CO 2 Retrieval over Clouds from the OCO Mission: Model Simulations and Error Analysis JE ´ RO ˆ ME VIDOT,* RALF BENNARTZ, AND CHRISTOPHER W. O’DELL 1 Atmospheric and Oceanic Sciences Department, University of Wisconsin—Madison, Madison, Wisconsin RENE ´ PREUSKER AND RASMUS LINDSTROT Institute for Space Sciences, Free University of Berlin, Berlin, Germany ANDREW K. HEIDINGER NOAA/NESDIS Center for Satellite Applications and Research, Madison, Wisconsin (Manuscript received 15 July 2008, in final form 4 November 2008) ABSTRACT Spectral characteristics of the future Orbiting Carbon Observatory (OCO) sensor, which will be launched in January 2009, were used to infer the carbon dioxide column-averaged mixing ratio over liquid water clouds over ocean by means of radiative transfer simulations and an inversion process based on optimal estimation theory. Before retrieving the carbon dioxide column-averaged mixing ratio over clouds, cloud properties such as cloud optical depth, cloud effective radius, and cloud-top pressure must be known. Cloud properties were not included in the prior in the inversion but are retrieved within the algorithm. The high spectral resolution of the OCO bands in the oxygen absorption spectral region around 0.76 mm, the weak CO 2 absorption band around 1.61 mm, and the strong CO 2 absorption band around 2.06 mm were used. The retrieval of all parameters relied on an optimal estimation technique that allows an objective selection of the channels needed to reach OCO’s requirement accuracy. The errors due to the radiometric noise, uncer- tainties in temperature profile, surface pressure, spectral shift, and presence of cirrus above the liquid water clouds were quantified. Cirrus clouds and spectral shifts are the major sources of errors in the retrieval. An accurate spectral characterization of the OCO bands and an effective mask for pixels contaminated by cirrus would mostly eliminate these errors. 1. Introduction The mixing ratio of atmospheric CO 2 has increased from 275 to ;380 ppm in the last 1000 yr. Whereas the first 50-ppm increase was reached in the 1970s, the second 50-ppm increase was achieved in about 30 yr, with a 19-ppm increase occurring between 1995 and 2005. This increase in CO 2 mixing ratios continues to yield the largest sustained radiative forcing of any ra- diative atmospheric component (Solomon et al. 2007). To better understand the relationship between the CO 2 radiative forcing and the human-induced climate change, the scientific community must face the following key question: How is the CO 2 released from fossil fuel combustion, cement production, and land cover change distributed among the atmosphere, oceans, and terres- trial biosphere? To answer this question, the scientific community needs to deeply investigate the carbon cycle (Houghton 2007). The latter issue requires knowledge of the temporal and spatial distribution of atmospheric CO 2 . The major concern in CO 2 flux estimation is placed at the surface where most of the interactions between the three main reservoirs (ocean, atmosphere, and land) take place. However, giving greater importance * Current affiliation: Laboratoire de Me ´te ´ orologie Physique, Universite ´ Blaise Pascal, Clermont-Ferrand, France. 1 Current affiliation: Department of Atmospheric Sciences, Colorado State University, Fort Collins, Colorado. Corresponding author address: Je ´ro ˆ me Vidot, Laboratoire de Me ´te ´ orologie Physique, Universite ´ Blaise Pascal, 24 avenue des Landais, 63000 Clermont-Ferrand, France. E-mail: j.vidot@opgc.univ-bpclermont.fr 1090 JOURNAL OF ATMOSPHERIC AND OCEANIC TECHNOLOGY VOLUME 26 DOI: 10.1175/2009JTECHA1200.1 Ó 2009 American Meteorological Society