On two potential sources of systematic error in

On two potential sources of systematic error in retrievals using limb-scattered sunlight Chris A. Mc. Linden (1) and Christopher E. Sioris (2) (1) Meteorological Service of Canada, Toronto, Ontario, Canada (chris. mclinden@ec. gc. ca) (2) Atomic and Molecular Physics Division, Harvard. Smithsonian Center for Astrophysics, Cambridge, Massachusetts, USA 14 -16 April 2003 International Limb Workshop, Bremen

Outline 1. Allowing for a varying atmosphere along the line-of-sight (LOS) when performing retrievals 2. Systematic errors in plane-parallel models, including so-called “pseudospherical” limb scattering models. 14 -16 April 2003 International Limb Workshop, Bremen

Varying atmosphere along the LOS (1) w Perform multiple plane-parallel calculations, each one for a particular atmosphere & SZA along the LOS w From these, calculate LOS sources function – integrate to obtain limb radiance w Using vector RT model of Mc. Linden et al. (Can. J. Phys. , 80, 375 -383, 2002) 14 -16 April 2003 International Limb Workshop, Bremen

Varying atmosphere along the LOS (2) - in practice, slices narrower near tangent point 2 N 14 -16 April 2003 1 International Limb Workshop, Bremen OS or SCIA

Varying atmosphere along the LOS (3) w Want to estimate magnitude of errors when atmosphere is not homogeneous along LOS, but assumed to be in retrieval (‘ 2 D’ forward, ‘ 1 D’ inversion) w If species diurnal, can a ‘ 2 D’ retrieval improve result for scans near terminator and SZA variations along LOS ? 14 -16 April 2003 International Limb Workshop, Bremen

Varying atmosphere along the LOS (4) 14 -16 April 2003 International Limb Workshop, Bremen

Varying atmosphere along the LOS (4) - Flittner algorithm used -‘ 2 D’ forward calculation, ‘ 1 D’ inversion 14 -16 April 2003 International Limb Workshop, Bremen

Varying atmosphere along the LOS (5) 14 -16 April 2003 International Limb Workshop, Bremen

Varying atmosphere along the LOS (6) -Sioris NO 2 inversion algorithm used -Uses an iterative onion peel technique SZA=27º, d. AZ=92º; Aug. , 25ºN, albedo=0 14 -16 April 2003 International Limb Workshop, Bremen

Varying atmosphere along the LOS (7) SZA=27º, d. AZ=92º; Aug. , 25ºN, albedo=0 14 -16 April 2003 International Limb Workshop, Bremen

Varying atmosphere along the LOS (8) SZA=27º, d. AZ=92º; Aug. , 25ºN, albedo=0 14 -16 April 2003 International Limb Workshop, Bremen

Varying atmosphere along the LOS (9) Comparison of ‘ 1 D’ and ‘ 2 D’ OSIRIS inversions with HALOE coincidence LOS scaling factors from photochemical box model 4 June 2002 lat=4ºN, lon=55ºW 14 -16 April 2003 International Limb Workshop, Bremen

Systematic Errors in “Pseudo. Spherical” Limb Radiances (1) - Comparison of several “pseudo-spherical” (PS) radiative transfer model calculations with 3 -D, spherical Monte-Carlo (MC) model calculations show systematic errors - These errors increase with (a) tangent height and (b) surface albedo 14 -16 April 2003 International Limb Workshop, Bremen

Systematic Errors in “Pseudo. Spherical” Limb Radiances (2) SZA=39º; albedo=0. 95 14 -16 April 2003 International Limb Workshop, Bremen

Systematic Errors in “Pseudo. Spherical” Limb Radiances (3) -PS calculations based on plane-parallel source function & integrated through spherical-shell atmosphere -These errors are driven by an overestimation of the surface reflected component – either the surface proper or the effective Rayleigh “surface” 14 -16 April 2003 International Limb Workshop, Bremen

Systematic Errors in “Pseudo. Spherical” Limb Radiances (4) -root of problem is how much of the surface a point in the atmosphere “sees”: -PP atmosphere: point sees 2 [str] -spherical atmosphere: point sees <2 [str]; this decreases with altitude 14 -16 April 2003 International Limb Workshop, Bremen

Systematic Errors in “Pseudo. Spherical” Limb Radiances (5) Direct Point under consideration Surface reflected 14 -16 April 2003 International Limb Workshop, Bremen

Systematic Errors in “Pseudo. Spherical” Limb Radiances (6) - try simple, near-transparent atmosphere correction factor using ratio of PS to spherical solid angles Where 14 -16 April 2003 is cos(SZA) and is surface albedo International Limb Workshop, Bremen

Systematic Errors in “Pseudo. Spherical” Limb Radiances (7) 800 nm SZA=39º 14 -16 April 2003 International Limb Workshop, Bremen

Systematic Errors in “Pseudo. Spherical” Limb Radiances (8) -Trace gas retrievals should be minimally impacted since there will be cancellation/removal of this effect - e. g. , the Flittner radiance-ratio method there will be a large degree of cancellation due to form of retrieval vector; - in DOAS it will not cancel but it would be included or removed in closure polynomial or high pass filter 14 -16 April 2003 International Limb Workshop, Bremen

Systematic Errors in “Pseudo. Spherical” Limb Radiances (9) -Aerosol retrievals may suffer systematic errors if the retrieval involves a modelmeasurement matching of radiances -If aerosol signal small, the PP errors can lead to amplified aerosol errors; unclear how “optimal”-type solutions would handle this 14 -16 April 2003 International Limb Workshop, Bremen

Systematic Errors in “Pseudo. Spherical” Limb Radiances (10) -If aerosol signal is: S = I(measure) – I(model w/o aerosol) -But P-P aerosol signal is: SPP = I(measure) – I(model w/o aerosol) / c Where c= 14 -16 April 2003 International Limb Workshop, Bremen

Systematic Errors in “Pseudo. Spherical” Limb Radiances 14 -16 April 2003 International Limb Workshop, Bremen

Systematic Errors in “Pseudo. Spherical” Limb Radiances (12) 14 -16 April 2003 International Limb Workshop, Bremen

Systematic Errors in “Pseudo. Spherical” Limb Radiances (13) 14 -16 April 2003 International Limb Workshop, Bremen

Conclusions & Future Work -systematic PS errors may be important in OS aerosol retrievals + other RT applications including photochemistry (& other planetary atmospheres) -attempt to obtain “universal” correction valid at all wavelengths, tangent heights, albedo 14 -16 April 2003 International Limb Workshop, Bremen

Systematic Errors in “Pseudo. Spherical” Limb Radiances (14) 14 -16 April 2003 International Limb Workshop, Bremen