The measurement requirement of the orbiting carbon observatory (OCO-2) is the determination of column average CO2 (i.e. XCO2) to 1 ppmv on a regional basis. Since CO2 is near 400 ppmv, this 0.25 %…

This is a research grant awarded by the National Aeronautics and Space Administration (NASA) to the University of Colorado, doing business as the Office of Contracts and Grants Division, to study the impact of 3D cloud and surface effects on the accuracy of carbon dioxide (CO2) measurements from the Orbiting Carbon Observatory-2 (OCO-2) satellite. The $543,955.44 grant will fund the development of algorithms to detect 3D cloud effects, quantify associated biases in retrieved CO2 data, and assess the impact of 3D surface reflectance heterogeneity. The researchers will analyze OCO-2 spectral data, perform 3D radiative transfer calculations, and work with the OCO-2 algorithm team to potentially incorporate the new methods into the operational data processing. No set-aside designation is mentioned for this grant.

The measurement requirement of the orbiting carbon observatory (OCO-2) is the determination of column average CO2 (i.e. XCO2) to 1 ppmv on a regional basis. Since CO2 is near 400 ppmv, this 0.25 % measurement is challenging. Our proposed research addresses physical effects (3D radiative transfer effects due to clouds and surface reflectivity) that are not in the current operational retrieval, yet directly impacts the biases and standard deviations in retrieved XCO2. Our previous work demonstrated that 3D cloud effects increase as the cloud heterogeneity, and associated radiance field heterogeneity, increases. XCO2 standard deviations increase from 0.98 ppmv for clear scenes upwards to 2.6 ppmv for cloudy scenes, demonstrating that 3D cloud effects add an important component to the XCO2 error budget. The quantification and mitigation of XCO2 biases that are associated with 3D cloud effects is a primary goal of this proposal. Our objectives are to a) develop a 3D cloud effect detection algorithm (based upon stand-alone OCO-2 radiance data), b) develop 3D cloud effect bias parameterizations (expressed as a function of 3D cloud field heterogeneity, molecular line structure, and viewing geometry metrics), c) carry out radiative transfer calculations that test our understanding of how 3D cloud effects arise at the molecular line structure level, and d) address the issue of 3D effects due to surface reflectance heterogeneity. The research will be based upon an analysis of OCO-2 "lite file" XCO2 data, 3D radiative transfer calculations, a thorough analyses of how 3D cloud effects are manifested in OCO-2 spectra as a function of wavelength-dependent optical depth (i.e. molecular line structure), and applications of the publicly available OCO-2 retrieval software. Several dozen target, glint, and nadir mode scenes will be analyzed, from clear sky scenes to cloudy scenes of increasing heterogeneity and various viewing geometries. 3D cloud heterogeneity metrics for each scene will be specified from separate analyses of modis and OCO-2 radiance fields, while 3D radiative transfer calculations will specify molecular line structure metrics. We will calculate OCO-2 radiance spectra with and without the 3D cloud effects for an observation, for all three OCO-2 spectral bands, by applying our OCO3DS 3D radiance simulator. This simulator is based upon the mcarats 3D radiative transfer code. The OCO-2 retrieval software will calculate two XCO2 values based upon the original observed spectra, and the original spectra modified by the two sets of OCO3DS spectra. The difference in the two retrieved XCO2 values then becomes one measure of the 3D cloud effect bias. We will interact with the OCO-2 algorithm team to determine if the 3D cloud effect detection algorithm and the 3D cloud effect bias parameterizations can be applied in the post-retrieval "bias correction" processing to improve OCO-2 lite file XCO2. Code will be delivered to the team such that the 3D cloud effect detection algorithm and bias correction parameterizations can be updated as improvements in the OCO-2 calibration and retrieval program are implemented.