need to be sufficiently identified in order to reduce the uncertainty of the assessment on GPP estimates. The plan is to develop a new soil flux map of COS based on the uptake of molecular hydrogen, which shares the common soil uptake term but lacks other major sinks on the continent. A joint inversion for CO2 and OCS for North America will be performed to optimize selected key parameters of the SiB model based on an ensemble Kalman filter method, which has the ability to simultaneously optimize GPP and respiration. The Stochastic Time-Inverted Lagrangian Transport (STILT) model will be used as the transport model to link the surface fluxes with atmospheric observations of CO2 and OCS. The customized WRF meteorology will be used to drive the STILT model. This work will improve knowledge not only of biospheric carbon fluxes, but also of the underlying processes that control the carbon fluxes, which in turn will enhance understanding of the changing climate system. NOAA ESRL CarbonTracker will benefit from this work by acquiring access to the a priori fluxes from an optimized biosphere model, and a framework to assimilate atmospheric measurements of OCS and CO2. The improved carbon fluxes as well as the intermediate product of the GPP-related plant uptake of OCS that will be made publicly available are important to assess and improve a variety of carbon cycle models.