Jet Propulsion Labratory

MEETING SUMMARY - written by George Aumann, AIRS Project Scientist

The theme of AIRS Science Team meeting on 12-14 February 2002 was "Science Data Systems and Software launch readiness". Click for the Agenda , Names/affiliation/email addresses of attendees of the meeting and a Picture of the AIRS Science Team (absent Staelin, Rizzi, LeMarshal).

AIRS/AMSU-A/HSB Launch Readiness Review

Tom Pagano was introduced as the new Deputy AIRS Project Manager. He presented the Project Status , AIRS TVAC Results , and AIRS Calibration Software . Most of his presentation summarized material presented during the formal AIRS data system launch readiness review held at JPL on 8 Febraury 2002. Many people and organizations have contributed to bring the AIRS hardware, the AIRS data systems and the software to the current state of readiness. Click for Logos of AIRS contributors . He also presented details of the On-orbit Activation Phase from L to L+90 days. First data will come from the microwave instruments at nominally L+12 days, followed by VIS/NIR data at L+15 days. Following the cooldown of the AIRS spectrometer several days of "first light" data will nominally be obtained at Launch + 39 days, with the first "golden day(s)" at Launch + 66 days. Plans have been developed for the "early evaluation of level 1b" using the first light and golden day data. Bjorn Lambrigtsen presented the AMSU-A/HSB TVAC Results and Calibration Software Status.

Science Data Systems Status

George Aumann , AIRS Project Scientist, presented an overview of the flow of data from the EOS Aqua via EOS-DIS and the GSFC/DAAC. From there the data flow to NOAA/NESDIS and to the TLSCF at JPL, where they will be available to the science team members for early evaluation, software upgrades and quantitative evaluation of the data product accuracy, referred to as validation, relative to the nominal retrieval accuracy of 1K/1km under clear and partly cloudy conditions. The data systems meet their requirements based on simulated data. All data systems are ready for launch.

Jeff Lubelcyk presented details of the EOS DIS downlink. Based on current latency estimates the data will arrive at the GSFC/DAAC 50 minutes after downlink to the ground, i.e. data from the start of the orbit will be 150 minutes old. At the GSFC DAAC the data flow is split into the NOAA/NESDIS path in support of the weather forecasting and the NASA path. The NASA path is used to transmit data to the JPL/TLSCF, and ultimately to the science investigators outside of the AIRS team. A data distribution policy and a data distribution implementation plan is being drafted by the AIRS Project at JPL.

Mitch Goldberg and Walter Wolf presented details of the NESDIS Near Realtime Processing System. The software used to convert the level 0 data to level 1b data will be identical to the software running at the TLSCF at JPL. The HDF output of the level 1b software for each granule (118MBytes/granule for all channels) will be reformated to BUFR (520Kbytes/granule for the 280 key channels). The entire process takes about 25 minutes using the 20 AIRS dedicated CPU's of the availble 23 CPU SGI Origins computer. The 25 minute data processing is added to the 150 minute latency on arrival, just satisfying the the 3 hour requirement for distribution to the weather forecasting centers. Simulated BUFR data generated from the daily NCEP analysis are currently routinely distributed. Distribution of real AIRS BUFR data will not start until the data flow from EOS Aqua has stabilized and the data have passed through some initial validation. This is expected to occure at about Launch + 5 months.

Steve Kempler described details of the data flow and data processing at the GSFC/DAAC. The JPL/TLSFC provided "launch ready" software for processing the data from level 0, to level 1A, to level 1b, and to level 2 has been installed. Actual data processing will ramp up as function of time. At launch there will be 6 of the 24 processors of a SGI Challenge XL available. This will increase to 20 of 32 processors of an SGI Origin 3000 at Launch + 6 months. The level 0 data will be transmitted to the JPL/TLSFC for detailed analysis within 24 hours. After the initial validation of the data at the JPL/TLSCF and the post-launch software upgrades, expected at Launch+3 months, Launch+7 months, and Launch+12 months, data will be distributed to outside science investigators through the DAAC, in accord with the AIRS data distribution policy.

George Serafino defined basic and advanced DAAC Data Support Services and their readiness for launch.The first test of the interface with users will use AIRS/AMSU-A/HSB data (15 December 2000) simulated by the TLSCF based on the NCEP analysis, and used for algorithm development and software performance testing.

The launch readiness of the Team Leader Science Computing Facility (TLSCF) at JPL was discussed by Steve Friedman, the AIRS Science Data System Manager. The TLSCF develops the data processing system (PGS) for the GSFC/DAAC and is responsible for the validation of the data. PGS version V2.1.6.14 is the last prelaunch PGS installed at the DAAC. The data system has been test with data from TVAC tests, tests at TRW and three days of simulated data. Based on the results of the TVAC and TRW data work has started on upgrading the PGS to V2.2.xx. Ed Olsen provided an overview of the AIRS level 1b and level 2 data products which will be validated by Launch + 12 months. Level 3 products are level 2 products average on a weekly or longer timescale and are presented on a standard global grid. The definition of L3 products has started and should be ready for e requirements review by the science in April 2002. Eric Fetzer discussed the data validation timeline. Data from AMSU-A, HSB and the VIS/NIR channels of AIRS will start flowing nominally at Launch + 15 days. First light data from AIRS should start flowing at Launch + 39 days. The initial AIRS instrument commisioning is scheduled to be completed by Launch + 90 days. The data from field validation of the level 2 products (T(p), q(p), TSurf, etc) using RAOB's, surface buoys and ARM/CART will be collected during a 6 months period starting at launch + 3 months. These basic data will be used during the 6 months of intensive core level 2 data product validation phase.

Sharon Ray is the JPL Outreach coordinator for AIRS. She discussed AIRS outreach products (primarily the AIRS Broshure) , PR products (AIRS banners, mugs, shirts etc.) and details of the upcoming EOS Aqua launch. Sharon is also the AIRS web page coordinator. The old AIRS web page at http://www-airs.jpl.nasa.gov is being replaced by a pure outreach (K-12) website http://www.jpl.nasa.gov/airs. The technical information for team members has been mooved to http://airsteam.jpl.nasa.gov.

AIRS Science Team Workshop

George Aumann, AIRS Project Scientist provided the introduction to the AIRS Science Team Workshop. The focus points of the workshop were: Early Level 1b Evaluation and Status/new developments with the level 2 software.

Early Level 1b Evaluation

Aumann: Early Level 1b Evaluation Assignments. The on-orbit activation plan for AIRS/AMSU-A/HSB includes active participation of the science team in the early evaluation of data, leading to the early update of the level 1b software and possible release of AIRS sample data. The first data to become available from AIRS are the VIS/NIR data, when the scan mirror starts rotating at about Launch + 15 days. This is the start of the Hofstadter/Gautier: Early Level 1b VIS/NIR evaluation. The "First Light" IR spectrometer data start to flow at about Launch + 39 days. This data will be used to evaluate the level 1b based on internal quality indicators (Pagano), evaluation of the radiometric calibration in window channels (Aumann/Hagan), spectral centroid calibration (Gaiser), spatial calibration (Gregorich) and the overall spectral calibration (Strow). The first data with the AIRS in a stable configuration is expected to be obtained at about Launch+70 days. This data is suitable for indepth evaluation. Goldberg will be evaluate the data using principle component analysis and start retrievals of cloud-free footprints using empirical regression. Susskind will evaluate the data for the noise characteristics and (calculated - observed) bias. The Revercomb/Tobin evaluation will focus on bias calculation relative to ARM/CART overpass RAOB launches, and McMillin will start the evaluation of bias relative to the routine RAOB launches.

Level 2 Product Generation Status

Although the first post-launch upgrade to the level 2 PGS at the DAAC is not scheduled until launch + 10 months, the development of the retrieval software is continuing. Susskind reported on the performance of his level 2 physical retrieval testbed algorithm. Chris Barnett reported on surface emissivity issues related to optimizing retrievals over land. Rosenkranz reported on upgrades to the Microwave Retrieval Surface Classification and Water condensation model. This has implications on the precipitation flag used by the cloud-clearing algorithm.

Other Reports

AIRS retrieval have to achieve radio sonde accuracy under partly cloudy conditions. This is achieved by combining microwave (AMSU-A) and AIRS (IR) data in the cloud-clearing process. Data to evaluate cloud-clearing using this process are potentially available using the microwave radiometer and infrared spectrometer on the National Atmospheric Sounder Testbed (NAST): NAST-M and NAST-I. Bill Blackwell and David Staelin have started to evaluate NAST-M/NAST-I data for cloud-clearing and reported on cloud analysis statistics using CAMEX 3 data.

In the December 2001 issue of IEEE-TGRS was an announcement for an EOS Aqua special issue. Abstarts for sixteen papers related to AIRS/AMSU-A/HSB have been received from AIRS science team members and their collegues. The full papers are due by 2 April 2002.

The next AIRS Science Team meeting is timed to coincide with the launch of the EOS Aqua on 18 April 2002, with meetings from 2pm to 8pm on 18 and 19 April 2002.