Validation of Solar Backscatter Radiances Using Antarctic Ice Glen Jaross and Jeremy Warner Science Systems and Applications, Inc. Lanham, Maryland, USA Outline Justification for using ice surfaces The technique, including necessary external information Error budget – where do we focus attention? Results for OMI, TOMS, MODIS, and SCIAMACHY

What products benefit most from scene-based calibration ? Cloud Fractions (λ-independent radiance errors) –Cloud studies –Energy balance and UV irradiance –Cloud height: errors are directly proportional to cloud fraction (low cloud amounts) –Gas vertical column amounts: Air Mass Factor errors directly related to cloud fraction. ( 3% - 5% column NO 2 error per 5% cloud error; low cloud amounts) Aerosol Properties (λ-independent and λ-dependent radiance errors) –0.015 error in single-scatter albedo per 1% radiance error –Optical Depth error per 1%/100 nm λ-dependent radiance errors

Where is scene-based calibration less effective ? Spectral fitting algorithms (e.g. DOAS) –Insensitive to low-order-in-λ calibration errors –Conversion from slant to vertical column still sensitive Gas abundances (slant column) –Need knowledge of abundance to calculate expected radiances, but gas abundance depends upon calibration Limb scattering and Occultation –Normalizing radiances at a reference height nearly eliminates sensitivity to underlying scene reflectance –Most instruments do not have a nadir view

TOMS Earth Probe 360 nm Reflectivity (1996) Antarctica is a good radiance calibration target High Reflectance > direct / diffuse TOA radiance ratio greatest > radiances least affected by clouds and aerosols Low Aerosol Loading Uniform Reflectance Over a Large Area Highly Repeatable (stable) Reflectance R (Lambertian) > %100 %

TOMS Earth Probe 360 nm Reflectivity (1996) Areas with Slope<0.005 radians Data Selection Region Region selected for low surface slope high reflectivity uniformity 90 %100 %

 cloud = 5 at 440 mb Are clouds an issue ? Modeled nadir-scene 357 mn 775 nm albedo ratio GOME (Jan. 2000) nadir-scene 357 mn 775 nm albedo ratio Either the cloud model is wrong, or … Clouds are statistically unimportant

Time Dependence of radiometric calibration Seasonal Cycle: Neglecting terrain height variations Surface reflectance non-uniformity TOMS Nimbus nm TOMS Earth Probe 360 nm OMI (Aura) 360 nm Greenland Antarctica

Comparison between sensors GOME / TOMS-EP Radiance Ratio Very early GOME calibration Comparisons need not be over the same time period 360 nm 331 / 360 nm

Validation of Absolute Radiometry 1.Develop a 2  steradian directional reflectance (BRDF) model for the Antarctic surface; independent of wavelength. 2.Combine BRDF with surface measurements of total hemispheric reflectance measurements; wavelength-dependent 3.Create a look-up table of sun-normalized Top-of-the-Atmosphere (TOA) radiances for all satellite observing conditions using a radiative transfer model 4.Process sensor sun-normalized radiance data from a region of Antarctica chosen for uniformity and low surface slope 5.Compute ratio between each measurement and table entries; average results

Warren et al. Reflectance anisotropy derived from data Spectral Albedo Measuremnts at South Pole, 1986 = 600 nm Sol. ZA = 80  BRDF probably the same: nm Surface properties based upon reflectance measurements by Warren et al. BRDF derived from parameterization of measured reflectance anisotropy

New Reflectance Measurements by Warren et al.  Funded by U.S. National Science Foundation and CNES will support radiometric validation for SPOT4 (Laboratoire de Glaciologie et Géophysique de l’Environnement) data not yet published  Measurements at Dome C, Spectral BRDF of surface 0.35 – 2.5  m Solar Zenith Angles 52  - 87  Measure spectral transmission of sunlight into snow Measurements used for inputs to models for effect of clouds on TOA radiances

Error Budget Surface BRDF model represents single largest error source

Surface BRDF model vs. Solar Zenith Angle SolZA=40  SolZA=60  SolZA=50  SolZA=85 

BRDF is most important at longer wavelengths Simulated Nadir-scene albedos Solar Zenith Angle = 75  Column Ozone = 325 DU BRDF plays bigger role as diffuse / direct ratio decreases Lambertian Non-Lambertian Non-Lambertian / Lambertian Radiance Ratio

OMI Results OMI L1b Data: 7 Dec – 4 Jan, 2004 Perfect model would yield flat SolZA dependence Perfect calibration would yield values = 1 at all wavelengths  Plot suggests probable radiative transfer errors – surface BRDF model – treatment of atmosphere  We believe that results obtained below SolZA = 70  fall within our 2.2% uncertainty estimate

OMI Full spectral range ice radiance results Flat spectral result gives us confidence that result is resonable 62  < SolZA < 68  83  < SolZA < 86  Spectral dependence is not realistic – consistent with BRDF error Apparent error increases at long  as predicted

Shadowing Errors Large scale structures (snow dunes) not captured by ground characterizations From Radarsat-1

Simple linear shadow model for testing errors Tune barrier height and separation to yield flattest SolZA dependence in data

Shadow study using MODIS / Aqua Comparison to RTM, without correction Comparison to RTM, with shadow correction Consistent with ~2% uncertainty estimate

RTM handles ozone poorly at < 330 nm Comparison between MODIS, OMI, TOMS and model radiances OMI / Aura MODIS / Aqua TOMS / EP O 2 O 2 Absorption RTM does not include Ring Effect or O 2 -O 2 abs.

Preliminary SCIAMACHY Results SCIAMACHY Level 1b ( v5.04 ) 18 – 24 Dec., 2004 Provided by R. van Hees, SRON } Ozone Absorption ignored Comparison with RTM over Sahara (from G. Tilstra, KNMI)

Summary Model calculations of TOA radiances over Antarctica are good to approximately 2% at low solar zenith angles (i.e. near Dec. 21) Radiometric characteristics of nadir-viewing sensors can be validated from ~330 nm to ~750 nm Wavelength-to-wavelength radiometry is better than 2%, but not useful for absorption spectroscopy We derive the following sensor calibration errors (preliminary) OMI / Aura: -2.5% (330 < < 500 nm) MODIS / Aqua : -0.5% ( < 500 nm) TOMS / Earth Probe : 0% (331 nm), -1% (360 nm) Future Work : Evaluate more sensors. SCIAMACHY, GOME 2 ? Refine BRDF for improved performance at high SolZA and SatZA

Spares

X-track dependence is mostly Lambertian near SolZA = 50  Results near 50  are least affected by BRDF errors BRDF surface slices at   67 

Same time and geographic location OMI radiances compared directly to MODIS / Aqua band 3 MODIS has broad bandwidth (459 < < 479 nm) which includes O 2 - O 2 absorption OMI MODIS

Developed a 2  steradian BRDF model  Existing parameterization from Warren et al. J. Geophys. Res., 103, 1998  s  50   o  67  all   Extrapolate function for use at  s > 50   Invoke reciprocity (  s > 67   o < 67  )  Fill remaining “hole” (  s < 67   o < 67  ) – assume    s 2 dependence for all  o < 67  – derive  (  s =0) at each  o < 67  from a quadratic parameterization of observed scattering phase fn. (Warren, et al., ibid) ss   (  s = 0) oo