A54D-05
The First Year of OCO-2 XCO2 Observations: Bias Correction and Comparison to Models.
Friday, 18 December 2015: 17:00
3012 (Moscone West)
Christopher O'Dell1, Annmarie Eldering2, Christian Frankenberg3, David Crisp2, Michael R Gunson4, Brendan Fisher2, Lukas Mandrake2, James L McDuffie2, Harold R Pollock2, Paul O Wennberg5, Debra Wunch6 and Gary B Doran Jr.2, (1)Colorado State University, Fort Collins, CO, United States, (2)NASA Jet Propulsion Laboratory, Pasadena, CA, United States, (3)NASA Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, United States, (4)Jet Propulsion Laboratory, Pasadena, CA, United States, (5)California Institute of Technology, Division of Engineering and Applied Science, Pasadena, CA, United States, (6)California Institute of Technology, Pasadena, CA, United States
Abstract:
Observations of atmospheric carbon dioxide from the Orbiting Carbon Observatory-2 (OCO-2) have the potential to be revolutionary in their impact on our understanding of carbon sources and sinks. For this to be achieved, however, requires the observations to have sub-ppm systematic errors; the large data density of OCO-2 generally means that random errors will be of lesser importance in terms of regional scale fluxes. In this presentation we report on results from the Atmospheric Carbon Observations from Space (ACOS) algorithm as applied to the first year of OCO-2 observations, with a particular focus on filtering and bias-correction of the OCO-2 data “Nadir” and “Glint” mode data. In general, we find the random errors to be low (0.5-2.0 ppm), especially for ocean glint retrievals, consistent with the higher signal-to-noise ratio and reduced effects of aerosols over ocean. Systematic errors are explored via comparison to TCCON, low-variability southern hemisphere data, and data taken over small spatial regions. These are used to form a multi-linear bias correction, similar that that implemented for ACOS/GOSAT observations. Additionally we present comparisons of the first year of bias-corrected OCO-2 observations to inverse model output. This comparison will shed light on potential retrieval biases still lurking in the OCO-2 data, such as from surface albedo, aerosol effects, and other error sources.
