I attended the NASA EO-1 SVT meeting in Hilo, Hawaii from November 18-23, 2002. This was my first SVT meeting and the last of the four held during the calibration/validation phase of the EO-1 mission. The purpose of this visit was, together with Alex Held from CSIRO Land and Water, to present the main outcomes from the Australian EO-1 validation sites. An important milestone reached during the Hilo meeting was the second anniversary of the EO-1 mission’s launch on 21st November 2000. The mission was initially planned to be a 12-month experiment, but was extended for some months, and currently the EO-1 data is open for purchase by the general public.

The meeting was attended by SVT members, EO-1 instrument teams and several non-SVT people. The week’s program consisted of paper presentations on Monday, Wednesday, Thursday and Friday, a full day field trip to volcanic lava flows on Tuesday, Hyperion cal/val workshop (Friday morning) and ALI cal/val workshop (Saturday morning). I presented a paper summarizing the activities and outcomes from the Australian EO-1 cal/val sites at Coleambally (Agriculture), Tumbarumba (Forestry), Kunoth (Rangeland ecology), and Panorama/Mt. Fitton (Geology and mineral mapping). Alex Held presented a paper on mapping of sugarcane biophysical variables with Hyperion at Cape Tribulation as well as results from Moreton Bay (Optical water quality) and Cape Tribulation (coral reefs). Both presentations went very well. I also gave a short presentation (based on notes provided by David Jupp) on the Hyperion spectral smile issue during the Hyperion discussion session.

There were many interesting papers presented during the week. All presentations were of a high quality and reflected the advances made in the processing and applications of Hyperion and ALI data over the last two years. I found the meeting very beneficial to my work since the presentations covered a wide range of current issues in Hyperion data processing, including atmospheric corrections, destriping, spectral smile, and noise reduction strategies. It was also good to see so many applications reported in the vegetation, soils and ecology areas. There was obviously more interest in Hyperion data than ALI, as reflected in the greater number of papers on Hyperion data.

There were several interesting papers on forestry applications of Hyperion data. These included papers by David Goodenough (sustainable forest production), Kathy Heidebrecht (canopy structure and chemistry), Phil Townsend (forest composition, structure and chemistry), Dar Roberts (forest fire danger assessment), Marie-Louise Smith (canopy Nitrogen). There were two papers on mapping weed species with Hyperion by Ken McGwire and Ralph Root. Other papers reported use of EO-1 data in landuse/landcover applications and ecosystem studies.

There were also some good results presented on optical qualities of water and coral reef mapping. S. C. Liew showed that by integrating Hyperion 10x10 pixels to a 300m resolution the signal to noise was sufficient to estimate chlorophyll, dissolved organic matter and suspended sediment concentrations. Similar results were also shown in Vittorio’s work that Alex held presented. Tsuneo Matsunaga showed some excellent results on coral, sea grass, and algae discrimination in shallow waters using Hyperion.

In the area of atmospheric correction most people are using ACORN and FLAASH software packages to atmospherically correct Hyperion data, and have reported good agreement between Hyperion reflectance and ground based spectroradiometer data and airborne (AVIRIS, HyMap) imaging spectrometer data. However, at present these packages do not account for the spectral smile in Hyperion data. There were two papers on new software packages being developed, which will have the option of correcting for spectral smile effects. Alex Goetz compared results obtained with the HATCH-2d atmospheric correction software with ATREM, FLAASH and ACORN derived reflectances. However, HATCH-2d results obtained on a column-by-column basis did not offer any significant advantages over using the average spectral calibration for Hyperion data. Mikhail Alexandrov described the ATMCOR package being developed, which will also correct for spectral smile effects.

The EO-1 instrument teams from MIT Lincoln Labs and TRW had several presentations on the on-orbit calibrations and performance of the ALI and Hyperion sensors. These sessions were very useful for the SVT members who are using EO-1 data. Data issues with the ALI sensor as outlined in Jeffrey Mendenhall’s presentation were important, especially the changes in ALI Calibration Coefficients in December 2001. Pre-December 2001 ALI data will be re-processed by the EO-1 Science team upon request only. An aperture selector malfunction on the ALI sensor on 5th July 2002 was also reported, but this has not affected the ALI data. An on board cloud cover detection algorithm for the Hyperion sensor was also described. The algorithm was scheduled for first on-board tests in December 2002, and once operational would regulate which scenes are to be transmitted to the ground for processing.

There was only one paper (by Fred Kruse) on geologic applications using Hyperion data. Hyperion appears to be of marginal value for mineral mapping due to the lower signal to noise ratio in the SWIR wavelengths and the lower spatial resolution of 30 m. The two papers on EO-1 observations of volcanoes and lava flows by James Crowley and Luke Flynn were of interest since most of the participants had been on the tour to the lava flows two days earlier.

The Hyperion discussion session at the end of the week was dominated by the Hyperion spectral smile issue. There are a number of groups working on the spectral smile using Hyperion data from different sites. While the shape of the smile has been shown to be consistent with the pre-launch model, there are vertical offsets between the smile model and smile plots from different sites and dates. The cause of these offsets is still under investigation.

CSIRO’s role in the EO-1 cal/val activity and in particular David Jupp’s contribution was acknowledged at the meeting. Steve Ungar (the meeting chair) made special mention of CSIRO’s contribution to the CAL/VAL activity and thanked David and all Australian Co-Investigators for their efforts, following the conclusion of my talk on the first day of the meeting. Hyperion data processing methods developed by CSIRO (in particular the local and global destriping methods) have been used and were referred to in quite a number of presentations. The Hyperion Workshop CDs developed by CSIRO EOC were also mentioned in various discussions in the corridors; obviously these materials have been circulating abroad!

This last SVT meeting was by no means the conclusion of the EO-1 Science Validation activities. The SVT teams have acquired substantial EO-1 and ancillary datasets for their sites, which will continue to be the source of ongoing research in hyperspectral science for years to come. It was decided at the meeting that the SVT members continue to meet and present their work at other major hyperspectral meetings. Towards this end a special EO-1 session would be included in the February 2003 AVIRIS workshop in Pasadena. The IGARSS 2003 in Toulouse, France during July 2003 is also expected to have many papers on EO-1 data.

Presentations from the meeting are available on the EOC website at http://www.eoc.csiro.au/, by following the links to the oz_pi area under the Hyperion pages.

Acknowledgements
The trip was funded by the EOC travel committee (50%) and the EOC Hyperspectral Project (50%). The Australian EO-1 Co-Investigators are thanked for their help, many of who provided material for the CSIRO presentations.