MISR L1 Georectified radiance product (MI1B2E, MI1B2T)
Radiometrically and geometrically calibrated, Earth projected multiangle imagery. Radiometric uncertainty estimates are ±4% absolute, ±1-2% relative. Radiances from all 4 bands and 9 angles are mapped to a Space Oblique Mercator (SOM) projection. Camera-to-camera co-registration meets the < 1 pixel (275 m) specification. Radiometric calibration and stability are verified and maintained through bimonthly deployments of the on-board calibrator. Ongoing assessments of image geolocation and co-registration are performed to ensure quality. Product is Stage 3 validated.
MISR L2 Stereo product (MIL2TCST): heights, winds
Stereoscopically-derived cloud top heights (CTH), aerosol plume-top heights, and atmospheric motion vector (AMV) winds. Geometric parallax and computationally fast image matching yields instantaneous CTH uncertainty of < 1 km and AMV uncertainty of a few m/sec. Stereo CTH measurements are insensitive to atmospheric temperature profile, emissivity, and radiometric calibration drifts, providing climate diagnostics independent of common error sources. The CTH and AMV are Stage 3 and Stage 2 validated, respectively.
MISR L2 Cloud masks and cloud classifiers product (MIRCCM, MIL2TCST, MIL2TCCL): radiometric, stereo, angular signature
Three independently-derived cloud masks. The Radiometric Camera-by-camera Cloud Mask (RCCM) uses spectral and spatial measures to provide a separate mask for each of the 9 MISR cameras. The Stereoscopically-Derived Cloud Mask (SDCM) detects clouds as a result of their elevation above the surface terrain, and is part of the Level 2 Stereo Product. The Angular Signature Cloud Mask (ASCM) is part of the Level 2 Cloud Classifiers Product, and is derived from a band differenced angular signature. SSM/I-based snow/ice mask, available from NSIDC, is used to generate climatologies to support production. The RCCM is Stage 3 validated (Zhao and DiGirolamo, 2004, 2006; Zhao, 2006; Yang, 2007; Yang et al., 2007). Over ice-free ocean it has a classification accuracy of 94% (96% excluding sunglint) for the nadir camera, and increases slowly with view angle to a classification accuracy of 96% (98% excluding sunglint) for the most oblique cameras. RCCM over snow-free land has a classification accuracy of 92% for all cameras, averaged over all land surface types. The SDCM/ASCM are Stage 2 validated (Shi et al., 2007; Di Girolamo and Wilson, 2003).
MISR L2 Top-of-atmosphere albedo product (MIL2TCAL); BRF, albedo
Spectral bidirectional reflectance factors (BRFs) and
albedos projected to cloud-top altitudes for cloudy scenes and to the surface elevation for clear scenes. MISR generates local albedos at 2.2 km resolution. Two coarse resolution albedos, defined over 35.2 km regions, are also defined. For comparative purposes, especially with products from CERES, the “restrictive” albedo entails angular integration of the observed BRFs over a given region only, and the “expansive” albedo integrates over all surrounding regions influencing the TOA radiative flux, simulating what a pyranometer would observe. MISR is unique in the sense that its high-resolution spectral BRFs and albedos cannot be compared directly with other products; therefore, progression to higher validation stages is limited by availability of independent data sources. Given the maturity level of the precursor data products (radiances and stereo heights), the BRFs are Stage 3 validated. Based on a global comparison
with CERES (including narrow-to-broadband conversions) for overcast ocean, spectral albedos are Stage 1 validated.
MISR L2 Aerosol product (MIL2ASAE): optical depth, particle properties
Ocean and land aerosol optical depth (AOD), Angstrom exponent, single scattering albedo (SSA), and optical depth due to small, medium, large, spherical, and nonspherical particles. Applications are to air quality and aerosol-climate forcing. Global comparisons of MISR and groundbased AERONET sunphotometer data shows that 63% of the mid-visible AODs fall within 0.05 or 20% * AOD of AERONET, and about 40% are within 0.03 or 10% * AOD. AOD is Stage 3 validated; particle properties are Stage 2 validated (Liu et al., 2004; Martonchik et al., 2004; Redemann et al., 2005; Abdou et al., 2005; Kahn et al., 2005; 2007, 2009, 2010).
MISR L2 Surface product (MIL2TCLS): albedos/BRFs, LAI
Atmospherically-corrected land surface spectral directional reflectance factors and associated hemispherical reflectances (albedos), bidirectional reflectance model (Rahman-Pinty-Verstraete) parameter fits to retrieved surface BRFs (which characterize surface reflectance angular anisotropy), leaf area index (LAI), and fractional amount of PAR (FPAR) absorbed by green vegetation. Compensation for atmospheric effects uses MISR aerosol retrievals. Validation shows typical albedo errors of ±0.005, and BRF errors of ±0.01. Surface albedos and BRFs are Stage 3 validated; other parameters are Stage 2 (Abdou et al., 2006; Lyapustin et al., 2006; Lavergne et al., 2006; Pinty et al., 2007; Pocewicz et al., 2007; Hu et al., 2007; Chen et al., 2008; Schaaf et al., 2008; Taberner et al., 2010).
This product includes, monthly, seasonal, and annual summaries of the aformentioned MISR Level 1 and Level 2 input data products, globally gridded at 0.5° resolution. Modelers use the global statistics to improve, tune, and validate their models, examples of which include those dealing with cloud-climate and aerosolclimate interactions, aerosol transport, and ecosystem productivity. As part of the verification process the MISR science team collaborates with members of the modeling community to ensure that product content, utility, and accessibility are consistent with their needs. Product maturity levels and user communities track the Level 1 and Level 2 inputs.
This product has been generated since 2007, and is a publicly-available repository for stereoscopically-derived smoke plume injection heights from MISR and MODIS radiative power to support wild fire, climate change and air quality studies. Plume heights are retrieved using the MISR INteractive eXplorer (MINX) tool. The database has now been expanded to include smoke data from Africa and Siberia, and ash plume data from the Eyfjallajökull volcanic eruption. Smoke results for Indonesia and Australia, more volcanic eruption data, and a dust plume climatology will be added.
The MISR Level 3 Cloud Fraction by Altitude product offers the highest vertical resolution cloud climatology available from a passive instrument, providing the frequency of cloud occurrence partitioned into different cloud top height bins at a global and monthly scale with a spatial resolution of 0.5° × 0.5° latitude/longitude and vertical resolution of 500 m. For each height bin, the frequency of cloud occurrence of a region over a time period is represented by the temporal mean of the spatial coverage of cloud tops. The spatial coverage of clouds is referred to as cloud fraction, which is defined in this document as the ratio of the number of cloudy pixels to the total number of cloudy and cloud-free pixels observed by the instrument. Clouds are assigned to height bins based on their top height as retrieved by the MISR stereoscopic technique. This product makes use of several unique strengths of multiangle observation: multiple cloud detection methodologies (dynamic radiance thresholding, stereoscopy, and angular signature); use of stereoscopic height retrieval, which is insensitive to radiometric calibration and atmospheric thermal structure; and
This product contains retrievals of cloud motion determined by geometrically triangulating the position and motion of cloud features observed by MISR from multiple perspectives and times during the ~7 minute overpass of Terra over each cloud scene. Cloud motion vectors (CMVs) are a valuable proxy observation of the horizontal atmospheric wind field at the retrieved altitude of the cloud. MISR CMVs have been and continue to be operationally produced as part of the publicly available Level 2 Stereo product. The CMVP provides users a complete global list of the highest quality CMVs extracted from the standard Stereo product, distributed as monthly, seasonal, and annual NetCDF files that are neither gridded nor averaged. The annual files, the largest of these, are a manageable 160MB, facilitating scientific applications requiring CMV information spanning multiple months or years. An ATBD has been released (Mueller, et al, 2010).