UMass Boston

Due to its three dimensional structure, the Earth’s surface scatters radiation in an anisotropic manner, especially at the shorter wavelengths that characterize solar irradiance. The Bidirectional Reflectance Distribution Function (BRDF) specifies the behavior of surface scattering as a function of illumination and view angles at a particular wavelength. The albedo of a surface describes the ratio of radiant energy scattered upward and away from the surface in all directions to the down-welling irradiance incident upon that surface. The completely diffuse bihemispherical (or white-sky) albedo can be derived through integration of the BRDF for the entire solar and viewing hemisphere, while the direct beam directional hemispherical (or black-sky) albedo can be calculated through integration of the BRDF for a particular illumination geometry. Actual clear sky albedo (blue-sky albedo) under particular atmospheric and illumination conditions can be estimated as a function of the diffuse skylight and a proportion between the black-sky and white-sky albedos. The Visible Infrared Imaging Radiometer Suite (VIIRS) BRDF/Albedo Science Data Products provide white-sky albedos and black-sky albedos (at local solar noon) as both spectral and broadband quantities (VNP43IA3,VNP43MA3, and VNP43DNBA3, where M stands for moderate resolution bands, I stands for imagery resolution bands and DNB stands for the Day/Night Band of VIIRS).

The VIIRS BRDF/Albedo Products also provide Nadir BRDF-Adjusted Reflectances (NBAR)—- surface reflectances corrected to a common nadir view geometry at the local solar noon zenith angle of the day of interest (VNP43IA4, VNP43MA4 and VNP43DNBA4). These anisotropically-corrected surface reflectances can serve as important inputs for studies using vegetation indices and for land cover classification efforts (they are used directly as the primary input to the VIIRS Phenology Product (PI, X. Zhang).

The BRDF specification itself is supplied to the scientific community as a separate product (VNP43IA1, VNP43MA1 and VNP43DNBA1) since it is useful in specifying a surface radiation scattering model for boundary layer parameterization of surface vegetation atmospheric transfer schemes in global climate models. With the model weighting parameters (fiso, fvol, fgeo) and a simple polynomial equation, black-sky albedo can be realistically estimated at any solar zenith angle a user may require. And, since the BRDF is an intrinsic property characterizing the structure of the surface, the parameters themselves may also provide biophysical information of interest.

Algorithm

Every day, the NASA standard VIIRS BRDF/Albedo algorithm makes use of 16 days’ worth of multi-date data from Suomi National Polar-orbiting Partnership (NPP) with a semiempirical kernel-driven bidirectional reflectance model to determine a global set of parameters describing the BRDF of the land surface (VNP43IA1, VNP43MA1 and VNP43DNBA1) (see Liu et al., 2017). The high quality input observations are weighted as a function of proximity to the day of interest and the observation coverage. The day of interest (the center date of the retrieval period) will be most heavily weighted. These gridded BRDF model parameters are then used to determine directional hemispherical reflectance (“black-sky albedo” at the solar zenith angle of local solar noon), and bihemispherical reflectance (“white-sky albedo”) for the spectral bands (I1, I2, I3 at a 500m scale, and M1- M5, M7-M8, M10-M11, DNB at a 1km scale) and three broad bands (0.3-0.7µm, 0.7-5.0µm, and 0.3-5.0µm). Note that the VNP43IA1,2,3,4 products do not refer to the individual I bands but provide the 4 BRDF albedo products (BRDF parameters, Quality, albedo quantities, and NBAR) for all three of the 500m I bands. Similarly with the VNP43MA1, 2, 3, 4 products, note that only the 1km scale has sufficient bands to estimate a broadband product while the 500m scale does not. Thus only at 1km scale, can the BRDF and albedos be generated for the three broadbands (0.3–0.7 visible, 0.7–5.0 NIR, and 0.3–5.0 nm shortwave) through the use of narrowband to broadband conversion coefficients. The daily Suomi-NPP VIIRS standard algorithm (Liu et al., 2017; Wang et al., 2013; Schaaf et al., 2011; 2002; Lucht et al., 2000) relies on a combination of the RossThick-LiSparseReciprocal kernels (Wanner et al., 1995; 1997)as the semiempirical model used to invert 16 days’ worth of cloud clear, atmospherically corrected, VIIRS directional surface reflectance data and to fit a BRDF to each land surface pixel. Broadband values (using Narrow-to-Broadband conversion factors) are computed as well. In the future, VIIRS data on board the Joint Polar Satellite System (JPSS) satellites could also be used. The semiempirical kernel-driven BRDF model (Roujean et al., 1992) represents the weighted sum of an isotropic parameter (fiso) and two functions (or kernels) of viewing and illumination geometry. One of these kernels (Kvol) is derived from volume scattering radiative transfer models (Ross, 1981), while the other (Kgeo) is derived from surface scattering and geometric shadow casting theory (Li and Strahler, 1987).

The BRDF parameters (fiso, fvol, fgeo) computed in the standard product are the spectrally dependent weights of each of these kernels used in forming the overall reflectance:

R = f_iso + f_vol K_vol + f_geo K_geo

When insufficient high quality reflectances are available (currently set to less than seven observations) or even more importantly, a poorly representative sampling of high quality reflectances is obtained (as indicated in the quality flags and determined through the weights of determination), it is not possible to perform a high quality full inversion. Instead, use is made of a database of a priori BRDF parameters to supplement the observational data available and perform a lower quality magnitude inversion. This database is currently updated from the latest high quality full inversion retrieved for each pixel.

The VIIRS BRDF/Albedo Science Data Products are provided in a Sinusoidal Grid (SIN) projection with standard tiles representing 10 degree by 10 degree (2400 by 2400 pixels for 500m scale and 1200 by 1200 pixels for 1km scale) on the Earth. While the projection becomes increasingly sheared poleward, the equal area properties of the SIN projection mean that it is a good data storage format, and it is possible to convert each tile to other, more common projections through the use of any one of a number of commercial or public software packages. These Level-3 VIIRS Land products are being released in version 5 Hierarchical Data Format - Earth Observing System (HDF-EOS5) for each of the 10 degree by 10 degree land tiles.

The standard product is associated with mandatory quality and additional extensive quality assurance information stored in VNP43IA2, VNP43MA2 and VNP43DNBA2, so that users can reconstruct the processing methodology used for each tile or pixel if they choose. At a minimum, all VIIRS Land products supply a mandatory per-pixel quality flag indicating whether the algorithm produced results or not for that pixel and if so, whether the result is of the highest quality or whether (due to some uncertainties in the processing) the user should check the extensive additional product-specific quality assurance to make sure the output is appropriate for their application. Note that the per-pixel data and the quality information are currently computed for all land, inland water and coastal areas. The products and quality flags are not currently computed for open ocean regions.

Data Flow

The Level 2 Surface Reflectance Product for VIIRS (VNP09) provides daily, cloud-cleared, atmospherically-corrected surface reflectances†. The data from the moderate resolution bands (M1–M5, M7–M8 and M10–M11), the imagery resolution bands (I1, I2, I3) and DNB are stored in Level 2Glite gridded SIN tiles. The M bands are gridded to 1km, while the I bands are gridded to 500m. This gridding and binning occurs on a daily basis and at the higher latitudes, all layers of valid data for each day will be used for the retrieval for each pixel. The data from sixteen days’ worth of VNP09 are then used as the primary input for the VNP43 BRDF/Albedo Product. VNP43M* is the 1km SIN product for the M bands, the VNP43I* is the 500m SIN product for the I bands and VNP43DNB* is the 1km SIN product for the Day/Night Band (DNB). The algorithm then fits the BRDF model to these directional surface reflectances and the parameters of the BRDF model (RossThick-LiSparseReciprocal) are provided to the community as Science Data Products (VNP43IA1, VNP43MA1 and VNP43DNBA1) for all the appropriate bands. These same parameters are used to compute the albedo measures provided in VNP43IA3, VNP43MA3 and VNP43DNBA3 and the NBAR values provided in VNP43IA4, VNP43MA4 and VNP43DNBA4. These BRDF parameters of the M bands are also produced in a 30 arc-second resolution Climate Modeling Grid (CMG) in a global geographic lat/long projection (VNP43D*). There are also 0.05 degree resolution Climate Modeling Grid (CMG) Products produced in a global geographic lat/long projection (VNP43C1 (BRDF parameters), VNP43C2 (snowfree BRDF parameters), VNP43C3 (albedo quantities), and VNP43C4 (NBAR)). Users are cautioned that the coarser resolution CMG quality flags only indicate the majority quality of the underlying 30 arc-second data and for that reason the 30arc second data is preferable for quantitative research.

†Please note that as of summer 2019, the VIIRS Land Surface Reflectance product (VNP09) seems to consistently flag a larger number of retrievals as ‘High Aerosol’ in the quality data set as compared to the heritage MODIS reflectance product generated using MODIS observations. This is observed globally but specifically impacts regions in the mid to high latitude, arid lands, and tropical vegetative areas. Inaccurately flagged high aerosol quality flags in VNP09 directly impact the quality of the downstream VIIRS BRDF Albedo product, causing the BRDF retrieval algorithm to incorrectly reject these high aerosol observations and switch from full inversion (high quality retrieval) to magnitude inversion (low quality retrieval) in the case of less than required number of observations to perform the full inversion. https://landweb.modaps.eosdis.nasa.gov/cgi-bin/NPP_QA/displayCase.cgi?esdt=NPP_SRFL&caseNum=PM_NPP_SRFL_19200&caseLocation=cases_data

Separately, there appears to be erroneous blocky pattern in input surface reflectance product in the vicinity of inland water bodies. The science team is aware of this problem and is currently implementing a correction for the next reprocessing.

References Cited

Li, X., and A. H. Strahler, Geometric-optical bidirectional reflectance modeling of the discrete crown vegetation canopy: Effects of crown shape and mutual shadowing, IEEE Trans. Geosci. Remote Sens., 30, 276-292, 1992.

Liu, Y., Z. Wang, Q. Sun, A. Erb, C. Schaaf, W. Zhang, M.O. Román, R. Scott, Q.Zhang, K. Novick, M. S. Bret-Harte, S. Petroy, M. SanClements, Evaluation of the VIIRS BRDF, Albedo and NBAR Products and an assessment of continuity with the long term MODIS record, Remote Sens. Environ. 2017 (in review). 

Lucht, W., C.B. Schaaf, and A.H. Strahler. An Algorithm for the retrieval of albedo from space using semiempirical BRDF models, IEEE Trans. Geosci. Remote Sens., 38, 977-998, 2000.

Ross, J., The radiation regime and architecture of plant stands, Dr. W. Junk, Norwell, MA, 392 pp, 1981.

Roujean, J. L., M. Leroy, and P.Y. Deschamps, A directional reflectance model of the Earth's surface for the correction of remote sensing data, J. Geophys. Res., 20, 455-468, 1992.

Schaaf, C. B., F. Gao, A. H. Strahler, W. Lucht, X. Li, T. Tsang, N. C. Strugnell, X. Zhang, Y. Jin, J.-P. Muller, P. Lewis, M. Barnsley, P. Hobson, M. Disney, G. Roberts, M. Dunderdale, C. Doll, R. d'Entremont, B. Hu, S. Liang, and J. L. Privette, First Operational BRDF, Albedo and Nadir Reflectance Products from MODIS, Remote Sens. Environ., 83, 135-148, 2002.

Schaaf, C. L. B., J. Liu, F. Gao and A. H. Strahler, MODIS Albedo and Reflectance Anisotropy Products from Aqua and Terra, In Land Remote Sensing and Global Environmental Change: NASA's Earth Observing System and the Science of ASTER and MODIS, Remote Sensing and Digital Image Processing Series, Vol.11, B. Ramachandran, C. Justice, M. Abrams, Eds, Springer-Cerlag, 873 pp., 2011.

Wang, Z., C. B. Schaaf, M. J. Chopping, A. H. Strahler, J. Wang, M. O.Román, A. V. Rocha, C. E. Woodcock,Y. Shuai, Evaluation of Moderate-resolution Imaging Spectroradiometer (MODIS) snow albedo product (MCD43A) over tundra, Remote Sensing of Environment, 117, 264-280, 2012.

Wanner, W., A.H. Strahler, B. Hu, P. Lewis, J.-P Muller, X. Li, C. Schaaf, and M.J. Barnsley, Global retrieval of bidirectional reflectance and albedo over land from EOS MODIS and MISR data: Theory and algorithm, J. Geophys. Res., 102, 17143-17161, 1997.

Wanner, W., X. Li, and A. H. Strahler, On the derivation of kernels for kernel-driven models of bidirectional reflectance, J. Geophys. Res., vol. 100, pp. 21077--21090, 1995.