Using CALIPSO to Explore the Sensitivity to Cirrus

Using CALIPSO to Explore the Sensitivity to Cirrus Height in the Infrared Observations from NPOESS/VIIRS and GOES-R/ABI Andrew Heidinger (GOVERNMENT PRINCIPAL INVESTIGATOR) NOAA/NESDIS/STAR/Co. RP/ASPB Requirement: : Element 7. 4. 1 - Improve weather forecast and warning accuracy and amount of lead time. Advances in Cirrus Microphysical Retrievals from VIIRS and GOES-R Example of the discrepancy between IR and solar- Determining the Sensitivity to Cirrus Height Science: What is the impact of the channel selection on NOAA’s imagers on the ability to estimate cloud height. Benefit: Increase the confidence and accuracy of NOAA’s imager cloud height data records. Improved knowledge of cloudiness forecasters and NWP. Introduction Motivation MODIS Cloud height is an important parameter derived from NOAA’s imagers. NOAA’s future images (NPOESS/VIIRS) and (GOES-R/ABI) offer very different spectral information for cloud remote sensing. This work aims to explore these differences in quantitative method that is independent of the particular algorithm specifics. GOES-R Statement of the Problem: The images on the right show the spectral information contained by the NASA EOS/MODIS sensor (top), GOES-R/ABI, the NPOESS/VIIRS and POES/AVHRR. Cloud top height sensitivity to cirrus (semi-transparent ice clouds) comes from IR absorption bands. As the images demonstrate, VIIRS (similar to AVHRR) provides no IR channels located in atmospheric absorption bands. GOES-R does provide one channel (13. 3 mm) on the edge of the 14 mm CO 2 absorption band. VIIRS Method • The previous analysis has shown that the lack of IR absorption channels has impaired the performance of VIIRS relative to ABI for cloud height estimation. • The following steps describe our analysis that couples real observations with clearsky radiative transfer models to uniquely determine the vertical space in which a cloud could be placed and match the observations used in the algorithm. • However, this insensitivity to cloud height does not mean that the microphysical parameters from the VIIRS window channels are not accurate. • We also compute the relationship between b-profiles of different channel pairs based on scattering theory. Goals of this Work: CALIPSO Analysis • MODIS provides 36 channels at 1 km (or better) resolution and provides nearly all of the IR channels with similar spectral widths provided by GOES-R/ABI and NPOESS/VIIRS. • CALIPSO provides LIDAR observations colocated with MODIS. The CALIPSO data provide the cloud vertical structure which is critical to proper interpretation of the passive observations. False color MODIS with a CALIPSO track overlaid • We define the solution space for cloud-pressure as the region in the atmosphere where the bprofiles are consistent with theory. In other words, any level where the observations match theory are considered levels within the solution space. In the figures above, this is where the blue and red lines intersect. • The larger the solution space, the larger the uncertainty in the cloud-top pressure estimate. • For the VIIRS channels, the uncertainty in cloud pressure is much larger than that for the GOES-R/ABI. Note the difference in regions where the blue and red lines cross. • In addition advances in fast IR models and ancillary data (surface emissivity) have greatly increased our ability to apply physically based methods in IR cloud remote sensing. • We therefore have unprecedented observations to test new and characterize old remote sensing concepts. • Data is this poster is from one AQUA/MODIS granule from the Tropical Pacific observed on August 10, 2006. Images on the right are from this data set. MODIS channel numbers are used throughout. Application to Imagers Current and Future • Knowledge of habit may help define new models that eliminate this discrepancy. This is critical if we ever want to use all spectral information for cloud remote sensing. Global distribution of the presence of the column ice crystal habit Results • The above analysis can be applied to the entire scene (shown in lower left). GOES-R VIIRS • The black curves are the CALIPSO boundaries plotted for reference. Selected publications: Heidinger, A. K. , M. J. Pavolonis, R. E. Holz, B. A. Baum, and S. Berthier (2010), Using CALIPSO to Explore the Sensitivity to Cirrus Height in the Infrared Observations from NPOESS/VIIRS and GOESR/ABI, J. Geophys. Res. , doi: 10. 1029/2009 JD 012152, in press GOES-NOP • Also included are the results for a GOES-NOP (ch 31, ch 33) and a AVHRR or GOES-IM sensors (ch 31, ch 32). • It interesting that VIIRS is more similar to AVHRR than GOESR/ABI in this context cloud-top pressure uncertainty. cross-section of some of the key IR terms in the retrieval • For example, the MYD 06 results from MODIS are derived from solar reflectances and give higher optical thickness and smaller particles than the IR. Variation of habit with temperature • In the figures to the right, the grey region denotes the cloudtop pressure solution space derived using the methodology described above. CALIPSO backscatter cross-section 1: 1 • Ice crystal habit defines the shape of the ice crystals. Knowledge of habit is critical since current ice crystal models often generate spectrally inconsistent results for solar reflectance techniques compared to IR techniques. Apply radiative theory to real observations to study the impact of these channel selections on the cloud height estimation for thin cirrus. AVHRR 2: 1 • From the VIIRS 8. 5, 11 and 12 mm channels, we have developed a technique that estimates cloud optical depth, cloud particle size and the dominant ice crystal habit at cloud-top. Bottom images show results of inferred habit distributions. • First we compute the profiles of cloud emissivity by placing a cloud at many levels in the atmosphere. Where the cloud emissivity for a given channel is between 0 and 1, is the vertical space where a cloud could reside and match the channel’s observation. The cloud emissivity profiles for a single pixel for the MODIS IR channels is shown to the right. • Second, for every channel pair, we can ratio the emissivities to give a b-ratio that is directly relative to cloud microphysics. We construct the b-profiles for the channel pairs used in the algorithm (GOES-R, VIIRS …) reflectance based estimates of cirrus cloud microphysics • As expected the uncertainty in cloud-top pressure is related to the cloud emissivity (how thick the cloud is). The bottomleft figure demonstrates this. Note that the growth in depth of the solution space is less with GOES-R then NPOESS/VIIRS. Center for Satellite Applications and Research (STAR) Review 09 – 11 March 2010 AVHRR Heidinger, Andrew K. and Pavolonis, Michael J. . Gazing at cirrus clouds for 25 years through a split window, part 1: Methodology. Journal of Applied Meteorology and Climatology, Volume 48, Issue 6, 2009, pp. 110 -1116. Science Challenges: Producing high quality cirrus cloud heights from NOAA’s imagers (past, present and future) Next Steps: Develop methods to extract the most information possible out of the NPOESS/VIIRS spectral information for cloud height. Transition Path: Implement these as NOAA- unique products during NPOESS and generate physically consistent GOES-R and NPOESS cloud heights. Deliver these products to NWP and forecasters in a manner consistent with GOES-R products.