CCSD3ZF0000100000001NJPL3IF0PDS20000000001 = SFDU_LABEL RECORD_TYPE = STREAM INSTRUMENT_HOST_NAME = ("NASA DC-8", "NASA C-130", "NASA ER-2" , "FIELD EXPERIMENT") INSTRUMENT_ID = ("ASAR", "ASAS", "AVIR", "AWND", "DAED", "GPSM", "HSTP", "PARB", "PFES", "REAG", "RMTR", "SHYG", "SIRS", "THRM", "TIMS", "WTHS") TARGET_NAME = EARTH FILE_TITLE = "GEOLOGIC REMOTE SENSING FIELD EXPERIMENT DATA" END Archive of Geologic Remote Sensing Field Experiment Data Release 1.0 May 13, 1991 Raymond E. Arvidson, Mary A. Dale-Bannister, Edward A. Guinness, Susan H. Slavney, and Thomas C. Stein Planetary Data System Geosciences Node Department of Earth and Planetary Sciences McDonnell Center for the Space Sciences Washington University St. Louis, Missouri With input from: Ronald Greeley, Arizona State University Nicholas Lancaster, Desert Research Institute, Reno, Nevada Lisa Gaddis, Arizona State University James R. Irons, Goddard Space Flight Center David Harding, Goddard Space Flight Center Don Deering, Goddard Space Flight Center James B. Garvin, Goddard Space Flight Center Diane L. Evans, Jet Propulsion Laboratory Thomas Farr, Jet Propulsion Laboratory Bruce Jakosky, University of Colorado Fred A. Kruse, University of Colorado William Farrand, University of Arizona Jeffrey J. Plaut, Washington University Shelley B. Petroy, Washington University Kathy Young, University of Colorado Robert Singer, University of Arizona James Conel, Jet Propulsion Laboratory Carol Bruegge, Jet Propulsion Laboratory i TABLE OF CONTENTS 1. INTRODUCTION ................................................... 1 2. GRSFE OBJECTIVES ............................................... 3 3. OVERVIEW OF GRSFE REQUIREMENTS AND IMPLEMENTATION .............. 7 3.A. SITE SELECTION ............................................... 7 3.B. AIRBORNE CAMPAIGN ............................................ 10 3.C. FIELD CAMPAIGN ............................................... 11 3.D. NARRATIVE OF FIELD CAMPAIGN ACTIVITIES ....................... 12 4. MODELING SITE DESCRIPTIONS ...................................... 17 5. UNIVERSITY OF COLORADO DIRECTIONAL EMISSIVITY EXPERIMENT ....... 21 6. FIELD DATA SETS ................................................ 25 6.A. PHOTOGRAPHS .................................................. 25 6.B. SAMPLES ...................................................... 26 6.C. VISIBLE AND REFLECTED INFRARED SURFACE DATA .................. 26 6.C.1. DAEDALUS SPECTROMETER ...................................... 26 6.C.1.1. INSTRUMENT DESCRIPTION ................................... 26 6.C.1.2. DATA SET DESCRIPTION ..................................... 27 6.C.2. PARABOLA ................................................... 28 6.C.2.1. INSTRUMENT DESCRIPTION ................................... 29 6.C.2.2. DATA SET DESCRIPTION ..................................... 31 6.C.3. SIRIS SPECTROMETER ......................................... 33 6.C.3.1. INSTRUMENT DESCRIPTION ................................... 33 6.C.3.2. DATA SET DESCRIPTION ..................................... 34 6.D. THERMAL INFRARED SURFACE DATA ................................ 36 6.D.1. PORTABLE FIELD EMISSION SPECTROMETER (PFES) ................ 36 6.D.1.1. INSTRUMENT DESCRIPTION ................................... 36 6.D.1.2. DATA SET DESCRIPTION ..................................... 38 6.E. ATMOSPHERIC DATA ............................................. 39 6.E.1. REAGAN RADIOMETER .......................................... 39 6.E.1.1. INSTRUMENT DESCRIPTION ................................... 39 6.E.1.2. DATA SET DESCRIPTION ..................................... 39 6.E.2. SPECTRAL HYGROMETER ........................................ 40 6.E.2.1. INSTRUMENT DESCRIPTION ................................... 40 6.E.2.2. DATA SET DESCRIPTION ..................................... 40 6.E.3. ARIZONA STATE UNIVERSITY WIND EXPERIMENT ................... 41 6.E.3.1. INSTRUMENT DESCRIPTION ................................... 41 6.E.3.2. DATA SET DESCRIPTION ..................................... 41 6.E.4. WEATHER STATION ............................................ 42 6.E.4.1. INSTRUMENT DESCRIPTION ................................... 42 6.E.4.2. DATA SET DESCRIPTION ..................................... 42 6.F. GEOPOSITIONAL SATELLITE PROFILES ............................. 42 6.G. PROFILES FROM HELICOPTER-BORNE STEREOPHOTOGRAPHY ............. 44 7. AIRBORNE DATA SETS ............................................. 44 7.A. ASAS ......................................................... 44 7.A.1. INSTRUMENT DESCRIPTION ..................................... 44 7.A.2. DATA SET DESCRIPTION ....................................... 46 7.B. AVIRIS ....................................................... 47 7.B.1. INSTRUMENT DESCRIPTION ..................................... 47 7.B.2. DATA SET DESCRIPTION ....................................... 48 7.C. TIMS ......................................................... 49 7.C.1. INSTRUMENT DESCRIPTION ..................................... 49 ii 7.C.2. DATA SET DESCRIPTION ....................................... 51 7.D. AIRSAR ....................................................... 52 7.D.1. INSTRUMENT DESCRIPTION ..................................... 52 7.D.2. DATA SET DESCRIPTION ....................................... 52 8. SOFTWARE ....................................................... 54 9. FLIGHT LINE LOCATOR MAPS ....................................... 54 10. GRSFE SAMPLER DATA SETS ........................................ 54 11. INDEX TABLES ................................................... 55 12. DATA SET LABEL AND FILE ORGANIZATION ........................... 56 13. DISK DIRECTORY STRUCTURE AND FILE NAMES ........................ 58 14. RELEVANT BIBLIOGRAPHY .......................................... 60 APPENDIX A - CD-ROM VOLUME, DIRECTORY AND FILE STRUCTURES .......... 67 A.1 VOLUME AND DIRECTORY STRUCTURES ............................... 67 A.2 FILE STRUCTURE ................................................ 67 A.2.1 FIXED-LENGTH FILES .......................................... 67 A.2.2 STREAM FILES ................................................ 67 A.2.3 EXTENDED ATTRIBUTE RECORDS .................................. 68 APPENDIX B - SYNTACTIC RULES OF KEYWORD ASSIGNMENT STATEMENTS ...... 69 B.1 INTEGER NUMBERS ............................................... 69 B.2 REAL NUMBERS .................................................. 69 B.3 DATES AND TIMES ............................................... 70 B.4 LITERAL VALUES ................................................ 70 B.5 TEXT CHARACTER STRINGS ........................................ 70 APPENDIX C - KEYWORD DEFINITIONS ................................... 71 APPENDIX D - GRSFE AIRBORNE AND FIELD INSTRUMENTS .................. 76 TABLE D.1 GRSFE AIRBORNE REMOTE SENSING INSTRUMENTS ............... 76 TABLE D.2 ATMOSPHERIC INSTRUMENTS DEPLOYED DURING GRSFE FIELD CAMPAIGN ....................................................... 76 TABLE D.3 GROUND TRUTH INSTRUMENTS DEPLOYED DURING GRSFE FIELD CAMPAIGN ....................................................... 77 TABLE D.4 AVIRIS FLIGHT LINES ..................................... 78 TABLE D.5 ASAS AND TIMS FLIGHT LINES .............................. 80 TABLE D.6 AIRSAR FLIGHT LINES ..................................... 85 TABLE D.7 AVIRIS CROSS-REFERENCE TABLE ............................ 88 TABLE D.8 AIRSAR CROSS-REFERENCE TABLE ............................ 89 TABLE D.9 ASAS CROSS-REFERENCE TABLE .............................. 89 TABLE D.10 TIMS CROSS-REFERENCE TABLE .............................. 90 TABLE D.11 GRSFE TEAM PARTICIPANTS ................................. 91 APPENDIX E - WASHINGTON UNIVERSITY EXPERIMENT -- THERMAL MEASUREMENTS AT LUNAR CRATER VOLCANIC FIELD .................... 93 TABLE E.1 LUNAR LAKE THERMISTOR DATA ............................... 93 TABLE E.2 LUNAR LAKE THERMISTOR DATA ............................... 93 TABLE E.3 TEMPERATURE AND RELATIVE HUMIDITY AT THERMAL TEST SITE ... 94 TABLE E.4 LUNAR LAKE THERMISTOR DATA FOR MANTLED FLOW SITE PITS .... 95 TABLE E.5 RADIOMETER DATA COLLECTED AT MANTLE FLOW SITE PITS ....... 95 TABLE E.6 RADIOMETER DATA COLLECTED DURING TIMS OVERFLIGHTS ........ 96 APPENDIX F - SAMPLES COLLECTED DURING GRSFE ........................ 98 TABLE F.1 SUMMARY OF SAMPLES COLLECTED DURING GRSFE CAMPAIGN ....... 98 TABLE F.2 SOIL ANALYSES ............................................ 99 APPENDIX G - EPPLEY PYRANOMETER DATA ACQUIRED TO SUPPORT PARABOLA DATA ANALYSES .................................................. 101 APPENDIX H - GRSFE CD-ROM ARCHIVE VOLUME AND DIRECTORY STRUCTURE.... 107 1 1. INTRODUCTION This document provides an overview of the objectives, requirements, implementation, data processing, and archiving for the Geologic Remote Sensing Field Experiment (GRSFE). It also describes GRSFE sampler data sets that provide examples of the products that can be generated from the airborne observations. GRSFE consisted of coordinated airborne remote sensing and field observations over geological targets in the Mojave Desert, California, and the Lunar Lake Volcanic Field, Nevada in July, September and October of 1989. GRSFE was funded by NASA's Planetary Geology and Geophysics Program, the Geology Program, the Planetary Data System, the Pilot Land Data System, the Mars Observer Project, and the Magellan Project. The GRSFE archive is maintained and updated by the Planetary Data System Geosciences Node at Washington University, in collaboration with the Pilot Land Data System. Release of Version 1.0 of the GRSFE archive on a set of CD-ROMs is meant to provide a well documented data set collection for research and teaching. Release on CD-ROMs also provides concrete examples of archived data using Planetary Data System and Pilot Land Data System standards (e.g. Martin et al., 1988). The intent of this document is to provide information needed for researchers to understand why and how GRSFE was implemented, the structure and content of the data sets that comprise the archive, and how the various observations can be used in a synergistic fashion to address geological remote sensing problems. We note that the GRSFE archive is dynamic, and release of Version 1.0 on CD-ROMs should be viewed as a convenient way to ensure wide distribution. A few data sets were not ready for this release. These data sets include: (a) ATLAS laser altimetric profiles of selected sites; (b) AIRSONDE balloon measurements of atmospheric temperature, pressure, and relative humidity, (c) PIDAS field spectral reflectance observations, (d) vegetation analyses of sites in the Lunar Crater Volcanic Field, Nevada; and (e) a number of topographic profiles extracted from helicopter-borne stereophotography. We hope to release these data sets in the future as an addendum to Version 1.0. We will also concurrently release an updated version of this VOLINFO.TXT document. GRSFE was possible because of the expertise and interest of the Science Steering Group (SSG), field participants, and personnel who have been involved in data reduction, processing, and documentation. These personnel are listed in Table 1.1 and are hereby thanked for their advice and work. Special thanks are extended to the Science Steering Group for their advice and work throughout GRSFE. We thank Joseph Boyce, Theodore Maxwell, and Miriam Baltuck, NASA Headquarters, for insight and support, and we also thank the High and Medium Altitudes Branches of the Ames Research Center for their efforts during the data acquisition phase of GRSFE. Finally, we thank Susan McMahon, Gail Woodward, and Jason Hyon, Planetary Data System, for their support and work, and Blanche Meeson, Pilot Land Data System, for her interest in the GRSFE archive. 2 GRSFE objectives and requirements are described in the next two sections of this document. Summaries of the airborne and field campaigns are then provided, followed by data set descriptions. This document ends with a discussion of the formats and the directory structure used for the data sets, and CD-ROMs, respectively. Appendices provide specific details on formats and keywords used on the CD-ROM, and also contain some data tables and ancillary information. TABLE 1.1 GRSFE PARTICIPANTS Arizona State University ------------------------- Ronald Greeley - Science Steering Group (SSG) Member Lisa Gaddis Nicholas Lancaster Goddard Space Flight Center ------------------ James R. Irons - SSG Member Don Deering - SSG Member James B. Garvin - SSG Member David Harding Jet Propulsion Laboratory --------------------- Diane Evans - SSG Co-Chair Elsa Abbott Carol Bruegge Fred Burnette James Conel Pascal Dubois Thomas Farr - SSG Member Rob Green Gordon Hoover John Holt Phil Hughes Howell Johnson Anne Kahle - SSG Member Mike Kobrick Joel Norris Jeff Plescia Jakob van Zyl - SSG Member Gregg Vane Michele Vogt Steve Wall - SSG Member Cathy Weitz Rich Zurek Stanford University ----------------------- Richard Simpson - SSG Member 3 USGS -------- John Dohrenwend University of Arizona ------------------------ Bill Farrand Paul Geissler Jeff Kargel Robert Singer - SSG Member University of California ------------------------ Susan Ustin University of Colorado ------------------------ Jose Aguirre Bruce Jakosky - SSG Member John Dietz Gary Finiol Alex Goetz Bradley Henderson Fred Kruse - SSG Member Kathy Young University of Nevada, Reno -------------------------- John Perry University of Washington ------------------------- Alan Gillespie Washington University ----------------------- Raymond E. Arvidson - SSG Co-chair Kristy Chamberlain Mary A. Dale-Bannister Glen Green Edward A. Guinness - SSG Member Shelley B. Petroy Jeffrey J. Plaut Susan H. Slavney Thomas C. Stein 2. GRSFE OBJECTIVES The primary GRSFE objectives were: (a) to acquire airborne remote sensing and field data in a coordinated campaign for key geological targets, e.g., alluvial fans, dunes, lava flows, volcanoes, etc.; (b) to reduce and document the data and deliver the archive to the Planetary 4 Data System (PDS) and the Pilot Land Data System (PLDS); (c) to take advantage of the multisensor approach, allowing cross-comparison of results acquired with different wavelengths and facilitating rigorous characterization of surfaces using multiple wavelength intervals; (d) to use results to test quantitative models for the extraction of surface property information from remote sensing data for Earth (e.g., Earth Observing System), Moon (e.g., Lunar Observer), Mars (e.g., Mars Observer), and Venus (e.g., Magellan); and (e) to use data in prototype EOS studies focused on the nature and ages of geological features and implications for regional climatic and tectonic histories. Table 2.1 shows relevant remote sensing instruments currently flying or scheduled to be flown on upcoming missions. The spaceborne missions will acquire new, detailed information about materials exposed on the surfaces of Venus, Mars, and Earth, respectively. For example, multiple incidence angle observations with the Magellan radar system and altimeter provide information about the Fresnel reflectivity and the roughness of the Venusian surface. The EOS SAR and its precursor, the Shuttle Imaging Radar (SIR-C), will provide multiple frequency, multiple incidence angle polarimetric data for the Earth's surface. The French ISM instrument on PHOBOS II measured the reflected infrared spectrum of Mars with approximately 22 km footprint widths for the equatorial latitudes. VIMS on some Mars mission, and HIRIS on EOS, are imaging spectrometers that will acquire detailed spectral reflectances in an image context. Similar capabilities will exist in the thermal infrared, with TES on Mars Observer. The EOS MODIS-T and MISR instruments will acquire multi-emission angle data to help in deciphering atmospheric and surface scattering and absorption. The TES system will also be used to acquire multi-emission angle data for given areas. The visible through microwave data to be acquired by the various spaceborne instruments will be used to extract information on mineralogy and composition (e.g., from locations, depths, shapes of absorption features due to electronic and vibrational processes) and on selected physical properties (e.g., grain size and degree of packing, macro-scale roughness) associated with surface materials. Appendices D-1, D-2, and D-3 contain detailed information about the airborne and field instruments used in the GRSFE campaign. GRSFE was designed to use, to the extent possible, instruments similar to those used or to be used on the space missions discussed in the last paragraph. Further, the GRSFE instruments acquired data for surfaces that are roughly analogous to those found on Mars and Venus, and that are likely to be regions of study with EOS data. Also, the GRSFE airborne and ground campaign was designed to ensure that the objective of testing quantitative data reduction models for regions with extensive ground truth would be met. For reference, Table 2.2 illustrates the types of analyses that the NASA-supported science community is currently pursuing using GRSFE data. The spaceborne missions of the 1990s will result in high volume, complex data sets that will present major archiving challenges. Thus, another major GRSFE objective was to explore the extent to which PDS and PLDS standards and guidelines could be used to facilitate archiving a 5 modestly complex set of data. The extent to which the Version 1.0 release of the GRSFE archive will be used by the research community will provide direct information on the utility of the standards and guidelines, and our overall archiving approach. TABLE 2.1 SELECTED EARTH AND PLANETARY REMOTE SENSING INSTRUMENTS Instrument Brief Description EOS HIRIS Earth Observing System Imaging (High Resolution Spectrometer with 192 bands covering 0.4 Imaging Spectrometer) to 2.5 micrometer with 10 nm sampling. 30 km swath with 30 m pixels. Pointable +60 degrees +/-30 degrees along track and 24 degrees cross track. EOS ASTER Broad band EOS thermal IR mapper. (Advanced Spaceborne Thermal Emission and Reflectance Radiometer) EOS MODIS EOS moderate resolution imaging (Moderate Resolution spectrometer. Two optical packages, Imaging Spectrometer) including MODIS-T (tilt) and MODIS-N (nadir view). MODIS-T capable of 60 degrees pointing along track. 52 bands between 0.4 to 12 micrometer. 1500 km width with 0.5 to 1.0 km pixels. To be augmented with MISR. EOS SAR EOS radar system operating at 3.5 cm (X- (Synthetic Aperture band), 5.66 cm (C-band), 23.98 cm Radar) (L-band). Incidence angles between 15 to 55 degrees. C, L bands have polarimetric capability. High resolution mode will have 20 to 30 m pixels, 30 to 50 km cross track widths. Magellan Radar S-band (12.6 cm) altimeter. Allows Altimeter estimates of elevation, quasi-specular roughness, and Fresnel reflection coefficient. Magellan Radar System S-band SAR with approximately 150 m best radar resolution. HH (or VV) polarization. Multiple incidence angle coverage during extended mission by 6 rotating spacecraft for given area. Incidence angle varies with latitude. TES Mars Observer instrument covering (Thermal Emission thermal IR with high spectral Spectrometer) resolution. Also broad band channel for VISIR radiance. Three km pixels. VIMS Imaging Spectrometer covering (Visible and Infrared approximately 0.3 to 5.0 micrometers Mapping Spectrometer) with high spectral resolution. Will fly on some Mars Mission. TABLE 2.2 EXAMPLES OF USE OF GRSFE DATA - To test multi-spectral radiative transfer models for scattering and emission from planetary surfaces, including: - Use data to evaluate procedures for separation of macroscale roughness (Hapke rms; grain size, packing; radar-derived roughness) from other surface properties (dielectric constant, refractive index, emissivity). - Use AVIRIS, ASAS, TIMS to understand mixed pixels (including shade and aeolian components) for studies to determine the composition of the upper 100 micrometer of an outcrop. - Use TIMS data collected four times in one diurnal cycle to understand the physics of thermal emission. - Use multi-incidence angle C, L, P polarimetric SAR to test models for radar backscatter from planetary surfaces. - Use data to help establish the paleoclimatic history of the arid S.W. United States. - Use data to understand remote sensing signatures of basalt flows of varying ages and depth of aeolian fill. - Use data to develop models for crater degradation (rille development, wall slumping, etc.). - Use SAR to pursue understanding of the correlation between radar roughness and aerodynamic roughness. - Use data to support landing site selection procedures for Mars Rover Sample Return (MRSR) Mission, concentrating on extraction of roughness information. 7 3. OVERVIEW OF GRSFE REQUIREMENTS AND IMPLEMENTATION 3.A. SITE SELECTION A major task of the SSG was the selection of GRSFE sites for the coordinated airborne and field campaign. A number of site requirements were imposed, including relevance to planetary surfaces. Clearly there are no sites on Earth that simulate surface properties of Mars or Venus in detail. The surface of Mars, for example, is highly desiccated and probably has been for millions, if not billions, of years. The 750 degrees K, 90 bar carbon dioxide conditions at the Venusian surface probably lead to unique surface properties. Rather, the approach was to select sites that are roughly analogous to what we expect on planetary surfaces and, as discussed in the last section, to use GRSFE data to test models for extraction of surface property information. Volcanic terrains, aeolian deposits (e.g., sand dunes), and fluvial landscapes exist on Mars, as do an abundance of craters. Certainly, volcanic terrains and craters abound on Venus. Aeolian deposits are also seen and ancient fluvial landforms are a possibility. Thus, reasonable analog sites for Mars and Venus include volcanic, aeolian, and fluvial surfaces. The terrestrial sites also needed to be as dry as possible, relatively free of vegetation, and accessible by aircraft and from the ground. Finally the sites needed to have relevance for EOS prototype tasks, which the SSG decided were best focused on neotectonic and paleoclimatic studies. Two specific site types were defined for GRSFE. Modeling sites were designated for concentrated airborne and field observations and resultant detailed modeling of GRSFE data. These areas were characterized on the ground in enough detail to provide first-order quantitative simulations of how electromagnetic radiation interacts with geological surfaces and materials in the visible and reflected infrared, thermal infrared, and microwave wavelengths. Surface topography was obtained at a variety of length-scales. Samples were collected so that composition, mineralogy, and physical properties could be characterized close enough together and at sufficient depth intervals to be useful in modeling interactions of radiation with the sites. Various in-situ field measurements were made of both the surface and the atmosphere at the same time as the airborne campaign was underway. The modeling sites thus needed to be where extensive field work and sample collection could be accomplished. Further, the collection of sites needed to be simple enough to allow modeling of remote sensing signatures from ground data. Table 3.1 gives an overview of all GRSFE sites. The modeling sites are located in the Lunar Crater Volcanic Field. Detailed locations are given in Table 3.2. The modeling sites are discussed in more detail in Section 4 of this document. Calibration sites were defined to be regions where ground measurements were obtained to be able to calibrate the airborne data independently of any instrument-specific procedures, e.g. pre-flight and post-flight AVIRIS instrument radiometric calibrations. Logistical and financial constraints limited both the number of calibration sites and 8 the ground measurements that could be acquired at each site. We focused on field spectrometer measurements to characterize the reflectance and emittance of various surfaces. Corner reflectors were also deployed to calibrate AIRSAR. The specific calibration sites are summarized in Table 3.2. TABLE 3.1 OVERVIEW OF GRSFE FIELD SITES Lunar Crater Volcanic Field, Nevada Location: 250 km northwest of Las Vegas, NV (38 deg. 15'N, 116 deg. W). Age: Middle to late Pliocene and Pleistocene (0.015 to 4.2 m.y.). Features: - The field contains about 95 vents and at least 35 associated lava flows within a northeast-trending zone, up to 10 km wide and about 40 km long. - Vents include cinder cones, elongate fissures, and at least two maar craters. - Lava flows range up to 1.9 km wide and 6.1 km long with thicknesses of less than 3 to as much as 25 m. - Progressive degradation of the cones and flows is very similar to that displayed by other basaltic volcanic fields in the southwest Basin and Range (including the Cima, Crater Flat, and Coso fields). - Many of the flows in the northeast and central parts of the field are veneered with various thicknesses of air-fall tephra. - In other areas, all but the youngest flows are mantled with extensive deposits of aeolian silt and fine sand. - Full range of igneous (volcanic) rocks present. Sites of interest: Lunar Lake (a playa) was used as modeling site because it has fan on one side, transition to cobble-strewn playa, then to playa, and on other side are volcanic materials of various ages and compositions. Death Valley, CA Location: 300 km northeast of Los Angeles, CA (36 deg 31'N, 116 deg 50'W). Age: Alluvial fan units are as old as 800,000 years. Features: A variety of rock types (metamorphosed Precambrian Paleozoic limestones, quartzites, and shales and Miocene volcanic rocks) are present. 9 Sites of interest: - Basaltic lava flows and fanglomerates have been exhumed to form a bouldery surface appearing much like the surface of Mars as observed by Viking Lander 2. Informally called "Mars Hill". - Sand dunes, about three kilometers across, rise to 50 m and are located to the northwest and east of Stovepipe Wells. - Ubehebe crater, approximately 700 m in diameter, and about 150 m deep, is located in the northern end of Death Valley. A small amount of basaltic tephra that erupted during the explosion that created Ubehebe blankets the area. Southern Mojave Desert, CA Location: Southeast of Baker, California (35 deg 15'N, 116 deg 45'W). Age: Variable. Features: Sand dunes, alluvial fans, basaltic volcanic flows. Sites of interest: - Kelso dune field: an extensive, complex dune field. - Cima volcanic field: basaltic flows and tephra cones ranging from recent to several million years in age. - Providence Mountains: alluvial fans of granitic and carbonate provenances. TABLE 3.2 GRSFE SITE SUMMARY Specific site names are given in left-hand column. The next two columns provide the approximate latitude and longitude of each site. Line and sample locations for each site are then given for ASAS, AVIRIS or TIMS images that cover the relevant locations. Line and sample coordinate origin is in the upper left of the image data; lines are rows and samples are columns in the image arrays. See Section 13 for a discussion of file naming conventions for the airborne products. See Section 3.D for discussion of calibration site descriptions. Lat. Lon. Finder Image, deg. N deg. W Line, Sample ----- ----- ------------ Modeling Sites Lunar Crater Volcanic Field, NV ASASLL05F.IMG - Playa 38.38 116.02 203, 301 - Disturbed Playa 38.38 116.02 205, 244 - Cobble Strewn Playa 38.38 116.02 242, 279 - Mantle Lava Flow 38.38 116.02 23, 105 10 Calibration Sites Death Valley, CA - Death Valley Dunes - Bright Target TIMDD03A.IMG 36.65 117.13 4628, 335 - Devil's Golf Course - Bright Target AVRDG04A.IMG 36.3 116.83 370, 315 - Trail Canyon Fans - Dark Target AVRDG04A.IMG 36.3 116.92 354, 550 - Ubehebe Maar - Dark Target TIMUB01A.IMG 37.0 117.45 2824, 363 Cima Volcanic Field, CA - Bright Target AVRCM05A.IMG 35.2 115.75 284, 336 Cima Volcanic Field, CA - Dark Target AVRCM05A.IMG 35.2 115.75 246, 341 Kelso Dune Field, CA - Dark Target AVRPV29A.IMG 34.9 115.73 173, 143 3.B. AIRBORNE CAMPAIGN The airborne component of the GRSFE campaign was conducted on July 17, 1989 for ASAS; September 28-29 and October 4, 1989 for AVIRIS; July 17, September 27-29, 1989 for TIMS; and September 13-14, 1989 for AIRSAR (see Table D.1 for brief instrument descriptions; Tables D.4 to D.6 summarize coverage by these instruments.). Schematic maps showing flight lines covered by each airborne instrument are provided in three image files in the LOCATOR directory on the Volume 1 CD-ROM. AVIRIS flew on the NASA ER-2; ASAS and TIMS both flew on the C-130 on July 17. TIMS flew alone on the C-130 during September. AIRSAR was on board the NASA DC-8 aircraft. The airborne campaign was organized by flight lines and runs. The lines denote the azimuthal direction and start and stop positions for acquisition. The runs indicate the number of acquisitions along a given line that occurred during a given data acquisition period. Section 11 of this document describes the index files that provide detailed tabular descriptions of the airborne data. Tables D.7 to D.10 provide conversions between GRSFE file names, lines, and runs, and information used by each instrument archive. These tables are needed to obtain data directly from the data producers, e.g. from the AVIRIS processing facility at JPL. However, remember that these data are also contained on the GRSFE CD-ROMs. The intent of the airborne campaign was to fly all four instruments at the same time over the modeling sites and within a day or so of one another for the calibration sites. In fact, ASAS and TIMS collected data throughout the day of July 17 until engine trouble on the C-130 forced cessation of operations. The ER-2 flew on July 17 with AVIRIS, but operational difficulties with the instrument precluded data collection. The DC-8 was unable to fly on July 17 because of scheduling problems, so it was not possible to get AVIRIS and AIRSAR coverage until September. It was decided to refly TIMS during September, but prior commitments for ASAS precluded its use during that month. Thus, despite 11 best intentions, and a field campaign focused on the July 17 date, instrumental difficulties forced the full airborne campaign to be spread out over three months. We also note that a sun photometer from the NASA Ames Research Center flew on the C-130 during the July deployment. However, it proved to be impossible to extract reduced data records. Thus, these data were not included in the GRSFE archive. For the modeling sites the intent was to acquire multitemporal, multiangle (incidence and emission) AVIRIS, ASAS, and TIMS data. In fact, during the July deployment, TIMS data were collected over the modeling sites at approximately 4:17 (pre-dawn), 8:05 (early morning), 12:16 (noon), and 13:55 (afternoon) Pacific Daylight Time (PDT). ASAS data were acquired during the latter three times. AVIRIS data were acquired over these sites at 9:44, 11:43, and 13:44 PDT on September 29, 1989. For ASAS, data were acquired both along the principal solar plane and perpendicular to it. The multiple angle ASAS and AVIRIS data for the modeling sites will allow tests of radiative transfer models for surface reflectance. The multitemporal TIMS data will allow pursuit of emissivity and thermal inertia studies. AVIRIS and TIMS data were acquired for the calibration sites during one pass. Finally, ASAS also acquired data for a portion of a lava flow north of the modeling sites and for the Ubehebe site. AIRSAR observations over the modeling sites were focused on multiple incidence angle coverage with values of approximately 25, 35, and 45 degrees. Perpendicular tracks were also acquired. A mix of single and multiple angle coverage was acquired for the calibration sites, as noted in Table D.6. The multiple incidence angle AIRSAR data will allow detailed scattering models to be evaluated. 3.C. FIELD CAMPAIGN The GRSFE field campaign supported the July and September airborne observations. A base camp was set up at Lunar Lake for the July campaign focused on the modeling sites. Personnel were also deployed at the various calibration sites listed in Table 3.2. Activities were divided into five functions: (a) atmospheric measurements at Lunar Lake; (b) detailed ground observations at the modeling sites; (c) thermal emission experiments at the Lunar Crater Volcanic Field in July by researchers from the University of Colorado; (d) calibration site measurements in the visible, reflected infrared, and thermal infrared wavelengths; and (e) deployment of corner reflectors at the microwave calibration sites. The calibration team focused on acquiring measurements for bright and dark (in VISIR) surfaces at each calibration site. The University of Colorado radiometry and thermistor experiments are described in Section 5 of this document. Table D.2 summarizes the atmospheric instruments used during the field campaign, and Table D.3 summarizes the other field instruments that were utilized, including field spectrometers and corner reflectors. Table D.11 summarizes the participants involved in various aspects of the field campaign. The locations of the modeling sites at the Lunar Crater Volcanic Field are given relative to an ASAS frame in Table 3.2. Table 3.3 provides 12 location information for the other activities that took place on Lunar Lake. TABLE 3.3 LOCATIONS OF SELECTED GRSFE ACTIVITIES ON LUNAR LAKE Locations are given relative to ASAS frame ASALL05F.IMG. Activity Line Sample -------- ---- ------ Lunar Base Camp 170 198 Univ. Colorado Directional Emissivity Experiment 178 221 Washington Univ. Thermistor and Radiometry Experiments 271 279 (see Section 3.D) Reagan Radiometer and Spectral Hygrometer 177 201 Weather Station 237 337 3.D. NARRATIVE OF FIELD CAMPAIGN ACTIVITIES This section is a chronological sequence of the field campaign activities. This master sequence, together with the other parts of this section, Sections 4 and 5, and the data set descriptions, should allow the reader to understand how the airborne and field campaigns were implemented and how the resultant data can be used in synergistic fashions. Times are given in PDT. July 15, 1989 (Lunar Lake) Greeley and Lancaster set up and began measurements of wind velocity with ASU wind velocity experiment. Arvidson and Plaut set up weather station, which began measuring air temperature, wind velocity, and direction. Weather station was located approximately 700 m ENE of cobble site. Wind velocity experiments (2 towers) were located NNE of cobble site. Deering (PARABOLA, GSFC sun photometer, pyranometer), Zurek and Norris (Reagan Radiometer and Spectral Hygrometer) arrived at Lunar Lake in PM to check out site and equipment. (Note: GSFC sun photometer data not included on version 1 release of GRSFE archive.) July 15, 1989 (Calibration sites) CALTEAM arrived at Kelso Dunes about 11 am. Found a "bright" area on the nominal AVIRIS Afton-Kelso and Providence fans lines at the first parking area on road to dunes. Appeared on TM image as eastern "lump" in road. Site name: Bright Target. (Note: Target not covered during airborne GRSFE campaign.) Took several SIRIS, Daedalus, and PFES spectra. Done about 1 pm. SIRIS, Daedalus moved at about 12:30 pm to a dark area at the pumping station near the intersection of Kelso Dunes Road with Kel-Baker Road. Site is Dark Target. Obtained several spectra. Surface samples collected at both sites. 13 Moved to Cima Volcanic Field at Cone I, measured tephra (dark). Started about 2 pm. Obtained SIRIS, Daedalus, PFES spectra and surface samples. Site called Dark Target. Found bright area on southeast flank of Cone I in a dry stream channel. Site called Bright Target. Done about 5 pm. July 16, 1989 (Lunar Lake) Four modeling sites staked out by Arvidson and Plaut. Disturbed playa site generated by driving automobile around in circles, churning up playa. July 17, 1989 (Lunar Lake) Note: ASAS, TIMS acquired data. Modeling sites characterized using random grid approach. Reagan Radiometer and Spectral Hygrometer obtained data. PARABOLA measurements of atmosphere and surface at cobble site obtained. SIRIS, Collins, Daedalus, PFES spectra were acquired over selected playa, volcanic, fan and vegetated surfaces at modeling sites. Performed SIRIS and Daedalus field spectrometer intercomparison at playa, varnished rock, and disturbed playa sites. Corner Reflector Team arrived at Lunar Lake with a truck full of corner reflectors. July 18, 1989 (Lunar Lake) PARABOLA obtained data at playa and mantled flow modeling sites. Irons examined sites and collected soil samples. July 18, 1989 (Calibration sites) Arrived Trail Canyon fan site about 11:45 am. Site name is Dark Target. SIRIS and Daedalus obtained data along traverses. PFES died from heat. Surface samples collected. Descended to first clean, salt covered flood plain area on Devil's Golf Course at about 1 pm. Obtained SIRIS and Daedalus measurements and surface samples. Site name is Bright Target. Moved to Death Valley dunes site. SIRIS spectra acquired at sand dunes picnic area, Daedalus spectra at main dunes off Stovepipe Road. Samples obtained. Called Bright Target. July 18, 1989 (Corner reflector sites) Corner Reflector deployment team deployed reflectors and acquired soil moisture and dielectric constant data. Reflectors were placed in a number of locations and in two orientations to cover the several parallel and one crossing swath. Note: Detailed field measurements were conducted in May 1988 at the Pisgah lava flow and adjacent Lavic Lake as part of airborne polarimetric SAR coverage associated with the Mojave Field Experiment. Analysis of these data show that topographic variations largely control variations in cross section. Spatial variations in dielectric constants were found to be much less important. Thus, the emphasis for the radar part of GRSFE was to deploy corner reflectors to allow calibration of the data, and to acquire helicopter stereophotographs. 14 The main objective of the Corner Reflector Team was to place trihedral corner reflectors within the image swaths for calibration of the aircraft SAR data. In this technique, the known phase characteristics and backscatter cross-sections of the corner reflectors are used to find the transfer function from image pixel values to surface backscatter cross-section. Typically, several corner reflectors are placed spanning the swath to determine the transfer function as a function of incidence angle. Near-surface samples were also collected for determination of soil moisture. July 17 - 18, 1989 (Lunar Lake Subsurface Temperature Experiments) Temperature data were collected from the playa and cobble modeling sites. This work was led by Petroy, Washington University, and is a separate experiment from the University of Colorado experiment described in Section 5. Eight (8) thermistors were buried 4 m apart by Arvidson and Plaut, each to 3 cm depth, in a continuous line starting in the cobble site and extending south into the adjacent playa. The emplacement was done by noon. A thermistor is an electrical resistor which makes use of a semiconductor whose resistance varies sharply in a known manner with temperature. Each probe is approximately 2 mm in diameter and is placed in direct contact with the soil, usually buried. The resistance across the surface of the thermistor is measured; that resistance is then connected to the temperature of the soil. The thermistor probes used in this experiment were manufactured by Yellow Springs Instruments (YSI) and are accurate to 0.2 degrees C. Temperatures were recorded by Petroy from each site every hour over a 25 hour period, beginning on July 17, at 6:00 pm and concluding on July 18, at 6:00 pm. Results are presented in Table E.1. In addition to these shallow test sites, at approximately 1 pm, thermistors were buried to greater depths in the playa just to the south of the cobble site, and at the southern edge of the cobble site, to measure the depth of the diurnal temperature wave. Thermistors were buried at 30, 10, and 3 cm depths at both locations. In all cases soil was tamped back in place and cobbles returned to their original positions. Temperatures were recorded at these sites every hour over the same 25 hour period. Results are shown in Table E.2. A thermometer and hygrometer were mounted approximately 0.7 m off the ground just to the north of the cobble site. Air temperature and relative humidity were recorded every hour over the same 25 hour period. Results are given in Table E.3. July 19, 1989 (Calibration sites) Arrived at Ubehebe crater about 10 am. Collected SIRIS and Daedalus spectra at Dark Target on north flank of Ubehebe Crater. SIRIS obtained spectra of small playa just to north of dark site, but it seemed too small for calibration, so SIRIS team chose another site where Ubehebe Crater Road crossed stream to east of crater. Clouds began moving in from east. Surface samples collected. July 20, 1989 (Corner Reflector sites) Reflectors were deployed at Death Valley. Four corner reflectors were placed at Trail Canyon fan: one at the base and three distributed 15 along the road going up the fan. Areas of desert pavement were chosen for their smooth background. At Mars Hill, near the exit of Artist's Drive, one corner reflector was placed south of the hill and one north. At Stovepipe Wells, four corner reflectors were distributed along Highway 190 from just east of Stovepipe Wells, east to sand dune picnic rd. At Ubehebe Crater, three corner reflectors were distributed along the crater access road from about 1.4 mile west of its junction with North Highway, west to near the crater. One additional corner reflector was placed about 1 mile down Racetrack Valley Rd. July 21, 1989 (Corner Reflector sites) Reflectors were deployed at Kelso. Four corner reflectors were placed along the Kel-Baker Rd.; two north of the railroad tracks and two south. July 26, 1989 (Corner Reflector sites) Checked orientation of corner reflectors at Kelso. July 27-28, 1989 (Corner Reflector sites) Checked corner reflectors at Death Valley, Ubehebe, and Lunar Crater. September 10, 1989 (Corner Reflector sites) Arrived Kelso about 3 pm. All four corner reflectors were OK. Arrived Death Valley about 18:00. Checked corner reflectors at Mars Hill. Corner reflector 1 was loose and off its base and its vertical sides were bowed inward. September 11, 1989 (Corner Reflector sites) Checked corner reflectors at Stovepipe Wells and found westernmost corner reflector (corner reflector 1) slightly off base. Corner reflector 2 needed minor adjustment. Corner reflector 3 was near a public turnout and was therefore most disturbed. Corner reflector 4 was OK. Proceeded to Ubehebe Crater about 10:30. Started with corner reflector 1 on Racetrack Rd. Corner reflector 1 was turned over due to soft ground. Corner reflector 2 (next to east) also blown over due to soft ground. Corner reflectors 3 and 4 OK. September 12, 1989 (Corner Reflector sites) Lunar Crater corner reflectors were checked and found to be OK. September 26, 1989 (Lunar Lake) A second thermistor-based experiment was conducted by Petroy, Washington University, during the September GRSFE field campaign at Lunar Lake. In addition, a radiometer was used to obtain brightness temperatures as part of this experiment. Petroy and Plaut dug two pits on mantled flow modeling site at 1 p.m. Three stakes left over from July campaign were found to be still standing on SE, NW, NE corners of site, as well as small stakes from random sampling. Pit A dug 25m south of site. Three samples obtained at different depths after which thermistors were placed at 10, 20, and 30 cm depths. Rock hammer point was used to make holes for the thermistors. Soil was tamped back into hole and rocks placed on top in approximately natural density. At 1:30 16 pm, pit B was dug at about 12 m from south boundary of site. Three samples were obtained at different depths, after which thermistors were placed at 3, 10, 20, and 30 cm depths. Soil was tamped back into hole and rocks placed on top in approximately natural density. Pictures were taken before, during, and after the activity. At 2:10 pm, visited disturbed playa modeling site. Surface was found to be back to normal, except for a few faint skid marks still visible. Clouds built up all morning to about 50% cover. Altimeter set for altitude of Lunar Lake: 5760 feet (from topographic map) = 1756 m. Barometric pressure based on altimeter at 9:00 a.m. was 622 mm. 10:00 a.m.: clouds from south, 20% cover. 10:05: Deployed weather station and began recording data. At 12:40, 50% cloud cover, dust kicking up on east side of lake. At 14:10, 55% cloud cover. At 16:00, 35% cloud cover. At 17:45, 40% cloud cover, wind dying slightly. Barometric pressure at 12:40 was 622 mm, at 14:10 it was 620 mm, at 17:45 it was 619 mm. September 27, 1989 (Lunar Lake) Note: TIMS data acquired on September 27 and 29, 1989. At 6:30, 75% cloud cover. At 8:05 barometric pressure was 621 mm with clouds to E, N, S, and wind picking up. By 11:00 there was 100% cloud cover. At 18:24, clouds clearing, wind dying. September 26-27, 1989 (Lunar Lake Surface and Subsurface Temperature Measurements) Temperature data were collected from the mantled flow modeling site using the thermistors buried to 30, 20, 10, and 3 cm depth. Temperature data were collected every hour over a 25 hour period beginning at 6:00 pm on September 26 and concluding at 6:00 pm on September 27. This period was originally to coincide with the TIMS overflight, however, weather precluded acquisition of good data for two days. It should be noted that during the 25 hour monitoring period, the sky was generally overcast and it was extremely windy (up 30 km/hr). Table E.4 provides the temperature data. In addition to the subsurface temperature data, surface temperature data were collected using a Raynger Radiometer at both pit sites during the same 25 hour period. Temperature measurements were made by holding the instrument vertically over the site about 1 meter above the surface (emission angle = 0 degrees), setting the emissivity variable on the instrument to 1.0, and collecting two temperature measurements. Then, the instrument was positioned vertically over an adjacent undisturbed surface and two more brightness temperature measurements were collected. Data are presented in Table E.5. Additional data: On the day of the TIMS overflight (September 29, 1989) surface temperature data were also collected on the playa at two sites - an undisturbed playa site and a cobble site (not the same sites as described during the July GRSFE) located within meters of Lunar base camp. Data were collected every hour for four (4) hours. At the playa 17 site, two sets of temperature data were collected - one set with the emissivity=1.0 and one set with the emissivity=0.95. Unless otherwise noted, all data collected with the Raynger instrument were collected with the emissivity=1.0. Also, one surface temperature measurement was collected at the mantled flow test sites at 1:00 pm during the day of the overflight. Table E.6 presents these data. September 28, 1989 (Lunar Lake) At 6:30, 75% cloud cover, cold, and no wind. At 7:25 barometric pressure was 624 mm; at 9:05 it was 625 mm, with clearing to E, S, 60% cloud cover. September 29, 1989 (Lunar Lake) At 6:30 it was clear and calm. Set up Reagan Radiometer. Sunrise at 6:50, started measurements at 6:52 a.m. with Reagan Radiometer and two Spectral Hygrometers. At 10:58 barometric pressure was 623.5 mm. At 11:00, wind strong, kicking up dust on lake. 4. MODELING SITE DESCRIPTIONS As noted in the previous section, four modeling sites were selected in the Lunar Crater Volcanic Field. The modeling sites are representative of the range of surface compositions and terrain types found in the area. The sites selected were: a) undisturbed playa, b) disturbed playa, c) cobble-strewn section of the playa, and d) a mantled lava flow. Locations of the sites are given in Table 3.2. The disturbed playa was generated by driving a vehicle around in circles until the playa material was thoroughly churned up. The cobble site is a mix of basalt cobbles and playa. The mantled flow site is an old flow that has been partly buried by aeolian debris and covered with a desert pavement of basalt cobbles and boulders. For each site, a 50 m by 50 m area was delineated using stakes, with boundaries approximately aligned with north-south and east-west directions. Each site was characterized in terms of its general surface composition and particle size, vegetation cover, soil moisture status, and compositional "end" members. Note: The following material was edited from contributions by Greeley and Lancaster. Sub-areas within each site were selected randomly within the 50 m by 50 m squares using the following method. Each 50 m by 50 m square was divided into 25 5 m by 5 m sub-sites. Five squares were randomly selected for sampling and characterization, and were located using the coordinate system below. Sub-site numbers refer to the squares numbered as below. NW NE 5 10 15 20 25 4 9 14 19 24 18 3 8 13 18 23 2 7 12 17 22 1 6 11 16 21 SW SE At each sub-site, photographs (black and white and color) were taken to show: a) the general nature of the sub-site, b) a close up of the surface, and c) a near vertical view of the surface. The particle size composition of each sub-site was estimated visually for a 1 m by 1 m quadrant. A 1200 g sample of the surface materials was taken for soil moisture determination, and placed in a sealed can. Field determinations of soil moisture were made by Hughes (JPL) at selected sub-sites at the disturbed playa, undisturbed playa and cobble sites, using a "Speedy" soil moisture tester. Samples of the compositional end members were collected at one sub-site location for each site. Detailed descriptions of each site follow. Times are in PDT. A. Undisturbed playa (July 17th, 10:05-10:40 am) General characteristics: This was a smooth, flat, silty-clay playa surface displaying major 20-30 cm polygons with fine cracks, and clay "sheen" to surface. Within 20-30 cm polygons were finer cracks that formed 2-3 cm polygons. A few scattered clusters of 2-3 cm basalt gravel were visible. There were also some very small craters on the surface in places caused by degassing or water escape. Particle size: silty clay Vegetation: one green Atriplex bush End member samples: Basalt pebbles; Silty clay playa surface Sub-site 3: There was smooth playa soil; a moisture sample was taken. Sub-site 5: The sub-site had smooth playa soil; a moisture sample was taken. Sub-site 7: Here there was smooth playa with scattered 1-2.5 cm basalt pebbles covering < 5% of surface. A soil moisture sample was taken, resulting in measurements of 1.38 wt.%, 1.56 wt.%. Sub-site 12: This area was smooth playa soil, and a moisture sample was taken. Soil moisture measurements were 1.73 wt.%, 1.88 wt.%. Sub-site 17: This sub-site was smooth playa. A soil moisture sample was taken, resulting in measurements of 1.64 wt.%, 1.83 wt.%. Sub-site 19: This area was smooth playa containing very scattered 1-2 cm basalt pebbles. End member samples of basalt and silty-clay were taken here. A soil moisture sample was also taken. B. Disturbed Playa (July 17th, 9:00-9:56 am) 19 General characteristics: Playa surface was disturbed by driving an automobile around the site for several tens of minutes to "roughen" surface and destroy original surface. End member samples : undisturbed playa surface within this site disturbed playa surface within this site Particle size: silty clay Vegetation: A total of six Atriplex bushes, each 0.4-1 m in diameter occurred at this site. Only two were green. Most bushes have accumulated silt and clay around their lower stems to form a small mound. Sub-site 3: There was well disturbed surface soil; a moisture sample taken. Field soil moisture measurements were 0.50 wt.%, 0.22 wt.%. Sub-site 10: End members and soil moisture samples were taken here. The surface was only lightly disturbed. Soil was largely silty clay, with 10-20 cm polygonal cracks and rare basalt pebbles. Soil moisture measured by Farr's instrument 1.75 weight %. Sub-site 18: This surface was well disturbed, mostly powder and a few clods up to 5 cm across; a soil moisture sample was taken. Sub-site 19: The surface was well disturbed, and a soil moisture sample was taken. Sub-site 22: This was a well disturbed surface. Soil moisture sample was taken, resulting in measurements of 1.59 wt.%, 1.78 wt.%. Sub-site 25: Very few car tracks here; the surface was virtually undisturbed. A soil moisture sample was taken. C. Cobble site (July 17th, 10:50 am -12:03 pm) General characteristics: The site was east of 4-5 m high basalt knob that formed an "island" in the playa. Also, the site had numerous blocks and clasts derived from basalt, together with silt matrix. There was poorly developed desert pavement in many areas. See table given below for particle sizes. Area percentage of: Sub-site Clay/silt Sand Gravel Cobbles ------------------------------------------- 5 75 5 15 5 8 27 3 65 5 9 78 5 15 2 14 37 10 45 8 16 15 2 75 8 23 18 10 70 2 20 Vegetation: Atriplex bushes were scattered about; they were up to 1.5 m across and 40 cm high. Some had silt-clay accumulations at their base. End member samples: Playa silts; Basalt pebbles; Basalt cobbles. Sub-site 5: This area had a silt-clay playa surface with scattered basalt gravel and cobbles; scattered Atriplex bushes measured 20 cm high, 50-60 cm across, spacing 2-3 m. A soil moisture sample was taken, resulting in measurements of 0.92 wt.%, 1.04 wt.%. End member samples were taken here. Sub-site 8: This sub-site consisted of gravel-sized basalt clasts resting on a silty substrate. This incipient desert pavement surface contained two dead 5 cm high Atriplex, 40 cm apart. A soil moisture sample was taken, resulting in measurements of 0.47 wt.%, 0.59 wt.%. Sub-site 9: This was a silty-clay surface containing scattered gravel. Two Atriplex bushes were there, 40 cm high, 1-1.5 m across, 1 m spacing. Sub-site 14: Silty-clay playa material and gravel clasts were at this site, but no vegetation. PARABOLA Sub-site was 10 m east of here. Sub-site 16: This was a gravel surface containing dead herbs, 10-15 cm high, 25 cm spacing, 10 cm diameter. A soil moisture sample was taken. This was PFES Sub-site. Sub-site 23: Here there was gravel having occasional cobbles; a silty surface was below basalt clasts, but no vegetation. A soil moisture sample was taken, resulting in measurements of 0.62 wt.%, 0.83 wt.%. D. Mantled Lava Flow (17th July 1:00-2:30 pm). General Characteristics: The site was north of Lunar Lake, and consisted of lava flow with surface of boulders to gravel with windblown silt cover, forming a well developed desert pavement surface. The flow formed a scarp approximately 12 m high to the southeast overlooking Lunar Lake. The basalts forming the flow were "layered", with each unit having a thickness of ~1.5 m. Basalt clasts on the surface were vesicular, with a well developed desert varnish on their surface. The sub-site is fairly flat, dipping 1-2 degrees to the southeast. There were some silt mounds around bushes, and some caliche chips on the surface. See table given below for particle sizes. Area percentage of: Sub-site Clay/silt Sand Gravel Cobbles Boulders ------------------------------------------------------------------------ 8 28 1 60 10 1 9 60 2 20 10 8 10 43 2 30 15 10 18 15 65 20 21 19 37 1 45 15 2 20 60 15 15 10 Boulders are all < 30 cm in diameter. Vegetation: Sage brush and rabbit brush. End member samples: Basalt cobbles, silt, caliche, red varnished basalt. Sub-site 10: This area had boulders and gravel having silty cover; a soil moisture sample was taken. Sub-site 9: Basalt gravel formed a desert pavement surface with large silt patches here. A soil moisture sample was taken. Sub-site 8: Here there was basalt gravel with smaller silt patches. A soil moisture sample was taken. Sub-site 18: This area had basalt gravel and cobbles having some silt and desert pavement formation. A soil moisture sample was taken. Sub-site 19: This site had basalt gravel and a cobble surface. A soil moisture sample was taken. Sub-site 20: Basalt boulders, cobbles, and gravel with silt were here; a soil moisture sample was taken. 5. UNIVERSITY OF COLORADO DIRECTIONAL EMISSIVITY EXPERIMENT Note: The following material was edited from text and data from Jakosky, Finiol, and Henderson. Thermal emission from geologic surfaces is known to be non- isotropic due to the presence of surface rocks and slopes (which have non-uniform kinetic temperatures) as well as emissivity effects arising from the non-uniform Fresnel emission from a flat or a rough surface. These effects have been observed on the Moon, Mars and other rough surfaces in the solar system. Observations of directional effects can tell us something about the surface roughness and structure of a planet's surface; because most spatially resolved thermal infrared observations of the earth or planets are done with near-nadir viewing only, the magnitude of such effects must also be known to properly interpret the diurnal energy balance (and derivation of thermal inertia) and emission spectra. As part of GRSFE we investigated the directional variations in thermal emission of different surfaces. Results obtained using hand- held thermistor probes (to determine local kinetic temperature) and a ground-based, hand-held infrared radiometer (to determine scene-averaged infrared brightness temperature from multiple view angles) are reported. The text of this document describes the experimental protocol and some 22 of the results and conclusions. The radiometer observations are provided in the file RMTLL001.TXT, in the DIREMISS directory on GRSFE Volume 1. The thermistor data are located in the file THMLL001.TXT in the same directory and volume. Field work took place in the Lunar Crater Volcanic Field. Three natural sites were selected, and four artificial sites were constructed. The natural sites included: (1) Dry playa. The selected site was relatively smooth and flat, with a very small number of interbedded rocks less than 0.5 cm in size. Although a number of dessication cracks were present, they occupied a small fraction of the radiometer field of view; observations at multiple viewing angles and directions suggests they are not important in the thermal emission from the ensemble surface. (2) Rocky playa. This surface consisted of dry playa material, with approximately 10 % of the surface covered by rocks with size 1-10 cm. (3) A'a lava flow. A 3-m square, vegetation-free surface was selected within the Black Rock Lava Flow; the surface was extremely rough, with 1-m variations in heights occurring. Four artificial sites were constructed on the Lunar Lake playa close to GRSFE base camp. The artificial sites consisted of a 1-m- square patch of material overlying undisturbed playa material. The sites were: (1) Smooth sand. Overturned playa material was covered to a uniform depth of about 10 cm with sand. (2) Smooth sand plus a single rock. A similar sand surface was constructed, and a single 13-cm cubical rock was placed on top. (3) Pebble surface. Smooth playa material was covered to a depth of about 10 cm with 1- to 3-cm rounded pebbles. (4) Rocky surface. Playa material was covered with a close- packed single layer of 15- to 30-cm slightly weathered and rounded rocks. After construction, each site was allowed to partially equilibrate with sunlight and ambient temperatures for 36 hours prior to beginning measurements. Infrared brightness temperature measurements for each surface were obtained with a hand-held 8- to 14-micron broadband infrared radiometer obtained from the Cole-Parmer Instrument Co. Manufacturer's specifications indicate an absolute calibration to about 3 K, with relative uncertainties between measurements of better than 2 K; field investigation suggests a relative calibration that was usually better than this over short time spans. As we were investigating variations in emission as a function of viewing geometry, the absolute calibration of the instrument was not a limiting factor. Although the field of view of the radiometer is small, an internal averaging function allowed the instrument to be swept over the entire site in a boustrophedonic pattern in order to obtain a reading of the brightness temperature of the ensemble surface. Measurements were made of each site at emission angles of 0, 30 and 60 deg, and, for the latter two emission angles, every 45 deg of azimuth; measurements of the sand and playa sites were obtained at additional emission angles. Experiments were performed to determine the radiometer field of view using adjacent surfaces which had differing temperatures; the field of view was sufficiently well-defined that no significant emission was thought to come from regions outside of the specific sites. 23 Measurements of the actual surface kinetic temperature were made with a hand-held thermistor probe. The probe was thermally connected to the surface only at the time of the measurement and insulated from the atmosphere by a molded piece of styrofoam. The probe itself had a time constant of 10 s in air, and was held in contact with each surface for up to 30 s to obtain a stable temperature. For the rocky and rough surfaces, temperatures were obtained for a representative sampling of surface orientations (typically, about 30), and the strike and dip of each local surface was recorded. In order to obtain measurements of all surfaces at the same local times, we obtained data over a span of three days, partially overlapping with other GRSFE field and aircraft investigations. Logistical and weather problems prevented our obtaining complete diurnal coverage of each site; 10 a.m. observations and variations with emission angle are explicitly discussed in the text of this report. Results and discussion: The angular variations in 10 a.m. brightness temperature for the pebble, rocky, and a'a lava flow surfaces show the same general trend of the warmest temperatures occurring on the side of the roughness elements which face toward the sun, as expected. The a'a lava flow, with the largest rock masses, shows the largest variation with viewing angle; the pebble surface, with the smallest rocks, shows the smallest. This is as expected since the timescale for energy to conduct through a rock is approximately 2 minutes for a 1-cm pebble, 3 hours for a 10-cm rock, and 1 day for a 30-cm rock (e.g., Carslaw and Jaeger, 1959); the pebble surface should be nearly isothermal. The radiometer observations of the a'a lava flow show a nearly 15 K variation in brightness temperature depending on viewing geometry. This variation is a significant fraction of the total diurnal variation because the sunlit faces have heated to nearly their maximum temperature while the shaded faces have not been heated except by conducted or radiant heat; clearly, viewing geometry will affect interpretation in terms of the thermal inertia of the surface. These observations at 10 a.m. local time should show the largest variation with viewing azimuth; temperatures should be more uniform at other times of day. The data set includes the thermistor-probe measurements of these same surfaces. Temperatures are included as a function of local solar incidence angle, as calculated from the known location of the sun and the orientation of each rock face. The data for the rocky surface also includes the temperatures measured at the center of the five visible faces of the rock cube. Scatter results partly from uncertainties in the probe measurement and partly from variations in actual temperature which result from differences in local rock thickness (and, hence, conductive loss) and radiant heating. As expected, the largest variation occurs with the large rock masses in the a'a lava flow, and the smallest with the pebble surface. Within the sand and playa surfaces, individual sand and playa particles should be isothermal, due to the rapid conduction time across 24 mm and smaller grains (less than 1 s) and the dominance of solid over radiative conduction through individual grains. Any variations of the brightness temperature seen at different emission angles should represent pure directional emissivity effects. Relative instead of absolute emissivity is used due to uncertainties in calibration of both the radiometer and thermistor; normal emissivities are expected to be approximately 0.9-1.0 (e.g., Conel, 1969). Relative emissivity is calculated by converting the measured brightness temperature to an energy flux, using a numerical integration over the instrument passband, and normalizing to the nadir-viewing flux; values greater than 1 result from uncertainty in the relative calibration. The data set also includes observations made of the playa and sand surfaces at additional times of day and at additional emission angles and azimuths. Each set of data was normalized separately, and the remaining scatter results from calibration uncertainties. Notice that data collected simultaneously at low and high phase angles produce the same trend of decreased flux at higher emission angles. Jakosky et al. (1990) analyzed the data discussed above. They show normalized emissivity for a surface which is smooth and flat at the scale of the wavelength, i.e., the Fresnel emissivity for transmission through a smooth dielectric boundary (taken as the average of the transmission for parallel and perpendicular polarizations). They also show the measured flux from a smooth pane of glass, which should be a Fresnel emitter. Both the playa and sand surfaces exhibit a smaller decrease in emissivity with increasing emission angle than the Fresnel surface, although both do show a decrease. This effect can be understood as a result of surface roughness at the scale of the individual sand and playa grains. Each grain is itself much larger than the wavelength, so that emission is generated within the grain and the emissivity is governed by the Fresnel relation at the grain surface. At high emission angles, the radiometer preferentially views those parts of the grains that are tilted toward it, so that the normal to the grain surface is also tilted toward it; the local emission angle is thus smaller than that calculated using the normal to the average surface, and the Fresnel emissivity is correspondingly higher. Jakosky et al. (1990) also show model results of the Fresnel emissivity of an isothermal surface which is rough at scales much larger than the wavelength. The model assumes a gaussian distribution of surface slopes (characterized by an r.m.s. slope), and involves a numerical integration over all slopes and azimuths. For each geometry, the Fresnel emissivity (again, the average of the two polarizations) is calculated for the local emission angle and weighted by the projected area of the surface facet and the relative probability of a facet having that geometry; the average or effective emissivity is then calculated from the sum over geometries. Scattering by grains comparable in size to the wavelength is ignored, as are multiple reflections. In Jakosky et al. (1990), r.m.s. slopes up to 20 deg are expected at scales larger than centimeters (McCollom and Jakosky, 1990), and values up to 50 deg might occur at the scale of the individual grains for the sand-covered surface. The model confirms our intuition that rougher surfaces have a 25 smaller decrease in emissivity toward larger emission angles than do smoother surfaces. Notice also that the sand surface has a smaller decrease in emissivity at high emission angles than does the playa surface. This is consistent with the formation of the playa in the presence of water and the presumed more-efficient packing of surface grains. Conclusions: The data suggest that all surfaces emit in a non- Lambertian manner. For surfaces which are rough at scales larger than the thermal conduction depth, such as the a'a lava flow, different faces of the surface will be at different temperatures; observations of the sunlit sides of rocks then will show higher temperatures than observations of the shaded sides. For surfaces which are rough at scales much smaller than the conduction depth, such as the sand surface, the surface will be isothermal, but emissivity will be governed by the roughness; oblique emission will be from grain faces tilted preferentially in that direction, such that the Fresnel emissivity from the grain will be at a lower local emission angle relative to the grain normal, and the average emissivity will be higher than the Fresnel emissivity calculated using the emission angle relative to the average surface. Interestingly, a surface composed entirely of grains much smaller than the wavelength of emission, such as the dust-covered regions on Mars (grain size ~ 1 micron) will emit as if from a uniform half-space, and should show a much more dramatic drop-off with emission angle. The next decade will see a large number of thermal infrared radiometers and spectrometers being flown on spacecraft. By acronym, these include ITIR on the Earth Observing System, PPR on the Galileo mission to the Jupiter system, TES on the Mars Observer, and TIREX on the Comet Rendezvous/Asteroid Flyby mission. In addition, the Magellan mission to Venus includes a microwave radiometer. Most of these will usually operate pointed nearly toward the nadir; clearly, additional useful information can be obtained if they do scans of the surface at multiple emission angles; this will be important in defining the structure of the surface, deriving accurate thermal inertias, and providing input for corrections for the non-isotropic character of the thermal emission in thermal emission spectroscopy. 6. FIELD DATA SETS 6.A. PHOTOGRAPHS It was not possible within existing resources to digitize the field photographs, most of which were acquired on 35mm film. Field photographs are archived at the Geosciences Node, Washington University. Orders for color slides and black and white prints may be placed with Mary Dale-Bannister, via telephone at 314-889-6652 or via electronic mail at WURST::DALE (NASA Science Internet). The cost for part or all of the photographs will be based on reproduction and distribution costs. 26 6.B. SAMPLES Appendix F.1 contains information about samples collected during the GRSFE field campaign. Subsets of this sample collection can be made available if needed. Contact Mary Dale-Bannister. In addition, Jim Irons took a few samples at Lunar Lake on July 18, 1989. They were sent to the Cornell Soil Characterization Lab. One sample was taken from the playa modeling site. A second sample was taken from the cobble modeling site, below the surface cobbles. In both cases samples were acquired within 15 cm of the surface. The other two samples were taken from a hole within the mantled flow modeling site. One sample was taken from the top 3 cm of soil, and the other sample was taken from a distinct B horizon at a depth greater than 3 cm. In all cases, analyses were performed on sieved samples for particle sizes less than 2 mm in diameter. Table F.2 consists of data provided by Cornell University under Jim Irons' auspices. The units for the particle size fractions, the Fe and Al contents, and the organic carbon content are percent by weight. The cation exchange capacity (CEC) is expressed in centimoles of charge per kg of dry soil. The carbonate content is expressed in centimoles of charge per kg of dry soil relative to the calcium carbonate equivalent (CaCO3). The results of the Fe and Al extractions are strange for the following reason. The first extraction was performed with heated citrate-dithionite-bicarbonate following pretreatment with pH 5.0 sodium acetate. This extraction is supposed to remove all of the free irons (crystalline plus amorphous or paracrystalline). The oxalate extraction is only supposed to extract amorphous or paracrystalline iron and aluminum. The oxalate extract, however, contains more iron and aluminum. Regardless, both methods indicate that the samples contained very little free iron. 6.C. VISIBLE AND REFLECTED INFRARED SURFACE DATA 6.C.1. DAEDALUS SPECTROMETER Note: The following material was provided by Guinness. 6.C.1.1. INSTRUMENT DESCRIPTION The Daedalus AA440 Spectrafax is a portable field spectrometer. The important features of the instrument are summarized in the table below. The Daedalus measures radiance between the wavelengths of 0.45 and 2.4 micrometers. The measured radiance values are digitized to values in the range of 0 to 1023 (10 bits). The instrument also has a variable gain setting to adjust the dynamic range. The gain scale is a logarithmic scale in which one step represents about a 16 percent change in gain. A change in gain of 4 units is approximately a factor of 2 change in the digitized output signal. Reflectance values are derived 27 by ratioing the signal of a sample to that of a pressed and bonded halon standard viewed at the same lighting angle. Note that a difference in gain setting for the standard and the target must be accounted for when computing reflectance. The derived reflectance values are radiance coefficients. Daedalus AA440 Spectrafax Instrument Features --------------------------------------------- Dynamic range: Radiance digitized to values between 0 and 1023 (10 bits). Variable gain setting. Wavelength range: 0.452-2.398 micrometers in 280 channels. Wavelength resolution: Varies from 0.01 micrometers (visible) to 0.04 micrometers (infrared). Field of view: 2.5 centimeters when instrument is about 1 meter above target (1.5 degree angular field of view). Detectors: Silicon detector sensitive to wavelengths of 0.4 to 1.1 micrometers; lead sulfide detector sensitive to wavelengths of 1.1 to 2.4 micrometers. Filters: Circular filter wheel with 360 optical interference filters in three overlapping wavelength segments. Overlapping segments must be edited to properly display a Daedalus spectrum. 6.C.1.2. DATA SET DESCRIPTION The Daedalus AA440 spectrafax data set consists of over 500 data files. The purpose of these measurements was to provide ground calibration for AVIRIS and ASAS data. A number of reflectance measurements were made at all sites to estimate the average reflectance of an area about 50 by 50 meters in size (i.e., the size of several AVIRIS pixels). In addition, measurements were made at Lunar Lake to characterize the reflectance of the spectral components at each modeling site. The detached PDS labels for each spectrum contain a brief description of the site and purpose of the measurement. The Daedalus instrument was used during the July 1989 field campaign. The reflectance of a bright and dark target area was characterized with the Daedalus instrument at Kelso Dunes and the Cima Volcanic Field. The reflectance of Trail Canyon Fan, the salt deposit at the Devil's Golf Course, Death Valley Dunes, and the tephra deposit at Ubehebe Crater were also characterized with the instrument. At each site measurements were made along 2 or 3 traverses. The reflectance of the natural surface was measured at points along the traverse. Samples were spaced about 3 meters apart with 10 to 12 samples per traverse. The spectra halon calibration standard was generally measured at the beginning and end of each traverse. Additional samples of vegetation and other significant components of the site (e.g., coated rocks) were also measured. The measurements were made with the operator facing 28 toward the sun and holding the spectrometer at a height of approximately 94 cm above the ground. Nearly 200 measurements were made at Lunar Lake for the purpose of calibrating the airborne data, characterizing the reflectance of end- member materials in the modeling sites, and comparing reflectance spectra made by the various spectrometers used during GRSFE. The procedure used to collect data for the calibration of airborne data was the same as describe in the previous paragraph. Data for this purpose were collected at the playa, disturbed playa, and cobble sites at about 8:00am, 9:30am, and 11:00am (local time). Daedalus measurements were also made for characterizing spectral end-members at the disturbed playa and the cobble site. Three grid locations were sampled at the disturbed playa site. The end-members that were measured were cloddy playa material, fine powder, and undisturbed playa material. One grid location was sampled at the cobble site. The end-members that were measured were silt, basalt fragments, brown-coated rocks, and red-coated rocks. The smooth playa site was not characterized because the instrument batteries wore down. However, the playa site is very uniform and there should be sufficient data from the airborne calibration experiment to characterized the spectral components of this playa. Also, there are no Daedalus measurements of the lava flow site, again because the instrument batteries wore down before visiting that site. Several measurements were made at Lunar Lake to compare spectra from the Daedalus and SIRIS instruments. In this experiment, both spectrometers measured the same calibration standard and then measured the same sample of undisturbed and disturbed playa, along with a large basalt cobble. Finally, a few Daedalus spectra were collected to support the thermistor experiment by measuring the reflectance of the surface near each thermistor. Measurements made with the Daedalus instrument during GRSFE were hand held with the instrument height only approximately fixed. Thus, there can be a small difference in the height used to measure the calibration standard and a given sample. Such a height difference will produce a systematic error in the absolute reflectance value. Each centimeter difference in height will produce about a 2% change in reflectance. During GRSFE the Daedalus instrument showed two operating problems. One problem was the result of saturation in the electronics that affected channels in the wavelength range of about 0.73 to 0.97 micrometers. This problem can be readily seen in the spectra of the calibration standard. The radiance values in the affected wavelength range are approximately constant and well below the values expected for these wavelengths. Data in this wavelength range should be considered bad. The second problem involved the filter wheel not properly rotating during data collection. Measurements with the filter wheel problem have been removed from the GRSFE archive. 6.C.2. PARABOLA Note: The following material was provided by Deering and edited by Michael Shepard, Washington University. 29 6.C.2.1. INSTRUMENT DESCRIPTION The Portable Apparatus for Rapid Acquisition of Bidirectional Observations of the Land and Atmosphere, or PARABOLA instrument was designed to overcome important limitations of previously existing field instruments for obtaining the adequate sampling needed to analyze the bidirectional reflectance angular distributions of earth surface targets and the sky. The key previous major instrumentation limitation was determined to be the difficulty in obtaining multiple viewing angle measurements of ground surfaces in a very short period of time to eliminate, or at least sufficiently minimize, the effects of changing sun position, sky conditions, and the vegetation's dynamic biophysical conditions during the sampling period. A second important previous limitation was the capability to measure the downwelling sky radiance distribution concurrent with measurements of the ground target. The PARABOLA is a battery-powered, two-axis scanning head three-channel (visible, near infrared, and mid-infrared; 0.650-0.670, 0.810-0.840, and 1.620-1.690 micrometers, respectively), motor-driven radiometer that enables the acquisition of radiance data for almost the complete (4 pi) sky- and ground-looking hemispheres in 15 degree instantaneous field-of- view sectors in only 11 s. The detectors, two silicon and one germanium, are temperature regulated (by cooling or heating) through thermoelectric proportional control circuits. Also, due to the tremendous range in target brightness that can be expected in scanning a 4 pi sr field of view, an auto-ranging amplifier is used to switch the gain levels back-and-forth by factors of 1, 10 and 100 to maintain maximum radiometric sensitivity. In designing the PARABOLA for maximum sensitivity for the range of reflectances typically experienced for earth surfaces, including vegetation and snow, the direct solar disk measurement had to be sacrificed. Thus, the direct solar beam saturates the PARABOLA detectors for the pixel at that view angle, and therefore a separate measurement of the sun is required for computations of direct and total irradiant flux. Neutral density filters have been designed for mounting over the detector cones of the PARABOLA for this purpose to obtain periodic measurements when the instrument is placed in its manual (or "calibration", as opposed to scanning) mode. The radiometer is then pointed directly at the solar disk to acquire the necessary measurement to complete the missing data point. Calibration: Radiometric laboratory calibration of the PARABOLA is performed at GSFC on a 1.8 m spherical integrator employing 12 200- W quartz halogen lamps (2950K at 6.5A). The number of lamps is varied to produce 12 radiance levels for calibration. Three separate calibration runs are made to fully calibrate the PARABOLA at the three gain levels of the instrument. Neutral density filters (0.1 and 0.01 density levels; precisely calibrated) are used for the two lowest gain settings. The voltage response to the radiance level relationship is linear in all three spectral channels for each gain setting with linear correlation coefficients of 0.999, and has been found to be very stable from year to year. Calibrations are typically performed at the beginning and at the end of each seasonal field campaign. 30 Data recording and processing: The PARABOLA data system was originally designed to transform the analog data and record the records onto digital cassette tapes, providing a "rugged" recording medium with unlimited data storage capability. The data were then transferred to computer diskettes via a microcomputer and serially linked tape reader. The original data system has been upgraded to enable direct transmittal of the data onto computer disks via a field-hardy laptop portable computer. The original cassette recording system is maintained as a redundant backup. Specially developed software programs are then used to apply geometric computations and system calibration coefficients to the data to generate an output file that can be handled by a variety of commercial software packages, as well as those designed in-house. Mounting system: The primary mounting device is a lightweight, collapsible boom apparatus, called the Transportable Pickup Mount System or TPMS, whose primary unit consists of an aluminum triangular truss that decouples as four 2-m long sections. At the top end resides a detachable, two-axis motorized PARABOLA radiometer head mounting and leveling head with a camera mounting attachment. All operations of the PARABOLA/TPMS, except for raising and lowering the boom, are controlled from the PARABOLA data system control panel. For the GRSFE study the complete PARABOLA/TPMS system was mounted on a large tripod, as the system is adapted for the field measurements where motorized vehicles either cannot go or where traversing the terrain would be potentially damaging to the pickup-mounted equipment (such as the rough lava site on the GRSFE study area). Field deployment: The PARABOLA design provides multidirectional viewing, but the geometry of the systems does not allow the same "spot" on the ground to be measured at each view direction. Thus, considerable care must be taken in locating the instrument at a field site to ensure that the data acquired will actually represent the kind of surface that is desired; and some assumptions about the homogeneity of the area must be made. One approach to increasing the sampling density, which has been found to particularly effective, was employed at the GRSFE site. This involves the rotation of the radiometer boom by + and - 7.5 degrees from the solar principal plane axis and acquiring two additional complete scans. The three scans are combined in a special program that has been written to analyze the bidirectional reflectance distribution characteristics of a site. This procedure also enables more accurate sampling of the "hot spot" effects and the aureole surrounding the solar disk. In more spatially heterogeneous sites three or more (n) complete PARABOLA replicate scans are occasionally acquired by moving the instrument within the site. The replicates are then averaged for each viewing angle position to minimize the within-field inhomogeneity effects. With proper site selection three to five subsamples have been found to be generally adequate for most herbaceous vegetation, since the measurement pixel instantaneous fields of view are large (ranging from approximately 2 meters squared at nadir to 5.7 meters squared at a 45 degree off-nadir angle) relative to the canopy spatial structure. The average of the n values is taken as the canopy radiance for a particular view direction. Directional reflectances are normally computed as hemispherical-directional reflectance factors using the PARABOLA 31 directional radiance measurements from the ground-looking hemisphere divided by the PARABOLA-derived incident irradiance as computed from the PARABOLA sky irradiance data or from a calibrated painted BaSO4 reference standard panel. Supporting Instrumentation: A pair of Eppley Pyranometers was used to obtain broad-band (visible, reflected IR) sky irradiance and the irradiance reflected from the ground. The instrument was deployed at PARABOLA height. These data are in Appendix G. 6.C.2.2. DATA SET DESCRIPTION PARABOLA data were obtained at three modeling sites in the Lunar Crater Volcanic Field. The playa site, the cobble site, and the mantled flow site. Each site was "imaged" by PARABOLA at a variety of solar incidence angles to characterize the scattering properties of the surface completely. This information can be used as a reference to correct airborne data from ASAS and AVIRIS from any lighting and viewing geometry, or as a base data set to test radiative transfer models. Six sets of data from the playa site were gathered primarily on July 17, 1989 with solar incidence angles that varied from 28 to 71 degrees. Five sets of cobble site data were gathered during the morning of July 18, 1989 with incidence angles from 19 to 57 degrees. Data from the mantled flow site was obtained six times during the afternoon of July 18 with incidence angles from 21 to 73 degrees. In all, over 1000 data points exist for each site to characterize its scattering properties. Additionally, an equal number of points exist for the sky above the sites. The data are presented in two formats. The first set of files contain the complete, though filtered, data set consisting of almost all the individual pixels from the three replicate scans of the same site. Because of the scanning pattern of PARABOLA, the pixels are not at equidistant angles in the off-nadir or azimuth viewing planes. These files are on GRSFE Volume 1 in directory PARABOLA, and have names in the format PRBxxxxF.DAT. The files have the following information in each record: 1. Hemisphere i.e. 'GR' or 'SK' 2. Sequence number of data record 3. Off-nadir (view zenith angle, or emission angle) value of the center of a bin assigned to the pixel 4. Azimuth (view azimuth angle) value of the center of a bin assigned to the pixel 5. Actual off-nadir view zenith angle (relative to normal) of pixel 6. Actual view azimuth angle in the 0-360 degree coordinate system, where the sun is at zero and backscatter direction is at 180 degrees 7. Actual view azimuth angle in 0-180 and -180-0 degrees coordinate system; negative azimuth refers to boomside where equipment or shadow of the instruments obstruct the scan 32 8,9,10. Radiance value (W/m2/ster/um) for channels 1,2,3, resp. For users who prefer average sets of data binned into even intervals, the second set of files were created. In these files, the "raw" data of the first set of files has been averaged into bins with centers every 15 degrees in off-nadir and 30 degrees in azimuth. In these files, the last column gives the number of points used to compute the average of the bin. If there are gaps due to instrument shadows, instrument "noise", or other causes, the value of the bin is taken from its mirror bin (this assumes symmetry of data across the solar principal plane). To identify these cases, the number of data points averaged is assigned a negative (mirror image) value. If for some reason, there is no data for substitution, an interpolated value is used. Since interpolated values are not real data, the user is cautioned by placing a zero in the last column. These files are found on GRSFE Volume 1 in the PARABOLA directory, and have names in the format PRBxxxxA.DAT. The files contain the following information in each record: 1. Hemisphere, i.e. 'GR' or 'SK' 2. Sequence number of data records 3. Azimuth (view azimuth angle) value of the center of a bin assigned to the pixel 4. Off-nadir (view angle zenith, or emission angle) value of the center of a bin assigned to the pixel 5. Actual off-nadir view zenith angle (relative to local normal) of the pixel 6. Actual view azimuth angle in the 0-360 deg coordinate system, where the sun is at zero and backscatter direction is at 180 deg 7. Actual view azimuth angle in 0-180 deg and -180-0 deg coordinate system; negative azimuth refers to boomside where equipment or shadow of the instrument obstruct the scan 8,9,10. Radiance value (W/m2/ster/um) for channels 1,2,3, resp. 11. Data "flag" - indicates the number of samples averaged (see above description) Throughout the days during which PARABOLA data were being gathered, a white Barium Sulfate reference plate was periodically measured by PARABOLA. This plate is known to be Lambertian within a correctable amount. The measured radiance of this plate was plotted as a function of the time of day (or solar zenith angle) and fit to a sinusoid. In this way, the radiance from a Lambertian reference is known at any time during the data gathering period. The Lambertian radiances at the time each data set was obtained are tabulated in Table 6.1. TABLE 6.1 LAMBERTIAN SURFACE RADIANCES Col Description --- ----------- 1. Site i.e. playa, cobble, or lava 2. Date of data acquisition 3. Original data file name 33 4. Solar zenith or incidence angle 5. Local time of data acquisition 6,7,8. Radiance of reference plate (corrected to Lambertian) for channels 1,2,3 respectively in (W/m2/ster/um). Dividing the radiance of any bin in the data files by the radiance of the reference plate (taken the same time of day) gives a value of reflectance, sometimes called radiance coefficient. GRSFE89 CORRECTED BASO4 ------------------------------------------------------------------------ SITE DATE FILE NAMES SZA TIME(PDT) CH 1 CH 2 CH 3 ------------------------------------------------------------------------ PLAYA 7/17 PRBLL23,24 27.8 14:32 408.29 275.31 67.70 " " PRBLL25,26 43.7 15:58 331.43 223.79 55.89 " " PRBLL27,28 53.8 16:41 268.48 181.59 46.22 " " PRBLL29,30 62.6 17:34 206.45 140.02 36.69 " " PRBLL31,32 70.9 18:17 143.21 97.64 26.98 " 7/19 PRBLL33,34 48.8 09:17 301.40 197.18 48.61 ------------------------------------------------------------------------ COBBLE 7/18 PRBLL09,10 57.6 08:29 210.98 164.49 38.71 " " PRBLL07,08 48.7 09:14 264.65 202.95 47.74 " " PRBLL06,05 37.7 10:14 317.77 241.02 56.67 " " PRBLL04,03 28.5 11:07 356.91 269.06 63.25 " " PRBLL02,01 19.3 12:14 382.71 287.55 67.58 ------------------------------------------------------------------------ LAVA-FLOW 7/18 PRBLL11,12 21.4 13:45 440.17 304.43 68.97 " " PRBLL13,14 32.6 14:56 402.20 278.11 62.93 " " PRBLL15,16 42.1 15:47 350.79 242.48 54.76 " " PRBLL17,18 52.2 16:45 289.27 199.83 44.98 " " PRBLL19,20 62.7 17:32 222.34 153.44 34.33 " " PRBLL21,22 73.2 18:27 144.58 99.54 21.97 ------------------------------------------------------------------------ 6.C.3. SIRIS SPECTROMETER Note: The following material was edited from a contribution by Kruse. 6.C.3.1. INSTRUMENT DESCRIPTION The Single beam visible/InfraRed Intelligent Spectroradiometer (SIRIS) is a field-portable grating spectrometer manufactured by Geophysical and Environmental Research, Inc. The system is controlled by a portable MS-DOS compatible computer and spectra are saved in digital format on 3 1/2" computer disks. Data acquisition can be custom programmed, however, for this application the instrument was used in the standard mode. The SIRIS measures radiance using three separate gratings on a single stepper motor driven mount. The standard operating mode (from 0.35 to 2.5 micrometers) consists of a scan from 0.35 micrometer to 1.08 micrometer using the first grating, from 1.08 micrometer to 1.8 micrometer with the second grating, and from 1.8 micrometer to 2.5 micrometer with the third grating. Grating 1 uses a silicon detector with sensitivity out to about 1.1 m. A PbS detector is 34 used for the 1.1 to 2.5 micrometer portion of the spectrum. Both detectors are temperature stabilized on the same thermal electric cooler. Three blocking filters are used to prevent 2nd order contamination effects. The first blocking filter cuts off at about 0.65 micrometer, the second at approximately 1.05 micrometer, and the third at approximately 1.7 micrometer. Analog output signals are processed through phase sensitive detector systems with narrow band electronic filtering (Collins, written communication, 1989). Analog to digital conversion in the microcomputer results in 12 bit digital data. Because the SIRIS is a single-beam instrument, it is necessary to measure reference and sample radiance spectra separately. Spectra are typically measured using solar illumination and HALON as a reflectance standard. HALON is a highly reflective material with no absorption features in the 0.4 to 2.5 micrometer range (Weidner and Hsia, 1981). The reference spectrum is measured first and stored on disk. Depending on sky conditions, a reference spectrum should be measured before each sample spectrum unless atmospheric conditions are clear and dry, in which case a new reference spectrum may not be needed each time. The sample is measured next and both the reference and the sample radiance spectra are automatically stored together in a disk file that can later be reduced to reflectance. The reflectance spectrum can be viewed on the MS-DOS computer prior to proceeding to the next measurement. The SIRIS spectral sampling interval varies continuously with wavelength. In all cases, the actual resolution (sampling interval x 2) is greater than the AVIRIS resolution of 10 nm. 6.C.3.2. DATA SET DESCRIPTION The SIRIS measurement objectives were: 1) to measure reflectance spectra of a bright and dark target at each GRSFE site for potential use in AVIRIS calibration; 2) to fully characterize selected sites at the GRSFE modeling site (Lunar Lake, NV) using visible/infrared reflectance; 3) to collect reflectance spectra for selected endmember materials at each of the GRSFE sites; and 4) to make reflectance measurements for inter-instrument calibration. Visible/Infrared field spectral measurements using the SIRIS were made on 15-19 July, as described in Section 3.C of this document. The SIRIS was set up with standardized geometry; a nadir view from 75 cm above the ground surface, facing towards the sun. Undisturbed surfaces were measured including rock outcrops, surface coatings where present, and soils. All measurements were made using the HALON standard, which makes conversion to absolute reflectance based on NBS standard HALON possible. This conversion was not routinely performed for the SIRIS data and all spectra provided for GRSFE are relative to HALON. No additional SIRIS measurements were made for September AVIRIS flights. SIRIS spectra were reduced to reflectance using software provided by GER. This procedure consists of simple division of the sample radiance spectrum by the reference radiance spectrum. The GER grating- match adjustment was applied to correct for offsets at the grating 35 boundaries and grating overlaps were removed by deleting overlap channels. Previous experience with the instrument and inspection of raw data plots indicated that noisy channels between 1.32 - 1.42 micrometers and 1.78 - 1.95 micrometers should be deleted. Deleted points were coded as 9.99999999. Spectra for some measurement sites were discarded completely because of excess noise and other problems. Measurements made with the SIRIS at the Lunar Lake modeling site for comparison of instrument response. These measurements were made using the JPL SPECTRALON plate as a reference standard. Additionally, several instrument artifacts (or atmospheric bands) are apparent in the SIRIS data. First, there are small instrument glitches near 0.6, 0.76, 0.93, 1.06, and 1.49 micrometers. Secondly, there are commonly prominent CO2 absorption features near 2.005 and 2.05 micrometers that are probably caused by changing atmospheric conditions between measurement of the reference and the sample spectra. Spectra measured with the SIRIS in the lab for the Lunar Lake materials do not show many of these features. Kelso, California, site Spectra were measured on 15 July 1989 along the road and parking lot south of the sand dunes in the center of the valley. These measurements were averaged to constitute the Bright Target. The Dark Target measurement is the average of two spectra taken of the gravel at the power station along the main KelBaker road. Cima, California, site Spectra were measured on 15 July 1989 for basalt and basalt gravel and averaged as the Dark Target. The average of two stream wash spectra constitute the Bright Target. Lunar Crater Volcanic Field, site Spectra were measured periodically throughout the day of 17 July, 1989 for the playa, disturbed playa surface, and cobble modeling sites. Ubehebe Crater, Nevada, site Spectra of a Dark Target (basalt gravel) and Bright Target (small playa surface) were measured on 19 July 1989. Spectra of additional targets were not measured because of rapidly changing, cloudy, weather conditions. Death Valley, California, site Spectra of the Trail Canyon fan, the salt flat in the center of Devil's Golf Course adjacent to the fan, and sand dunes near Stovepipe Wells were measured on 18 July 1989. 36 6.D. THERMAL INFRARED SURFACE DATA 6.D.1. PORTABLE FIELD EMISSION SPECTROMETER (PFES) Note: The following material was provided by Petroy. 6.D.1.1. INSTRUMENT DESCRIPTION The PFES was built for the purpose of measuring the ambient spectral thermal emission of geologic materials in situ, thus avoiding the disturbance of the natural setting which occurs during sample collection. This instrument was designed and built at the Jet Propulsion Laboratory (JPL). The spectral range covered is 5 to 14.5 micrometers and the resolution is approximately 1.5% of the wavelength (0.1 - 0.2 micrometers). The PFES system consists of two main parts, the sensor head and the data recorder. Accessories include a gas bottle, signal monitoring box, reference hot and cold blackbodies, and connecting cables. The sensor head contains the optics, the detector, and the preamplifiers. During operation, the sensor head is connected by a hose to a tank that supplies high pressure argon gas for cooling the detector. The data recorder consists of a lap-top computer, signal processing circuitry, scan sequencing logic circuits, and the power supply. The analyzing element of the spectrometer is a filter wheel containing three filter segments (continuously variable multilayer interference filters). The detector is a MCT (mercury-cadmium- telluride) photoconductor. Spectral measurements of the target are made relative to the spectrum of a reference blackbody at ambient temperature. The kinetic temperature of this blackbody is measured using an attached thermistor (Hoover and Kahle, 1987). Usually about two to three spectra of each target are measured to provide a check of consistency and to reduce noise by averaging. Each spectrum takes about thirty seconds to scan. The digital output from the detector, N, is related to the input by the following simplified equation: N = K[L(t)T(l) + L(l)] (1) where: N = digital output K = constant L(t) = radiance from the target L(l) = radiance from the lens T(l) = transmission of the lens Applying this relationship to measurements of the target and the external blackbody: N(t) = K[L(t)T(l) + L(l)] (2) 37 N(bb) = K[L(bb)T(l) + L(l)] (3) where the variable L(bb) represents the radiance from the external blackbody. Since the two measurements (surface and external blackbody) are made close together in time, the radiance contribution due to the lens is essentially constant. Taking the difference between Equations 2 and 3: N(t) - N(bb) = KT(l) [L(t) - L(bb)] (4) from which L(t) (the target radiance) can be derived: L(t) = [L(bb) + (N(t) - N(bb))] / KT(l) (5) In this equation L(bb) is calculated from the Planck formula and the KT(l) divisor is drawn from a table which has been generated from measurements made on blackbodies at different temperatures. The instrument response function is checked during each operation by measuring a second blackbody at a much lower temperature (the "cold" blackbody) and comparing this measured radiance against calculated radiances at that temperature from the tables of instrument responses (Hoover and Kahle, 1986; 1987). Emissivity can be calculated by dividing the target radiance by the radiance of a blackbody at the same temperature. Target kinetic temperature is determined as the lowest temperature which produces a calculated blackbody curve that is greater than the curve of the target spectral radiance and tangent to the target spectral radiance curve in the region between 7.0 and 7.5 micrometers (in general, most silicates closely approach blackbody behavior in this wavelength region). However, in the case of field spectroscopy, in addition to radiance from the surface, the instrument is also sensitive to the emitted skylight radiance. The skylight radiance is reflected from the surface and scattered into the instrument's field of view. Also, the signal from the surface is modified by attenuation, scattering and emission during transit through the atmosphere (Vincent and Thompson, 1972a). These complications can be modeled as: L(tot) = [e(t)L(bb,T)+r(t)L(sky)]trans(A) + L(A) (6) where: L(tot) = the total spectral radiance received by the detector e(t) = emissivity of the target surface L(bb,T) = the spectral radiance of a blackbody at the same temperature (T) as the target surface r(t) = the reflectivity of the target surface L(sky) = the spectral radiance from the sky, incident on the rock surface trans(A) = the spectral atmospheric transmissivity 38 L(A) = the spectral radiance from atmospheric emission and scattering in the path between the detector and the target surface L(A) and trans are generally considered negligible in field spectroscopy measurements because of the relatively short distances between the target surface and the detector (usually less than one meter). However, the effect of reflected skylight radiance (L(sky)) can not be dismissed. To correct for the reflected skylight term, the radiance of a texturized aluminum plate (e=0.01 to 0.05) is measured to obtain L(sky)), with the assumptions given: L(tot) = e(t)L(bb,T) + [1-e(t)]L(D) (7) where L(D) is the measured spectral radiance from the texturized aluminum plate. The radiance measured from the aluminum plate is representative of the skylight radiance sensed by the detector. Using Kirchhoff's Law, 1-e(t) yields the reflectance of the target. Rewriting equation (7) to solve for target emissivity: [L(tot) - L(D)] / [L(bb) - L(D)] = e(t) (8) Note: the emissivity of the aluminum plate is high for wavelengths shorter than 8 micrometers. Thus, the technique discussed above is only applicable for the 8 to 12 micrometer wavelength range. 6.D.1.2. DATA SET DESCRIPTION PFES data were acquired on July 15 and 17, 1989. A total of 31 measurements were collected for GRSFE. Of these measurements, 13 were calibrations and 18 were of representative surfaces. The data were collected primarily to support the calibration of the TIMS data and to assist in interpreting spectral mixing in the mid-infrared. Sites were selected for calibration that covered a range of emissivities. On July 15, PFES data were collected at Kelso Dunes and the Cima Volcanic Field as part of the Calibration Team effort. Daedalus and SIRIS data were collected over the same sites. For the PFES data, the Kelso Dunes Bright Target site represented the silica-rich endmember and the Cima basalt tephra Dark Target site represented the more silica-poor endmember. On July 17, PFES data were collected at two of the modelling sites at Lunar Lake (the bright and the cobble sites). Several spectra were also collected at the playa surface next to the Lunar Lake thermistor site. PFES data files in this release of the GRSFE archive consist of tables of wavelength, emissivity, measured radiance, gain of internal Blackbody calibration source (not used), Blackbody radiance fit to data, and measured radiances of external cold and hot Blackbodies. Note that emissivities have not been corrected for atmospheric effects. However, data acquired over the aluminum target (called CALIBRATION READING in PDS labels) are provided for users who want to do the corrections. 39 6.E. ATMOSPHERIC DATA 6.E.1. REAGAN RADIOMETER Note: The following material was edited from contributions by Bruegge. 6.E.1.1. INSTRUMENT DESCRIPTION In order to characterize the optical depth over a given site and at a given time and wavelength, a spectrally filtered, solar-pointing radiometer is typically used. The Reagan Sunphotometer, operated by JPL, was used during GRSFE. This photometer was built under the supervision of Dr. John Reagan, University of Arizona, Tucson. The characteristics are summarized in the table given below. It has 10 spectral channels, each about 10 nm in bandwidth, fast response, low noise, and an internal filter wheel. The output voltages of the instrument are proportional to the incident irradiances within a given band, although no calibration relating voltage to physical units is required. Features of the JPL Reagan Sunphotometer ---------------------------------------- Number of Channels 10, internal filter wheel Wavelengths (um) 0.37, 0.40, 0.44, 0.52, 0.61 (Ozone), 0.67, 0.78, 0.87, 0.94 (Water Vapor), 1.03 Bandpass (nm) ~10 Field-of-view 2 degrees, full-field Detector Photodiode, temperature stabilized to 40+/- 0.5 degrees C Response Time < 1 Second Output 2.0-0.2 V Tracking Manual, tripod mounted To prepare the Reagan sunphotometer for data collection, the instrument was mounted to a tripod, the detector heater turned on, and the detector temperature monitored to verify stability at 40 +/- 0.5 degrees C. Operator time clocks were set to the nearest second using a portable radio tuned to WWV. The instrument was then aligned such that the solar disk passes through a site at the front of the optical barrel and onto a cross-hair reticle. Alignment was verified by observing that the output voltage was maximized for this position. Next the start time was recorded to the nearest second, the filter wheel sequenced through its ten positions, a gain and voltage recorded for each channel, and a final stop time recorded. Data were then obtained repeatedly during the observing runs in July and September. 6.E.1.2. DATA SET DESCRIPTION Total instantaneous optical depth was obtained from the data through use of Beer's Law, as follows. The equation in this section are given in Fortran-like notation. 40 V = V0 * exp(-(tau/cos i)) where V = instrument voltage for one channel V0 = equivalent voltage that would be obtained above atmosphere tau = optical depth i = solar incidence angle Taking the natural logarithms of both sides and plotting ln V versus sec i (i.e. Langley Plot) for the set of observations for one channel allows one to solve for V0 = intercept and tau = slope. Instantaneous optical depths were then derived by using this intercept and solving for tau using pairs of voltage values acquired close together in time. The GRSFE data files consist of tables of time and optical depth values computed for each of the 10 Reagan Radiometer channels. The data were also further processed to separate aerosol optical depth from the total, but results are not included in this GRSFE archive release. 6.E.2. SPECTRAL HYGROMETER Note: The following material was edited from contributions by Bruegge and Conel. 6.E.2.1. INSTRUMENT DESCRIPTION Two spectral hygrometers were used simultaneously at Lunar Lake base camp. Their serial numbers are SH015 and SH004. The spectral hygrometers provide the ratio Rsh = L(935)/L(880), where L(935) is irradiance at 935 nm and L(880) is the irradiance at 880 nm. The measurement is acquired by aiming the optics at the sun and keeping the sun's image near the center of a ground-glass screen while the ratio is read from the display. Note that 935 nm is in the middle of an atmospheric water band, whereas 880 nm is not. The principle of operation is that the greater the water vapor abundance, the lower the ratio. 6.E.2.2. DATA SET DESCRIPTION The conversion from spectral hygrometer ratio to column water abundance was calibrated prior to GRSFE for column water abundance using more than 100 radiosonde observations. The line of sight precipitable water, WLOS, in centimeters, is: WLOS = W + (W/6.8)**3 where W = 13.7/((10**1.38)*Rsh) The uncertainty is +/- 10%. To adjust the expression for site elevation: WLOS(corr) = WLOS * SQRT (P0/P), where P0 = pressure at sea level and P = pressure at site elevation. 41 The Spectral Hygrometer data files consist of data acquired on September 29, 1989. Columns are time in PDT, ratio of L(935)/L(880) for instrument SH015, time in PDT and corresponding ratio for instrument SH004. For the July campaign, the Reagan Radiometer data could be used to extract WLOS, since: Rrr = L(940)/L(870) and Rsh = 0.08703 + 1.07992 * Rrr with r**2 = 0.9999 where Rrr = Reagan Radiometer ratio r = correlation coefficient 6.E.3. ARIZONA STATE UNIVERSITY WIND EXPERIMENT Note: The following material has been edited from contributions by Greeley and Lancaster. 6.E.3.1. INSTRUMENT DESCRIPTION Boundary layer wind profiles were measured using field-portable anemometer masts with a total height of 9.8 m. Cup anemometers (Tradewinds) were placed at heights with a logarithmic spacing of 0.75, 1.25, 2.07, 3.44, 5.72, and 9.5 m. Pairs of AD590 temperature sensors were placed in a shielded and ventilated mounting at heights of 1.3 and 9.6 m. Wind directions were measured with WD1 instruments from Remote Measurement Systems at heights of 9.7 m and 1.5 m. Data were recorded using an ADC-1 analog to digital converter and a Tandy 102 computer as a data logger. A recording interval of 20 minutes was adopted, so all data represent a 20 minute average of wind speed and temperature conditions. Data were stored in the field on cassette tapes and transferred to a Macintosh computer at Arizona State University. 6.E.3.2. DATA SET DESCRIPTION Near surface winds were studied at two sites on Lunar Lake Playa: a) Lunar Lake North (LLN) is at the northeast end of the playa, on a smooth clay-silt surface; and b) Lunar Lake South (LLS) is toward the center of the playa, just east of the area of prominent basalt gravel bars on the playa. In addition to providing data on surface winds to GRSFE, these investigations will provide input to studies of the relationships between the radar backscatter and aeolian roughness characteristics of desert surfaces (Greeley et al. 1988). Data were recorded at Lunar Lake North between 13:32 on 16 July 1989 and 17:16 on 18 July 1989 for a total of 52 hours; and at Lunar Lake South between 16:51 on 16 July 1989 and 16:13 on 18 July 1989 for a 42 total of 48 hours. During these periods, wind speeds recorded at a height of 9.6 m varied between 1.36 and 8.16 m/sec. The 20 minute averages of wind speed recorded do not reflect the very unsteady and gusty wind conditions that were observed during GRSFE. Wind directions during periods of winds above 4m/sec were mostly from south to southeast. Note: wind data are presented in degrees clockwise from magnetic north. Temperatures at a height of 1.5 m above the surface ranged from a minimum of 9.25 degrees C around 05:00 on 17 July, to a maximum of 35.62 degrees C on the afternoon of 18 July. Unfortunately, a malfunction in the analog to digital converter resulted in no temperature or wind direction data being recorded at Lunar Lake South. The wind experiment data set is divided into two files, one for each wind tower. The data are in tabular form, with column descriptions and listed results. 6.E.4. WEATHER STATION 6.E.4.1. INSTRUMENT DESCRIPTION The weather station was a portable station designed and built by Meteorological Research, Inc. Cup anemometer and thermistor thermometer measurements of wind speed and direction, dry-bulb air temperature, and rainfall were monitored continuously on a strip chart recorder. The instrument was mounted on a tripod approximately two meters above the surface. 6.E.4.2. DATA SET DESCRIPTION The weather station data consist of wind velocity, direction, and air temperature collected at Lunar Lake during the July and September field campaigns. Data were digitized at 30 minute intervals from stripchart recordings. 6.F. GEOPOSITIONAL SATELLITE PROFILES Note: The following material was edited from a contribution by Garvin. A field differential GPS survey team which included Jim Garvin, Jack Bufton, Bill Krabill, and Earl B. Frederick deployed a pair of Motorola Eagle II GPS receivers to the southern flanks of the feature known as Mars Hill (an alluvial boulder field superimposed on a major lobe of alluvial and colluvial material in Eastern Death Valley) on Oct. 19, 1989. The objective was to measure the 5-20 cm scale microrelief of the boulder field at pixel scales (30-50 m long transects), with vertical control to the few cm level. These microterrain profiles were to be used to help calibrate radar scattering models, and to compare with helicopter stereo data for the same location. The GLOTAS technique, as implemented at the time of GRSFE, involves recording GPS carrier phase data for at least 4 satellites simultaneously with two 8-channel receivers. In this approach, one 43 (master) receiver/antenna is fixed at a reference point (which can be surveyed in using GPS itself), and after a suitable calibration period, the second GPS receiver/antenna system is moved from its position beside the fixed master in whatever pattern is desired to develop a topographic profile or grid. The mobile GPS receiver/antenna permits positional data in all three axes (x,y,z) to be recorded at a rate of once per second, which can then be cross-referenced to the fixed master, and later tied to an absolute geodetic positional reference system. The primary interest was in relative horizontal position and in relative relief at each position. Hence, an inverter was attached to a jeep battery to power a COMPAQ 386 computer to record all the GPS tracking data from both receiver/antenna systems. Data were collected by cabling the mobile GPS system to the computer by means of a 150-foot cable, which limited profiling length to just under 50 m. The system was deployed to the southern flanks of Mars Hill, and a profile position was randomly chosen with a bearing of 040 degrees from true magnetic N (measured from the base of Mars Hill using a Brunton). A 100 m long tape measure was used to demark the measurement positions, which were spaced every 5 cm except for flat inter-boulder regions, at which time measurements were made every 10 or 20 cm (our "fractal" assumption). The initial profile extended for 37 m on a heading of 040 degrees, and the resampled average measurement spacing was ~ 10 cm. Measurements were made at 10 sec intervals (10 samples) at each data point, and moved the mobile GPS unit as rapidly as possible between ground sampling points. The antenna was mounted on a fibreglass pole and attached on a circular plate at the top of the 6-foot long pole to minimize interference due to intervening people etc. The first profile took approximately 50 minutes to collect, and the error analysis suggests that 2-3 cm RMS vertical precision was achieved over the length of a 37 m long GLOTAS profile (w