Samuel Goward SE590 Documentation OTTER Data Description INVESTIGATOR: Samuel N. Goward Associate Professor Laboratory for Global Remote Sensing Studies Department of Geography University of Maryland College Park, MD 20742 (301) 405-4055 Contacts: Ms. Jingli Yang, Mr. Karl F. Huemmrich, both with LGRSS, UMCP, (301) 405-4071. Requested Acknowledgement: Co-Author if primary analysis based on these measurements (otherwise you will never learn what we did wrong) or citation if used to support analysis of a related problem. In any case, interested users should contact measurement team. INTRODUCTION: Objective: The objective of this study is to characterize the sources of spectral diversity in the OTTER sites, for evaluation of biophysical phenomena which may be inferred from remotely sensed observations. The integration tool is a "turbid media" radiative transfer model (SAIL) which requires knowledge of element spectral properties in a scene. Parameters: The diversity of landscape components within an Otter site was defined in situ, by visual inspection. In general, these include litter, soil, bark, scrubs and grasses, leaves of tree species. In all cases the bidirectional spectral reflectance factor of these materials, in the 400 to 950 nanometer range was parameter sought. Discussion: The spectral reflectance factor of landscape elements should be relatively insensitive to variations in irradiance. Variations in observed remotely sensed observations should originate from interactions between these reflectance sources, incident irradiance and the 3-d geometry of the landscape elements. If variable, the spectral diversity in the landscape elements will be of primary importance in understanding remotely sensed measurements. In 1990 two measurement campaigns were undertaken by the LGRSS/UMCP staff; once in June and a second time in October. The enclosed data come from the October campaign, carried out between 10/6/90 and 10/11/90. The team. consisting of Dr. Goward and Ms. Yang and accompanied by Dr. Waring, travelled to Sisters, Oregon for measurements at Santiam, Metolius and the Juniper Site for the first 2.5 days. The remainder of the time was spent in Corvallis, conducting day trips to Cascade Head, Scio and Warings Wood. Most of the measurements were collected in situ. However, this was of only limited success in Cascade Head and of no success in Scio. In the latter case, samples were collected and brought back to Corvallis for measurement in the parking lot of the Forest Science Building, OSU. The measurements of bark, litter and other "background materials was successful. The branch needle and scrub measurements are representative of in situ conditions but are not controlled of density relative to other spectral sources and therefore must be considered unique to the observation conditions. The understory measurements, particularly at Warings Wood display wide variance and unrealistic numbers. They should be considered suspect. In general, as the variance of the measurements increases their reliability decreases. THEORY OF MEASUREMENTS The underlying assumption in this measurement campaign is that there are selected sources of spectral reflectance in landscapes which combine to produce observed remotely sensed observations. This spectral reflectance factor should be relatively invariant, at least over short periods of times. The spectrometer used in this study was designed to measure spectral reflectance. Because any radiometer measures only reflected radiance it is necessary to independently measure irradiance. A panel of painted barium sulfate was used as the transfer standard. The source of illumination is unobscured solar radiation Most materials are not perfect diffuse or "lambertian" reflectors. It is therefore quite possible to derive quite different spectral reflectance factors in solar irradiance and sensor sun relations are not controlled. First, observations were only collected when the sun was greater than 30 degrees above the horizon. Numerous approaches to collecting field spectra have been attempted in the past, including quite elaborate boom mounts on trucks. Such complex approaches are both costly and difficult to use on field settings. Use of tripods or hand-held instruments have alternate problems such as small surface IFOV and the "sneaker factor" - looking at your own shoes when the instrument is at arms length. Indeed the reflectance of the tripod and light colored clothing can be a problem. To overcome these various problems we have defined field procedures which balance between the desire for near-nadir measurements, large IFOV and the sneaker factor. For all measurements (with the exception of selected bark and water tower readings) we align the instrument head at 90 degrees from the principle plane of the sun and at 45 degree from vertical. From a 1.2 meter high tripod, this places the IFOV 1.2 meters away from the operator and the equipment. The portion of the bidirectional field observed by the instrument approximates the nadir view by observing at right angles to the strong bidirectional patterns that occur in the principle plane of the sun. For those cases were this is not possible, such as the water tower sides and the tree bark of mature species., the measurements are collected with the sun slightly to one side but behind the operator. Each measurement is paired with a panel reading, held in the same position as the object being measured. Multiple samples (10 in this case) are collected of each material. The mean of the measurements is taken to represent the material. The variance of the measurement is a quality check. If the variance is large, we reject the mean as a representative measurement and plan another measurement campaign to redo the measurements. EQUIPMENT: Instrument Description Platform: ground measurements from tripod. Key Variables: Bidirectional Spectral Reflectance Factor Principles of Operation: Dispersive grating spectroradiometer Instrument Measurement Geometry: The spectrometer head was mounted on a tripod approximately 1.2 meters above the ground. Measurements were made with the spectrometer head aimed at a 45 degree elevation angle to the target, and perpendicular to the solar plane. For bark and branch measurements the spectrometer head was held facing the targets with the sun behind to get maximum illumination. The same procedure was used with the calibration panel. % Manufacturer of Instrument: Spectron SE590 ( Spectron Engineering, Inc., 1990) equipped with a CE390 VIS/IR head from LGRSS, University of Maryland. The instrument was manufactured by: Spectron Engineering, Inc. 255 Yuma Ct. Denver, CO 80223, (303) 733-1060 Calibration Specifications: Barium Sulfate panel, covered when not in use. A bit worn but still intact. No characterization of specific reflectance properties, either spectral or bidirectional have been carried out on this panel. Based on this failure, Dr. Goward requested that OSU purchase a laboratory standard for future work. The reflectance spectra presented here assume that the Barium Sulfate is a perfectly diffuse 100% reflector. We hope to improve this assumption in later work. During the field measurements, calibration readings were taken every 15 -20 minutes during measurements. Generally the reading nearest the time of the measurement is used for calibration. Only a single scan of the panel was used in October. This error was recognized during the winter and later studies will used an 8-scan average of the panel measurement for calibration in the future. Unfortunately the LGRSS/UMCP spectrometer was not available in June 1990 when the NASA Ames staff conducted the cross- calibration study. This instrument has been periodically evaluated at the calibration facilities at NASA Goddard Space Flight Center. WE hope to compile a report of this work in the near future. A quick cross-calibration with the OSU instrument suggests that we are within three channels of the OSU spectral calibration. PROCEDURE and DATA MANIPULATION: All measurements were taken in the field under solar illumination. The instrument is aimed at a 45 degree elevation angle and perpendicular to the solar plane in order to get the most uniform solar illumination and to avoid getting the tripod or investigator's feet in the field of view. For each landscape component, a measurement of the calibration panel was taken first, and then several measurements were taken of the landscape component to measure the variability of the measurements. At the end, another measurement of the calibration panel was taken to monitor the variability of the solar illumination condition. Averages and standard deviations of the landscape measurements for each wavelength were determined. Reflectances were then calculated by ratioing the average of the landscape measurements for each wavelength and the average of calibration panel observations bracketing the data. Both the averages and the standard deviations of the landscape reflectance for each wavelength are reported. ERRORS: The lack of a calibration standard for the barium sulfate panel is a problem. Errors of as much as 10-15% relative are expected from this. Failure to scan average on the calibration readings increases noise in the measurements. This is estimated to be 1%-5% relative. Sample selection is not always successful in isolating material reflectance properties. High variance measurements should be rejected. Spectral measurements below 400 nanometers and above 900 nanometers should not be used. Validation: measurements collected in June 1990 and June 1991 attempt to replicate observations. NOTES We have already discovered mislabeling (e.g., bitter root should be bitter brush) and found that the water tower and bark samples were processed incorrectly. We are currently working on revisions of these files. We have attempted to carefully produce the measurement set we thought we were collecting in the field. Inevitability, some errors creep in. If you notice any we sure would like to know about them. REFERENCES: Petzold, D. E. and Goward, S. N. (1988), Reflectance spectra of Subarctic lichens, Remote Sens. Environ. 24 : 481-492. Goward, S. N. and Huemmrich, K. F. (1991), Vegetation canopy PAR absorptance and the normalized difference vegetation index: An assessment using the SAIL model, Remote Sensing of Environment (submitted, June 1991) : Spectron Engineering Inc., SE590 Field-Potable Data-Logging Spectroradiometer Operating Manual