The NASA Clouds and the Earth’s Radiant Energy System (CERES) project relies on top-of-atmosphere (TOA) broadband fluxes derived from geostationary (GEO) satellite imagery to account for the diurnal flux variations between the CERES observation intervals, and thereby produce a synoptic gridded (SYN1deg) product based on continuous temporal observations. Consistent broadband flux derivation depends on accurate radiative property measurements and cloud retrievals, which largely determine the radiance-to-flux conversion process. Therefore, it is important to ensure a high quality of cloud property input in order to maintain a reliable broadband flux record. In Edition 4 of the CERES SYN1deg product, a robust automated image anomaly detection algorithm based on inter-line and inter-pixel differences, spatial variance, and 2-D Fourier analysis has been successful in identifying imagery with linear artifacts, but the line-by-line inspection and cleaning process must still be performed by a human. Therefore, further automation of this quality assurance process is warranted, especially considering the excessive amount of additional cleaning necessitated by the GOES-17 Advance Baseline Imager (ABI) cooling system anomaly. As such, this article highlights advancement of the CERES GEO image artifact cleaning approach based on a convolutional neural network (CNN) for classification of bad scanlines. Once trained, the CNN approach is a computationally inexpensive means to ensure greater consistency in cloud retrievals, and therefore broadband flux derivation, based on GOES-17 measurements.
For accurate cloud ceiling information, a data fusion approach is proposed that utilizes satellite data to extend surface station information to much wider areas. Cloud base height (CBH) retrieved from satellite observations provides for much larger spatial coverage and higher resolution. The direct comparison of GOES-16 CBH with surface station ceiling yields a local bias that has to be corrected for in the initial GOES-16 cloud base information. This sparsely sampled bias correction presents an irregular 2D mesh of control points, which is then interpolated by constructing a continuous smooth field using polyharmonic splines. The influence of remote stations is restricted by grouping the control points into clusters depending on an effective distance. This cluster-based approach allows for constructing separate spline surfaces corresponding to physically different clouds. The obtained continuous bias correction function is then applied to the entire GOES-16 pixel level CBH except for areas far away from surface stations in data sparse regions such as offshore. The described method is currently being tested using daytime-only observations over the central and eastern United States. Overall, this approach has potential to provide more accurate, high spatial resolution cloud ceiling information for the aviation community.
The Deep Space Climate Observatory (DSCOVR) enables analysis of the daytime Earth radiation budget via the
onboard Earth Polychromatic Imaging Camera (EPIC) and National Institute of Standards and Technology Advanced
Radiometer (NISTAR). Radiance observations and cloud property retrievals from low earth orbit and geostationary
satellite imagers have to be co-located with EPIC pixels to provide scene identification in order to select anisotropic
directional models needed to calculate shortwave and longwave fluxes.
A new algorithm is proposed for optimal merging of selected radiances and cloud properties derived from multiple
satellite imagers to obtain seamless global hourly composites at 5-km resolution. An aggregated rating is employed to
incorporate several factors and to select the best observation at the time nearest to the EPIC measurement. Spatial
accuracy is improved using inverse mapping with gradient search during reprojection and bicubic interpolation for pixel
resampling.
The composite data are subsequently remapped into EPIC-view domain by convolving composite pixels with the EPIC
point spread function defined with a half-pixel accuracy. PSF-weighted average radiances and cloud properties are
computed separately for each cloud phase. The algorithm has demonstrated contiguous global coverage for any
requested time of day with a temporal lag of under 2 hours in over 95% of the globe.
The Clouds and the Earth’s Radiant Energy System (CERES) has incorporated imagery from 16 individual geostationary (GEO) satellites across five contiguous domains since March 2000. In order to derive broadband fluxes uniform across satellite platforms it is important to ensure a good quality of the input raw count data. GEO data obtained by older GOES imagers (such as MTSAT-1, Meteosat-5, Meteosat-7, GMS-5, and GOES-9) are known to frequently contain various types of noise caused by transmission errors, sync errors, stray light contamination, and others. This work presents an image processing methodology designed to detect most kinds of noise and corrupt data in all bands of raw imagery from modern and historic GEO satellites. The algorithm is based on a set of different approaches to detect abnormal image patterns, including inter-line and inter-pixel differences within a scanline, correlation between scanlines, analysis of spatial variance, and also a 2D Fourier analysis of the image spatial frequencies. In spite of computational complexity, the described method is highly optimized for performance to facilitate volume processing of multi-year data and runs in fully automated mode. Reliability of this noise detection technique has been assessed by human supervision for each GEO dataset obtained during selected time periods in 2005 and 2006. This assessment has demonstrated the overall detection accuracy of over 99.5% and the false alarm rate of under 0.3%. The described noise detection routine is currently used in volume processing of historical GEO imagery for subsequent production of global gridded data products and for cross-platform calibration.
Spatial cross-talk has been discovered in the visible channel data of the Multi-functional Transport Satellite (MTSAT)-1R. The slight image blurring is attributed to an imperfection in the mirror surface caused either by flawed polishing or a dust contaminant. An image processing methodology is described that employs a two-dimensional deconvolution routine to recover the original undistorted MTSAT-1R data counts. The methodology assumes that the dispersed portion of the signal is small and distributed randomly around the optical axis, which allows the image blurring to be described by a point spread function (PSF) based on the Gaussian profile. The PSF is described by 4 parameters, which are solved using a maximum likelihood estimator using coincident collocated MTSAT-2 images as truth. A subpixel image matching technique is used to align the MTSAT-2 pixels into the MTSAT-1R projection and to correct for navigation errors and cloud displacement due to the time and viewing geometry differences between the two satellite observations. An optimal set of the PSF parameters is derived by an iterative routine based on the 4-dimensional Powell’s conjugate direction method that minimizes the difference between PSF-corrected MTSAT-1R and collocated MTSAT-2 images. This iterative approach is computationally intensive and was optimized analytically as well as by coding in assembly language incorporating parallel processing. The PSF parameters were found to be consistent over the 5-days of available daytime coincident MTSAT-1R and MTSAT-2 images, and can easily be applied to the MTSAT-1R imager pixel level counts to restore the original quality of the entire MTSAT-1R record.
The Moderate Resolution Imaging Spectroradiometer (MODIS) is a unique source of reach spectral information useful for many applications. It provides observations in 36 spectral bands ranging in wavelengths from 0.4μm to 14.4μm with a spatial resolution from 250m to 1km. The standard MODIS data processing system and products cover the basic operational needs for a number of products and applications. Implemented globally they, however, cannot always make the best use of MODIS 250m and 500m land channels required for terrestrial monitoring and climate change applications. To address the need of regional users in enhanced MODIS data, especially in terms of spatial resolution, an independent technology for processing MODIS imagery has been developed. It uses MODIS level 1B top of the atmosphere swath data as input. The system includes the following steps: 1) fusion (downscaling) of MODIS 500m land channels B3-B7 with 250m bands B1-B2 to obtain consistent 250m imagery for all seven bands B1-B7; 2) re-projection of 250m bands into standard geographic projection; 3) scene identification at 250m spatial resolution to obtain mask of clear-sky, cloud and cloud shadows; 4) compositing clear-sky pixels over 10-day intervals; 5) atmospheric correction; 6) landcover-based BRDF fitting procedure. The fusion technique is designed to work with MODIS/TERRA data due to known problems with band-to-band registration accuracy on MODIS/AQUA. The developed method is applied to generate MODIS clear-sky land products in the Lambert Conformal Conical (LCC) projection for Canada and the Lambert Azimuthal Equal-Area (LAEA) projection for the North America and the Arctic circumpolar zone. The novel clear-sky compositing approach is proposed that significantly reduces impact of BRDF effect on raw composites by separation of pixels into two ranges of relative azimuth angle within 90°-270° and outside of this interval.
Among all trace gases, the carbon dioxide and methane provide the largest contribution to the climate radiative
forcing and together with carbon monoxide also to the global atmospheric carbon budget. New Micro Earth
Observation Satellite (MEOS) mission is proposed to obtain information about these gases along with some
other mission's objectives related to studying cloud and aerosol interactions. The miniature suit of instruments is
proposed to make measurements with reduced spectral resolution (1.2nm) over wide NIR range 0.9μm to
2.45μm and with high spectral resolution (0.03nm) for three selected regions: oxygen A-band, 1.5μm-1.7μm
band and 2.2μm-2.4μm band. It is also planned to supplement the spectrometer measurements with high spatial
resolution imager for detailed characterization of cloud and surface albedo distribution within spectrometer field
of view. The approaches for cloud/clear-sky identification and column retrievals of above trace gases are based
on differential absorption technique and employ the combination of coarse and high-resolution spectral data. The
combination of high and coarse resolution spectral data is beneficial for better characterization of surface
spectral albedo and aerosol effects. An additional capability for retrieval of the vertical distribution amounts is
obtained from the combination of nadir and limb measurements. Oxygen A-band path length will be used for
normalization of trace gas retrievals.
A new technology has been developed at the Canada Centre for Remote Sensing (CCRS) for generating North America continental scale clear-sky composites at 250 m spatial resolution of all seven MODIS land spectral bands (B1-B7). The MODIS Level 1B (MOD02) swath level data were used as input to circumvent the problems with image distortion in the mid-latitude and polar regions inherent to the sinusoidal (SIN) projection utilized for the standard MODIS data products. The new data products are stored in the Lambert Conformal Conical (LCC) projection for Canada and the Lambert Azimuthal Equal-Area (LAEA) projection for North America. The MODIS 500m data (B3-B7) were downscaled to 250m resolution using an adaptive regression algorithm. The clear-sky composites are generated using scene identification information produced at 250m resolution and multi-criteria selection which depends on pixel identification. Cloud shadows were also identified and removed from output product. It is demonstrated that new approach provides better results than any scheme based on a single compositing criterion, such as maximum NDVI, minimum visible reflectance, or combination of them. To account for surface bi-directional properties, two clear-sky composites for same time period are produced for the relative azimuth angles within 90°-270° and outside of this interval. Comparison with Landsat imagery and MODIS standard composite products demonstrated advantages of new technique for screening cloud and cloud shadow and providing the high spatial resolution. The final composites were produced for every 10-day intervals since March 2000. The composite products have been used for mapping albedo and vegetation properties as well as for land cover and change detections applications at 250m scale.
A method is proposed to derive spatially enhanced imagery for all seven Moderate Imaging Spectroradiometer (MODIS) land spectral bands at 250 m spatial resolution. Originally, only bands B1 and B2 [visible (VIS) at 0.65 μm, and near-infrared (NIR) at 0.85 μm] are available from MODIS at 250 m spatial resolution. The remaining five land channels (bands B3 to B7) are observed at 500 m resolution. The adaptive regression is constructed for each individual MODIS L1B granule of 500 m spatial resolution by splitting the area into smaller blocks and generating nonlinear regression between bands B3 to B7 and B1, B2 and NDVI. Once a set of regression coefficients is generated based on 500 m image, it is then applied to 250 m data containing only channels B1 and B2 to produce five intermediate synthetic channels (B3 to B7) at 250 m spatial resolution. The final step involves normalizing the generated 250 m images to original 500 m images to preserve radiometric consistency. It is achieved in two stages and ensures that downscaled results are unbiased relative to original observations. The developed method was applied to generate Canada-wide clear-sky composites containing all seven MODIS land spectral channels at 250 m spatial resolution over the area of North America 5700 km by 4800 km.
A novel algorithm to address the reprojection of MODIS level 1B imagery is proposed. The method is based on the simultaneous 2D search of latitude and longitude fields using local gradients. In the case of MODIS, the gradient search is realized in two steps: inter-segment and intra-segment search, which helps to resolve the discontinuity of the latitude/longitude fields caused by overlap between consecutively scanned MODIS multi-detector image segments. It can also be applied for reprojection of imagery obtained by single-detector scanning systems, like AVHRR, or push-broom systems, like MERIS. The structure of the algorithm allows equal efficiency with either the nearest-neighbor or the bilinear interpolation modes.
Surface bi-directional reflectance distribution function (BRDF) and albedo properties are retrieved over the Atmospheric Radiation Measurement (ARM) Program Southern Great Plains (SGP) area. A landcover-based fitting approach is employed by using a newly developed landcover classification map and the MODIS 10-day surface reflectance product (MOD09). The surface albedo derived by this method is validated against other satellite systems (e.g. Landsat-7 and MISR) and ground measurements made by an ASD spectroradiometer. Our results show good agreements between the datasets in general. The advantages of this method include the ability to capture rapid changes in surface properties and an improved performance over other methods under a frequent presence of clouds. Results indicate that the developed landcover-based fitting methodology is valuable for generating spatially and temporally complete surface albedo and BRDF maps using MODIS observations.
Information about the surface bi-directional reflectance distribution function (BRDF) and albedo is required as a boundary condition for radiative transfer modeling, aerosol retrievals, cloud retrievals, and atmospheric modeling. The typical spatial resolution provided by MODIS and MISR standard surface products (~1km) is insufficient to measure the BRDF of the pure surface types, because most pixels at this scale correspond to mixed classes. We present an approach for the retrieval of the basic surface BRDFs from the observations of MODIS/Terra and MISR using an angular unmixing method. Our analysis is focused on the Atmospheric Radiation Measurement (ARM) Program area in the Southern Great Planes (SGP) region, which is a predominantly agricultural area with a few major crop types. Pure surface classes were identified using high-resolution (30m) Landsat imagery and results of a ground survey.
Assuming that the reflectance for each coarse pixel is a linear superposition of reflectances of basic surface types, it is possible to estimate the original BRDF parameters for each landcover type. In our case, three dominant classes were selected: wheat, grass, and baresoil. In the case of wheat and grass, the dispersion of the results is smaller than in the case of soil. This can be explained by the relatively low fractional coverage of the soil class within large pixels and by the significant variability of soil reflectance depending on wetness, soil type (sand, clay, etc.), and other factors. The correlation between the BRDF shape factors and the normalized difference vegetation index (NDVI) has also been analyzed. There is a high degree of correlation between the NDVI and BRDF isotropic factor (r0 in the case of MISR), while the correlation with other BRDF parameters was found to be smaller. In general, the NDVI can be used as a crude proxy for the BRDF shape.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.