The 14th Clouds and the Earth's Radiant Energy System (CERES) Science Team meeting was hosted by David Randall of Colorado State University on September 11-15, 1996, in Fort Collins, CO. The focus of the meeting was CERES instrument status, algorithm development, and validation plans. CERES schedules are being driven by the Tropical Rainfall Measuring Mission (TRMM) launch, now planned for November 17, 1997. The Science Team guides the definition of the CERES instrument and science studies to provide a climate data set suitable for examining the role of clouds in the radiative heat balance of the climate system.
Bruce Wielicki, CERES Co-Principal Investigator, opened the meeting with an Earth Observing System (EOS) program status report. He also called attention to the cover article on the CERES project in the May issue of the Bulletin of the American Meteorological Society. This article describes CERES as an investigation to examine the role of cloud/radiation feedbacks in the Earth's climate system. The CERES broadband scanning radiometers are an improved version of the Earth Radiation Budget Experiment (ERBE) radiometers. The CERES instruments will fly on several NASA EOS satellites starting in 1998 and extending over at least 15 years. The CERES science investigations will provide data to extend the ERBE climate record of top-of-atmosphere (TOA) shortwave (SW) and longwave (LW) radiative flux. CERES also combines simultaneous cloud property data derived using EOS narrowband imagers to provide a consistent set of cloud/radiation data, including SW and LW radiative fluxes at the surface and at several selected levels within the atmosphere. CERES data should provide radiative fluxes with a factor of 2-to-3 less error than the ERBE data.
CERES Instrument Status
Robert B. Lee III (LaRC), Lou Smith (LaRC), and Kory Priestley (Virginia Tech) presented the instrument status report. The CERES Proto-Flight Model (PFM) instrument was successfully integrated on the TRMM spacecraft. Instrument weight and power are slightly below the TRMM allocations. Mechanical and electrical integration on TRMM were completed in March, and full comprehensive functional tests verified instrument operation. Electro-Magnetic Compatibility (EMC) tests were successfully completed in July. Spacecraft level tests indicated no adverse effects on CERES science data with anticipated signal strengths of the spacecraft and the other TRMM instruments. Separate tests with the Precipitation Radar indicated no effect on the CERES data. The azimuth gimbal experienced a stall in one particular region when the instrument was cantilevered out from the spacecraft. The stall occurred randomly, and did not occur in the flight-like orientation. A procedure is in place to test extrication from the stall should it reoccur. The azimuth encoder showed a preset azimuth position upon power up. The problem was isolated in the control logic electronics and was solved via an instrument reset.
EOS-AM Flight Model 1 (FM1) assembly is complete and FM2 assembly is continuing. FM1 detector offset and drift problems were resolved. Process changes (de-airing of epoxy) were made to avoid a delamination of the flake from the substrate due to trapped air bubbles. All mechanical and structural mechanisms survived vibration testing without problems. Electronics damage occurred when two boards came in contact. Board spacing was increased to correct the situation. Several changes were made to sensor electronics to reduce susceptibility to electromagnetic radiation.
The TRMM spacecraft level testing continues. Thermal/vacuum tests began September 7; vibration testing begins in January 1997. The spacecraft will be shipped to Japan in April 1997 and will be ready for launch by August 1, 1997.
Data Management System
Jim Kibler presented the CERES Data Management System status report. He summarized Release 1 integration and test activities and status of Release 2 code development. Level 0 data from the TRMM Sensor Data Processing Facility (SDPF) have been processed on Science Computing Facilities to produce Bi-Directional Scan (BDS) and Instrument Earth Scans (IES) data products. This successful test demonstrated many of the processing steps which must be operational for the TRMM launch. Release 2 data products and architecture are mostly defined, and code development is in progress. In the near term, mission simulation tests will be conducted with live CERES data from TRMM, CPU-intensive subsystems will be optimized, and Release 2 algorithms will be finalized, coded, and tested. If the TRMM launch slips, additional design and code reviews will be considered.
Algorithm Theoretical Basis Documents (ATBDs)
The Release 2.1 CERES ATBDs have been completed for all subsystems and placed on a WWW page at LaRC. Major changes from Release 1 include updating subsystem input data, data product catalogs, data flow charts, and product size and processing time estimates; adoption of the 1deg. equal-angle EOS grid system; deleting surface products from the ERBE-like processing; adding a subsystem for gridding geostationary narrowband radiances; improved algorithms for cloud mask and cloud optical property retrievals; algorithm improvements for the Surface and Atmospheric Radiation Butget (SARB) subsystem; and use of International Satellite Cloud Climatology Project (ISCCP) B1 data for temporal interpolation. A peer review of the ATBDs is now underway through the EOS Project Science Office and will be followed by an oral panel review on November 19-21, 1996. Release 2 will be finalized following the review process.
Validation
The CERES Validation Plans are currently being developed. Following a peer review in late 1996, the Validation Plans will be made available on the WWW in the Spring of 1997. In the interim, a separate document containing the Validation Plan Summary Charts for each major subsystem was developed to accompany the ATBDs. This CERES Validation Plan Summary provides the basic elements of each plan in a common format for all the subsystems. This includes a concise overview of the data to be validated, the validation techniques and approach to be applied, and data sources to be used for validation. There are five major data types to validate: instrument broadband radiances, TOA radiative fluxes, surface radiative fluxes, atmosphere radiative fluxes, and cloud properties. Validation will be accomplished by seven major techniques: on-board calibration, theoretical sensitivity studies, pre-launch satellite data surrogates, internal consistency checks, surface observations, other satellite data, and field campaigns.
Because of the emphasis placed on validation, Tom Ackerman (Penn State) gave an invited presentation on "Ground-Based Remote Sensing of Cloud Properties." He discussed various ground-based sensors including cloud radar and lidar systems, narrowband radiometers, sky imagers, and broadband hemispheric-view radiometers. Observable cloud properties using ground-based sensing include cloud boundaries (top and bottom), cloud amount, mean particle size, particle phase, SW and LW optical depth, and liquid water path. He outlined retrieval techniques for cloud microphysics focusing on bulk retrievals, use of Doppler radar pulse pairs, and use of Doppler radar spectra. An aircraft/surface radar experiment at Penn State in October will provide data to verify the surface-based cloud property retrievals.
Wielicki led a discussion of the various aircraft, remotely piloted vehicles, Atmospheric Radiation Measurement (ARM) instrument packages, and other instruments that will be used to validate cloud properties. He concentrated on identifying cloud regimes that do not have sufficient validation sources in the current plan, namely subtropical oceans, deserts, high-latitude oceans (storm tracks), and tropical land. A need for additional lidar/radar/radiometer combinations was recognized. Tom Ackerman discussed Penn State's long record of radar-based effective radii retrievals as a function of height. CERES personnel will participate in their October field experiment to acquire satellite data to complement aircraft- and surface-based observations for use in cloud property retrievals.
Working Group Reports
Cloud Working Group: Decided to run both daytime and nighttime cloud algorithms when the solar zenith angle is between 78o and 80o. This will allow some ability to pick and choose between algorithms and do some consistency and bounding checks to ascertain if the results from the daytime are reasonable.
Surface and Atmospheric Radiation Budget (SARB) Working Group: Made changes to a list of parameters in the Single Satellite Footprint (SSF) product, which has the imager-based cloud properties and broadband measurements. Discussed the characteristics of the ancillary meteorological data to be used for all CERES processing, which are to be acquired from the Data Assimilation Office (DAO) at the GSFC. Man-Li Wu (GSFC) described the characteristics of the data set currently available from the DAO, and the significant enhancements to be made before the TRMM and EOS launches. Tom Charlock inquired if water vapor channel IR radiances from satellites will be assimilated directly in the DAO analysis for water vapor; such assimilation is done by the European Center for Medium-Range Weather Forecasts (ECMWF) and is claimed to substantially improve the water vapor profile. The water vapor profile is critical for the determination of diabatic heating by CERES. Wu indicated that the water vapor IR radiances will not be used for such direct assimilation by DAO, but she was confident that the water vapor retrievals for AM-1 and PM-1 will be much better than the present ones. This remains a major area of concern for cloud/radiation studies.
Time Interpolation and Spatial Averaging (TISA) Working Group: Discussed algorithm changes, code development, validation plans, and ongoing temporal and spatial averaging studies. Takmeng Wong compared results from a DAAC run of the ERBE-like TISA code with ERBE. Slight discrepancies were seen in the monthly means from the two data sets, and possible causes were discussed. All TISA subsystems are on target for delivery to meet TRMM schedules.
Investigator Presentation Highlights
Bruce Wielicki (LaRC): Commented that most algorithms that deal with cloud overlap are considered "high risk" and will likely not be implemented until a later release. These algorithms are still in the development stage. Most also require Moderate Resolution Imaging Spectroradiometer (MODIS) wavelengths and/or resolution, rather than the Visible and Infrared Scanner (VIRS) products, so implementation is not really necessary until EOS AM/PM.
Bryan Baum (LaRC): Showed that timings on an SGI R10000 single processor are down to 0.5 hours per hour of VIRS or Advanced Very High Resolution Radiometer (AVHRR) data for the CERES cloud property retrieval subsystem. IBM Explorer is an effective visualization tool, especially for global data. The CERES-developed Satellite Imagery Visualization System (SIVIS) used for cloud algorithm testing has been distributed to the team.
Tom Charlock (LaRC): Compared surface SW fluxes computed with the Fu-Liou code and measurements from sites in north-central Oklahoma and at the Boulder Atmospheric Observatory (BAO) tower. Computed fluxes are consistently higher than measured fluxes for clear-sky conditions. There are significant differences among the many surface flux measurements in the ARM Extended Shortwave Experiment (ARESE). Similar discrepancies noted by the ARM team have been reported in the on-line CAGEX (CERES/ARM/Global Energy and Water Cycle Experiment, GEWEX) activity: http://snowdog. larc.nasa.gov/cagex.html. Further work is needed to clarify absolute accuracy of the different surface radiation sensors.
Jim Coakley (Oregon State Univ.): Completed new spatial coherence routines for retrieving cloud structures. The new methodology requires only a single pass through the observations to analyze the cloud properties.
Dominique Crommelynck (Belgium): Presented the results of a paper by S. Dewitte et al. in which radiation budget estimates were derived at the highest possible time/space resolution from Meteosat images and compared to ERBE results. This work explores combining broadband radiation data from satellites such as ERBE, Scanner for Radiation Budget (ScaRaB), and CERES with narrowband geostationary satellite data to enhance spatial and temporal resolution.
Jennifer Francis (Rutgers Univ.): Discussed the need for 4.0, 4.46, and 4.52 um channels from MODIS in retrieving nighttime polar clouds. MODIS channels will include these wavelengths, some new wavelengths for SARB, and other channels needed for cloud retrievals.
Richard Green (LaRC): Showed the final validation results of a new set of ERBE-like angular distribution models (ADMs) constructed from Nimbus-7 data and the Radiance Pairs Method (RPM). The new ADMs eliminated albedo growth to the limb, reduced latitudinal bias, and reduced cloud cover growth with viewing zenith. Applying the new ADMs to a month of ERBS data changed the daily global SW flux by 5%, depending on the local time of the orbit (i.e., solar zenith angle sampling). However, the monthly global flux was almost unchanged.
Qingyuan Han (S. Dakota School of Mines & Technology): Emphasized the need to explicitly define effective radius and diameter when doing microphysical retrievals/validation. At least 6 different mathematical definitions are in the literature. He also discussed the effect of varying particle shapes on scattering/measured radiances and presented results from a simulation of bullet rosettes and hexagonal columns and their respective scattering behavior.
Anand Inamdar (representing V. Ramanathan, Scripps): Presented a new LW surface flux parameterization for cloudy skies that uses a matrix for cloud overlap structure based on synoptic cloud climatology reports. Cloud heights are assumed in accordance with the International Cloud Atlas. Cloud radiative forcing at the surface is uncorrelated with that at the TOA, except for distinct cloud patterns. For each cloud type (low, mid, and high), the cloud radiative forcing at the surface shows distinct correlations with the total precipitable water. Nearly 60-70% of the cloud forcing at the TOA and about 70-90% of the cloud forcing at the surface occurs in the window region.
Michael King (GSFC): Briefed the team on the Arctic Radiation Measurements in Column Atmosphere-surface System (ARMCAS), a small, focused field campaign designed to better understand radiative processes in the Arctic and to serve as a pilot study for the Arctic First ISCCP Regional Experiment (FIRE)-III and Surface Heat Budget of the Arctic Ocean (SHEBA).
Norman Loeb (Oregon State Univ.): Reported on the applicability of 1D radiative transfer theory for retrieving cloud properties. He noted significant differences between the 1D theory representation of the cloud reflectance field and actual observations, even for overcast marine stratus cloud layers. The 1D model underestimated reflectances at nadir and systematically overestimated (by 20-30%) reflectances in the forward scattering direction at moderate and low sun elevations. In the backscattering direction, the 1D model provided an accurate (within 10%) representation of the observed reflectance.
Pat Minnis (LaRC): Reported on cloud algorithm development using the 12 um split window technique. This wavelength helps to eliminate ambiguous solutions that can occur in the backscatter direction when using the Visible, Infrared, Near-infrared Technique (VINT). The backscatter problem seems to occur in thinner clouds. Thin cloud over thick cloud situations can also be explained using information gained from the split window technique.
Dave Randall (Colorado State University): Presented data and rationale to promote the use of the diurnal cycle as a testbed for large-scale models. The forcing of the diurnal cycle, i.e., the diurnal variation of solar radiation at the TOA, is external to the climate system, is well understood, and can be specified very accurately. The response to the diurnal cycle is often quite strong such as in the case of diurnal convection and precipitation over tropical land masses. The diurnal cycle is periodic, so that compositing can be performed in a highly objective manner. Finally, the period of the diurnal cycle is short, so that many realizations can be captured by a relatively brief simulation or observational record.
David Rutan (LaRC): Presented surface optical property maps that will be part of the SSF. "Point and click" maps for a priori global surface spectral reflectance and broadband emissivity are maintained by CERES on the WWW at: http://tanalo.larc.nasa.gov:8080/surf_ htmls/SARB_surf.html. CERES is seeking advice on (a) seasonal interpretation of the surface reflectance maps, (b) the solar zenith angle dependence of reflectance, and (c) spectral bi-directional reflectance functions (BDRFs). The 17 surface types used in the International Geosphere-Biosphere Program (IGBP) will be adopted for SARB processing, and optical properties will be derived for each surface type in all Fu-Liou bands and for the broadband and window in the LW region. He also showed comparisons with CAGEX data over Oklahoma. Charles Whitlock (LaRC) has begun a helicopter-based program to determine spectral BDRFs (FieldSpec FR 350-2500 nm) and broadband albedo.
Lou Smith (LaRC): Presented results of ScaRaB validation studies. He analyzed the errors resulting from computing daily or monthly means of outgoing LW on the basis of restricted temporal sampling. He compared ERBE and ScaRaB results to quantify regional sampling errors and developed methods to deal with similar sampling problems that could be encountered on CERES.
Shi-Keng Yang (NOAA): Compared CAGEX LW fluxes with corresponding fluxes from NOAA's medium-range forecast model and regional spectral model. The data were also compared with NCEP Reanalysis. Comparisons for both outgoing and downward LW were good for clear skies, but not as good for cloudy skies.
David Young (LaRC): Presented new ERBE-derived monthly, regional clear-sky LW thresholds for improving the CERES ERBE-like clear-sky processing. Reprocessing one month of ERBS data with the new thresholds substantially reduced the number of regions without a clear-sky classification.
Ming-hua Zhang (representing Bob Cess, SUNY Stony Brook): Discussed the atmospheric SW absorption bias in the NCAR CCM and possible causes. The CCM underestimates atmospheric SW absorption by 20 Wm-2 (global mean) compared to collocated TOA ERBE data and Global Energy Balance Archive (GEBA) surface observations. Direct aircraft measurements of SW absorption during ARESE indicate that the cause is not a clear-sky problem, but rather one of cloud absorption.
Science Team Logistics
The next CERES Science Team meeting is scheduled for April 18-21, 1997, at the Langley Research Center. Major topics will include instrument status for the TRMM launch and for the EOS satellites, the review of Release 2 ATBDs and validation plans, and software development and testing.