OMCLDRR README FILE

Overview

This document presents a brief description of the OMCLDRR data product. OMCLDRR contains effective cloud pressures and fraction along with ancillary information generated using the OMI global mode measurements. In this mode, each file contains the pole-to-pole sunlit portion of a single orbit that is 2600 km wide in the cross-track direction and consists of 60 ground pixels across the track. OMCLDRR retrieves cloud pressures from an amount of filling in of Fraunhofer lines caused by rotational Raman scattering in the atmosphere.

You may refer to Release Specific Information about OMCLDRR for details about software versions and known problems.

Algorithm Description and Validation

For a description of the algorithm used in deriving OMCLDRR please refer to the Algorithm Theoretical Basis Document (ATBD) on http://eospso.gsfc.nasa.gov/eos_homepage/for_scientists/atbd/viewInstrument.php?instrument=13, which also contains other algorithm related documents. There are several journal papers related to the most recent algorithm updates available on the web. These include Joiner et al. (2004) J. Geophys. Res., 109, #D01109, Vasilkov et al. (2004) Geophys. Res. Lett. 31,#L20109 and Joiner and Vasilkov (2006) IEEE Trans. Geosci. Rem. Sens, (in press). The latter reference contains a description of a soft-calibration procedure that is used to remove scan position-dependent biases (i.e.. striping) from the retrieved cloud pressures and comparisons with MODIS data. Further information regarding recent validation can be found in the OMCLDRR validation document volume 1. This algorithm is one of the two algorithms that derive cloud information from OMI data. The other algorithm uses O₂-O₂ absorption near 477 nm and its product is named OMCLDO2.

Data Quality Assessment

Users should be aware that both the OMCLDRR cloud pressure and fraction are effective, meaning that the cloud fraction may not represent true geometrical cloud fraction and the cloud pressure may not represent the true physical cloud-top pressure (especially in the case of multiple cloud layers). Specifically, it is not possible to derive a sub-pixel cloud fraction using OMI radiances. The effective cloud fraction is based on gross assumptions about the cloud and ground reflectivities. A preliminary comparison with MODIS geometrical cloud fractions shows that the effective cloud fraction derived from OMI is approximately 20% lower than the geometrical cloud fraction derived from MODIS. The standard deviation is about 20% and the correlation coefficient is ~0.82. The OMI underestimate of cloud fraction as compared with MODIS can be explained by the fact that OMI is not very sensitive to thin cirrus, whereas MODIS is. The effective cloud fraction is intended for use in conjunction with the effective cloud pressure such that the combination of the two produces the amount of observed Raman scattering.

The cloud pressures are representative of pressure levels reached by back-scattered photons averaged over a weighting function.

The algorithm uses the concept of Lambertian-equivalent reflectivity (LER) in which a surface (cloud or ground) is assumed opaque and Lambertian. Scattering and/or absorption from within and below a cloud or between multiple cloud decks can be accounted for by using a pressure higher than the physical cloud top. The derived effective cloud pressures are therefore typically higher than (i.e., lower in altitude) than cloud-top pressures such as those derived from thermal infrared measurements and cloud lidars. Thus, it is difficult to fully validate these products using correlative measurements. Based on preliminary comparisons with MODIS, we find the effective cloud pressures to be on average about 250 hPa higher than the physical cloud-top, with a standard deviation of ~100 hPa and a correlation coefficient of ~0.72. These numbers are consistent with previous studies using different instruments (e.g., GOME/ATSR).

The original (prelaunch) estimates of the accuracy and precision of the effective cloud pressure retrieval were 100 and 30 hPa, respectively. Our preliminary comparisons with MODIS (Joiner and Vasilkov, 2006) are consistent with radiative transfer calculations that show a large enhancement in scattering from multiple cloud decks (this occurs frequently) and significant light penetration into deep convective clouds. MODIS and OMI cloud pressures agree well when there is a well-defined single-layer cloud deck. We can estimate the precision of the cloud pressure retrieval with a linear error analysis as in Joiner et al. (2004). Using a conservative estimate of 0.5% for the normalized radiance precision, we find that a cloud pressure precision of better than 30 hPa is obtained at all solar zenith angles when the cloud fraction is 100% and there is a single cloud layer. Based on these comparisons and considerations, we believe that our original error estimates are reasonable.

Algorithm Features:

1. In the current version of OMCLDRR, we use the spectral range 392-398 nm. This spectral range is also sensitive to Raman scattering in the ocean. Fortunately, over water surfaces for scenes with reflectivities greater than about 0.25, most of the backscattered light comes from the clouds and atmosphere and has not seen the surface. Thus, under these conditions, the effects of ocean Raman scattering are insignificant. For lower reflectivities, the effects can be significant, especially in very clear waters such as in the remote Pacific. We are currently using a theoretical model with climatological values of chlorophyll content to estimate the effect. However, under some conditions, the climatology will not be accurate and the result will generally be cloud pressures with a positive bias (i.e., pressures too high). It was originally envisaged that a larger spectral range would be used to simultaneously estimate cloud pressure and chlorophyll content. However, using the current spectral range, it is not possible to simultaneously derive cloud pressure and chlorophyll content over oceans. We are currently investigating the feasibility of expanding the spectral range to improve accuracy for broken cloud conditions over ocean.
2. Under similar conditions to those described in (1) for ocean Raman sensitivity, sea glint can cause erroneously high values of retrieved reflectivity and low values of cloud pressure. Sea glint primarily affects the West side of swath at low and mid-latitudes. The sea glint possibility flag is contained in bit 4 of the ground pixel quality flag.
3. Over snow/ice, the processing quality flag bit 5 is set to 1, and the cloud fraction is assigned to 1. Therefore, the effective cloud pressure for these pixels is representative of an average scene pressure (i.e.,. the LER pressure of a pixel that produces the observed amount of Raman scattering). This is done in order to more positively identify the existence of thick clouds over snow/ice which is of interest for the retrieval of ozone and other trace gases as well as the calculation of surface UVB. The snow/ice information comes from the Near real-time Ice and Snow Extent (NISE) product created using passive microwave data. It is provided by the National Snow and Ice Data Center (NSIDC) and is included in the level 1b data set.
4. As the cloud fraction tends to zero, the error in retrieved cloud pressure increases rapidly. These errors can be seen in some cases where cloud fractions are very low (approaching 20%) and low cloud pressures (<200 hPa) are retrieved. Therefore, for cloud fractions <20%, we do not attempt a cloud pressure retrieval. Instead, an effective scene pressure is reported for diagnostic purposes only. These cases are indicated where bit 13 of the processing quality flag is set to 1.
5. Transient events due to radiation hits on a detector may produce striping in the cloud pressures (e.g., anomalously low or high values at one scan position). They may last only for a short period or may continue until elevated dark currents are corrected for in the calibration (these adjustments currently are made once per month). Transient data are currently flagged in the level 1b data set. OMCLDRR has the option of checking this flag. However, the default is currently not to check the flag (because this causes errors when reprocessing with early versions of level 1b data). When the transient flag is checked, the algorithm disregards affected transient pixels as well as pixels affected by other types of error within the fitting window. In practice, we found that the transient flags are set very infrequently and our internal quality control checks are able to detect affected pixels most of the time. When any type of warning or error occurs for pixels within the fitting window for radiance or irradiances, bits 9-12 of the processing quality flag are set as appropriate.
6. Absorbing aerosol can affect the OMCLDRR cloud pressures. In general, it will reduce the cloud pressures. The presence of absorbing aerosols is currently not flagged in the OMCLDRR file. The aerosol index flag in the OMTO3 file can be used to check for the existence of absorbing aerosol within a pixel.

Product Description

A 2600 km wide OMI scan contains 60 pixels. Due to optical aberrations and small asymmetry between the instrument optic axis with the spacecraft nadir, the pixels on the swath are not symmetrically aligned on the line perpendicular to the orbital plane. However, the latitude and longitude provided with each pixel represent the location of each pixel on the ground to a fraction of a pixel.

The OMCLDRR product is written as an HDF-EOS5 swath file. For a list of tools that read HDF-EOS5 data files, please visit this link: http://disc.gsfc.nasa.gov/Aura/tools.shtml.

An OMCLDRR file, also called a granule, contains two effective cloud pressures and fractions. The data fields called CloudPressure and CloudFraction contain the retrievals using ground (clear sky) and cloud reflectivities of 11% and 40% respectively. This product, which existed in the previous Version 0.9.40 software, provides effective cloud fractions, which are closer to MODIS IR-derived cloud fractions. The second product is called CloudPressureforO3 and CloudFractionforO3 and provides retrievals using ground and cloud reflectivities of 15% and 80%, respectively. Those assumptions are used in the OMTO3 total column ozone algorithm. The second product is specifically intended for use with the total ozone PGE - OMTO3. Users must determine which product is most appropriate for their particular application. Further questions should be directed to the contacts provided below.

The output file also contains associated information retrieved from each OMI pixel from the sun-lit portion of an Aura orbit. The data are ordered in time sequence. The information provided on these files includes: Latitude, longitude, solar zenith angle, satellite zenith angle, relative azimuth angle, reflectivity at 394.1 nm and a large number of ancillary parameters that provide information to assess data quality. By far the most important of these parameters is the processing quality flag. Most users should accept data where the processing quality flag bits 0,1,2,3,4,6,7,13,14,15 are set to zero. In addition, data with bit 5 set to 1 should be used with caution as these are data over snow/ice where the cloud fraction has been set to unity. A cloud mask is provided in the data set but should not be used, as its quality has not yet been assessed.

For a complete list of the parameters and bit settings for quality control flags, please read the OMCLDRR file specification document. In addition, we intend to make OMCLDRR data available in a geographically-ordered (rather than time-ordered) format that can be more easily subsetted and manipulated on-line prior to ordering. Please check the Goddard Earth Sciences (GES) Distributed Active Archive Center (DAAC) website for current information on these products where the usual standard time-ordered level 2 products can also be found.

Full OMCLDRR data, as well as subsets of these data over many ground stations and along Aura validation aircraft flights paths are also available through the Aura Validation Data Center (AVDC) website to those investigators who are associated with the various Aura science teams. B. R. Bojkov is the point of contact at the AVDC.

Questions related to the OMCLDRR dataset should be directed to the GES DAAC.

Users interested in these parameters, or having other questions regarding the OMCLDRR dataset are advised to contact Alexander Vasilkov (alexander_vassilkov@ssaihq.com) with a copy of your e-mail to Joanna Joiner (Joanna.Joiner@nasa.gov), who has the overall responsibility for this product.