Home

SSR195 - Standards for reflectance spectral measurement of

1. CS01_YYYY_MM_DD_es1 jpg East Sky 1 Walk onto road and face east No zoom on camera Include horizon for reference Standard lens is small focal length so top of photo is approx 45 deqrees above horizon Figure A 11 Photograph es1 107 2006 Omi 13 2 CS01_YYYY_MM_DD_es2 jpg point camera about 30 degrees upward no zoom and snap Include tree tops for reference and for ability to match with lower east sky shot Figure A 12 Photograph_es2 CS01 YYYY MM DD ws1 jpg West Sky 1 No zoom on camera Include horizon and trees for reference Figure A 13 Photograph ws1 108 CS01 YYYY MM DD ws2 jpg West Sky2 Facing west camera at 30 degree angle to horizon No zoom on camera Include tree tops for reference Figure A 14 Photograph ws2 2006 9 12 13 27 CS01 YYYY MM DD z1 jpg Zenith Facing west looking directly overhead No zoom Figure A 15 Photograph z1 109 The number and types of photographs that can be collected for one site alone eg Figure A 16 take small amounts of additional time when compared with the data record they provide The photographs can be linked with illumination and viewing geometry metadata including environmental conditions and information on the target along with the spectral information When many data are recorded over time and processing of spectra are not immediate or the value of spectral records are given a new application in time the photo record
2. 2 1 Spectral database application remote sensing for minesite assessment and monitoring The role of the Supervising Scientist Division is to ensure protection of people and the environment from the effects of uranium mining and to encourage best practice in wetland conservation and management in an area known as the Alligator Rivers Region ARR see Figure 1 The ARR is centred about 220 km east of Darwin in the Northern Territory covering an area of about 28 000 sq km The ARR includes all of Kakadu National Park and the western boundary of Arnhem Land Well known uranium sites in the Region include the operational Ranger mine the rehabilitated Nabarlek mine in Arnhem Land the lesser known abandoned mines of the upper South Alligator River valley such as Coronation Hill and the Jabiluka mineral lease Remote sensing technologies offer synoptic data to reduce the inherent sampling limitations of traditional ground based methods and have the advantage of contributing information to a variety of closure criteria Closure criteria are often site specific but general measures such as the creation of a stable land surface free of excessive soil erosion or sedimentation botanical succession low maintenance vegetation that blends in with the surrounding environment and a post mining landscape that is non polluting are common rehabilitation objectives Hannan amp Bell 1993 Mifsud 1996 Waggitt amp McQuade 1994 Bell 1996 Mineral
3. 7 0 cm diameter IFOV The pistol grip mounted to the tripod is fitted with a laser pointer low watt is fine for the lab to ensure the focus point is centred The standard panel dimensions which are marked on the bench should be used whenever the WR panel is being measured to ensure that the panel is in the centre FOV and that measurements are consistent Note that the WR panels are intended to be measured from the bench surface The panels should be carefully taken out of their boxes and placed on the bench Be very careful not to contaminate the surface of the panels with your fingers or any other material The panels do not lie flat on the bench because of the small step 1 cm on one of the underneath sides of the panel Two circular plastic pieces at 1 cm tall are placed on the underneath side of the panel opposite to the step so that the panel lies flat on the bench lifted by 1 cm from the surface of the bench These plastic pieces are left on the bench so that they are easy to locate Prior to spectral measurements ensure curtains are pulled to block out all light sources from the laboratory environment and turn off fluorescent lights during spectral measurements 89 A 2 3 Switch on the HgAr lamp to warm up Connect the HgAr lamp to the mains power and warm up the HgAr lamp for a minimum time of 10 minutes A 2 4 Switch laptop on The spectrometer is already switched on and running Turn the controlling laptop computer on
4. Data 2008 Field Data 2008 CSIRO 2008_03_19 El 2500 1750 2000 2250 A 1 15 Adjusting the measurement configuration fore optics and spectral averaging Open the Control Adjust configuration Alt C C RS 64664 Display Control GPS Help Take Dark Current measurement F3 De 2390 7o Initialize Radiometric measurement F9 Dic Carvent Take White Reference measurement F4 None Taken setings curo 0 Abort Spectrum Collection Ctra Parabolic Correction measurement CtrkP Spectrum Save White Reference ViewSpec Pro None Taken Spectrum Save lab 003 Optimize Parms Vnir IT 17 ms Swirl G 500 0 2048 Swir2G 500 0 2048 X a o 500 750 1000 1250 1500 1750 2000 2250 2500 Wavelength nm Latitude Longitude Elevation Fore optic selection in the pull down menu box next to the integration time set the fore optic to 8 Spectral averaging selection Spectrum averaging is the number of samples taken per observation Check the software to see the configuration for the number of samples is correct For field measurements Spectrum 25 Dark current 25 and White reference 10 The interface should appear similar to the following one RS 646 d Display Control GPS Help DAD Dt e ADDIS IAD Current instrument Configuration Foreoptic Ctl F Dark Current White Relerence 79 V Absolute Reflectance K cae Optimize Parms X Vow
5. Gomez 2001 and Specchio Hueni amp Kneub hler 2007 Reference spectra from public domain spectral libraries are often not appropriate for image matching techniques in remote sensing applications primarily because the spectra represent the reflectance response of a single specimen with a unique chemical and physical make up recorded at a particular point in time and under given 15 experimental conditions Further wavelength errors may be common with uncalibrated spectrometers Although geological materials are generally more spectrally stable than biological materials their optical properties are affected directly or indirectly by many factors such as chemical constituents scale moisture content organic matter content associated induced interferences of some minerals such as Mn and Fe and roughness and texture of the material The concept of a spectral standard becomes prohibitive given the numerous spectral measurements required to capture these spectral variations given the potential change in reflectance magnitude absorption feature position width and or depth The basis of geological remote sensing has formed from knowledge of electronic processes and their associated cause eg charge transfer crystal field affect or vibrational transitions and the resultant location of absorption features at specific wavelengths The accessibility of laboratory spectra through public domain spectral libraries has provided a basis for absorption fe
6. is saved A WR average of 10 is then saved The stabilisation pole is then swung to 90 from the panel and the electronics are allowed to adjust to the target surface by waiting for two screen refreshes Target spectra of 25 averages are then saved This step is then repeated at 60 from the panel and at 30 from the panel Finally the stabilising pole is swung back to the WR panel the electronics are allowed to adjust to the WR surface by waiting for two screen refreshes and another average of 10 readings are saved Spectral measurements begin with the DC optimisation average for each new plot site or whenever illumination conditions change The standard panel spectra are not only saved for post processing but also used as visual in situ checks If errors such as a non stable signal or spectral steps are observed the data is eliminated and saved once a stable signal is observed If any deviation from the near 100 line occurs steps or slopes another WR is collected Note that the spectrometer archives the next spectrum measurement not the one on the screen Salisbury 1998 found that the largest deviation from the averages of individual spectra was the first spectrum and that this is so common that researchers should be prepared to discard the first spectral average The reason for this is probably that users are not waiting for the spectrometer electronics to adjust to the new measurement surface or in that the operator is not realisin
7. it is possible that the measurement is considered a true representation of the target despite a contribution from external sources Absorption features from atmospheric gases increase in intensity as the atmospheric path length of the incoming solar radiation increases Clouds smoke and haze also attenuate solar irradiance by absorption which results in scattering that contributes to the secondary source of illumination and variable irradiance as a result of changing conditions between target and standard measurements Chang et al 2005 High level cloud may be invisible to the naked eye Milton amp Goetz 1997 but short term changes in irradiance caused by invisible patches of water vapour can be identified by ratioing a reflectance panel spectrum of a clear atmosphere to others in the series Milton amp Goetz 1997 The attenuation of solar irradiance degrades the signal to noise especially in the SWIR region Salisbury 1998 Fortnightly temporal measurements necessitate sampling in sub optimal environmental conditions When conditions are limiting optimisation and WR readings are saved before each measurement Metadata recording is essential to correlate the atmospheric conditions with the spectral response SSD account for environmental conditions during spectral measurement by acquiring quantitative measurements of temperature relative humidity wind speed and direction documented with a portable weather station Kestral 4000 Pocket weath
8. measurements of accuracy and precision are provided The remainder of this report provides a review of the factors affecting spectral measurements highlights those issues to which consideration can be given makes recommendations on measurement methods to minimise the influence of these factors and documents standardised procedures to maximise a true spectral response Section 5 focuses on spectra collected with a single beam instrument like that of the FieldSpec Pro FR ASD Inc 4 2 SSD s spectrometer Revegetation applications require data of high spectral resolution measured at narrow sampling intervals contiguously across the visible to shortwave infrared The spectral instrument needs to be portable easily operatable in the field environment have a low Noise equivalent Radiance NEdL and have demonstrated accurate repeatability Here we refer specifically to the FieldSpec Pro FR FieldSpec Pro FR instrument characteristics are provided in Table 4 see ASD 1999 amp 2002 for details The instrument utilises three integrated spectrometers In the VNIR 350 1050 nm the spectral sampling interval of each channel is 1 4 nm but the spectral resolution FWHM is approximately 3 nm at around 700 ASD 1999 The sampling interval for the SWIR regions 900 1850 and 1700 2500 is 2 nm with spectral resolution varying between 10 12 nm The spectral information from the three spectrometers is subsequently corrected within software for baselin
9. spatially and temporally dependent Figure 3 and complicated by atmospheric and phenological changes over time It has therefore been difficult to make recommendations on the most suitable data for revegetation assessment or rehabilitation assessment more generally The vegetative spectral response is controlled by the chemical make up of the target which is compositionally similar for many species but not identical the physiology of the plant the architecture of the plant and external factors such as localised climatic conditions and soil type or growth medium If the external factors can be controlled to be similar a measurement of a plant s spectral response will be indicative of that plant for a point in time at a particular location If measurements are carried out in an identical way for a period of time then it is possible to build up a spectral library of a plant s signature over time If this method is repeated for a number of species a database of spectral signatures over time for a number of species can be collated At that point the data can be integrated and similarity and dissimilarity measures can be undertaken both for between species and within species separability With such information it would then be possible to resample signatures to wavelengths of existing and future sensors and make recommendations on the suitability and limitations of using such data for the vegetation application It is only once such data are colle
10. 04 12 October 01 10 2007 12 00 00 44 30 32 76 54 50 November 01 11 2007 12 00 00 104 43 54 82 24 37 December 01 12 2007 12 00 00 138 48 49 TT 2T 05 52 Milton and Goetz 1997 performed field experiments to determine the spectral significance of short term changes in irradiance under clear blue skies and found little variation on first glance but significant difference with the coefficient of variation s d mean 100 calculation Anderson et al 2003 undertook a field experiment to investigate the hypothesis that the nadir reflectance of calibration surfaces asphalt and concrete remain stable over a range of time scales and found measurable differences in spectral reflectance factors over periods as short as 30 minutes despite clear atmospheric conditions Between the highest position of the Sun and that of the Sun lying low in the horizon irradiance varies but the reflectance of a Lambertian surface is independent of the position of the Sun for the same viewing angle Solar zenith angle can become a critical measurement parameter because the column density of water vapour in a given atmosphere increases rapidly as zenith angle increases from its minimum at vertical either with time of day or season because as water vapour absorption increases solar irradiance decreases and this results in lower signal to noise for the same integration time and greater difficulty in detecting spectral features throughout the SWIR but e
11. 2 11 Spectral measurement WR laboratory panel Change the spectrum average to 25 dark current average of 25 and WR of 10 Note that we are actually using the WR as both the standard and the target Change the base name by date this is the laboratory 25 4 x 25 4 cm panel Allow the spectrometer to adjust to the new surface by waiting for two screen refreshes Optimise the spectrometer Take a WR and spectrum average and save these results A 2 12 Spectral measurement WR field panel Carefully replace the laboratory 25 4 x 25 4 cm panel with the field 25 4 x 25 4 cm panel Check the field panel is in centred at the focus point Change the base name by date this is the field 25 4 x 25 4 cm panel Allow the spectrometer to adjust to the new surface by waiting for two screen refreshes Optimise the spectrometer Take a WR and spectrum average and save these results A 2 13 Spectral measurement WR small field panel Carefully replace the field 25 4 x 25 4 cm panel with the field 5 x 5 cm panel Check the small field panel is in centred at the focus point Change the base name by date this is the field 5 x 5 cm panel Allow the spectrometer to adjust to the new surface by waiting for two screen refreshes Optimise the spectrometer Take a WR and spectrum average and save these results 93 A 2 14 Identification of spectral degradation Compare the measurements of the field panels with the previous measurements by date Note th
12. CtrhA Parabolic Correction measurement CtrhP Spectrum Save Se ee None Taken Spectrum Avg 0 S Spectrum Save lab 003 Optimize Parms Vnir IT 17 ms Swirl G 500 0 2048 Swir2G 500 0 2048 1250 1500 Wavelength nm 500 750 1000 Latitude Longitude Elevation to 1750 2000 2250 2500 Tab down to Path Name C and ensure the correct working folder is marked as the target folder for all data If not click on the box with the three dots at the end of the Path Name box and navigate to desired folder Tab to the Base Name and put in the correct format for data The correct format for data is site CS BF or CP plot number eg 01 and begins at 000 eg CS01 000 Note that the software will only allow a maximum of eight alphanumeric characters in a file name The interface should look similar to the one below RS 64664 3833 nm NJIJ Dark Current None Taken Current Spectrum Save White Reference y None Taken Path Name Base Name CSIRO y Starting Spectrum Num pog Spectrum Avg Number of Files to save 0001 Interval between saves 00 00 00 Comment Spectrum Save lab 003 Optimize Parms Vnir IT 17 ms Swirl G 500 0 2048 Swir2G 500 0 2048 1250 1500 Wavelength nm 500 750 1000 Latitude Elevation Longitude Click OK or press ALT O letter o 82 IC
13. Lambertian provides a good approximation to the true bi directional reflectance factor of the subject needs to be assured by defining that the near Lambertian properties of the panels are maintained To do this we measure the spectral response of the Spectralon panels in the laboratory under the standard laboratory setup During intensive fortnightly vegetation surveys prior to each field campaign the panels are also assessed fortnightly Spectra from the two 25 4 x 25 4 cm 10 x 10 Spectralon panels and a smaller 5 x 5 cm Spectralon panel are measured One of the large panels remains in the controlled laboratory environment Like the measurements for the Mylar panel the laboratory standard panel is positioned with the focus point centred Standardised averages are a spectrum average of 25 dark current of 25 and WR of 10 The laboratory measurements are used to indicate the stability of the panels whereby a relatively flat nearly perfect reflectance should be shown Any deviation from previous measurements may indicate deterioration in the condition of the standard panel that may not yet be apparent by visual inspection These reference spectra stored by date can be queried and correlated with reflectance measurements The spectral response of the laboratory panel should not change over time and any change identified may indicate an issue with the measuring instrument that needs investigation Any variation in spectral response of the f
14. Phinn SR 2003 Hyperspectral analysis of chlorophyll content and photosynthetic capacity of coral reef substrates Limnology and Oceanography 48 489 496 Jupp DLB 1997 Issues in reflectance management CSIRO Earth Observation Centre Discussion Draft August 1996 Updated April 1996 http www eoc csiro au millwshop ref cal pdf Web site accessed May 18 2005 Kimes DS amp Kirchner JA 1982 Irradiance measurement errors due to the assumption of a Lambertian reference panel Remote Sensing of Environment 22 145 158 Kindel BC Qu Z amp Goetz AF 2001 Direct solar spectral irradiance and transmittance measurements from 350 2500 nm Applied Optics 40 21 3483 3494 Kokaly RF 2001 Investigating a physical basis for spectroscopic estimates of leaf nitrogen concentration Remote Sensing of Environment 75 2 153 161 Kokaly RF 2005 View SPECPR Software Installation Procedure and User s Guide Version 1 1 U S Department of the Interior US Geological Survey Open File Report 2005 1348 Kokaly RF amp Clark RN 1999 Spectroscopic determination of leaf biochemistry using band depth analysis of absorption features and stepwise linear regression Remote Sensing of Environment 67 267 287 Kumar L amp Skidmore AK 1998 Use of derivative spectroscopy to identify regions of differences between some Australian Eucalypt Species In Proceedings 9th Australasian Remote Sensing and Photogrammetry Conference Sydney Australia 20 24
15. Pub 92 2 Gu XF Guyot G amp Verbrugghe M 1992 Evaluation of measurement errors in ground surface reflectance for satellite calibration International Journal of Remote Sensing 13 2531 2546 Guanter L Alonso L amp Moreno J 2005 First Results from the PROBA CHRIS Hyperspectral Multiangular Satellite System Over land and Water Targets IEEE Geoscience and Remote Sensing Letters 2 3 250 254 July 2005 Hannan JC amp Bell LC 1993 Surface rehabilitation In Australasian Coal Mining Practice eds Hargraves AJ amp Martin CH The Australasian Institute of Mining and Metallurgy Victoria 260 280 Hapke B 1993 Theory of reflectance and emittance spectroscopy Cambridge University Press UK Hatchell D 1999 Fundamentals of spectroradiometry In Analytical Spectral Devices Inc 1999 Technical Guide 4th edn Analytical Spectral Devices Inc Boulder CO USA Section 13 1 to 14 7 Herold M Roberts DA Gardner ME amp Dennison PE 2004 Spectrometery for urban area remote sensing Development and analysis of a spectral library from 350 to 2400 nm Remote Sensing of Environment 91 304 319 Hick P 1999 The role of remote sensing for measuring mining impact on the Australian arid environment and the link to ecological function In On the threshold Research into practice Proceedings of the Second AMEEF Innovation Conference Western Australia 4 5 August 1999 Australian Minerals and Energy Environment Foundation Melbourn
16. Ranger site and weed species not Eriachne shultziana Psuedopogonantherum irritans Schizachyrium fragile Spermacoce sp present but important if found EWL Sciences 2005 Weeds of most concern to the Ranger site Andropogon gayanus Gamba grass Calopognium mucunoides Calopo vine Pennisetum polystachion perennial Mission grass Pennisetum pedillatum annual Mission grass Themeda quadrivalvis Grader grass Other important species at RUM are Crotalaria goreensis Gambia pea or Rattlepod Stachytarpheta spp Snake weeds Sida acuta Spinyhead sida Hyptis suaveolens Hyptis Ipomoea spp Morning glory vine Macroptilium atropurpureum Siratro vine and M lathyroides Phasey bean vine Senna alata Candle bush Senna obtusifolia Sicklepod Senna occidentalis Coffee senna Passiflora foetida Wild passionfruit vine Stylosanthes hamata Carribean stylo Stylosanthes humilis Townsville stylo Stylosanthes scabra Shrubby stylo Cenchrus cilaris Buffel grass Important species if found at RUM are Cenchrus echinatus Mossman River grass Urochloa mutica NT W A Para grass Urochloa maxima NT W A Guinea grass 21 3 1 3 Species of the Nabarlek area A project was commenced by eriss in mid 2003 at the Nabarlek minesite to quantitatively assess revegetation performance since 1995 and to develop survey methodologies applicable to the future rehabilitation of the RPA Canopy cover and ground cover vegeta
17. When chlorophyll absorption decreases so does the overall width of the absorption feature This change results in the edge shifting to shorter wavelengths Clark et al 1995 Horler et al 1983 Pu et al 2003 Collins 1978 reported that as crop vegetation approaches maturity the position of the chlorophyll absorption edge shifts towards longer wavelengths and that the red shift is a means of assessing the maturity of vegetation particularly if narrow bands around 750 and 780 nm are available 2 4 3 Senescing and stressed vegetation The spectral difference between green and dying or chloritic leaves occurs primarily in the region of 400 to 800 nm Elvidge 1990 as absorption of incident light by chlorophyll decreases As chlorophyll content begins to decrease with the occurrence of stress leaf reflectance increases initially at the chlorophyll absorption band 610 nm and then at 690 nm and 710 nm and additional chlorophyll must be lost before reflectance will increase significantly at wavelengths where chlorophyll or other pigments are strongly absorbed such as 420 or 670 nm Carter 1994 Stressed vegetation can result in water loss and a breakdown of pigments which may also lead to yellowing of leaves and a subsequent rise in blue and red reflectance wavelengths as an overall loss of chlorophyll absorption Yellowing leaf spectra show intense pigment absorptions being retained in the blue and an increase in green and red reflectance Chlorit
18. affected by a variety of technical and environmental factors However there are no national or international standards for the collection of in situ spectra The quality of spectra is dependent on the technique s used to collect spectra but these methods are rarely reported Environmental conditions in the field influence the spectral response but these factors are often not given appropriate consideration nor does appropriate metadata routinely accompany the recorded spectra This project relies on a review of the factors that affect the quality of spectral measurements and the development of an appropriate methodology to reduce extraneous variation in spectral response Data collection is coupled with the development of a spectral database that links appropriate metadata with spectra Research effort has focused on describing the factors that may contribute to a spectral response and in designing a sampling method and metadata record to both reduce and account for extraneous factors in spectral sampling Because many ground covers provide the initial stabilising component of a newly revegetated surface and considering that introduced ground covers particularly weedy grasses are threats to the rehabilitated minesite the initial focus has been on acquiring high quality time series spectra for ground cover The developed standards for spectral acquisition have provided the basis for acquiring these spectra 2 Literature review and research context
19. battery allowing spectral sampling to begin on arrival at the field destination Battery power is not an operational limitation at SSD as three NiMH spectrometer batteries and chargers are available For laboratory measurements where there are no operational considerations preventing warm up time the spectrometer should be allowed to warm up for 90 minutes connected to mains power to ensure thermal equilibrium The warm up period should also be documented in the spectral metadata so that if spectral degradation is identified a lack of warm up time can be excluded as a contributing factor 4 5 2 Standard laboratory set up at SSD There are a number of reasons why measurements are made in the laboratory and these include e Measurements to indicate the spectral stability of the spectrometer in the VNIR and SWIR such as irradiance measurements using a Hg Ar lamp or transmission measurements of a Mylar panel e Standard panel measurements and e Measurements of target spectra themselves such as soils SSD undertake measurements in the laboratory for these reasons and therefore require a standard laboratory setup to ensure consistency when measuring and recording spectral data 41 The spectral laboratory is a dark room to eliminate unwanted light sources from the laboratory environment Fluorescent lights are not used as these have their own spectral response from 350 800 nm ASD 1999 The positions of equipment for the s
20. combine small pixel size preferably combined with high spectral resolution and the capability to capture new images soon after disturbances occur now provide such continuous coverage and in contrast to intensive ground based methods over much smaller sample areas are cost effective Experience at SSD has shown that very high resolution VHR satellite data and airborne hyperspectral systems are suitable data choices for the mine environment 12 2 5 1 Very high resolution VHR remotely sensed data for revegetation applications VHR data such as that captured by DigitalGlobe and SpaceImaging presently measure up to 8 bands in the visible near infrared region at the VHR of 2 5 m pixel sizes Pan sharpening algorithms provide multispectral data down to 50 cm ground resolution elements However reducing the spatial resolution does not always ensure that vegetation covers will be detectable Lass et al 2005 because a vegetative species may have similar spectral reflectance to other vegetation or may be mixed with other vegetation Shafii et al 2004 in Lass et al 2005 SSD utilises VHR DigitalGlobe Quickbird products at 60 cm resolution to map vegetation cover including weedy distributions and to monitor the effects of disturbances such as fire at the Nabarlek minesite Northern Territory The Nabarlek minesite was rehabilitated in 1995 and has not yet met original closure criteria with the threats to revegetation success including in
21. doubtful how the absolute measurement represents the value of the quantity being measured and uncertainty must be evaluated based on any valid statistical method for treating data and based on scientific judgement using all relevant data available including previous measurement data experience general knowledge specifications data provided in calibration reports and uncertainties assigned to reference data Schaepman 1998 4 5 1 Spectrometer warm up time The spectrometer must be given ample warm up time prior to the collection of spectral data This period is required so that the three spectrometer arrays reach an equivalent internal instruement temperature A lack of appropriate warm up time will decrease the quality of spectral data and increase errors associated with detector overlap regions ie 1000 and 1800 nm ASD recommend a warm up period of 90 minutes Beal 1999 Taylor 2004 and the NERC FSF recommend a warm up time of 30 minutes MacArthur 2007a b amp c Phinn et al 2008 However Phinn et al 2008 suggest after 10 minutes there is little fluctuation in measurements SSD Approach e The warm up time should not become a limiting factor in the time or power available for spectral measurements For field sampling the spectrometer warm up period can begin while the field equipment is being loaded into a vehicle connected to the mains power o During transport to the sampling site the spectrometer is powered by a
22. for ERA Ranger Mine Draft Report 2 October 2008 EWL Sciences Pty Ltd Darwin Dawson TP 2000 The potential for estimating chlorophyll content from a vegetation canopy using the Medium resolution Imaging Spectrometer MERIS nternational Journal of Remote Sensing 21 10 2043 2051 Dawson TP amp Curran P J 1998 A new technique for interpolating the reflectance red edge position International Journal of Remote Sensing 19 2133 2139 Deering DW 1989 Field measurements of bidirectional reflectance In Theory and applications of optical remote sensing ed G Asrar Wiley New York 14 61 68 DiPietro DY 2002 Mapping the invasive plant Arundo donax and associated riparian vegetation using hyperspectral remote sensing Unpublished Master s Thesis University of California Davis May 2002 Duggin MJ amp Philipson WR 1982 Field measurement of reflectance Some major considerations Applied Optics 21 15 2833 2840 Dungan J 1998 Spatial prediction of vegetation quantities using ground and image data International Journal of Remote Sensing 19 2 267 285 Dury S Jia X Turner B amp Dibley G 2000 From leaf to canopy Determination of nitrogen concentration of Eucalypt tree foliage using HyMap image data In Proceedings 10th Australasian Remote Sensing and Photogrammetry Conference Adelaide 21 25 August 2000 875 891 Dymond JR Shepherd JD amp Qi J 2001 A simple physical model of vegetation reflectance for sta
23. high site frequencies e There were several herbaceous species with high frequency amongst the Ranger sites and the schist sites eg Heteropogon triticeus Schizachrium fragile Mnesithea formosa Alloteropsis semialata and Ipomoea graminea However only Heteropogon triticeus was noted as having high biomass on both site types The differences in species identified by Hollingsworth and Meek 2003 with Brennan 2005 can be attributed to the sites surveyed and more importantly the method of data reporting For example Hollingsworth and Meek 2003 identified many of the same grass species as Brennan 2005 but many of these grass species do not feature as a candidate species because of the criteria used to determine a candidate ie those that occur in more than one replicate plot Species important for the rehabilitation of the RPA can be derived from the Ranger revegetation strategy of the trial landform which include the following understory species Daws et al 2008 e Aristida hygrometrica e Aristida holathera e Eragrostis sp 20 Daws et al 2008 probably deliberately exclude high biomass covers like Sorghum spp and Heteropogon spp in an attempt to reduce fire on the landform Species may also have been selected for ease of germination and seed collection In addition Energy Resources of Australia ERA held a workshop on the Weeds at Ranger and defined those weeds of most concern to the Ranger Site important weed species to the
24. in stabilising soils and preventing erosion during the initial phase of revegetation Introduced ground cover species of weeds have the potential to impact on species diversity and abundance and affect the frequency and intensity of fire disturbances Further reports will address the spectral characterisation combined with appropriate metadata of geological materials including waste rock soil and ore outcrop trial landform characterisations vegetation measurements along environmental gradients processed mine materials and potentially aquatic components These spectra will also be documented in the SSD Spectral Database using laboratory and in situ measurements Further reports will document the analysis of the spectral data acquired from these sources 65 7 References Abdou WA Bruegge CJ Helmlinger MC Conel JE Pilorz SH Ledeboer W Gaitley BJ amp Thome KJ 2002 Vicarious calibration experiment in support of the multi angle imaging SpectroRadiometer EEE Transactions on Geoscience and Remote Sensing 40 7 1500 1511 Adams ML Philpot WD amp Norvell WA 1999 Yellowness index an application of spectral second derivatives to estimate chlorosis of leaves in stressed vegetation International Journal of Remote Sensing 20 18 3663 3675 Analytical Spectral Devices ASD Inc 1999 Technical guide 4h edn ed Hatchell DC Analytical Spectral Devices Inc Boulder CO USA Section 0 1 to 26 8 Available online http www asdi com tg
25. of declared weeds weeds of concern and ground covers of importance to RUM and Nabarlek are outlined in Table 2a c The list focuses on weeds and native grasses only and further information on native herbs and sedges can be found in Hollingsworth and Meek 2003 Brennan 2005 and Bayliss et al 2004a amp b Table 2a Summary of target weedy grass species important for Ranger Nabarlek and weeds declared of concern Genus Species Ranger Nabarlek Weeds declared of concern Andropogon gayanus v v v Cenchrus cilaris v Cenchrus echinatus v Chloris inflata v v Chloris gayana v Chloris virgata v v Cynodon dactylon v v Echinochloa colona v Weedy Melinis repens v v grasses Paspalum plicatulum v Pennisetum pedicellatum v v v Pennisetum polystachion v v v Setaria sp v Sporobulus sp v Themeda quadrivalvis v v Urochloa humidicola v Urochloa maxima v v Urochloa mutica v v identified by Brennan 2005 Hollingsworth and Meek 2003 EWLS 2005 and Daws et al 2008 identified by Bayliss 2004a amp b 23 Table 2b Summary of target weedy herb and vine species important for Ranger Nabarlek and weeds declared of concern Genus Species Ranger Nabarlek Weeds declared of concern Acanthospermum hispidum v Aeschynomene americana v v Alysicarpus va
26. on growing season with environmental conditions and scale dependence including background reflectance also affect the spectral response of plant material Particularly for ecological applications the relationships between EMR biophysical features illumination geometry and viewing geometry are increasingly complex when compared to static inorganic materials Given both the number of spectral samples and metadata required it is therefore not surprising that there are few standardised vegetation spectral references in the public domain It is unfortunate that much field based research effort is not transferable due to poor research methods and a lack of considered metadata There are therefore several causes of inherent variability in replicated measurements of the reflectance of vegetation in the field These include 1 geometric differences of both object structure and illumination viewing configurations 2 variations in performance of the actual field sampling protocol and 3 true temporal variations in the individual samples collected for analysis or natural variability It is necessary to determine whether the observed reflectance differences between plant species are not only statistically significant but consistent eg over different seasons or in different habitats before they can be generally applied with success in remote sensing species mapping Fyfe 2003 16 2 6 2 A methodical approach for collecting field based spectra a r
27. optic at the Spectralon panel and press the WR button or F4 A reflectance curve with a near horizontal line at a value of 1 should appear if the illumination and viewing geometry set up is correct Allow two screen refreshes you can wait longer and observe the reflectance line confirming that illumination conditions are not changing because the line is quite stable and if the white reference reading is stable press the spacebar to record the WR It will have the file suffix_ 001 A 1 19 Taking measurements target Swing the horizontal bar of the stabilising pole by 90 over the vegetation plot Wait for two screen refreshes Press the spacebar to save File suffix 002 Repeat this step at 60 and 30 Save the target spectra File suffix 003 004 A 1 20 Taking measurements repeat white reference Swing the probe back over the centre of the Spectralon panel Wait for two screen refreshes 84 Take another WR reading press the WR button or F4 A reflectance curve with a near horizontal line at a value of 1 should appear if the illumination and viewing geometry set up is correct Save the Spectralon as a target by pressing the spacebar File suffix 005 Close white panel box to prevent airborne contaminants settling on white panel surface If during the measurements saturation occurs then optimise again and repeat the measurements with the steps as described above steps 16 to 20 A 1 21 Recording enviro
28. quadrant is given The laptop and weather station see Section 4 7 5 are synchronised to the Australian Central Standard Time Azimuth and altitude are calculated post field at the Geoscience Australia Compute Sun and Moon elevation site http www ga gov au geodesy astro smpos jsp The latitude and longitude coordinates degrees and minutes combined with the time zone recorded in the spectral header are entered to obtain the Sun s position and also the solar azimuth and altitude WR and solar radiance spectra are used to assess these factors both by visual in situ assessments and during post processing of spectra Although non Lambertian reflectance with respect to global radiation of Spectralon panels may occur at very large solar zenith angles above 60 zenith angle or equivalent to 30 solar elevation angle Rollin 1999 this is not an issue for spectral sampling around the wings of solar noon in the tropics from April through October see Table 6 53 4 6 5 Atmospheric conditions clouds smoke haze humidity wind and temperature Illumination contributions from diffuse and hemispherical sources are another potential variable in obtaining reference spectra because reflectance spectra measured under solar illumination are strongly modified by the absorbing molecules in the atmosphere Goetz 1992 in Schaepman 1998 and accounting for solar geometry and atmospheric fluctuation can increase accuracy Milton et al 1995 Radi
29. reduced in the spectral signature by spectral averaging as truly random noise will be reduced by an amount proportional to the square root of the number of spectra averaged together ASD 2000 SSD s sample average of 25 is adopted and three sets of 25 spectra for each target are measured which can be averaged during post processing Integration timing and sequential measurements are discussed in Section 4 7 3 Stray light is significantly greater than the lowest level random noise and is indicated by the appearance of a spectral reflectance signal in spectral regions of zero illumination energy eg the atmospheric water absorption bands around 1400 nm and 1900 nm The ultraviolet and blue wavelengths where illumination energy is extremely low are also susceptible to stray lightt Stray light may affect the accurate detection of features including chlorophyll a and b electron transitions at 430 nm and 460 nm respectively water O H bend at 1400 nm lignin C H stretch at 1420 nm starch O H stretch C O stretch at 1900 and water lignin protein nitrogen starch and cellulose O H stretch and O H deformation at 1940 nm Curran 1989 If these effects are noted these measurements and deviated products should be regarded with care In the field environment a solar radiance W m7 steridan nm measurement made over the WR is recorded prior to collecting each averaged reference spectra to provide an estimate of irradiance This spectrum
30. reflectance derivative spectroscopy band depth analysis of absorption features and stepwise linear regression Kumar amp Skidmore 1998 Adams et al 1999 Lamb et al 2002 Buschmann amp Nagel 1993 Kokaly amp Clark 1999 Price 1992 1994 Portigal et al 1997 Pu et al 2003 Buschmann amp Nagel 1993 Filella amp Penuelas 1994 64 6 Conclusion No matter what the application spectral data must be collected in a well designed and consistent manner Common practice should be to collect and document metadata associated with the spectral response Minimum metadata requirements described for the SSD Spectral Database have been outlined and these always accompany the spectral information Extreme caution should be placed on using reference spectra without such metadata Any spectral data sharing necessitates the supply of both spectral and metadata information The benefit of obtaining validated spectral data outweighs the small additional investment in time required for metadata collection A rigorous spectral and metadata collection protocol can reduce systematic bias and minimise variability by accounting for extraneous factors It is then possible that such data are useful for expediating application development due to the role of these spectra for remote sensing feasibility studies 6 1 Further work and reporting Ground based vegetation was the initial focus of this work because grasses herbs and sedges often have immediate importance
31. respect to species diversity Weeds of National Significance including wetland examples eg Salvinia and Hymenachne are currently not a high threat to minesite rehabilitation in the ARR due to the small areal extent of post mining wetland features Priority has been given in establishing plots of a range of species rather than establishing plots of species replication It is acknowledged that a limiting factor of the method is that there may be species that potentially confound the spectral response of the species range but for which no knowledge will be acquired Nevertheless the project scope will provide a knowledge base far greater than that ever obtained for vegetation spectra with respect to species numbers frequency of sampling duration of sampling and method and metadata documentation Where possible replicates of species are sampled at different locations For some species such as Heteropogon contortus which is very variable morphologically Sharp amp Simon 2002 replication is considered vital 3 4 2 Growth medium and environmental conditions A vegetative species spectral response is a function of a variety of factors ranging from soil type and soil condition to local meteorological conditions The spectral response also varies over scales ranging in time diurnal to seasonal and space It is acknowledged that the spectral response measured is a function only of a point in time for a particular vegetation sample Ideally a n
32. sciences soil sciences and limnology and oceanography since the 1960s Awareness of spectroscopy principles has moved beyond the specialist applications and into the general remote sensing community because airborne hyperspectral applications are increasing and higher spectral dimensions from satellite data are now available eg data from Hyperion travelling on EO1 Earth Observation 1 and CHRIS Compact High Resolution Imaging Spectrometer With the advent of sensors capable of collecting high spectral resolution radiance data has come the expectation that if measurements are made with sufficient spatial resolution to avoid spectral mixing most types of rock soil and vegetation should be remotely identifiable Cochrane 2000 Satellite platforms are currently inferior to airborne platforms with respect to spectral range and resolution and radiometric stability The major operational limitation with airborne platforms is a lack of affordable data at required frequencies For smaller scale applications the required number of flight runs due to small swath widths may be another limitation of airborne platforms Due to the dynamic nature of the mine environment it is not feasible to perform a cost benefit analysis of remotely sensed data due to a lack of data with suitable frequency and specifications At best a conceptual matrix of factors and suitable sensors can be developed similar to the Coastal Remote Sensing Toolkit University of Queensland
33. size and distribution of the target element and the orientation of the sun azimuth relative to any preferred orientations of the target Deering 1989 For in situ ground cover measurements a consistent 2 metre height above the ground providing an approximate 28 cm diameter GFOV Figure 22 is used Note that the IFOV is actually slightly larger than the 28 cm due to the point spread function of the optics however this is not a limiting factor given all plots are typically greater than 2 m2 and represent a dense and homogenous plot of the target of interest Vegetation height obviously needs to be taken into account 20c Coin Ll Figure 22 Direction position and FOV 48 4 6 2 WR standard panel and target measurements in the field 4 6 2 1 WR panel measurements The WR panel is housed in a wooden case on the shelf of the buggy 1 m from ground level shown in Figures 22 and 23 In situ WR measurements are made positioned on the side of the target point opposite the sun from a height of 2 m above ground level providing an approximate 14 cm diameter IFOV given the 1 m difference between FOV and panel The bubble leveller pistol grip attached to a stabilising pole and the laser pointers are attached to the pistol grip are utilised to locate the centre point on the WR panel Prior to the acquisition of the laser pointer a weight was strung from the pistol grip which was used to cast a shadow at nadir and highlight the focus point
34. stabilising pole is rotated 90 degrees to sample the target from a height of 2 metres Two additional sets of spectra are obtained by rotating the stabilising pole 60 and 30 degrees sequentially at a horizontal distance of 1 metre from the stabilising pole These three sets of target spectra are saved to measure the presence of inter target spectral differences and to compare these data for similarity A decision on the sampling height of target spectra was made during the design phase of the project One option was to sample the target from varying heights at a fixed distance dependant on the maximum height of the vegetative sample This option would have required a height adjustable stabilising pole and accurate measurements of the vegetation height defined by some criteria to account for height variation such as mean height This method would have given a consistent GFOV but would have required a change in setup for each target measurement The second option and that which was adopted was to maintain a consistent measurement height of 2 metres This method allows for efficient deployment of the stabilising pole and quicker sampling of sequential sites compared with the first option This method does mean that the GFOV of the target will vary as the plant grows Typically heights of vegetation covers sampled range from ground habits to that of Andropogon gayanus Gamba grass which can reach a growth height of 4 m Smith 1995 amp 2000 A
35. the maximum plant height mean density or the height at which most biomass occurs Describe the phenology of the sample with terms such as green growth flowering seeding senescing or drying Record any disturbances that are visualised such as trampling Record the side of the plot the measurement is recorded from eg western side of plot This position will be the side opposite the sun and can be calculated given the GPS position date and time of day recorded in the spectral header if required 85 A 1 23 Take standard reference photos buggyl photographed five paces from the site Includes buggy and fore optics in relation to the site Sl site 1 photographed five paces from site no zoom on camera Captures site and surrounds _s2 or site 2 taken from same location as s1 but with the camera zoomed to photograph the site only obnl oblique looking North 1 taken standing on the southern edge looking north with camera pointed 45 degrees at the plot _obn2 oblique looking North 2 taken at the same position as obn1 but with the camera held level to image taller vegetation _obsl oblique looking South 1 taken standing on the northern edge looking south with the camera pointed 45 degrees _obs2 oblique looking South 2 taken standing on the northern edge looking south with the camera held level to image taller vegetation nl n2 and n3 nadi
36. use the optical fibre cable must be kept loosely rolled to avoid any slight or permanent bends or kinks which will distort the light field reaching the spectrometer The user should be particularly careful to ensure that the cable is well within the case and does not get caught when closing the case When taking measurements care should be given to ensure the fibre cable is loose but not in the way of measurement activity where the cable could be stepped or kneeled on Velcro straps are used to secure the cable loosely to the stabilising pole Any length of cable not required for sampling is loosely rolled and secured with velcro attached to the spectrometer by the spectrometer handle Ensure the cable is capped when not in use Despite the use of the cap there is probably some dust still accululating on the fibre optics It has been suggested that the fibre optics are cleaned with normal lens cleaning tools for future use Note that if kinks are visible in the cable it is likely that the light field reaching the spectrometer is distorted and if this is not obviously apparent distortion can be checked by the standard laboratory measurements Kinks in the fibre optic usually mean that the instrument must be sent back to ASD Inc for testing recalibration and probably the replacement of the fibre optic cable 98 A 4 Set up of the spectrometer Note that this field set up is designed so that one person is capable of setting up and recording spec
37. with another Instructions for standardised photographic recording are provided in Appendix A 6 4 6 8 Information on the target The nature of the target in the localised environment must be documented with meaningful descriptions The site code must be documented for the vegetation plots that are sampled temporally CSIRO Berrimah Farm and Crocodylus Park have been given abbreviations of CS BF and CP and the site is followed by a 2 digit number eg CS02 referring to CSIRO site 2 Refer to Pfitzner and Bollh fer 2008 for a summary of the status of the vegetation plots including the site codes at CS BF and CP It is also important to describe the side of the plot that spectral measurements are made This is because the measurement side will change with sampling occurring at different times of the day and year due to the sampling side being opposite the Sun For any spectral vegetation sampling campaign documentation for vegetation includes e species name if confirmed or labelled sample for identification by the herbarium e homogeneity monoculture or mixed community described by percent cover of each component including a break down of any cover of leaf litter or soil interspace as well as differences in phenology of the target species For example a plot could contain 90 green cover of Hyptis 5 drying cover of Hyptis 3 dead leaves as leaf litter and 2 exposed lateritic gravel 57 single layer or multiple layer e t
38. 00 milliseconds One 512 element Si photodiode array 350 1000 nm VNIR Two separate TE cooled graded index InGaAs photodiodes 1000 2500 nm SWIR 1 and SWIR 2 1000 nm between VNIR and SWIR 1 1800 nm for SWIR 1 and SWIR 2 Optional fore optics available UV VNIR 1 4 x 10 9 W cm 2 nm sr 700 nm NIR 2 4 x 10 9 W cm 2 nm sr 1400 nm NIR 8 8 x 10 9 W cm 2 nm sr 2100 nm 15 8 Ibs or 7 2 kg Wavelength reflectance radiance irradiance All calibrations are NIST traceable radiometric calibrations are optional Standard ASD fibre optic cable is 1 m in length SSD s ASD has a 5 m fibre optic cable 35 4 3 Considerations with single Field of View FOV instruments It is beyond the scope of this report to review the physics of propagation of EMR in free space or the interaction of EMR with matter Extended summaries of the relationship with laws of radiation absorption and emissivity the physics of measuring extended sources in the field and the relationship of bidirectional reflectance distribution function or BRDF with reflectance measurements can be found in numerous references such as Nicodemus et al 1977 Horn and Sjoberg 1978 Silva 1978 Robinson and Biehl 1979 Duggin and Philipson 1982 Baumgardner et al 1985 Milton 1987 Deering 1989 Pinter et al 1990 Hapke 1993 Milton et al 1995 Jupp 1997 Schaepman 1998 Hatchell 1999 Rees 2001 and Schaepman Strub et al 2005 Note that
39. 00cm 4 TS r tan 8 2 x 100 7 02 cm A nr a 7 02Y 154 8cm The area A sampled from a height of Im is 0 0155m Note that the sensitivity across the FOV is not uniform and therefore the size of the area that is to be measured should be large relative to the GFOV of the sensor MacArthur et al 2006 demonstrated that areas outside the theoretical FOV influence the reflectance recorded and therefore the homogenous portion of the target should be larger than the anticipated GFOV The FieldSpec pistol grip is available with both a sighting scope and levelling device SSD also use two remote controlled laser pointers that are attached on either side of the pistol grip and these accessories allow the user to view the spot where the fore optic is pointed while oriented in nadir viewing geometry Because of the need to orient the FOV geometry in a stable way measurements are performed using the fore optics mounted on a tripod The small size of the fore optics greatly reduces error associated with instrument self shadowing but the instrument as well as other objects including the operator should be placed as far as possible from the surface under observation as even when the area viewed by the fore optic is outside the direct shadow of the instrument the instrument still blocks some of the illumination either diffuse skylight or light scattered off surrounding objects that would normally be striking the surface under observation ASD 19
40. 07 WE La j ka rar ow og etae ITE JE EE Zn T D Continuing Discontinued New Established 2007 Figure 7 Location of Crocodylus Park Sites Western Paddock April 2007 28 Continuing Discontinued New Established 2007 Figure 8 Location of Berrimah Farm sites April 2007 Examples of established plots are illustrated in Figure 9 Dm Digitaria milanjiana Jarra Grass Brachiaria humidicola Tully Grass Stylosanthes humilis Townsville 2008 01 23 2008 01 23 stylo 2008 03 19 Digitaria swynnertonii Arnhem Grass Melinis repens red natal grass 2008 Aeschynomene Americana 2008 03 Pennisetum pedicellatum 2008 03 04 03 19 04 Figure 9 Examples of vegetation plots used to record the spectral reflectance of selected species over time 29 Selected photographic and spectral examples for one plot of Digitaria swynnertonii Arnhem Grass is displayed in Figure 10 RJ a 2 Reflectance Reflectance Reflectance 9 iy f i Boy ad 1 1 Y T 500 1000 500 2000 25i 500 7060 500 2000 nm M 0 25 Wavelength nanometers Wavelength anman a 25 500 2007 04 11 2007 04 23 2007 05 10 d Reflectance Reflectance Reflectance m iy 2 a 1000 SUD 2000 25 500 1000 500 2000 Wavelength ranometers Wavelength Ra D meis 2007 06 12 2007 07 18 Endmember Collection Spectra Reflectance Reflectance Reflec
41. 09 06 06 20 06 06 03 07 06 LLL ALTE Rana Sample nadir Zenith Average spectra 50 samples R 51 samples Figure 27 Fortnightly temporal ground cover spectra accompanied by selected metadata for Stylosanthes humilis over 4 months 60 Soil inter space was also measured The standards described were implemented for each observation period This example shows changes in date time position sun azimuth sun altitude temperature humidity cloud cover and type homogeneity cover phenology and localised conditions The standard and averaged spectra also change The spectra show a similar overall shape and position of absorption features the depth and width of absorption features and the magnitude of reflectance changes as the sample senesces over time The depth of water absorption features also change over time Whether or not these changes are a result of biophysical changes of the target or attributable to the illumination conditions can only be assessed by an increased length of sampling record It is only with accurate spectral and metadata collection that both averaged reference spectra and any significant temporal change in spectral response can be identified 5 1 Data storage and processing Pfitzner et al 2008 describe the development of SSD s Spectral Database that is used to reference categorise and manage our spectral data and metadata so that suita
42. 1 17 ms Swwl S00 U 2048 SwwZ S00 0 2048 4000 4250 1500 1750 2000 Wavelength nm Select OK to accept the details and close the window 83 A 1 16 Taking measurements optimisation Given suitable sampling conditions ensure the fore optic is pointed at the centre of the WR panel Open the white panel lid to expose the white panel Press the Opt button or CTRL O You will see the profile changing while the instrument is adjusting The different regions of the three detector arrays will be visible with obvious separation around the 1000 and 1800 nm region You may notice a clicking sound when the optimisation process is complete Note that there should be no movement of the fore optic during spectral measurement which is obtainable with the fore optic mounted in the standard set up The operator must ensure he she is on the side of the computer panel and target away from the sun and that their presence is not interfering with the spectral measurement in terms of contributing shadowing or scattering components A 1 17 Taking measurements irradiance After optimising and collecting a dark current the graph will display measurements in radiance raw digital numbers and plot them against wavelength in nm This is the incoming solar spectrum Press the space bar to save the averaged spectrum It will have the file suffix_ 000 A 1 18 Taking measurements white reference Continue pointing the fore
43. 2006 To make recommendations on the most suitable remotely sensed data for a given application an understanding of target separability at a given spatial and spectral resolution is required Targets are rarely spectrally static over time and acquiring knowledge on target differentiation therefore requires investment in the collection of temporal spectra Feasibility studies on the Reflectance spectral differentiation of vegetation species over time and at different resolutions can only be achieved cost effectively by the collection of in situ spectral data due to the frequency of samples required The advantage of this approach is that results are transferrable to a variety of applications where information on land cover separability at different scales is required 2 4 The generalised spectral response of vegetation A review of the spectral characteristics of vegetation can be used to suggest possible causes of changes in vegetation spectral responses over time This section describes typical spectra of healthy green vegetation and the changes that occur for stressed and dried vegetation The descriptions can be visualised with Figure 4 w D 0 6 2 a T Reflectance e T 2 Reflectance o N T 29 E 9 a 9 a 500 7500 2000 25 500 j 1500 2000 25C 500 1000 1500 2000 25 TO velen ti nm 000 Wavelength nm Wavelength nm a Green spectrum b Senes
44. 23390 b Initialize Radiometric measurement F9 Take White Reference measurement F4 Dark Current 9 Adjust Configuration None Taken Optimize instrument settings Ctro Abort Spectrum Collection CtrhA Parabolic Correction measurement CtrhP None Taken Spectrum Save lab 003 Optimize Parms Vnir IT 17 ms SwirlG 500 O 2048 Swir2G 500 0 2048 500 750 1000 1250 1500 1750 2000 2250 Wavelength nm Latitude Longitude Elevation Tab down to Path Name C and ensure correct working folder is marked as the target folder for all data as described in the section above If not click on the box with three dots at the end of the Path Name box and navigate to desired folder Tab to Base Name and put in correct format for data by date Note that the software will only allow a maximum of eight alphanumeric characters in a file name The default starting spectrum is 0 and this is fine Check the Starting Spectrum is set to 0 Set the Number of Spectra to be Saved option to 1 and click OK 91 A 2 9 Spectral measurement HgAr lamp spectra Select the menu Control and the submenu Adjust Configuration and set the fore optic to bare fore optic and raw DN file Change the spectrum average to 30 and dark current average to 25 and WR average to 10 After the standard warm up times are reached 90 minutes for the spectrometer and at least 10 m
45. 7 0 23 3 38 8 41 3 2 5 75 1 31 10 5 34 9 87 2 93 0 5 8 100 1 8 14 1 46 6 155 0 165 3 10 3 110 1 9 15 5 51 3 188 6 201 1 12 5 150 2 6 21 1 70 0 349 0 372 0 23 0 200 3 5 28 1 96 3 620 6 661 6 41 0 250 4 4 35 1 116 5 969 8 1033 8 64 0 300 5 2 42 2 139 9 1396 0 1488 2 92 2 350 6 1 49 2 163 2 1900 3 2025 8 125 5 400 6 9 56 2 186 5 2482 4 2646 3 163 9 500 8 7 70 3 233 2 3878 2 4134 2 256 0 4 5 Spectral stability of the equipment Key sources of error are the standards to calibrate spectrometer devices as well as the laboratory equipment used for calibration Schaepman 1998 Routine quality assurance tests can be performed to ensure that any change in the performance or accuracy of the spectrometer or standard panels can be identified quickly Such changes may be a result of damage to the spectrometer or panels or as a result of long term drift in the instrument or standard panel stability Kindel et al 2001 found that the ASD FR instrument shows excellent radiometric stability over a nine month period of measurement better than 196 for virtually the entire wavelength regions and better than 0 5 for wavelengths beyond 1000 nm Schaepman 1998 provides 40 an extensive discussion on the calibration and characterisation of spectrometers and identifies all possible sources of uncertainty during characterisation and calibration of spectrometers Even if all sources of errors are identified and an uncertainty associated with each it is still
46. 99 In addition to the bare fibre optic 25 SSD also have an 8 and 1 degree lens 39 Table 5 provides a summary of the diameter of the FOV given for selected heights using a 1 8 and 25 lens The field and laboratory measurements made at SSD are undertaken with an 8 foreoptic so that the angle of acceptance is less than 20 full angle Baumgardner et al 1985 Deering 1989 Milton 1987 For practical purposes the FOV can be considered circular in shape The FOV will be elliptical if the viewing angle is off nadir or the target is not a flat plane eg the target is not flat and or textured Table 5 shows the difference in area for a circular and elliptical FOV using an 8 lense The area of an ellipse is slightly greater than the area of a circle and because a target will not usually be planar then it is best to exaggerate the required GFOV to ensure that it is only the homogenous target that is being measured in the FOV Table 5 Calculations at 90 nadir of diameter for varying FOV lenses and the difference between a circle and ellipse for an 8 FOV example A 8 of A8 of difference b n circle Height cm d1 cm d8 cm d25 cm circle cm ellipse cm and ellipse of 8 cm 5 0 1 0 7 2 3 0 4 0 4 0 0 10 0 2 1 4 4 7 1 6 1 7 0 1 15 0 3 2 1 7 0 3 5 3 7 0 2 20 0 4 2 8 9 3 6 2 6 6 0 4 25 0 4 3 5 11 7 9 7 10 3 0 6 30 0 5 4 2 14 0 14 0 14 9 0 9 35 0 6 4 9 16 3 19 0 20 3 1 3 40 0 7 5 6 18 6 24 8 26 4 1 6 50 0 9
47. 99 found the absorptions from different plant materials are similar and overlapping so a single absorption band could not be isolated and directly related to chemical abundance of one plant constituent while Wessman et al 1988 found each constituent eg cellulose protein of a complex organic mixture has unique absorption properties in the near infrared region 700 2500 nm of the spectrum Although the general shape of the spectral curve may be similar for all green vegetation changes in reflectance occur through variations in amplitudes of the curve For example Gates et al 1965 report that visible absorptance substantially increases from lighter to darker coloured leaves and for thick leaves reflectance drastically increases in the near infrared Differences in chlorophyll and water absorption positions and reflectance magnitude differences across regions of the spectrum may occur both between and within species Vegetation stress senescing and desiccation all produce changes in the spectrum 2 4 2 Chlorophyll and red edge changes Changes in the chlorophyll content of plants can be used as an assessment of nutritional and environmental stresses The chlorophyll a absorption band is centred at 680 nm Elvidge 1990 Datt 1999a amp 2000a Clark et al 1995 However as the absorption is intense the chlorophyll absorption band minima will not change much with increased or decreased absorption but the wings of the absorption will change
48. A 4 s site 1 is photographed five paces from the site with no zoom on the camera and this photograph captures the site and surrounds Figure A 5 s2 or site 2 is taken from same location as sl but with the camera zoomed to photograph the site only Figure A 6 _obn oblique looking North 1 is taken standing on the southern edge looking north with camera pointed 45 degrees at the plot Figure A 7 obn2 oblique looking North 2 is taken at the same position as obnl but with the camera held level to image taller vegetation obs oblique looking South 1 is taken standing on the northern edge looking south with the camera pointed 45 degrees Figure A 8 obs2 oblique looking South 2 is taken standing on the northern edge looking south with the camera held level to image taller vegetation n n2 and n3 nadir is taken from nadir with the camera held at shoulder height moving across the site from west to east Figures A 9 and A 10 n4 n5 and n6 is taken from nadir with the camera held at a 1 meter height or as the vegetation height will allow with camera on full zoom moving across the site from western edge to centre and then to eastern side es and es2 east sky is taken of the eastern sky at horizon and at 45 degrees respectively Figures A 11 and A 12 wsl and ws2 west sky is taken of the western sky at horizon and at 45 degrees respectively Figures A 12 and A 13 Note th
49. ASCII format Ensure that you do not plug the GPS into the laptop before starting up the laptop as if the GPS is connected prior to starting the computer the computer will falsely recognised the GPS as an input mouse Adjust the setup of spectral measurements in RS3 Note that spectrum averaging is the number of samples taken per observation and that the more samples taken the higher the signal to noise ratio and the longer the time taken for a target measurement A balance must be met between obtaining good signature averages in a period that will not expose a change in illumination conditions In RS3 go to the top menu list and select Control and then Instrument Configuration For field measurements set the number of sample configurations to Sample spectrum 25 Dark current 25 and White reference panel 10 In the pull down menu box next to the integration time set the fore optic to 8 The exception here would be during the Hg Ar lamp readings where the bare fibre optic tip is used and the fore optic should therefore be set to 25 The output spectrum type should be set to reflectance by selecting the pull down menu box next to the fore optics selection This parameter can also be set to raw DN and radiance when required Spectra are saved in a binary format It is essential to adopt a file management system for acquiring spectral data SSD s management of data is to set up a folder structure on the C drive of the laptop Each
50. ASD 1999 and Salisbury 1998 provide a glossary of terms for NIR terminology Simply the amount of the reflected power gathered by the sensor is proportional to the square of the FOV the sensor aperture area the irradiance the irradiance angles the sensor view angles the bidirectional reflectance distribution of the target optical transmission quantum efficiency and wavelength dependency 4 3 1 The reflectance factor RF The fundamental property governing reflectance behaviour is its Bidirectional Reflectance Distribution Function BRDF Nicodemus 1982 in Deering 1989 which cannot be measured directly Nicodemus et al 1977 but approximated if multidirectional field radiance measurements are made Deering 1989 The term bidirectional reflectance factor BRF relates the reflectance from a target surface to the reflectance that would be observed from a Lambertian surface located at the target BRF is considered the standard reflectance term as defined fully by Nicodemus et al 1977 to describe the field reflectance measurement one direction being associated with the viewing angle usually 0 from normal and the other direction being the solar zenith and azimuth angles Robinson amp Biehl 1979 Silva 1978 R of standard panel 0 0 where 0 and 0 are the zenith and azimuth angles of the incident beam and reflected beam respectively In reality the BRF can only be estimated using dual field of view goniomete
51. Clark et al 1995 Datt 1999a found that reflectance near 710 nm showed maximum sensitivity to chlorophyll content and that the reflectance near 550 nm was a less sensitive indicator in Eucalyptus sp leaves The reflectance near 550 and 700 nm shows maximum sensitivity to a wide range of chlorophyll contents Buschmann amp Nagel 1993 There are two primary red edge optical parameters red edge position REP and red well position RWP Pu et al 2003 The combined effects of strong chlorophyll absorption and internal leaf scattering cause this abrupt change Dawson amp Curran 1998 Horler et al 1983 Estimates of the spectral range of this red edge region differ slightly from author to author including 680 730 nm Clark et al 1995 690 740 nm Lamb et al 2002 and 680 750 nm Horler et al 1983 Miller et al 1991 Munden et al 1994 Filella amp Penuelas 1994 Belanger et al 1995 Datt 1999a Pu et al 2003 The red edge has been used to indicate changes in the chemical and morphological status or vitality of plants Clark et al 1995 Collins 1978 Boochs et al 1990 Dawson amp Curran 1998 Datt 2000a Barret amp Curtis 1992 Elvidge 1990 Pu et al 2003 Belanger et al 1995 found that seasonal chlorophyll values for trees expressed on an area basis tend either to increase to a short lived maximum and then to decline or to rise to 10 relatively steady state value during much of the season depending on the species
52. Exhibition Copenhagen Denmark 7 10 July 1997 vol II Ann Arbor MI ERIM 789 797 Price JC 1992 Variability of high resolution crop reflectance spectra International Journal of Remote Sensing 14 2593 2610 Price JC 1994 How unique are spectral signatures Remote Sensing of Environment 49 181 186 Pu R Gong P Biging GS amp Larrieu MR 2003 Extraction of red edge optical parameters from hyperion data for estimation of forest leaf area index IEEE Transactions on Geoscience and Remote Sensing 41 4 916 921 Rathmore CS amp Wright R 1993 Monitoring environmental impacts of surface coal mining International Journal of Remote Sensing 14 6 1021 1042 Rees WG 2001 Physical principles of remote sensing 2 edn Cambridge University Press UK Robinson BF amp Biehl LL 1979 Calibration procedures for measurement of reflectance factor in remote sensing field research Measurements of Optical Radiations SPIE 196 16 26 Rollin EM 1999 NERC FSF USER NOTE Sun angle Correction Factors for Spectrolon Reference Panels Natural Environment Research Council Field Spectroscopy Facility Available online http fsf nerc ac uk resources post processing post processingV2 pdf pp spec angular pdfH 75 Rollin EM Emery DR amp Milton EJ 1995 The design of field spectroradiometers A user s view In Remote sensing in action eds Curran PJ amp Robertson YC Proceedings of the 21st Annual Conference of the Remote Sens
53. Goetz AFH 2001 Field Spectrometry techniques and instrumentation In Technical guide 4 edn ed Hatchell DC Analytical Spectral Devices Inc Boulder CO USA Section 12 1 to 12 10 Available online http www asdi com Field 20Spectroscopy screen pdf Curtiss B amp Ustin SL 1989 The remote detection of early stages of air pollution injury in coniferous forests using imaging spectrometry In Proceedings European Joint Research Centre Remote Sensing of Forests Workshop Ispara Italy September 1988 Danson FM 1995 Developments in the remote sensing of forest canopy structure In Advances on environmental remote sensing eds FM Danson amp SE Plummer John Wiley Chichester 53 69 Datt B 1999a Visible near infrared reflectance and chlorophyll content in Eucalyptus leaves International Journal of Remote Sensing 20 2741 2759 Datt B 1999b Remote sensing of water content in Eucalyptus leaves Australian Journal of Botany 47 909 923 Datt B 2000a Identification of green and dry vegetation components with a cross correlogram spectral matching technique International Journal of Remote Sensing 21 2133 2139 Datt B 2000b Recognition of Eucalyptus forest species using hyperspectral reflectance data In Proceedings Geoscience and Remote Sensing Symposium 2000 IGARSS 2000 IEEE 2000 International Volume 4 1405 1407 Daws M Firth R Lu P Poole P Pugh L amp Zimmwemann A 2008 Methodology Trial landform revegetation and monitoring
54. Harper amp Row New York Buschmann C amp Nagel E 1993 In vivo spectroscopy and internal optics of leaves as a basis for remote sensing of vegetation nternational Journal of Remote Sensing 14 4 711 722 Buschmann C Nagel E Szabo K amp Kocsanyi L 1994 Spectrometer for fast measurements of in vivo reflectance absorptance and fluorescence in the visible and near infrared Remote Sensing of Environment 48 18 24 Campbell JB 1996 An introduction to remote sensing 2 4 edn The Guilford Press New York Carlson HW Lass LW amp Callihan RH 1995 Detection of yellow hawkweed with high resolution multispectral digital imagery Weed Technology 9 477 483 Carter GA 1994 Ratios of leaf reflectances in narrow wavebands as indicators of plant stress International Journal of Remote Sensing 15 3 697 703 Chang J Clay SA Clay DE Aaron D Helder D amp Dalsted K 2005 Clouds influence precision and accuracy of ground based spectroradiometers Communications in Soil Science and Plant Analysis 36 1799 1807 Chewings V Pickup G Bastin GN amp Pearce G 2000 The potential for hyperspectral data mapping in Australian arid zone vegetation In 70th Australasian Remote Sensing and Photogrammetry Conference Adelaide 21 25 August 2000 851 862 Clark RN 1993 SPECtrum Processing Routines User s Manual Version 3 US Geological Survey Open File Report 93 595 Clark RN King TVV Ager C amp Swayze GA 1995 Initial ve
55. It is a vague parameter and gives no indication as to the responsivity of the system to light from different angles within the FOV Most data are collected with the sensor mounted vertically over the surface nadir view Robinson amp Biehl 1979 Silva 1978 Rollin et al 2000 Baumgardner et al 1985 Milton et al 1995 but some spectral libraries contain data measured in other configurations such as along the solar principal plane maximum anisotropy or at the anti solar peak or hotspot Milton et al 2009 Rollin et al 1997 Here we refer specifically to data collected at nadir 38 The area of ground from which spectra are recorded or ground field of view GFOV is controlled by the angular FOV a of the lens attached to the fibre optics and the height H that the instrument is held above the target The FOV must be appropriate to integrate and represent the geometric features of the target The FOV is an ellipse that is approximately circular at nadir The geometry can be considered as a cone intersecting a plane that is perpendicular to the cone To estimate the area or GFOV covered from a certain height r tan 0 2 x H where Spectrometer FOV r radius of the circular FOV with area A H height the spectrometer is held above the target surface a angular FOV for spectrometer A nr where A area sampled For example to establish the area A sampled with a Figure 14 Obtaining the GFOV 8 and H 1
56. July 1998 CD ROM Labsphere undated Important information Care and handling Guidelines Available online http www asdi com spectraloncare pdf Lamb DW Steyn Ross M Schaares P Hanna MM Silvester W amp Steyn Ross A 2002 Estimating leaf nitrogen concentration in ryegrass Lolium spp pasture using the chlorophyll red edge theoretical modelling and experimental observations International Journal of Remote Sensing 23 18 3619 3648 Lass LW Prather TS Glenn NF Weber KT Mundt JT amp Pettingill J 2005 A review of remote sensing of invasive weeds and example of the early detection of spotted knapweed Centaurea maculosa and babysbreath Gypsophila paniculate with a hyperspectral sensor Weed Science 53 242 251 Lewis MM 2000 Discrimination of arid vegetation composition with high spectral resolution imagery The Rangeland Journal 22 141 167 72 Lillesand TM amp Kiefer RW 1994 Remote sensing and image interpretation 3 edn Wiley Inc New York MacArthur A 2007a Field Guide for the ASD FieldSpec Pro White Reference Mode Natural Environment Research Council Field Spectroscopy Facility NERC FSF Version 2 November 2007 Available online http fsf nerc ac uk resources guides pdf guides asd guide v2 wr pdf MacArthur A 2007b Field Guide for the ASD FieldSpec Pro Raw DN Mode Natural Environment Research Council Field Spectroscopy Facility NERC FSF Version 2 November 2007 Available online ht
57. MW 1977 Leaf optical system modelled as a stochastic process Applied optics 16 1151 1157 Underwood E Ustin S amp DiPietro D 2003 Mapping non native plants using hyperspectral imagery Remote Sensing of the Environment 86 150 161 University of Queensland 2006 Coastal Remote Sensing Toolkit http www gpem uq edu au crssis tools rstoolkit index html Ustin SL DiPietro D Olmstead K Underwood E amp Scheer G 2002 Hyperspectral remote sensing for invasive species detection and mapping In International Geoscience and Remote Sensing Symposium 24th Canadian Symposium On Remote Sensing Toronto Canada 3 1658 1660 Waggitt PW amp McQuade CV 1994 Mine close out criteria Present guidelines and future trends in Australia In The AusIMM Annual Conference Darwin 5 9 August 1994 Warren P amp Hick P 1996 Tracking our performance Airborne remote sensing at Comalco s bauxite mine Weipa North Queensland In 3rd International and 21st Annual Minerals Council of Australia Environmental Workshop 2 61 78 Wessman CA Aber JD Peterson DL amp Mellio JM 1988 Foliar analysis using near infrared reflectance spectroscopy Canadian Journal for Research 18 6 11 Wilson WJ 1963 Estimation of foliage denseness and foliage angle by inclined point quadrats Australian Journal of Botany 11 95 105 77 Appendix A SSD s standards for collecting field reflectance spectra SSD s spectral measurement standards have been de
58. Milton et al 1995 define errors in field spectroscopy specifically referring to diffuse irradiation non simultaneous sampling of target and reference panel and time delay between successive samples Curtiss and Goetz 2001 and ASD 1999 and 2001 outline the importance of appropriate ancillary data with respect to sources of natural illumination atmospheric transmission presence of clouds and wind timing of data acquisition sampling strategy and viewing geometry These issues must be considered because they have a potential effect on the accuracy of spectral measurements Ultimately field spectral measurements are both accurate and precise with uncertainty estimates for a constant integration Accuracy refers to confidence in the correlation between measurements in one location and another or between a measurement and a recognised standard whereas precision implies careful measurement under controlled conditions that can be repeated with similar results and measured with confidence Deering 1989 Error is defined as the difference between the measured value and true value of the entity and can result from random or systematic sources Milton et al 1995 Reflectance spectra measured under field conditions are subject to several sources of error but well designed field spectrometers and careful experiment design can minimise some of these Milton amp Goetz 1997 The sources of information pertinent to the issues affecting spectral m
59. NIR region integration time ASD 2000 DC measurements are made by clicking on the DC pull down menu button This operation closes a shutter on the spectrometer entrance aperture and measures the response of the system to no external input ie due to internal electrical current This reading is then subtracted from all subsequent readings until another dark current measurement is made The DC measurement is taken whenever the user instructs the software to do so by either pressing the DC button on the toolbar when taking a WR measurement or during optimisation Not accounting for integration time whenever these measurements are made the DC is subtracted so that it is a negligible contributor assuming DC calibrations are performed on a fully warmed instrument ASD 1999 Although dark current systematic noise is sensitive to temperature SSD s minimum standard warm up time of 30 minutes accounts for internal thermal equilibrium The operator should be aware that the external ambient temperature fluctuations may also cause dark drift although it is less significant than during the start up period ASD 1999 External ambient temperature is recorded as metadata for each target reading see Section 4 7 5 4 Note that the ASD INI file should never be altered by the user as this is where Dark Current Correction measurement is stored Optimisation results in automatic settings of gains and offsets for the SWIR detectors an integration time value for t
60. Remember that it is important to always have the spectrometer running before the laptop computer is powered and the laptop computer should be switched off prior to shutting down the spectrometer A 2 5 Check that the date and time on the PC is correct Check that the date and time on the PC is correct Australian Central Standard Time These fields will be recorded in the spectral header A 2 6 Create a path to store the spectral data Through Windows Explorer create a path to store the spectral data The standard root directory is C Data 20 Calibration Files eg C Data 2007 Calibration Files The following folders should then be in place Hg Ar lamp Laboratory panel Uncleaned field panel Cleaned field panel 5 x 5 panel Circular panel Mylar panel Create a new path for new target materials such as soils A 2 7 Start High Contrast RS3 instrument software Start RS3 to obtain an interface like that illustrated below 90 RS 64664 Display Control GPS Help JIDD u JIJE oe Dark Current None Taken Current White Reference 7 None Taken Optimize Parms Vnir IT 17 ms Swirl G 500 0 2048 Swir2G 500 0 2048 500 750 1000 1250 1500 1750 2000 2250 Wavelength nm Latitude Longitude Elevation A 2 8 Spectral measurement setup saving data Go to Menu Control Spectrum Save or press Alt S RS 64664 Display Control GPS Help gt Take Dark Current measurement F3 29999
61. Revegetation of Nabarlek minesite Seasonal comparison of groundcover vegetation on the minesite and adjacent natural reference areas September 2003 amp May 2004 Internal Report 491 October Supervising Scientist Darwin Unpublished paper Beal D 1999 Wavelength and radiometric calibration methods In Technical guide 394 edn Appendix B Analytical Spectral Devices Boulder CO USA Belanger MJ Miller JR amp Boyer MG 1995 Comparative relationships between some red edge parameters and seasonal leaf chlorophyll concentrations Canadian Journal of Remote Sensing 21 1 16 21 66 Bell LC 1996 Rehabilitation of disturbed land in environmental management In The Australian minerals and energy industries Principles and practices ed Mulligan DR University of New South Wales Press Sydney 227 261 Birch CJ Hammer GL amp Rickert KG 1998 Improved methods for predicting individual leaf area and leaf senescence in maize Zea mays Australian Journal of Agricultural Research 49 249 262 Boochs F Kupper G Dockter K amp Kuhbauch W 1990 Shape of the red edge as vitality indicator of plants International Journal of Remote Sensing 11 10 1741 1753 Brennan K 2005 Quantitative descriptions of native plant communities with potential for use in revegetation at Ranger uranium mine Internal Report 502 August Supervising Scientist Darwin Unpublished paper Brooks RR 1972 Geobotany and geochemistry in mineral exploration
62. Scientist Darwin Unpublished paper Pfitzner K Bollh fer A amp Carr G 2006 A standard design for collecting vegetation reference spectra Implementation and implications for data sharing Journal of Spatial Sciences 52 2 79 92 Pfitzner K amp Carr G 2006 Design and implementation of vegetation reference spectra Implications for data sharing In Proceedings Workshop on hyperspectral remote sensing and field spectroscopy of agricultural crops and forest vegetation 10 February 2006 University of Southern Queensland Toowoomba Queensland 21 22 Pfitzner K Esparon A amp Bollh fer A 2008 SSD s Spectral Library Database Proceedings of the 14 Australasian Remote Sensing and Photogrammetry Conference Darwin 29 September 3 October 2008 Pfitzner K Bollh fer A Esparon A Bartolo R amp Staben G 2010 Standardised spectra 400 2500 nm and associated metadata an example from northern tropical Australia In Proceedings 2010 IEEE International Symposium on Geoscience and Remote Sensing July 25 30 2010 Honolulu Hawaii USA 2311 2314 viii 1 Introduction 1 1 Project definition This report presents the development and implementation of a robust method for collecting reflectance spectra of ground covers particularly with respect to vegetative ground covers Development and implementation of such a method ensures the measured spectral response is representative of the target given the immediate phenological co
63. The fore optic would then be adjusted until it was positioned in the centre of the case Figure 23a amp b The weight was drawn back so that it did not influence the spectral response and the lid of the case was opened for immediate WR sampling Figure 23a amp b Weighted plumb line ensures sampling is obtained from central position of white panel The operator waits for two screen refreshes before recording any data to allow the electronics of the spectrometer to adjust to the WR surface With the FOV centrally positioned over the WR panel the spectrometer is optimised including DC A solar radiance spectrum is measured and saved The WR is measured and saved immediately afterwards For all measurements the data is only saved once a stable signal is realised If errors such as a non stable signal or spectral steps are observed the data is eliminated and new data saved only when a stable signal is achieved The solar radiance spectrum is characterised by most points greater than 1 with the maximum radiance value reaching around 40 000 digital numbers An accurate WR spectrum is characterised by most points close to a value of 1 4 6 2 2 Target measurements Averaging multiple measurements of a target is good practice to compensate for heterogeneity which may be too subtle for the eye to note and also so that scans with spectral artefacts can be discarded Salisbury 1998 Milton et al 1995 49 Immediately after the WR reading the
64. VNIR SWIR1 Stray light smile significant noise in SWIR1 SWIR2 standards effect in the UV delete this sentence Strong water absorption bands are evident at 1400 and 1900 nm Figure 19 Standard Spectralon panel measurements are essential metadata for reflectance spectra Note that SSD s spectrometer has a 6 m fibre optic cable which results in signal loss at wavelengths greater than 2400 nm 4 6 Viewing and illumination geometry in the field The ideal procedure for spectral sampling with single FOV instruments is the acquisition of near simultaneous measurements of the WR and target spectra under exactly the same viewing geometries and under perfect illumination conditions In practice this theoretical procedure for spectral sampling is impossible Our method for temporal spectral sampling of vegetation plots necessitates the violation of the ideal spectral measurement method The factors that affect spectral signatures are considered and the method of accounting for and documenting these factors are described Recommendations for both the field design and accompanying metadata are made so that the accuracy of spectral measurements are maximised and any environmental variation can be accounted for 47 4 6 1 The FOV and Instantaneous Field of View IFOV The FOV must be appropriate to integrate and represent the geometric features of the target The measurement diameter at the surface is equal to the height of the spectrometer abo
65. Vines e Calopognium mucunoides Calopo vine e Centrosema molle Centro vine e Ipomoea spp Morning glory vine e Macroptilium atropurpureum Siratro vine and M lathyroides Phasey bean vine 18 Merremia aegyptia Hairy merremia vine and M dissecta White convolvulus creeper Passiflora foetida Wild passionfruit vine Acanthospermum hispidum Goat s head Starburr Crotalaria goreensis Gambia pea or Rattlepod Hibiscus sabdariffa Rosella Hyptis sauveolens Hyptis Horehound Stylosanthes hamata Carribean stylo Stylosanthes humilis Townsville stylo Stylosanthes scabra Shrubby stylo Grasses Andropogon gayanus Gamba grass Cenchrus cilaris Buffel grass Cenchrus echinatus Mossman River grass Chloris inflata Purple top chloris Chloris virgata Feathertop rhodes grass Cynodon dactylon Couch grass Hymenachne amplexicaulis Olive hymenachne Melinis repens Red Natal grass Pennisetum polystachion perennial Mission grass Pennisetum pedillatum annual Mission grass Themeda quadrivalvis Grader grass Urochloa humidicola Brachiara humidicola Tully grass Urochloa mutica NT W A Para grass Urochloa maxima NT W A Guinea grass Shrubs Aeschynomene americana 3 1 2 Species of the Ranger Project Area The Primary Environmental Objectives for rehabilitation of the RPA are to revegetate the disturbed sites of the RPA using local native species similar in density and abundance to those exi
66. al to become weeds were identified at the Darwin Berrimah Research Farm These grasses were represented as dense and homogenous patches and were located within a short walking distance making spectra of these species easily obtainable in a few hours of sampling The opportunity to obtain temporal readings of these following four pastoral species was taken Brachiaria humidicola Tully grass Digitaria eriantha Pangola grass Digitaria milanjiana Jarra grass and Digitaria swynnertonii Arnhem grass 3 2 Fortnightly measurements of ground cover The reflectance signatures of weedy and native ground covers are to be sampled from plots The plots aim to represent dense and homogenous covers of the plant species of interest A fortnightly sampling period is both logistically feasible and designed to capture distinct phenological change excluding the micro and macroscopic chemical and physical changes continually occurring within plants The Top End is suitable for high frequency spectral readings apart from the wet season Variations in atmospheric conditions eg sun angle humidity and haze from bushfires do have to be accurately measured and recorded with the spectral response The fortnightly measurements are then correlated with meteorological data measurement metadata and cover descriptions 3 3 Sites A challenge in the project design phase was to locate sites with homogenous dense cover that were unlikely to be disturbed from threat
67. ance reflected back to the spectrometer is defined directionally whereas irradiance received by the surface is hemispheric The incident diffuse irradiance depends on the height of the Sun and relative direct and scattered irradiance proportions that typically vary throughout the day and with conditions By dividing the target signal by the reference all multiplicative parameters are ratioed out however diffuse illumination and scattered light may significantly influence the total measured signal Curtiss amp Goetz 1995 Pinter et al 1990 Rollin et al 2000 Anderson et al 2003 As a result spectral campaigns are advised to be undertaken only when the weather is fine and stable Taylor 2004 although consistency is impossible with fortnightly temporal measurements The environmental factors affecting reflectance measurements include atmospheric attenuation and scattering from gases water vapour ozone carbon monoxide carbon dioxide methane nitrous oxide and oxygen Salisbury 1998 atmospheric particles wind and temperature Suggested approaches to reduce these effects on spectral measurements have been documented Salisbury 1998 Curtiss amp Goetz 2001 Where these factors are present during spectral measurement the condition must be documented in the spectral metadata so that any loss in signal identified in the post processing can be attributed to relevant factor and if appropriate the measurement discarded Without spectral metadata
68. and environmental changes from microscopic to community scales Issues such as timing and frequency of data collection spatial scale of the field measurement target viewing and illumination geometry and the collection and documentation of metadata must be considered Although there have been significant advances in the technical performance of field spectrometers the same cannot be said of the methodologies of field spectroscopy Milton et al 1995 and it is both technological and research limitations that have prevented these applications from becoming fully commercialised Phinn University of Queensland 2006 pers comm Standards are a must if spectral libraries are to be populated with useful data Gomez 2001 To advance the spectral knowledge base of the broader remote sensing community it is essential that a considered and documented method is undertaken and that appropriate metadata accompany spectral data SSD has developed and implemented a standardised method of data collection This ground based spectral information will provide a knowledge base for feasibility studies and be used to determine the most appropriate scaling up method to airborne or satellite platforms 17 3 Plant species and sites 3 1 Target species Priority species for sampling were determined with stakeholders These species were identified as important species for minesite revegetation success and included native framework species weeds of the wet dry tropi
69. and future remote sensing platforms and reanalysed for similarity and separability The following research questions and associated objectives address the project aim presented in Section 1 1 to build knowledge on the spectral response of vegetation species and background targets important for land condition assessment and monitoring 1 2 1 What are the temporal changes of spectral responses of ground cover species e Develop document and implement minimum spectral and metadata requirements for temporal spectral studies e Measure and detect the fortnightly spectral response of ground cover vegetative species representing a particular phenological condition at a given point in time e Analyse the spectral response of vegetation species over time and relate these responses to environmental conditions 1 2 2 Can ground cover species be distinguished using ground based reflectance spectra and if so what spectral resolution spectral selectivity or full width half maximum FWHM and spectral sampling interval is required e Develop document and implement minimum spectral and metadata requirements for temporal spectral studies e Measure and detect the fortnightly spectral response of ground cover vegetative species representing a particular phenological condition at a given point in time e Determine the range of spectral resolution at which a vegetation species is separable over time 1 2 3 At what phenological stage is maximum spe
70. and saved as a new sheet by date of measurement These reference spectra stored by date can be queried and correlated with reflectance measurements and used to compare and document the response over time Should degradation in spectral performance be identified from the laboratory measurements all subsequent field spectra can be flagged until such a time that the spectrometer is recalibrated through ASD 4 5 4 Standard panel measurements in the laboratory The major uncertainty with secondary standards such as a Spectralon reflectance standard is instability over time For this reason the reflectance of the standard panels are regularly measured in the laboratory and their reflectance compared over time This method is used as a warning system to determine if there is degradation in the RF The standard panels are returned yearly to ASD for remeasurement along with the spectrometer and fore optics and the panels replaced if degradation is realised that cannot be rectified by the panel cleaning process SSD has three Spectralon panels Two panels are 25 4 x 25 4 cm 10 x 10 in size and housed in wooden boxes when not in use One panel is clearly marked laboratory panel and this panel must remain in the laboratory The other is for use in the field environment A third smaller Spectralon panel 5 x 5 cm is for use under non standard conditions such as data collection from a helicopter The assumption that a calibrated panel near
71. and the reflectance properties of the reference surface are known Deering 1989 Robinson amp Biehl 1979 Milton 1987 Of these assumptions the one that is always violated in the field situation is the absence of sky light which results in measurements of BRF being made under an irradiance distribution that may be significantly different from the slender elongated cone referred to In general terms radiance is a directional quantity and reflectance is defined as the ratio of the reflected radiation to the total radiation falling upon the surface However field spectral measurements are integrated over time finite wavebands and solid angles Terms such as hemispherical conical reflectance factor Deering 1987 Milton 1987 Schaepman Strub et al 2005 hemispherical directional reflectance factor Abdou et al 2002 and directional anisotropic hemispherical reflectance factor Milton et al 1995 have been used to emphasise that the reflected radiance is measured over an angle that is not strictly directional and these terms are more appropriate for field measurements Because a single beam instrument violates the assumptions of BRF ie the conditions of illumination will not be exactly the same the numerous variables that factor into reference spectra must be carefully considered The objective is to obtain the measurements that are nearly independent of the incident irradiation and atmospheric conditions at the time of measurement Robin
72. and the spectra are saved as raw DN files To collect a HgAr spectrum the fibre optic tip is inserted into the lamp and optimised using a spectrum average of 30 dark current of 25 and white reference WR of 10 Refer to dark current measurements in Section 4 4 5 When collecting a Mylar spectrum the illumination lamps are allowed to warm up for 30 minutes prior to spectral sampling using the viewing and illumination geometries of the standard laboratory setup The laboratory standard panel is positioned with the focus point on the centre of the panel An 8 fore optic is used A WR spectrum is taken and saved The Mylar card is placed directly on the Spectralon panel which provides near perfect two way transmittance G Fager pers comm ASD Inc 2006 The transmission spectrum is measured and saved A spectrum average of 60 dark current of 25 and WR of 10 are used To confirm that each spectrograph registers specific wavelengths accurately the HgAr and Mylar spectra can be compared to the the Noise Equivalent Radiance NEdL values supplied by ASD using the bse ref and Imp ill radiance measurements 43 On request ASD supplies a calibration spreadsheet where the emission and transmission spectral values from the HgAr lamp and Mylar panel can be pasted against the responding wavelength A linear regression fit of the data is used to compare and document the response of the VNIR and SWIR regions over time The spreadsheet can then be updated
73. at if the east and west sky are obscured photographs of the north and south sky are taken instead labelled as nsl ssl etc zl zoom 1 is taken towards zenith angle with the camera held vertically with no zoom Figure A 14 and provides a record of the atmosphere around the Sun h1 height 1 is taken of the height of plant with measuring ruler in view if species is clumped Any additional images are named add1 2 3 etc Wherever possible the measuring pole is included in images 103 CS01 YYYY MM DD buggy jpg 1is ERGGGiEph five paces Or the site including buggy and fore optics location in relation to the site Figure A 4 Photograph buggy CS01 YYYY MM DD s1 jpg F Five paces vigi North west far ticket slightly to the left of near picket No zoom on camera Note measuring pole for reference is always in view arrow on pole is up bottom of red segment is height of 1 metre Figure A 5 Photograph s1 104 near picket Zoom on camera close up of site Figure A 6 Photograph s2 n CS01 YYYY MM DD obn1 jpg facing Figure A 7 Photograph obn1 105 5 CS01 YYYY MM DD mera at head height h zoom Usually take 3 images at this setting from one side of site to other Pole lies east west Figure A 9 Photograph n1 106 setting from one side of site to other Try to get pole at top or bottom of photo Figure A 10 Photograph_n2
74. at the instrument is in reflectance mode with 100 percent reflectance being obtained from the lab reference to compare the field reference Any deviation from previous measurements may indicate deterioration in the condition of the standard panel that may not yet be apparent by visual inspection A 2 15 Cleaning of field panels and spectral remeasurements if required If contamination has occurred the panel needs to be cleaned following recommendations by Labsphere undated and ASD 2000 If the material is lightly soiled it may be air brushed with a jet of clean dry air or nitrogen do not use Freon For heavier soil the material is cleaned by sanding under running water with a 220 240 grit waterproof emery cloth until the surface is totally hydrophobic water beads and runs off immediately Blow dry with clean air or nitrogen or allow the material to air dry Always wear clean gloves when handling the material A 2 16 Retake the spectral measurement of the field panel The standard field panel measurements are repeated in the laboratory if the field panel needs to be cleaned The spectra are remeasured and the cleaned panel should then be compared to the last reading of the cleaned panel to ensure consistency of the RF of the panel s A 2 17 Post processing The emission values from the HgAr spectrum and Mylar panel are pasted against the responding wavelength in the spreadsheet supplied by ASD A linear regression fit of the data
75. ation with this spectrum are flagged Highlighting a detector array issue or atmospheric influence in a white reference spectrum is crucial for data analysis and these anomalies would be very difficult to detect visually with data volumes described here Accurate metadata is required during the data analysis stage to ensure that environmental conditions such as solar azimuth are not influencing the spectral response particularly for temporal spectral measurements Photographic records help to interpret and determine the data quality for temporal data by supporting quantitative and qualitative measurements of the hemispheric component SSD Spectral Metadata Site ID 1 Target veg O Dee oO Soil g e Details Date dd mm yyyy Time 24 hour format 11 2006 12 32 Pfitzner Carr Mineral Other Data Collectors Site setup spec sty humi Family FABACEAE Plant Height m Homogeneity target Homeogeneity other Description Phenology Flush Mulitple jravel soil 2 95 Layers Eastern sky 5 Ground cover 96 Flush green regenerating 5 Fe gravel soil Environmental and Illumination Conditions 35 95 Ambient Temperature C Wind Speed km hr Relative Humidity 96 Wind Direction SE Cloud Cover 96 Sun Alt Degrees 79 Air Pressure hPa 03 Sun Azimuth Degrees 149 Cloud Type Atmospheric Conditions High thick sirrus Data Collection 2gon Scientific High cloud cover slow moving Hu
76. ature matching in geological applications and probably initiated much interest for further in situ spectral studies Unlike minerals all vegetation is composed of a limited set of spectrally active compounds chlorophyll accessory pigments liquid water starches proteins sugars and lignin The causes of absorption features in vegetation are the electron transitions of molecules and the bending and stretching of chemical bonds particularly O H C H and N H The spectral response of vegetation is influenced by the plant structure or architectural arrangement of the plant components and this response is scale dependent eg scales of leaf branch crown or canopy Micro and macro scaled changes are continually occurring within plants and the spectral responses of plants also vary over time These changes include short range diurnal variations eg water balance responses chlorophyll concentration short term seasonal changes phenological states and associated chemical changes and biophysical differences plant architecture density and homogeneity chemical compounds present in the vegetation at a particular phenological stage Early spectral research eg Tucker 1977 identified that asymptotic spectral reflectance or unchanging spectral reflectance occurs as vegetation density increases to the point where additional increases in leaf area index or biomass do not causes a change in the spectral reflectance Within species variability dependence
77. becomes valuable and can be the difference between usable and non usable spectral data 1_ss1 JPG Figure A 16 An example of the number and types of photographs collected for one site 110
78. ble data can be queried and analysed A conceptual user interface showing the metadata elements of the Spectral Library Database is presented in Figure 28 5 1 1 SSD s Spectral Database The database structure has been custom designed to maximise cross referencing between spectra photos and metadata SSD required a system to account for and link the spectral and metadata standards implemented A SQL server is used as a data warehouse to store all information The spectra and photos are stored as binary files within the database The metadata table contains information about the conditions at the time spectra and photos were taken Metadata include a unique code site and date date of spectral measurement atmospheric conditions smoke haze temperature humidity air pressure wind direction wind speed and description cloud level and cover probe height from ground plant height from ground level and ground description by cover and phenology Searches can be performed on the fields and individual records displayed Figure 29 illustrates an example spectrum metadata page with associated photographs Each photograph has a description of the photographer s location and camera settings and has nomenclature in the database For example the first photograph in Figure 43 is named BF04 2007 04 11 buggyl jpg and describes the site date and photo type and all photographs described as Buggy 1 are taken 5 paces from the plot and incl
79. cing spectrum c Drying spectrum with soil Digitaria milanjiana Jarra digit Digitaria milanjiana 2007 05 22 interspaces grass 2007 04 11 Digitaria milanjiana 2007 10 04 Figure 4 Illustration of a green b senescing and c drying spectra of Digitaria milanjiana Jarra digit grass taken in the months of April May and October respectively in the Top End of Australia 2 4 1 Healthy green vegetation The reflectance of light from a vegetated ground surface is determined by several factors such as leaf and canopy geometry morphology plant physiology plant chemistry soil type solar angle and climatic conditions Barret amp Curtis 1992 Vegetation reflectance is primarily influenced by the optical properties of plant materials including proteins lignin cellulose sugar starch which are composed largely of hydrogen carbon oxygen and nitrogen The absorption bands observed in vegetation arise from vibrations of C O O H C H and N H bonds as well as overtones and combinations of these vibrations It is well known that the visible spectrum 400 to 700 nm represents the photosynthetically active region of the electromagnetic spectrum In the visible wavelength regions leaf pigments control reflectance Campbell 1996 particularly chlorophyll a and b carotenoids and xanthophylls Tucker amp Garrett 1977 Consequently healthy green vegetation is characterised by low reflectance of blue and red light absorbed by chlorophyll for photos
80. component and it is therefore not surprising that the unique spectral identification of many materials has proven difficult due to numerous problems present in real world measurements Cochrane 2000 The surface texture of the material being measured affects the relative proportion of the various sources of illumination and background radiance is particularly important for vegetation applications A surface with a rough texture will tend to have a higher proportion of illumination from the diffuse and scattered surrounding sources relative to the direct solar illumination when compared with smooth surfaces Light returned from plants is a complex mixture of multiple reflected and or transmitted components Curtiss amp Ustin 1989 in ASD 1999 and the BRF of vegetation is generally assumed to be determined by the proportions of 56 different scene components sunlit leaves shaded leaves sunlit background and shaded background presented to a sensor Milton 2001 While dense and homogenous plots of vegetation cover were established the texture of plants may still contribute to sources of hemispheric illumination by adjacency effects Further as the plants senesce over the growing season plots may become heterogeneous Descriptions of cover combined with photographic recording therefore become essential metadata with vegetation applications Further averaged spectra are collected from a stationary position at three different locations within
81. cs and species that increase the risk of fire due to fuel load Species of concern at the Ranger Project Area RPA and or the rehabilitated Nabarlek minesite were also targeted for sampling and these included species that may potentially threaten the ecosystems of the country surrounding minesites 3 1 1 Weeds Declared Weeds of the Northern Territory which must be managed according to NT legislation were identified as important target species and include the following Herbs e Hyptis suaveolens Hyptis Grasses e Pennisetum polystachion Mission grass e Themeda quadrivalvis Grader grass Shrubs e Lantana camara Common lantana e Senna alata Candle bush e Senna obtusifolia Sicklepod e Senna occidentalis Coffee senna e Sida acuta Spinyhead sida e Sida cordifolia Flannel weed e Stachytarpheta spp Snake weeds While focus was not on aquatic forms of weeds it is envisaged that the project will expand to include aquatic plants in the future Declared aquatic species of weeds affecting the NT include Hymenachne amplexicaulis Olive hymenachne and Salvinia molesta Salvinia which are species particularly relevant to the rehabilitation of Ranger uranium mine that may contain water features post rehabilitation Further weeds of concern Further weeds of the Wet Dry tropics that are found in the Alligator Rivers Region and have the potential to impact minesites in the region include those outlined by Smith 1995 amp 2002
82. cs as prior information for Bayesian classification of yellow starthistle Weed Science 52 948 953 76 Sharp D amp Simon BK 2002 AusGrass grasses of Australia ABRS Identification Series Australian Biological Resources Study Canbera ABRS and Environmental Proection Agency Brisbane EPA Silva LF 1978 Radiation and instrumentation in remote sensing In Remote sensing the quantitative approach eds Swain et al McGraw amp Hill New York 121 135 Slaton MR Hunt ER amp Smith WE 2001 Estimating near infrared leaf reflectance from leaf structural characteristics American Journal of Botany 88 2 278 284 Smith NM 1995 Weeds of natural ecosystems A field guide to environmental weeds of the Northern Territory Australia Environment Centre NT Smith NM 2002 Weeds of the wet dry tropics of Australia A field guide Environment Centre NT Taylor F 2004 Field guide for the ASD FieldSpec Pro FR Raw DN Mode Version 1 1 Natural Environment Research Council NERC Field Spectroscopy Facility FSF Available online http fsf nerc ac uk resources guides pdf guides asd guide v1 dn pdf Ticehurst C Phinn S Held A amp Edmonds T 2003 Mapping an invasive weed pond apple in the wet tropics using multi and hyperspectral image data In Proceedings of Spatial Sciences 2003 Canberra Tucker CJ 1977 Asymptotic nature of grass canopy spectral reflectance Applied Optics 16 5 1151 1156 Tucker CJ amp Garrett
83. cted collated and analysed that we can pursue vegetation remote sensing at the minesite scale from the research realm into operational management Figure 2 Multitemporal hyperspectral data covering the Nabarlek minesite From left to right Airborne Multispectral Scanner June 2004 4 5 m pixels 96 bands HyMap September 2002 5 m pixels 126 bands and CASI July 2002 1 m pixels 16 bands False colour images IR R G 1 Quickbird data May 2004 60 cm CASI data July 2002 1 m 16 bands AMS then DeBeers June 2004 4 bands 4 m 96 bands Figure 3 Subset of the Nabarlek minesite covering the rehabilitated plant run off pond area Results are sensor specific and spatially and temporally dependent 2 2 Reflectance spectrometry basic terminology Field ground in situ or handheld spectrometry spectroscopy and reflectance spectrometry are interchangeable terms used to describe measurements of spectral properties usually made under solar illumination in the natural environment Spectral measurements in the laboratory use an artificial light source such as halogen lamps Here reflective optical radiation is defined as propagating electromagnetic energy with characteristic wavelengths between 400 nm and 2500 nm including the visible portion of the spectrum and the infrared or IR When optical radiation interacts with a surface a portion of that radiation 1s either absorbed in the material below the surface or is transmitt
84. ctral separability and is there a phenological stage when spectra of different species cannot be distinguished e Measure and detect the fortnightly spectral response of ground cover vegetative species representing a particular phenological condition at a given point in time e Analyse data to determine the phenological stage s that maximum spectral separability is found between species and if there is a stage or stages where species are spectrally confounding 1 2 4 What are the implications for use of remote sensing imaging throughout the year e Develop a database of land cover end members that can be used to make recommendations on the most appropriate monitoring strategy for minesite rehabilitation assessment e Analyse data to determine the change in spectral response of vegetation species over time between species and at different sensor wavelengths to make recommendations on timing of data capture for vegetation assessment To answer these research questions the research design needs to ensure that the spectral response is not confounded by extraneous factors such as localised changes in atmospheric conditions 1 3 Background concepts A spectral database of land cover end members pertinent to remote sensing for minesite rehabilitation assessment is being developed End members include vegetative species introduced weedy and native vegetation geological materials including minerals and soils aquatic components and infrastru
85. cture mine related features including infrastructure Measurements are made with a portable FieldSpec Pro FR spectrometer Analytical Spectral Devices Inc across 350 2500 nm at full width half maximum FWHM resolution of 3 nm for the region 350 to 1000 nm and 10 nm for the region 1000 to 2500 nm The principal objective is to create a database of temporal spectral responses that can be assessed to make recommendations on the most appropriate remotely sensed monitoring method for land cover and condition assessment with particular application to minesite rehabilitation Remote sensing technologies offer advantages over traditional field based monitoring methods However compared with the use of remotely sensed data for applications in the natural landscape there are additional challenges for the disturbed mine environment These include the need to identify and discriminate subtle variation in land cover over short distances and variable frequencies Suitable remotely sensed data currently available for rehabilitation assessment include airborne hyperspectral and very high spatial resolution VHR satellite data A cost benefit analysis of these approaches compared with ground measurements is difficult because results are dependent on both the sensor specifications and the localised environmental conditions such as seasonality or occurrence of disturbances like bushfires The accuracy of radiance irradiance reflectance and transmittance spectra is
86. curate and precise representations of the target condition There are many factors that can affect the spectral response obtained Some of these factors are dependent on the experimental design The environmental conditions as well as the response of the spectrometer and reference panel used may also influence the spectral measurements However there are no national or international standards for the collection of in situ spectral data While many field spectral campaigns may be undertaken the effort expended in ground based spectral collection is often only applicable to a single point and time This is because few samples are acquired accurate metadata are not recorded the data are not stored in a manner that is easily retrievable the method of data collection is not described and the data represent targets whose spectral response varies spatially and temporally In order to gain quality reference spectra of objects of interest it is vital that careful consideration be given to the way in which spectral data are obtained The sample size and times series of spectra must be appropriate Importantly metadata describing what was measured how the measurement was taken and what the conditions were like during spectral measurement must accompany the spectral data Factors that affect spectral measurements including environmental factors must be documented so that any external spectral influences can be accounted for Photographic records can be a
87. d Beal 1999 ASD Inc uses Mercury Argon HgAr source lamps to measure and cross calibrate the monochromator emission values in the VNIR region Figure 16 and well defined absorption features of a material such as Mylar panels for the SWIR region Figure 17 Wavelength calibrations are checked using a 1 nm range when compared with published NIST wavelength values G Fager 2006 pers comm Beal 1999 The NIST values need to be adjusted based on the spectral resolution of the instrument and ASD Inc supply a spreadsheet so that calculations of the wavelengths using an HgAr lamp and Mylar panel can be made and monitored G Fager 2006 pers comm Note that SSD returns the spectrometer and fore optics for calibration yearly Mercury Argon Emission Spectrum Mylar Transmission Unit FSFR 650 ne peee eaer i i i ransmission L3 TERT HIN pee WM wa Figure 16 Mercury Argon Emission Spectrum Figure 17 Mylar transmission Spectrum Source ASD 2000 71 Source ASD 2000 72 SSD also monitors the calibration performance of the spectrometer regularly under the standard laboratory setup Ideally measurements are made at fortnightly intervals Suggested instructions on collecting HgAr and Mylar spectra were provided by J Brady pers comm ASD Inc 2005 and these have been adopted The HgAr lamp is warmed up for 10 minutes after the standard 90 minute spectrometer warm up time is reached No fore optic is used
88. d Field panel M eg g C Data YYYY Laboratory measurements WUncleaned Field panel date YYMMDD 000 95 5 WR small field panel WR of Spectralon laboratory panel path C Data Laboratory measurements 5 x 5 panel X eg g C Data YYYY Laboratory measurements 5x5 panel date YYMMDD 000 6 Does the field panel need Y N cleaning 7 Post cleaning repeat readings NA WR of Spectralon laboratory panel path C Data Laboratory measurements Field panel Nic eg g C Data YYYY Laboratory measurements Cleaned Field panel date YYMMDD 000 96 A 3 Care and transport of spectrometer The spectrometer is sensitive to electrical current and therefore the spectrometer should always be running prior to switching the laptop on To ensure this the spectrometer should be running before the parallel cable of the laptop is connected to the spectrometer The laptop should be turned off or the parallel cable disconnected prior to switching off the spectrometer The spectrometer is sensitive to high ambient temperatures The spectrometer should not be left in direct sunlight and should always be shaded by the shaded buggy see Figure A 1 or carried in the ergonomic Propack Considering that high ambient temperatures can cause dark drift Section 4 2 spectral measurements taken under tropical conditions require that optimisation is performed prior to measurements of each new target of interest The shaded buggy is only suitable for field work ove
89. d cover need to be quantified Details on how to describe clouds are provided in Appendix A 5 Hean nunber of cloudy days Jan Feb Har Apr May Jun Jul fug Sep Oct Nov Dec Honth E 014015 Mean number of cloudy days Australian Government Bureau of Meteorology Mean m dad days 240 216 197 113 63 37 34 26 32 56 115 201 Figure 25 Mean number of cloudy days Darwin Airport Source BOM http www bom gov au climate averages tables cw 014015 shtml 4 6 5 2 Smoke and haze descriptions Smoke and haze are recorded as either present or not present and if present altitude descriptions are described similar to cloud altitude levels of high mid low BOM use two laser devices situated at Darwin Airport to record the level of atmospheric particulate matter Smoke or haze is measured in units of distance visibility km Visibility of 30 km indicates very clear conditions while this reduces to 5 km in the presence of smoke or haze Extremely smoky conditions may see visibility reduced to 200 m 55 Since the sampling areas are relatively close to Darwin Airport these readings can be used to characterise spectral sampling conditions Archival figures are available on the Internet at http australianweathernews com archives capcity WR and solar radiance spectra are also used to indicate the effects of skylight as scattering by aerosols will increase skylight and the higher the concentration the greater the sky
90. d et al 2003 measured 80 individual ground based reflectance spectra in conjunction with AVIRIS 224 bands 4 m pixels data to detect invasive species of iceplant with a presence absence accuracy of 97 Lass et al 2005 classified 57 of known spotted knapweed Centaurea maculosa and 97 of known babysbreath Gypsophila paniculate using hyperspectral data 48 bands 2 m pixels DiPietro 2002 mapped giant reed Arundo donax using AVIRIS data with 71 95 accuracy with results dependant on the mapping method used Hunt et al 2004 and Parker Williams and Hunt 2004 mapped leafy spurge using AVIRIS data with an accuracy of 95 Hunt et al 2004 found the distribution and abundance of leafy spurge can be determined with hyperspectral AVIRIS data but not with multispectral data Ustin et al 2002 reported invasive species mapping using AVIRIS data Goel et al 2003 used CASI 72 bands 2 m pixels to evaluate detection of weed infestations with 91 accuracy for detected weeds against weed free crops Ticehurst et al 2003 collected in situ spectra of the weed Pond apple Annona glabbra and other vegetation to assess the potential of different remote sensing technologies to discriminate the weed and found that Landsat TM Hyperion and HyMap data identified pond apple stands but also erroneously included non pond apple vegetation Emery et al 1998 measured field reflectance data of heathland canopies over a range of ages over the course of a gro
91. da holathera Heteropogon triticeus Sehima nervosum Dicanthium fecundum Alloteropsis semialata Thaumastochloa major and Ectrosia agrostoides Brennan 2005 undertook a quantitative description of native plant communities for potential use in revegetation at Ranger uranium mine His research was undertaken on natural plant communities on hills both schists and sandstones in the region and those on the Koolpinyah surface on the Ranger lease He measured herbaceous plants quantitatively at 13 sites For the herbaceous component he found that there were natural plant communities on hills in the region that were very similar floristically to the vegetation in eucalypt woodlands on the Ranger lease A summary of the herbaceous flora findings by Brennan 2005 include e Ranger sites and all sandstone hills were dominated by Sorghum brachypodum Sorghum accounted for almost 60 of the total seasonal production of herbaceous biomass at the Ranger sites The species was absent on schist hills e A further 20 of the total annual productivity at the Ranger sites was added by other grasses 22 species Of these Heteropogon triticeus and Alloteropsis semialata were high biomass species but the short grasses Schizachyrium fragile Eriachne agrostidea Eriachne ciliata Thaumastochloa major Digitaria gibbosa Aristida holathera Brachiaria holosericea Mnesithea formosa Sporobolus pulchellus Pseudopogonatherum irritans and Yakirra nulla each had
92. data although not essential The heights of clouds are defined as high middle and low level clouds If possible the heights of clouds are further defined by the type of cloud but this is not considered to be essential information High level clouds are composed solely of ice crystals and include cirrus cirrocumulus and cirrostratus types Medium clouds are usually composed of water droplets or a mixture of water droplets and ice crystals and include altocumulus altostratus and nimbostratus types Low clouds are usually composed of water droplets though cumulonimbus clouds include ice crystals and include stratocumulus stratus cumulus and cumulonimbus 101 High Level Clouds above 6 km usually composed solely of ice crystals no precipitation Cirrus white tufts or filaments Cirrocumulus small rippled elements Cirrostratus transparent sheet or veil halo phenomena Middle Level Clouds 2 5 to 6 km composed of water droplets or a mixture of water droplets and ice crystals Altocumulus layered cloud rippled Altostratus grey sheet thinner layer Nimbostratus thicker darker and elements generally white with some allows sun to appear as through lower based sheet Precipitation shading Precipitation May produce ground glass Precipitation rain or heavier intensity rain or snow light showers snow Low Level Clouds below 2 5 km Stratocumulus layered cloud series Stratus layer or mass grey uniform Cumulus
93. e 115 124 Hick P Smith R Ong C amp Honey F 1994 Vegetation assessment and monitoring using airborne digital multispectral videography Implications for geobotanical prospecting and quantitative pre and post mining measurement In 9th Annual environmental workshop 1994 Karratha WA 9 14 October 1994 Australian Mining Industry Council Canberra 178 189 Hill GJE amp Phinn SR 1993 Revegetated sand mining areas swamp wallabies and remote sensing North Stradbroke Island Queensland Australian Geographical Studies 31 1 3 13 Horler NNH Dockray M amp Barber J 1983 The red edge of plant leaf reflectance International Journal of Remote Sensing 4 273 288 70 Horn BKP amp Sjoberg RW 1978 Calculating the Reflectance Map Massachusetts Institute of Technology http people csail mit edu bkph AIM AIM 498 OPT pdf Web site accessed May 18 2005 Howard JA Watson RD amp Hessin TD 1971 Spectral reflectance properties of Pinus ponderosa in relation to the copper content of the soil Machite mine Colorado In Proceedings of the 7th Symposium on Remote Sensing of the Environment Ann Arbor Michigan 285 296 Hueni A 2006 Field spectroradiometer data Acquisition organisation processing and analysis on the example of New Zealand Native Plants Unpublished Masters of Philosophy in Environmental Science Thesis Massey University New Zealand Hueni A amp Tuohy M 2006 Spectroradiometer data structuring pre p
94. e electrical signal dark current and then interpolated to a 1 nm sampling interval over the wavelength range Fyfe 2004 The FieldSpec Pro FR collects light passively by means of a fibre optic cable The standard fibre optic cable length of the FieldSpec Pro FR is 1m Longer cables result in signal attenuation particularly beyond 2000 nm D Hatchell ASD Inc pers comm 2004 Figure 12 illustrates the loss in signal short of 500 nm and particularly at wavelengths greater than 2200 34 1 5M gt 10M Performance Relative to a Standard FieldSpec 500 1000 1500 2000 2500 Wavelength nm Figure 12 Attenuation versus length of permanent FR fibre Source http support asdi com Document Viewer aspx id 56 A trade off between the future need for ease of measurement of shrubs and trees against a drop in the NEdL in the far infrared region was made so that the SSD FieldSpec Pro FR is characterised by a 5 m fibre optic cable Table 4 The fibre optic conical view subtends to a full angle of 25 and fore optics may be attached to the cable to limit the lens angle 1 and 8 Table 4 FieldSpec Pro FR product specifications Spectral range Spectral resolution Sampling interval Scanning time Detectors Transition splice position Input Noise Equivalent Radiance NEdL Weight Calibration Fibre optic cable 350 2500 nm 3 nm 700 nm 10 nm 1400 2100 nm 1 4 nm 350 1050 nm 2 nm 1000 2500 nm 1
95. easurements are fragmented Further there are no such documents or manuals that synthesise all the factors influencing spectral measurements and the methods used to minimise and account for extraneous factors in spectral measurement Issues to be considered when designing a spectral library database have been summarised Pfitzner et al 2005 and are conceptualised in Figure 11 The factors that affect standardised measurements can be summarised to include environmental eg wind speed and direction cloud cover and type temperature humidity aerosols viewing geometry fore optic degree and the field of view or FOV and instantaneous field of view or IFOV fore optic height above target and ground illumination geometry date time position and sun altitude azimuth and orientation smoke and haze properties of the target physical and textural chemical and structural make up and BRDF properties integration and measurement timing calibration of the instrument and reference standard and general experimental design 32 ISSUES TO CONSIDER WHEN OBTAINING REFERENCE SPECTRA EXPERIMENTAL DESIGN Timing of data collection CALIBRATION Method of data collection including geometry and scale Spectralon panel Number of samples consideration of spatial amp temporal variability Spectrometer ILLUMINATION AND VIEWING GEOMETRY INSTRUMENT SETTINGS Date IFOVIFOV Dark current integration Time Fore optic degree Position Fore opt
96. ed through the bulk of the material through another surface into another medium The reminder of the radiation 1s said to be reflectance from the surface and in general terms the ratio of the reflected radiation to the total radiation falling upon the surface is defined as reflectance Baumgardner et al 1985 Modern field spectrometers such as the FieldSpec Pro FR manufactured by Analytical Spectral Devices ASD Inc were developed in the late 1980s to mid 1990s and are capable of measuring spectra with high precision and accuracy and are portable and easy to use Terms such as photometers and spectroradiometers refer to instruments that collect data over only a range of wavelengths and multiband radiometers collect data in a few broad wavebands only Milton et al 1995 Spectral signatures represent the relationships between electromagnetic radiation EMR and the physical and chemical properties of the object of interest The signature is a result of radiance irradiance reflectance or transmission of light from a remote target by translating light energy into electrical current Fyfe 2004 The interactions of photons with the surface of a target occur on a molecular scale Photons may be transmitted reflected emitted or absorbed Molecules have discrete energy levels and can only absorb specific amounts of energy and the diagnostic regions of the reflectance spectrum are usually defined by absorption features at specific wavelengths Terms
97. el Figure 24 Absorption minima and maxima at the atmospheric water absorption regions combined with metadata on meteorological conditions Figure 25 Mean number of cloudy days Darwin Airport Figure 26 Effects of wind on mobile targets Figure 27 Fortnightly temporal ground cover spectra accompanied by selected metadata for Stylosanthes humilis over 4 months Figure 28 SSD s Spectral Database metadata records Figure 29 An example metadata page with associated photographs Figure 30 The spectral data associated with photographs in Figure 29 Figure A 1 The spectrometer buggy setup Figure A 2 Scaled set up standard photograph Figure A 3 Typical examples of the 10 Main Cloud Types Figure A 5 Photograph s1 Figure A 6 Photograph s2 Figure A 7 Photograph obn1 Figure A 8 Photograph obs1 Figure A 9 Photograph n1 Figure A 10 Photograph n2 Figure A 11 Photograph es1 Figure A 12 Photograph es2 Figure A 13 Photograph ws1 Figure A 14 Photograph ws2 Figure A 15 Photograph z1 Figure A 16 An example of the number and types of photographs collected for one site 33 35 38 39 42 43 43 47 47 48 49 55 55 56 60 62 63 64 9f 99 102 104 105 105 106 106 107 107 108 108 109 109 110 Executive summary The collection of ground based radiance irradiance and reflectance spectra is a critical and common exercise for many environmental applications The resulting measurements need to be ac
98. equirement Complicating the transfer of ground based spectra from one researcher to another are both the variance in techniques used to collect spectral information and the localised environmental conditions The many different techniques used in obtaining field spectra have resulted in problems of data comparability between studies which compromise the long term value of such data Milton 1987 Further the lack of appropriate ancillary data sets often makes previously collected data unusable for new applications Curtiss amp Goetz 2001 ASD 2001 Limited access to spectrometers and a narrow time frame for spectral data collection 1e the need to coincide spectral measurements with a remotely sensed overpass may be reasons why a considered and consistent method of spectral data and metadata collection have not been adopted by spectral scientists Further there has not generally been a practice of data sharing and one of the reasons for this is the missing standardisation of the sampling process Hueni amp Kneub hler 2007 While many types of spectral measurements will prove useful for a given application there is a need for data which may be compared from site to site independent of atmospheric conditions Robinson amp Biehl 1979 The field campaign must be calibrated with introduced uncertainty and validated reproducible for both the measurement equipment used and the ground target to compensate for spatial and temporal variability
99. er most of the 1300 2500 nm region all dry plant materials exhibit similar absorption features caused by lignin and cellulose Datt 2000a Murphy and Wadge 1994 describe a field spectrum of dead grass with no absorption in the blue or red and two prominent absorption features between 2050 and 2140 nm due to lignin absorption and absorptions at 2260 nm and 2330 nm due to both lignin and cellulose absorption 2 5 Remotely sensed data for vegetation assessment and monitoring with particular application to the mine environment The ability to map vegetation cover and discriminate species remotely offers significant advantages over traditional ground based field measurements Underwood et al 2003 McGowen et al 2001 However monitoring vegetation using remote sensing is challenging 11 because of the variations of vegetation reflectance with sun zenith angle view zenith angle terrain slope Dymond et al 2001 contribution from atmospheric noise humidity shadow and soil Price 1994 orientation of leaves age differences of plants and variation in leaf area index Joshi et al 2004 Isolated examples of the application of aerial photographs videography and broadband satellite data for mining applications exist Game et al 1982 Phinn et al 1991 Evans amp Williams 1995 Hill amp Phinn 1993 Hick et al 1994 Rathmore amp Wright 1993 McCall et al 1995 Hick 1999 Warren amp Hick 1996 Mueller et al 1997 Schmidt amp Glaesser 1998 Da
100. er station Clouds smoke and haze are given a semi qualitative description and further documented by digital photographs Along with the quantitative and semi qualitative environmental metadata and photographic recordings the WR readings are useful in combination as sources of information to check the quality of data measured Figure 24 shows two different in situ WR readings Figure 24a shows significant water absorption affecting the 1400 and 1900 nm regions as well as a low S N ratio in the SWIR compared to Figure 24b that shows much less atmospheric water absorption 54 o o 8 8 E 2 E T ec a 500 1000 1500 2000 250 500 1000 1500 2000 250 Wavelength nm Wavelength nm a Significant atmospheric water absorption b Atmospheric water absorption 1400 and 1400 and 1900 nm and effects in the SWIR 1900 nm Figure 24 Absorption minima and maxima at the atmospheric water absorption regions combined with metadata on meteorological conditions are useful documentation on illumination conditions at the time of sample measurement 4 6 5 1 Cloud descriptions Figure 25 shows the mean number of cloudy days for Darwin Airport averaged over a 54 year period and shows there are fewer cloudy days in the sampling period of low solar azimuth angles Table 6 between April and October While sampling is not undertaken on a cloudy day spectral sampling is undertaken on days when periods of cloud cover occur and the cloud type an
101. erational requirements needed for the SSD Spectral Database Project The SSD Spectral Measurement Database has been developed to take into account spectrometer metadata and performance data of the standard Spectralon panels including temporal laboratory Hg Ar Mylar panel and Spectralon spectra and associated metadata images of the target at nadir scaled set up horizon photographs and hemispherical photographs subject information classification condition appearance physical state subject background scene background information similar to subject data measurement information instrument mode date local time data collector s fore optics number of integrations reference material height of measurement from target and ground viewing and illumination geometry environmental conditions general site description specific site location geophysical location sun azimuth and altitude ambient temperature relative humidity wind speed and direction weather instrument used and sky conditions and of course reflectance spectrum and averaged reflectance data This information is stored and available for data retrieval through the SSD Spectral Database The standards are transferable to other researchers and applications The only difference required may be that of the fore optic height and target field of view It is envisaged that this report provides not only a reference manual for spectral measurements but will also play a key role
102. g of Environment 113 92 109 http www sciencedirect com science article B6V6V APSJT3R 1 2 1f6cf30222e1e5b467c6161e6eb6817d Minerals Council of Australia 2002 Sustainable Development Report 2002 Minerals Council of Australia ACT Available online http www minerals org au downloads pdf sd report02 pdf Moody ME amp Mack RN 1988 Controlling the spread of plant invasions the importance of nascent foci Journal of Applied Ecology 25 1009 1021 Mueller A Strobl P Lehmann F amp Reinhaeckel G 1997 Case studies of airborne remote sensing for the assessment of mining impacts In Third International Airborne Remote Sensing Conference and Exhibition Copenhagen Denmark Volume I 257 264 Munden R Curran PJ amp Catt JA 1994 The relationship between red edge and chlorophyll concentration in Broadbalk winter wheat experiment at Rothamsted International Journal of Remote Sensing 15 7 1459 1470 Murphy RJ amp Wadge G 1994 The effects of vegetation on the ability to map soils using imaging spectrometer data International Journal of Remote Sensing 15 1 63 86 Nicodemus FE 1982 Reflectance nomenclature and directional reflectance emissivity Applied Optics 9 6 1474 1475 Nicodemus F Richmond J Hsia J Ginsberg I amp Limperis T 1977 Geometrical considerations and nomenclature for reflectance NBS Monograph 160 US Department of Commerce Parker Williams A amp Hunt ER 2004 Accuracy assessment of leafy sp
103. g that it is the next spectrum measurement that is saved 4 6 4 Direct solar illumination sun angle and position Direct solar illumination is assumed to be the dominant illumination component when sampling is undertaken at high solar angles under ideal atmospheric conditions low cloud cover humidity smoke and haze Atmospheric conditions for spectral sampling are quite predictable in the tropics but rarely are optimal conditions realised Table 6 shows the solar azimuth and altitude for a 12 month period calculated for Darwin city and shows that the highest solar angle occurs during the wet season between October and April when cloud cover and humidity are typically at their peak In the dry season May to September combined with a lower solar angle smoke and haze from bushfires are common Table 6 Example sun azimuth and altitude measurements for Darwin Lat 12 27 00 Long 130 50 00 for the 1 of the month over a one year period Month Dd mm yyyy hour min sec Azimuth Altitude January 01 01 2007 12 00 00 133 27 43 7T4 05 56 February 01 02 2007 12 00 00 110 03 39 74 42 02 March 01 03 2007 12 00 00 73 35 49 74 42 45 April 01 04 2007 12 00 00 37 40 24 68 59 34 May 01 05 2007 12 00 00 21 57 07 60 33 11 June 01 06 2007 12 00 00 17 37 38 53 53 31 July 01 07 2007 12 00 00 19 09 48 52 20 38 August 01 08 2007 12 00 00 23 25 18 56 44 20 September 01 09 2007 12 00 00 29 42 46 66
104. getation species and senescence stress mapping in the San Luis Valley Colorado using imaging spectrometer data In Summitville Forum Proceedings 1995 eds Posey HH Pendleton JA amp Van Zyl D Colorado Geological Survey Special Publication 38 64 69 Available online http speclab cr usgs gov PAPERS veg1 vegispc2 html Clark RN King TVV Klejwa M Swayze GA amp Vergo N 1990 High spectral resolution reflectance spectroscopy of minerals Journal of Geophysical Research 95 B8 12653 12680 Clark RN Swayze GA Gallagher AJ King TVV amp Calvin WM 1993 The US Geological Survey Digital Spectral Library Version 1 0 2 to 3 0 microns US Geological Survey Open File Report 93 592 67 Cochrane MA 2000 Using vegetation reflectance variability for species level classification of hyperspectral data International Journal of Remote Sensing 21 10 2075 2087 Collins W 1978 Remote sensing of crop type and maturity Photogrammetric Engineering and Remote Sensing 44 43 55 Collins W Chang S Raines G Canney F amp Ashley R 1983 Airborne biogeophysical mapping of hidden mineral deposits Economic Geology 78 737 749 Curran PJ amp Milton EJ 1983 The relationships between the chlorophyll concentration LAI and reflectance of a simple vegetation canopy International Journal of Remote Sensing 4 2 247 255 Curran PJ 1989 Remote sensing of foliar chemistry Remote Sensing of Environment 30 271 278 Curtiss B amp
105. ginalis v Calopognium mucunoides vine v v Centrosema molle vine v Crotalaria goreensis v v Euphorbia heterophylla v Euphorbia hirta v Hibiscus sabdariffa v Hyptis suaveolens v v v Ipomoea graminea v Ipomoea spp vine v v v Macroptilium atropurpureum v v v Weedy Macroptilium lathyroides v v v herbs and vines Passiflora foetida vine v v v Senna alata v v Senna obtusifolia v v Senna occidentalis v v Senna alata v Sida acuta v v v Sida cordifolia v Sida rhombifolia v Stachytarpheta spp v v Siylosanthes hamata v v v Siylosanthes humilis v v Siylosanthes scabra v v Siylosanthes viscose v Tridax procumbens v identified by Brennan 2005 Hollingsworth and Meek 2003 EWLS 2005 and Daws et al 2008 identified by Bayliss 2004a amp b Weedy ground covers with an emphasis on grasses were identified by stakeholders as the priority species for spectral measurement The spectral identification and discrimination of these species is important to minesite applications because declared weeds in any location must be managed weedy covers do not feature in the surrounding ecosystem of the RPA and the expanse of weedy covers at Nabarlek has hampered revegetation attempts and increased the threat of fire affecting framework species The spectral identification and discrimination of weedy covers maybe relevant to other landscape applications Native species that were co located with dense and homogenous patches were ta
106. gs by McGowen et al 2001 were that phenological changes in spectral response were found for serrated tussock during flowering and in mid spring the reflectance was similar to that of many native pasture species Cochrane 2000 found potential for separation of eleven forest species based on foliar reflectance spectra Miller et al 1991 measured leaves from ten tree species at weekly samples over a period of 150 days and found short term variations in spectral response attributed to rainfall and temperature events Implementing a ground based reflectance feasibility study prior to initiating a remote sensing based mapping project may provide a knowledge base on the likelihood of adequate detection of target species This aspect of spectral research includes the identification of key stages of growth flowers green up senescence plant pubescence architecture shadowing growth forms to determine if and when species can best be discriminated from other vegetation over time With such a spectral knowledge base it may become cost effective to commission airborne overpasses at times of greatest likeliness of species separability 2 6 1 Spectral libraries and in situ spectral measurements Spectral libraries particularly for geological materials are available in the public domain for example Grove et al 1992 Clark et al 1993 and Satterwhite and Henley 1990 and there are proposals for a Web based Spectral Library Information System WSLIS
107. he VNIR detector and the dark current measurement Optimisation values depend on the response to light in a particular spectral region and a well optimised instrument will display between 20 and 35 thousand digital numbers ASD 1999 A Spectralon amp panel is used when optimising and when taking a white reference WR measurement Optimisation 1s required before any data is collected and the instrument must be re optimised after any change in temperature or lighting conditions SSD approach e SSD s standards when collecting spectra in the field are to optimise the spectrometer and therefore obtain a DC prior to the WR measurement for every new target measurement in order to adjust the sensitivity of the instrument s detectors according to the specific illumination conditions at the time of measurement In the laboratory and in the field a WR spectrum is taken for every new sample In the field a WR spectrum is also taken and saved whenever irradiance conditions change to ensure that changing levels of down welling irradiance do not cause the detectors to saturate If there is a change in atmospheric conditions such as cloud movement between optimisation and spectral measurement optimisation WR and spectral readings are redone The optimisation and WR function in the ASD software gets new reflectance values for the white reference panel and saving these spectra allows any change in irradiance to be identified 45 Noise can be
108. he controlling software needs to be told where to save data for the next site Adjust the working folder 86 A 1 26 Spectral measurement setup saving data Go to Menu Control Spectrum Save or press Alt S RS 64664 Display Controt GPS Help h Take Dark Current measurement F3 2939 DIC Initialize Radiometric measurement F9 Dark Current Take White Reference measurement F4 f Adjust Configuration None Taken 0 Abort Spectrum Collection CtrhA Parabolic Correction measurement CtrhP i UTTETSESUTESN White Reference M ERE ViewSpec Pro Optimize instrument settings Ctro None Taken Spectrum Avg Y J 0 25 m wf Spectrum Save Y icm NN T lab 003 Optimize Parms Vnir IT 17 ms Swirl G 500 0 2048 a fel Swir2G 500 0 2048 H H J T 500 750 1000 1250 1500 1750 2000 2250 2500 Wavelength nm amp Latitude Longitude Elevation Tab down to the Path Name C and ensure that the correct working folder is marked as the target folder for this plot by clicking on the box with three dots at the end of the Path Name box and navigate to desired folder Assuming you are measuring more than one plot at each site the change will only be to the plot number For example CS02 000 to indicate CSIRO plot 2 Click OK or press ALT O letter o You are now ready to repeat the spectral measurements at the next site including optimisati
109. ic height above target calles dais Becca ae Sun altitude Fore optic height above ground Sun azimuth ENVIRONMENTAL CONDITIONS Wind speed and direction PHOTOGRAPHS Cloud cover and type Site setup Nadir Azimuth Temperature and humidity Eastern sky Target Aerosols smoke haze Western sky Air pressure TARGET e g vegetation Species Layering Health Homogeneity Cover Form Localised conditions Phenology Texture specular diffuse ANALYSIS Identification of outliers using spectra amp metadata REFERENCE SPECTRUM For a point in time ANALYSIS Separability and similarity studies Figure 11 Conceptual diagram of the factors affecting spectral measurement The field analyst and experimental design can be used to control to an extent the viewing manner of the reference and target to reduce erroneous results due to poor illumination geometry and transition conditions the timing of data collection including integration and spectrum averaging spatial scale of measurement and the calibration procedures to minimise variability in the spectral response such as white reference monitoring Consideration and documentation of each of these components are essential in obtaining meaningful spectra in the field but rarely are these reported Lack of consistent field methodologies appropriate metadata collection associated with spectral data consideration of spatial and temporal variation in spectral response of the senso
110. ic leaves show a shift of the red edge to shorter wavelengths Horler et al 1983 Elvidge 1990 and the reflectance peak normally centred at 550 nm broadens towards the red Adams et al 1999 In stressed vegetation both the absorption efficiency of chlorophyll and the infrared reflectance decreases due to changes in the cell structure of the plant Adams et al 1999 Dawson and Curran 1998 and Datt 2000b found the red edge correlated strongly with foliar chlorophyll content and so provided a sensitive indicator of vegetation stress Nutrients and toxic metals may cause chloritic leaves as these elements tend to move toward the actively growing cells of green foliage observed as a variation in the shape and position of the chlorophyll absorption bands Collins et al 1983 Several authors have found that the red edge shifts towards shorter wavelengths for trees growing over copper mineralisation Howard et al 1971 Collins et al 1983 in Horler et al 1983 This shift is also found in geochemical anomalies of high Ag Cu Pb Zn and Au in the upper soil Collins et al 1983 and various other metallic elements Milton amp Mouat 1989 2 4 4 Dry vegetation Dry vegetation lacks chlorophyll and intense water absorptions although absorption wings may be present between the 400 900 nm regions Elvidge 1990 Dry vegetation leaves such as Eucalyptus species lack a 680 nm absorption and have a diagnostic absorption feature near 1730 nm Datt 20002 Ov
111. ida Arthrostylis aphylla Cyperus iria Fimbristylis composita Fimbristylis dichotoma Fimbristylis furva Fimbristylis pauciflora Fimbristylis phaeoleuca Fimbristylis squarrulosa Leptocarpus spathaceus Rhynchospora longisetis Tricostularia undulata Xyris cheumatophila Scleria brownii Scleria novae hollandiae A total of 121 ground cover species were recorded during the wet season survey Of these 34 28 were grasses 73 herbs 60 and 14 12 sedges There were 11 32 weed grasses and 17 23 weed herbs Weeds comprised 48 of all species on the minesite Twice as many native grass species were found on reference sites than minesites Overall five times more weed herb species were found on minesites compared with reference sites However there were three times more native sedge species on reference sites compared with minesites in both seasons Reference sites remain largely free of grass weeds that typify the minesite The cover of native grasses on reference sites was about five times that of minesites and the cover of native grasses approximately doubled in both locations Similar dominance ratios for biomass were found as for percentage ground cover ie grasses gt gt herbs gt gt sedges There was 5 5 times more native grass biomass on reference sites compared with minesites and in contrast 318 times more weed grass biomass on minesites compared with reference sites 3 1 4 Priority target species A summary
112. ield panel relative to the lab reference panel indicates that contamination has occurred Note we cannot assume that a change in the field panel only is an indication of contamination as a change in reflectance could be a result of a change in illumination by the lamps The panel is cleaned if contamination is realised following recommendations by Labsphere undated if the material is lightly soiled it may be air brushed with a jet of clean dry air or nitrogen For heavier contamination the material is cleaned by sanding under deionised running water with a 220 240 grit waterproof emery cloth until the surface is totally hydrophobic water beads and runs off immediately The panel is then blow dried with clean air or nitrogen or the material is allowed to air dry The standard panel measurements in the laboratory are repeated if the field panel has been cleaned and the reference spectra stored with metadata documenting the date and method of panel cleaning 44 4 5 5 Accounting for dark current and noise random noise and stray light The measured signal and computed reflectance are defined as Measured signal true signal dark current random noise stray light ASD 1999 A certain amount of electrical current generated by thermal electrons as a result of the spectrometer electronics false data is always added to that generated by incoming photons called dark current DC a property that varies with temperature and in the V
113. in and that any change in measurement would more likely introduce errors for the current application 4 6 3 Integration timing and sequential measurements The user can modify the number of optimisations WR and spectrum averages and averaging measurements will increase precision and reduce random error Milton et al 1995 Rollin et al 1995 However errors can arise from sequential measurements Deering 1989 Duggin amp Phillipson 1982 Milton amp Goetz 1997 Rollin et al 1995 so replication of measurements must be weighed up against the time taken and accuracy implications Statistical representative numbers of sample sizes are between 30 40 measurements Schaepman 1998 with 10 the minimum ASD 2002 The FieldSpec FR has a scan time of 0 1 seconds so the time difference to measure the reference compared to the target of interest is more a limiting factor than the number of integrations of reflectance measurement under a stabilised atmosphere Milton 1987 suggests that replication of each measurement and careful data screenings are safeguards against short term irradiance fluctuations between the target and reference The sequential method follows that described in Section 4 7 2 for optimisation WR readings and target readings 51 In summary the electronics are allowed to adjust to the panel surface by waiting for two screen refreshes Once a stable signal is realised optimisation is made and the solar radiance curve 25 averages
114. in enabling data comparisons by ensuring the quality consistency and portability of spectral signature measurements Apart from improved measurement quality compared with most ad hoc spectral campaigns the design and implementation of these spectral standards will also limit lost time due to poor measurements enable the measurements and associated uncertainties to be independent of the technician undertaking the measurements provide confidence that the operating equipment is performing as expected and accelerate the training of new staff members vii Importantly the standards facilitate measurement comparisons and improved measurement accuracy through identification and reduction of primary sources of uncertainty It is only once this level of rigor is applied to spectral measurements that ground based spectral feasibility studies will advance the use of spectral remote sensing beyond the short term project specific research realm and into practical cost effective tools for long term operational management The data compiled from this project form a knowledge base of spectral information suitable for data sharing particularly with respect to remote sensing feasibility studies The data collected to date will result in a knowledge base far greater than that ever obtained for vegetation spectra with respect to the range of species sampled the frequency of sampling duration of sampling and method and metadata documentation Further protocols o
115. individual cells vertical of rounded rolls generally white base if ragged referred to as rolls or towers flat base Precipitation drizzle fractostratus Precipitation drizzle Precipitation showers Cumulonimbus very large cauliflower shaped towers to 16 km high often anvil tops Phenomena thunderstorms lightning squalls Precipitation showers Figure A 3 Typical examples of the 10 Main Cloud Types Source http www bom gov au weather services about cloud cloud types shtml and http www bom gov au info clouds 102 A 6 Instructions for standardised photographic recording At each site a set of photos is obtained Each photo has a description of the photographer s location and camera settings and is given a formal nomenclature in the database XXNN_YYYY_MM_DD jpg Standard photos each have a naming format as follows e XX two capital letters designating the location where CP Croc Park CS CSIRO and BF Berrimah Farm e NN two number code for each individual site where a single digit 1s to be preceded by Zero e YYYY for number year e MM two number month and e DD two number day e These codes are all connected by underscores Eg CSO1 2007 01 01 sl jpg Each image has the following image codes added to the code described above connected by an underscore buggy is photographed five paces from the site and includes the buggy and fore optics location in relation to the site Figure
116. ing Society Nottingham UK 555 562 Rollin EM Emery DR amp Milton EJ 1997 Reference panel anisotropy in field spectroscopy In Seventh International Symposium on Physical Measurements and Signatures in Remote Sensing eds Guyot G amp Phulpin T Balkema Rotterdam 1997 Available online fsf nerc ac uk calibration Accessed May 2003 Rollin EM Emery DR amp Milton EJ 1998 Reference panel anisotropy in field spectroscopy International Journal of Remote Sensing 21 15 2799 2810 Rollin E Milton E amp Emery D 2000 Reference panel anisotropy and diffuse radiation some implications for field spectroscopy International Journal of Remote Sensing 21 15 2799 2810 Rowan LC Goetz AFH amp Ashley RP 1977 Discrimination of hydrothermally altered and unaltered rocks in visible and near infrared multispectral images Geophysics 42 522 535 Rueda CA amp Wrona AF 2003 SAMS Spectral Analysis and Management System Version 2 User s Manual Centre for Spatial Technologies and Remote Sensing Department of Land Air and Water Resources University of California Davis Available online http sams casil ucdavis edu Salisbury JW 1998 Spectral Measurements Field Guide Earth Satellite Corporation April 23 1998 Defence Technology Information Centre Report No ADA362372 Salvaggio C Smith LE amp Antoine EJ 2005 Spectral signature databases and their application misapplication to modelling and exploitation of multis
117. inter space are systematically sampled and recorded with a repeat of the above procedure Standardised averages are a spectrum average of 25 dark current of 25 and WR of 10 4 6 2 5 Other viewing geometries Phinn et al 2008 suggest a spectral data collection approach that varies with solar azimuth and zenith angle to minimise BRDF effects and maximise measurement of colour properties of vegetation cover They use an elevation angle of fore optics at 57 5 from the horizontal plane and at an azimuth angle of 90 to the plane of the sun The magic elevation angle is optimised for plant canopy observation and is derived from relationships between measurements of leaf area index LAT of foliage and observation angle The 58 degree angle is where the variability of LAI estimation to leaf angle distribution is minimised Wilson 1963 or put another way the solid angle of foliage viewed from this angle ie ratio of foliage to background for plants with a low LAI is more consistent between plants with variation in canopy structure Apparently this angle does not take into account any illumination effects it merely provides a more consistent solid angle of leaf area when observing different plant canopies particularly if sparse foliage P Daniel CSIRO pers comm 28 04 08 in Phinn et al 2008 SSD considered this method and decided that maintaining a 58 angle for vegetation habits up to 2 m high would be too difficult to accurately mainta
118. interpretation of remotely sensed data of vegetation indicate the need for reliable ground measurements of the physiological state of plants Buschmann et al 1994 McGowen et al 2001 Many weedy species are indistinguishable from other native plants particularly during vegetative growth Fitzpatrick et al 1990 Price 1994 McGowen et al 2001 and several species may have quantitatively similar spectra due to the spectral signature variation present within a species Price 1994 An important factor for distinguishing a particular species is obtaining data at the appropriate phenological stage usually during flowering Hunt et al 2003 Ticehurst et al 2003 Unique spectral differences may be apparent if the plant has an early green up or senescence phase a late senescence phase or a unique architecture or growth form Few feasibility studies exist using ground based reflectance spectra scaling up to remotely sensed data McGowen et al 2001 undertook field spectral studies on a range of pasture and weedy plants across a growing season to investigate the potential of Landsat TM for mapping serrated tussock Nassella trichotoma and Scotch thistle Onopordum acanthium Both species change colour distinctly from other species and make them appropriate for remotely sensed analysis In this example scotch thistle and serrated tussock were mapped with 80 and 72 of infestations being identified at a reliability of 97 and 87 respectively Further findin
119. into the three pin plug on the back plate of the spectrometer Always turn the spectrometer on before the laptop to prevent irreparable damage to the spectrometer array Warm up the spectrometer connected to the mains power for 90 minutes prior to the collection of laboratory spectra Record the time the spectrometer was turned on in data sheet so that the length of warm up time can be documented in the spectral metadata Connect the spectrometer and controlling laptop computer to the parallel ports using the parallel cable A 2 2 Ensure equipment conforms to the standard setup design Ensure that the tripods are located in the correct position indicated by the markers on the laboratory bench Two pro lamp assemblies should be positioned on a tripod each and fixed 100 cm from the surface at an angle of 30 degrees from the surface with a horizontal distance of 50 cm between the lamps Whenever a lamp bulb needs to be changed ensure that both bulbs are replaced at the same time Also ensure that the lamps have been switched on for a minimum of 30 minutes prior to laboratory readings This is required to maintain an even light source The spectrometer fore optics are mounted on a tripod at a height of 51 cm with the collecting optics of the spectrometer nadir to the sample A height of 51 cm is used so that the target can be lifted 1cm from the bench surface providing an approximate distance of 50 cm An 8 FOV lens is used providing an
120. inutes for the HgAr lamp insert the bare fibre optic cable into the lamp Draw the block out lined curtains and turn off the fluorescent lights Allow the spectrometer to adjust to the new surface by waiting for two screen refreshes Optimise the spectrometer Collect and save the HgAr a spectrum Exit the controlling RS3 software A 2 10 Spectral measurement Mylar card Warm up the tungsten filament lamps attached to the mains power for 30 minutes Check lamp tripods are positioned at the marked locations on the laboratory bench Check the illumination lamps are positioned on tripods at 1 m height and an angle of 30 from the bench Mount the spectrometer fore optics to the tripod at a height of 51 cm with the collecting optics of the spectrometer nadir to the focus point the focus is marked on the bench Ensure an 8 of FOV lens is attached to the fore optics Using the laser or weight attached to string from the fore optic pistol grip ensure that the focus point is in the centre of the marked position on the bench Adjust the fore optics if required ensuring the fore optics are maintained at a height of 51 cm nadir to the bench With clean washed hands locate the laboratory 25 4 x 25 4 cm standard panel the field 25 4 x 25 4 cm standard panel the field 5 x 5 cm standard panel the circular standard panel and the Mylar reference card Leave the panels housed in their protective cases Ha
121. is close to a Lambertian assumption and therefore insensitive to BRDF over the full wavelength range insensitive to contamination weathering and ageing and 100 reflectivity over all wavelengths Obtaining reflectance spectra of a standard provides a good approximation to the true BRF of the subject because the irradiance is dominated by its directional component the reference is nearly Lambertian and the BRF of the subject is not radically different from Lambertian Robinson amp Biehl 1979 For a true Lambertian reference the panel reflectance factor is assumed to be 1 0 and must be closely monitored and assessed for the panel to maintain its Lambertian behaviour Jupp 1997 and assure a valid reflectance factor data Jackson et al 37 1987 However in the field the panel is illuminated by a combination of direct and diffuse flux distributed non uniformly Milton et al 1995 When well maintained Labsphere Spectralon panels are relatively flat over the 250 2500 nm region providing near perfect reflectance 98 99 and thermal stability Schaepman 1998 Spectralon amp is a sintered polytetrafluoroethylene based material that has emerged as the preferred reflectance material for field reference panels Rollin et al 1997 1998 The Spectralon Calibration Certificate states the uncertainty of each panel and is often less than 0 005 for the spectral range 300 2200 nm however it should be realised that laboratory calibration conditio
122. is used to compare and document the response of the VNIR and SWIR regions over time The spreadsheet is then updated and saved as a new sheet by date of measurement These reference spectra stored by date can be queried and correlated with reflectance measurements 94 Table A2 Record sheet of laboratory naming conventions SSD s Standard Laboratory Measurements Date Time spectrometer switched on ACST Time HgAr lamp switched on ACST Time tungsten filament lamps switched on Path C Data Laboratory measurements eg C Data YYYY Laboratory measurements 1 HgAr Lamp Path C Data Laboratory measurements HgAr lamp eg g C Data YYYY Laboratory measurements HgAr lamp date YYMMDD 2 Mylar Card Optimisation Path C Data Laboratory measurements Mylar panel EN eg g C Data YYYY Laboratory measurements Wylar panel date YYMMDD 000 WR path C Data Laboratory measurements Mylar panel eg g C Data YYYY Laboratory measurements Mylar panel date YYMMDD 001 Mylar spectrum path C Data Laboratory measurements Mylar panel A eg g C Data YYYY Laboratory measurements Wylar panel date YYMMDD 002 3 WR laboratory panel WR of Spectralon laboratory panel path C Data Laboratory measurements Laboratory panel e eg g C Data YYYY Laboratory measurements Laboratory panel date YYMMDD 000 4 WR field panel WR of Spectralon laboratory panel path C Data Laboratory measurements Uncleane
123. is viewed and saved to document stray light interferences and checked to show zero reflectance at 1400 and 1900 nm atmospheric water bands Figure 18 Even though random noise signals are extremely small they graphically show vertical lines that shoot upward from the last wavelength channel with a non zero measured signal eg a random noise signal at 1900 nm of 3 and 6 radiance values for the reference and target respectively would equal 200 reflectance at 1900 nm Entire spectra of noise values may be calculated with the standard deviation from the mean of 25 or more spectra collected of a known source In the field environment solar radiance and WR standard spectra are recorded for each sample to indicate instrument and atmospheric stability systematic and random noise Figures 18 amp 19 Figure 19a shows a WR spectrum collected under near perfect sampling conditions with 0 cloud cover low humidity still wind and stable ambient temperature Compared with 19a Figure 19b illustrates systematic noise as a result of inadequate spectrometer warm up time and steps between the VNIR and SWIR 1 detectors This step is also a function of input radiance Hueni 2009 pers comm Maier 2009 pers comm Figure 19c shows an unstable atmosphere in the water absorption bands 1400 amp 1900 nm as well as significant random noise in the SWIR 1 SWIR 2 arrays Computed reflectance stability is assessed in situ on screen where an unstable atmosphere is i
124. lant species respond to geochemistry conditions of surface and near surface environments Milton amp Mouat 1989 Some geobotanical associations are related to regional lithologic variations whereas others are specifically related to anomalous concentrations of metals where changes in plant biomass towards less dense stunted vegetation or even barren ground may occur Brooks 1972 in Goetz amp Rowan 1981 For information derived from remotely sensed data to be beneficial targets of interest must be discriminated with accuracy and precision the data must be cost effective when compared with traditional methods and the information extractable and deliverable in a timeframe that is suitable for decision making McGowen et al 2001 Ticehurst et al 2003 Appropriate data for current revegetation applications at minesites are typically limited to hyperspectral airborne platforms due to the ground resolution of satellite hyperspectral sensors Collaborative airborne missions in the Top End of the Northern Territory are few and decommissioning costs prohibit customised data acquisition due to uncertainty of results in a changing ecological environment SSD has acquired opportunistic airborne hyperspectral data such as CASI HyMap and Airborne Multispectral Scanner AMS data Figure 2 but the different acquisition dates and variable sensor characteristics make a quantitative comparison and cost benefit analysis impossible Results are sensor specific
125. light intensity Salisbury 1998 4 6 5 3 Humidity and wind descriptions Humidity is measured using a Kestrel Pocket Weather Station Humidity is measured to accuracy of 0 1 Indirectly humidity can be assessed with the water absorption features in the WR spectra refer to Figure 24 Wind affects mobile targets eg leaves and can change target geometry During even slight breezes it can be difficult to maintain a steady fore optic but the stabilising pole minimises the variation in spectral averages associated with wind Figure 26 Wind speed and direction is measured using a Kestrel Pocket Weather Station Wind is recorded in km hr to an accuracy of km hr 4 6 5 4 Temperature Because DC systematic noise is sensitive to temperature ambient temperature is measured with a Kestrel Pocket Weather Station After turning the instrument on and waiting for the thermometer instrument to stabilise sometimes taking up to 2 minutes a reading to an accuracy of 0 1 degrees Celsius is recorded Figure 26 Effects of wind on mobile targets a gentle breeze b no wind All spectra are 5 replicates times 10 averages 4 6 6 Hemispheric contribution and scattering target texture surrounds and operator In addition to the viewing and illumination geometry and atmospheric conditions the texture of the target diffuse or specular shadows the surrounds and the operator of the instrument may also contribute to the hemispherical
126. literature review of the growth form and height of species was undertaken and it was found that most targeted species do not reach a maximum growth height of 2 metres and this was considered an operationally feasible measurement height It was decided that should a species encroach the 2 metre height of the stabilising pole then the height of measurement would be altered for that reading and that this change would be noted in the spectral metadata Vegetation height as well as senescence maturation are variables measured and listed in the metadata The target sampling height of 2 metres means that the height difference between the WR and GFOV of the target spectra vary as the growth form varies It is therefore essential that the height of the target be accurately measured discussed further in Section 4 7 8 4 6 2 3 Repeat WR panel measurements After the three target spectral samples have been measured the stabilising pole is swung back over the WR panel and another WR reflectance measurement is saved These last WR data can be assessed against the WR measurements taken prior to the target spectra to monitor unrealised solar changes during target sampling The resulting target spectra would be flagged of this solar change occurrence 4 6 2 4 Violation of the BRF assumption The viewing angle and height of measurement for the target and WR are not the same but any differences are minimised while maintaining an operationally feasible field cam
127. mid Zenith Number of samples taken 5 Foreoptic degrees Number of averages per sample 10 IFOV Diam in cm Dark Current Intergrations 25 Western sky White Reference Intergrations Foreoptic height above plant m Foreoptic height above ground m White Reference Source 10 3 2 5 3 2 10 x 10 Spectralon Panel Figure 28 SSD s Spectral Database metadata records 62 Main Option m i i f S n Photos FileName Code Description Date Taken Comments ViewBFO4 2007 04 11 bugsylJPG l buggylJPG 1 buggylJPG 1 buggylJPG ViewBF04 2007 04 11 nLJPG jewBFO4 2007 04 11 nslJPG View BFO4 2007 04 11 obnlJPG View BFO4 2007 04 11 obn2JPG jewBFO4 2007 04 11 obslJPG ViewBFO4 2007 04 11 obs2JPG View BFO4 2007 04 11 s1JPG iewBFO4 2007 04 11 samplelJPG 1 samplelJPG1 samplel JPG 1_samplel JPG ViewBFO4 2007 04 11 sample2JPG ViewBF04 2007 04 11 sample3JPG 1 sample3JPG1 sample3JPG1 sample3 JPG ewBFO4 2007 04 11 ss1JPG ewBFO4 2007 04 11 wsLJPG View BF04 2007 04 11 z1JPG Figure 29 An example metadata page with associated photographs 63 Main Options Date UniqueCode BF04_2007_04_11 Tarned on Sample Site BFO4 on None Temperature Hare None Humidity Distubances None Air Pressure Probe Height 2 Wind Direction Max plant height 1 Wind Min 1st Sample Time 12 13 00 PM Wind Max Ground Description 1 Green leaves Wind Description Cloud level Ground Description 2 Gro
128. mp Lenhoff CJ 1971b Visible and near infrared spectra of minerals and rocks IV Sulphides and sulphates Modern Geology 3 1 14 Hunt Jr ER Everitt JH Ritchie JC Moran MS Booth DT Anderson GL Clark PE amp Seyfried MS 2003 Applications and research using remote sensing for rangeland management Photogrammetric Engineering and Remote Sensing 69 6 675 693 Hunt ER McMurtrey JE Parker Williams AE amp Corp L A 2004 Spectral characteristics of leafy spurge leaves and flower bracts Weed Science 52 4 492 497 Hunt ER amp Rock BN 1989 Detection of changes in leaf water content using near and middle infrared reflectances Remote Sensing of Environment 30 43 54 71 Hunt ER Rock BN amp Nobel PS 1987 Measurement of leaf relative water content by infrared reflectance Remote Sensing of Environment 22 429 435 Irons JR Weismiller RA amp Peterson GW 1989 Soil reflectance In Theory and applications of optical remote sensing ed G Asrar Wiley Interscience New York 66 106 Jackson RD Moran MS Slater PN amp Biggar SF 1987 Field calibration of reference reflectance panels Remote Sensing of Environment 22 145 158 Joshi CM de Leeuw J amp van Duren IC 2004 Remote sensing and GIS applications for mapping and spatial modelling of invasive species In SPRS 2004 Proceedings of the XXth ISPRS congress Geo imagery bridging continents 12 23 July 2004 Istanbul Turkey Comm VII 669 677 Joyce KE amp
129. n the analysis of these data will follow this report and will document any change in spectral pattern for a given species the regions of the spectrum that provide the richest information for species discrimination the possibility to discriminate species at a particular point in time and over time in the hyperspectral feature space any optimum phenological stage to enhance the spectral separability of species and provide the most appropriate processing techniques Also further reports will detail other aspects of the project such as soil spectral measurements made in the laboratory Note Some of this work draws upon previously published materials Pfitzner K 2005a Ground based spectroscopy do we need it In Applications in Tropical Spatial Science Proceedings of the North Australian Remote Sensing and GIS Conference 4 7 July 2005 Darwin NT CD Pfitzner K 2005b Remote sensing for minesite assessment examples from eriss In Applications in Tropical Spatial Science Proceedings of the North Australian Remote Sensing and GIS Conference 4 7 July 2005 Darwin NT CD Pfitzner K Bartolo RE Ryan B amp Bollh fer A 2005 Issues to consider when designing a spectral library database In Spatial Sciences Institute Conference Proceedings 2005 Melbourne Spatial Sciences Institute ISBN 0 9581366 2 9 Pfitzner K amp Bollh fer A 2008 Status of the vegetation plots for the spectral library project Internal Report 546 Supervising
130. nd The laboratory spectral standards are transferable and should be used for all applications Prior to the planned field trip ensure that the battery packs for both spectrometer x 3 and controlling computer x 3 are charged Note that it takes about 4 5 h to charge a totally discharged battery In the laboratory stand the spectrometer securely on the supplied base unit and plug the AC adapter into an AC outlet and connect the cable from the power supply into the three pin plug on the back plate of the spectrometer A 1 1 Turn on the spectrometer Always turn the spectrometer on before the laptop to prevent irreparable damage to the spectrometer array Turn on the spectrometer connected to the mains power so that the spectrometer can warm up while the equipment is packed and loaded into the vehicle Note that the spectrometer must be running longer than 30 minutes and ideally warmed up for 90 minutes prior to the collection of spectra Note the time that the spectrometer was turned on so that the length of warm up time can be documented in the spectral metadata A 1 2 Pack equipment Pack equipment into the vehicle Use the Required field equipment checklist A 1 3 Pack spectrometer Pack spectrometer into the vehicle once all other equipment is packed and the operator is ready to leave The spectrometer can be packed in the pelican case and secured in the tray of a station wagon The spectrometer is sensitive to high ambient
131. ndardising optical satellite imagery Remote Sensing of Environment 75 350 359 Elvidge CD 1990 Visible and near infrared reflectance characteristics of dry plant materials International Journal of Remote Sensing 11 1775 1795 Emery DR Milton EJ amp Felstead R 1998 Optimising data collection for heathland remote sensing Developing International Connections In Proceedings of the 24th Annual Conference of the Remote Sensing Society Remote Sensing Society Nottingham UK 483 489 Evans MC amp Williams AM 1995 Integration of satellite imagery and GIS for natural resource management in the mining industry In NARGIS 95 Proceedings of the 2nd North Australian Remote Sensing and Geographic Information Systems Forum Darwin 18 20 July 1995 Supervising Scientist and the Australasian Urban and Regional Information Systems Association Inc Monograph no 11 AGPS Canberra 118 127 EWL Sciences 2005 ERA Ranger Weed Workshop Proceedings 25th October 2005 Filella I amp Penuelas J 1994 The red edge position and shape as indicators of plant chlorophyll content biomass and hydric status International Journal of Remote Sensing 15 7 1459 1470 Fitzpatrick BT Hill GJE amp Kelly GD 1990 Mapping and monitoring of weed infestations using satellite remote sensing data In Proceedings 5th Australasian Remote Sensing Conference Perth Western Australia 8 12 October 1990 598 601 Fyfe SK 2003 Spatial and temporal variati
132. ndicated by variability In addition to the solar irradiance and WR spectra data on the environmental conditions are recorded and these are discussed further Note that the operator must wait for two screen refreshes as the internal averaging cycles are completed before saving any information so that the electronics are allowed to adjust to the measurement surface Also note that the spectrometer archives the next spectrum measurement not the one on the screen Salisbury 1998 Measurement of the System Noise and Detector Dark Current at the beginning of a spectral campaign can be measured and saved and the peak and standard deviation of the spectral noise used to indicate current performance to historical performance 46 Fraunhofer lines ibi o ii ii a 1 4 j p a E c R S 5 o 5 o o c S 5 E 4 S o o if Y A absorption N iii ET UT ST QNS ER ee NA N ae RE DON L Jc A 1400 1600 1800 2000 2200 2400 2600 2800 3000 Wavelength nm Figure 18 Solar radiance spectrum measured in the field Stray light zero reflectance at atmospheric water bands illustrated Source ASD Inc 2001 Reflectance Reflectance Reflectance 500 TORR eier SN 2000 25 500 ORD cient nes 2000 250 5 TORE enka 2000 250 a WR spectrum under optimal b Detector array step in the c Detector array overlap and sampling and appropriate
133. ndition To exclude or at least minimise the effects of extraneous factors they must be known and then documented A method to link spectra and metadata must then be established The Supervising Scientist Division SSD of the Department of Sustainability Environment Water Populations and Communities SEWPaC aims to build knowledge on the spectral response of vegetation species and background targets important for land condition assessment and monitoring Temporal measurements were made of both weedy and native species from homogenous plots as well as measurements along environmental gradients This information was organised in the SSD Spectral Database This study was initiated to forward remote sensing technologies for the mine environment from the research realm into operational status by addressing the uncertainty in the spectral separability of land cover components over time 1 1 1 Purpose of the report The main purpose of this report is to develop and document the SSD standards for collecting field reflectance spectra This report summarises conceptualises and links existing references on aspects of spectral collection in the field and laboratory environments A specific protocol to acquire spectra potentially useful for revegetation assessment and monitoring is described The need and issues for acquiring robust in situ spectral data the factors affecting field spectral measurements and the importance of the SSD Spectral Database concep
134. ndle the Spectralon panels and Mylar card carefully do not touch the surface Touching only the sides and bottom of the laboratory 10x10 standard panel carefully lift the panel out of its case and place it on the marked panel position on the bench Use the yellow circular plastic pieces that are 1 cm high underneath the panel to obtain a level surface of the panel The laser light should fall in the centre of the panel Ensure the data directory has been created Start RS3 high contrast software Select the menu Spectrum Save Navigate to the data directory to where data will be saved Identify an appropriate File Base name Set the Starting Spectrum to 0 92 Set the Number of Spectra to be Saved option to 1 and click OK Select the menu Control and the submenu Adjust Configuration and set the fore optic to 8 reflectance mode Change the spectrum average to 60 dark current average to 25 and WR to 10 Draw block out lined curtains and turn off the fluorescent lights Allow the spectrometer to adjust to the new surface by waiting for two screen refreshes Wait for a stable signal Optimise the spectrometer Save this spectrum by pressing the spacebar Take and save a WR spectrum by pressing the spacebar Carefully place the Mylar card centred directly on the Spectralon panel Measure and save the transmission spectrum Carefully remove the Mylar panel leaving the Spectralon panel in place A
135. new target is given a separate file name and is given the nomenclature of site and date For example measurements at Crocodylus Park taken on 18 April 2007 would be stored in a root directory of CP 2007 04 18 All measurements taken at a plot including radiance WR and target measurements would be given a prefix For example measurements taken at plot 2 would be named CP02 2007 04 18 Sequential spectral files are saved from extension 000 The path name eg CACP 2007 04 18 and base file name eg CP02 can be established by going to the top pull down menu and selecting Spectrum Save and entering the file information By pressing the space bar at any stage the spectrum acquired will be saved into a binary file on the PC Ensure the spectrometer has been running for a minimum of 30 minutes and ideally 90 minutes before taking any spectral measurements Record the time that the spectrometer was switched on in the metadata recording sheets Whth the spectrometer and weather station running and the equipment set up as described above the setup is complete and ready for spectral measurements subject to the discussion following Note that when packing up and shutting down the PC should be switched off prior to switching off the spectrometer Disconnect all fore optics and replace the fibre optic cap to protect the spectrometer fore optic tip Ensure that the fibre optic cable is only loosely coiled and stored 100 A 5 Cloud descriptions Clo
136. ng Note that the times of measurements are recorded in the spectral header Table 7 Cover metadata collected in the data sheet and linked to the spectra in SSD s Spectral Database Ground cover 80 dead hyptis leaves and stalks 100 Calopo 80 dead SW leaves 95 para grass description 5 water cover 20 bare earth 10 mission grass seeds 10 Dead calopo Phenology Dead and dry Green and healthy Green and healthy Green and healthy Growing vigorously with small white flowers starting to emerge Additional No change since last None Banteng cattle are None comments sampled Plot will locked up and no need weeding soon longer grazing on para grass Ludwigia plants starting to grow on banks 58 Table 8 Site metadata collected in the data sheet and linked to the spectra in SSD s Spectral Database Site Code CP01 CP05 CP06 CP20 Species Hyptis Calopo Para Grass Snake Weed Date 30 11 06 30 11 06 30 11 06 30 11 06 Samples taken set 3 3 3 3 of 25 averages Spec turned on 8 30am 8 30am 8 30am 8 30am Solar spectrum Y Y Y Y WR prior to target Y Y Y P4 measurement WR end of target Y Y Y P4 measurement Temp 35 3 35 4 35 9 34 8 Humidity 49 6 49 4 49 2 50 4 Air pressure 1007 5 1007 4 1007 5 1007 2 Wind direction WNW WNW WwW Ww Wind min km 0 0 2 2 Wind max km 0 4 6 10 Wind description Still Very Gentle Ge
137. ng area or create a trip hazard A 1 7 Load and turn on weather station A 1 8 Check the viewing geometry Orientate the white panel box to ensure that the open lid will not cast a shadow on the panel Swing the arm of the stabilisation device around to the WR panel and adjust the pram and or white panel so that the probe is directly over the panel box Use the laser or small level and plumbline to ensure the probe is pointing to the centre of the white panel box Move the tripod and or WR panel on buggy as necessary Once the FOV is centred in the middle of the WR panel swing the arm of the stabilisation device back over the vegetation target at 90 60 and 30 from the WR panel to ensure that the probe will be measuring the vegetation within the plot Once satisfied that the viewing geometry setup is correct swing the arm of the stabilisation device around to the WR panel ready for spectral measurements A 1 9 Switch laptop on The spectrometer is already switched on and running Turn the controlling laptop computer on 80 A 1 10 Check that the date and time on the PC are correct Check that the date and time on the PC are correct Australian Central Standard Time These fields will be recorded in the spectral header A 1 11 Create a path to store the spectral data Through Windows explorer create a path to store the spectral data The correct working folder is based on C Data 20 Field Data 20 Location Croc Park Be
138. nmental considerations and physical considerations of the target itself Common to these variables are the semi controllable factors of viewing and illumination geometry The factors that may influence a spectral response and associated spectral metadata entries are interrelated refer back to Figure 11 and essential for accurate processing of spectral averages Accounting for varying incident solar irradiance atmospheric conditions meteorological conditions reflectance properties of the surface and sensor viewing conditions are fundamental in the experimental design and therefore require appropriate documentation Metadata need to capture information for each spectral measurement that can aid in both determining data quality and interpreting data averages Photographic recording of sky conditions and the state of the ground target at the time of the radiometric measurements can often be very helpful particularly if an explanation is required on a change in reflectance factors if a biophysical variable measurement shows no change Deering 1989 Milton et al 1995 Appropriate metadata enable outliers due to extraneous factors to be identified attributed and then excluded when processing averages of spectra in order to maximise a true representative reflectance spectrum Figure 27 illustrates a time series of ground cover reflectance spectra accompanied by selected metadata for Stylosanthes humilis Stylosanthes humilis 08 03 06 17 05 06
139. nmental metadata When the spectral data have been obtained record the environmental conditions on the data sheets The temperature relative humidity and wind speed and direction can be read from the Kestrel weather station An estimation of the cloud cover or oktas is recorded If the operator is confident with cloud descriptions the cloud types can be defined Provide a qualitative estimate of smoke and haze cover described by visibility in kms The sky will also be documented by photographs A 1 22 Record vegetation metadata Record the site code and species name Record the pattern of distribution where even distribution describes a uniform cover of vegetation over the ground and clumped describes vegetation that presents as distinct clumps across plot Estimate and record the amount of layering within the vegetation plot where single describes a layer of vegetation where all plant components are at the same level and little scattering would occur and multiple describes those vegetation that grow in layers as either different components of the plants or as different growth heights of individual plants Nearly all vegetation types will have multiple layering Estimate and record the cover homogeneity as cover of the target vegetation Ideally all plots will have a 100 cover of the target species At times cover may include a 96 component of exposed soil interspace leaf litter or an alien species Measure and record
140. ns are very different from the field environment Note that the panel reflectance is not uniformly high at all wavelengths as shown in Figure 13 and that there is a 6 absorption band near 2150 nm and a falloff in reflectance to longer wavelengths Salisbury 1998 Spectralon is an optical standard and although the material is very durable care should be taken to prevent contaminants such as finger oils from contacting the material s surface The surface of the panel should never be touched Every effort must be made to keep the panels clean and scratch free as the calibration precision and accuracy depends on a calibrated clean panel and the slightest cover can alter the reflectance properties Spectralon panels should be housed in their respective case and only opened for the time when an actual measurement is required Once the measurement is complete the case housing the panel should be closed to prevent contamination from particles including those that may be too small to be visualised such as ash and dust M g a i i p S F1 gt I amp 1000 1250 1500 Wavelength nm Figure 13 Typical 8 Hemispherical reflectance of a 99 calibrated Spectralon reflectance panel Source Labsphere 4 4 Spectrometer FOV and ground field of view GFOV Field of view FOV is used to define the solid angle through which light incident on the input or fore optics will enter the detector system
141. ntle Gusty Cloud level None None Mid Mid Thin High Cirrus Smoke Slight High Level Slight High Level Slight High Level Slight High Level Cloud cover 0 0 15 1 2 oktas 40 3 oktas Haze None None None None Disturbances None None None None Pattern of Even Even Even Clumped distribution Layering M S M M single multiple Homogeneity 96 100 100 100 100 cover of target vegetation Probe height m 2 2 2 2 Max plant height 1 8 0 1 0 4 0 7 Mean density height 0 4 0 1 0 3 0 4 of most biomass Nadir ruler height On ground On ground 0 4 On ground Position of measurement eg western side of plot Even uniformly covering ground Clumped distinc clumps across plot 59 5 Reflectance spectra and metadata A database approach Metadata are important for the interpretation of scientific data quality assessment and long term usability of data Hueni et al 2007 If detailed metadata describing the collection geometry do not accompany spectral data the comfort level of users of this data should decrease drastically Salvaggio et al 2005 A number of reflectance spectra management issues have been outlined here to maximise the collection of representative samples of calibrated and validated field spectra Issues include aboratory standard set up and measurements of the spectrometer and calibration panel performance optical considerations scale considerations local enviro
142. of the target point opposite the sun The setup side for measurement of the plot may differ depending on the time of day and the season With the vertical pole of the wooden stabilisation structure held upright by the metal tripod lift the arm of the 2 m pole so that it is horizontal and at a 90 angle from the vertical pole Secure chain onto hooks so that the arm is held in place Swing the arm over the vegetation plot to check the position of the tripod and probe over the desired target A 1 6 Attach the pistol grip and laser Screw the pistol grip that holds the fore optics into the end of the horizontal pole Remove the fibre optic cap and store in a secure place so that the fibre optic can be recapped at the end of measurement collection Unscrew the crimp on the pistol grip and carefully feed the fibre optic cable through the crimp and gently through the pistol grip until the tip of the fibre optic can be seen protruding through Tighten the crimp so that the cable is held in place but be careful to not over tighten as this will damage the fibre optic cable Remember to be careful not to kink or step on the cable and keep the cable only loosely rolled Screw on the 8 FOV lens attachment onto the fore optic and attach the laser pointers to the pistol grip Secure the fibre optic cable to the wooden tripod with Velcro straps so it runs along the horizontal pole and down the vertical pole and does not fall or cast shadow within the sampli
143. olo R Carr G Esparon A amp Bollh fer A 2011 Standards for reflectance spectral measurement of temporal vegetation plots Supervising Scientist Report 195 Supervising Scientist Darwin NT The Supervising Scientist is part of the Australian Government Department of Sustainability Environment Water Population and Communities Commonwealth of Australia 2011 Supervising Scientist Department of Sustainability Environment Water Population and Communities GPO Box 461 Darwin NT 0801 Australia ISSN 1325 1554 ISBN 978 1 921069 16 1 This work is copyright Apart from any use as permitted under the Copyright Act 1968 no part may be reproduced by any process without prior written permission from the Supervising Scientist Requests and enquiries concerning reproduction and rights should be addressed to Publications Inquiries Supervising Scientist GPO Box 461 Darwin NT 0801 e mail publications ssd Qenvironment gov au Internet www environment gov au ssd www environment gov au ssd publications The views and opinions expressed in this report do not necessarily reflect those of the Commonwealth of Australia While reasonable efforts have been made to ensure that the contents of this report are factually correct some essential data rely on references cited and or the data and or information of other parties and the Supervising Scientist and the Commonwealth of Australia do not accept responsibility for the accuracy currency or c
144. ompleteness of the contents of this report and shall not be liable for any loss or damage that may be occasioned directly or indirectly through the use of or reliance on the report Readers should exercise their own skill and judgment with respect to their use of the material contained in this report Printed and bound in Darwin by Uniprint NT Contents Executive summary 1 7 Introduction 1 1 Project definition 1 2 Hypothesis and research objectives 1 3 Background concepts Literature review and research context 2 1 Spectral database application remote sensing for minesite assessment and monitoring 2 2 Reflectance spectrometry basic terminology 2 3 Spectral remote sensing 2 4 The generalised spectral response of vegetation 2 5 Remotely sensed data for vegetation assessment and monitoring with particular application to the mine environment 2 6 The need for the collection of in situ spectra Plant species and sites 3 1 Target species 3 2 Fortnightly measurements of ground cover 3 3 Sites 3 4 Project limitations Factors affecting spectral reflectance measurements 4 1 Introduction 4 2 SSD s spectrometer 4 3 Considerations with single Field of View FOV instruments 4 4 Spectrometer FOV and ground field of view GFOV 4 5 Spectral stability of the equipment 4 6 Viewing and illumination geometry in the field Reflectance spectra and metadata A database approach 5 1 Data storage and processing Conclusi
145. on 6 1 Further work and reporting References Appendix A SSD s standards for collecting field reflectance spectra vi OQ A Oo On A 14 18 18 26 26 31 32 32 34 36 38 40 47 60 61 65 65 66 78 Tables Table 1 Plant species found on transects in the late wet season May 2004 22 Table 2a Summary of target weedy grass species important for Ranger Nabarlek and weeds declared of concern 23 Table 2b Summary of target weedy herb and vine species important for Ranger Nabarlek and weeds declared of concern 24 Table 2c Summary of target native grass species important for Ranger and Nabarlek 25 Table 3 Species sampled for the database during 2006 07 27 Table 4 FieldSpec Pro FR product specifications 35 Table 5 Calculations at 90 nadir of diameter for varying FOV lenses and the difference between a circle and ellipse for an 8 FOV example 40 Table 6 Example sun azimuth and altitude measurements for Darwin for the 1 of the month over a one year period 52 Table 7 Cover metadata collected in the data sheet and linked to the spectra in SSD s Spectral Database 58 Table 8 Site metadata collected in the data sheet and linked to the spectra in SSD s Spectral Database 59 Table A1 Required field equipment 78 Table A2 Record sheet of laboratory naming conventions 95 Table A3 Cloud cover 101 Figures Figure 1 The Alligator Rivers Region ARR 4 Figure 2 Multitemporal hyperspectral data co
146. on incoming solar radiation WR target spectra repeat WR spectra and metadata recording including photographic records A 1 27 Taking spectral measurements additional plots For every additional plot the following steps need to be repeated Adjusting the path to where the spectral files will be saved Optimisation Saving incoming solar radiation Saving WR spectra Saving the target spectra at 90 60 and 30 Saving an additional WR spectrum Recording metadata and photographs 87 A 1 28 Returning from spectral sampling On returning from field trip it is important to back up data immediately to avoid loss or damage to data The field data should be copied from the laptop to the server The metadata should be added to the spectral metadatabase as soon as possible Field data is entered and saved onto the server and then stored in a folder by date Field Notes 20 in the laboratory Images that are stored on the flashcard in the camera need to be copied to the server and given the appropriate filename A card reader is stored in the laboratory 88 A 2 Standards for collecting laboratory measurements A print out of Table 11 is used to record the file names used during the laboratory measurements This record is stored in the laboratory folder A 2 1 Turn on the spectrometer Stand the spectrometer securely on the supplied base unit and plug the AC adapter into an AC outlet and connect the cable from the power supply
147. on in spectral reflectance are seagrass species spectrally distinct Limnology and Oceanography 48 464 479 Fyfe S 2004 Hyperspectral studies of New South Wales seagrasses with particular emphasis on the detection of light stress in Eelgrass Zostera caprirorni Unpublished PhD Thesis University of Wollongong 2004 Game M Carrel JE amp Hotrabhavandra T 1982 Patch dynamics of plant succession on abandoned surface coal mines A case history approach The Journal of Ecology 70 3 707 720 Gates DM Keegan HJ Schleter JC amp Weidner VR 1965 Spectral properties of plants Applied Optics 4 1 11 20 69 Goel PK Prasher SO Landry JA Patel RM amp Viau AA 2003 Hyperspectral image classification to detect weed infestations and nitrogen status of corn Transactions of the American Society of Agricultural Engineers 46 2 539 550 Goetz AFH 1992 Imaging spectrometry for earth remote sensing In Imaging Spectroscopy Fundamentals and Prospective Applications ECSC EEC EAEC Brussels and Luxembourg 1 19 Goetz FH amp Rowan LC 1981 Geologic remote sensing Science 211 781 791 Gomez RB 2001 Spectral library issues in hyperspectral imaging applications In 5th Joint Conference on Standoff Detection for Chemical and Biological Defence 24 28 September 2001 Williamsburg Virginia Grove CI Hook SJ amp Paylor II ED 1992 Laboratory reflectance spectra of 160 Minerals 0 4 to 2 5 Micrometers Jet Propulsion Laboratory
148. otential weed Calopognium mucunoides CP05 Weed vine Centrosema molle Weed vine Chloris inflata CP12 Weed grass Chloris virgata Weed grass Crotalaria goreensis CP16 36 Weed shrub Crotalaria pallida CP34 Weed shrub Cynodon dactylon CP13 Weed grass Digitaria bicornis CP25 Native grass Digitaria eriantha BF05 Pasture grass potential weed Digitaria milanjiana BFO1 Pasture grass potential weed Digitaria swynnertonii BF04 Pasture grass potential weed Heteropogon contortus CP14 Native grass Hibiscus sabdariffa CP35 Weed shrub Hyptis suaveolens CP01 03 Declared weed NT Ipomoea spp CP23 Weed vine Melinis repens Cs12 Weed grass Panicum mindanease CP10 11 Native grass Passiflora foetida vine CS06 Weed vine Pennisetum pedicellatum CP04 CS07 Weed grass Pennisetum polystachion CS05 Declared weed NT Schizachrium spp CP15 Endemic native grass Senna spp CP18 Declared weed NT Sida cordifolia CP08 33 Declared weed NT Sorghum stipodeum 02 04 11 Native grass Stachytarpheta australis CP09 32 Declared weed NT Stachytarpheta cayennensis 07 20 31 Declared weed NT Stylosanthes hamata CP17 Weed herb Stylosanthes humilis CP02 CS01 03 Weed herb Urochloa maxima CS09 Weed grass Urochloa mutica CP06 Weed grass CP Crocodylus Park BF Berrimah Farm 27 7 A as e E E a Continuing Low intu aN s Discontinued MA 20 4 e t E N A 1 La Figure 6 Location of CSIRO Sites April 20
149. paign Despite the change in viewing geometry this set up allows almost simultaneous sampling of the WR panel and the target because the stabilising pole can be repositioned in a matter of seconds Importantly for temporal measurements the measurement method is consistent While operating in WR mode the variability in sky conditions can be checked by measuring a spectrum from the reflectance panel with any variation from a spectral reflectance of 1 0 indicating a change in the spectral irradiance since the panel was first measured Milton amp Goetz 1997 The spectral solar radiance result and surrogate global irradiance measurements are not usually reported This is surprising given these measurements may be used to ensure 50 that an appropriate RF is achieved and that the spectral readings are not influenced by stray light or random noise We consider the standard panel spectral sequence necessary to determine whether sufficient accuracy has been acquired and to assess that environmental factors are not limiting Simply a flat spectrum with near 100 reflectance indicates stable conditions whereas an unstable atmosphere is indicated by a computed reflectance that varies over time showing absorption minima or maxima If illumination conditions change within the sets of target spectral measurements the optimisation and WR readings are repeated before spectral averaging of the target are repeated For heterogeneous covers soil and or litter
150. pectral hyperspectral data In Proceedings of the SPIE Sensor Data Exploitation and Target Regognition Algorithms and Technologies for Multispectral Hyperspectral and Ultraspectral Imagery XI Volume 5806 March 28 through April 1 2005 Satterwhite MB amp Henley JP 1990 Hyperspectral signatures 400 2500 nm of vegetation minerals soils rocks and cultural features Laboratory and field measurements US Army Corps of Engineers Engineer Topographic Laboratories Fort Belvoir Virginia 22060 5546 Schaepman ME 1998 Calibration of a field spectroradiometer Calibration and characterisation of a non imaging field spectroradiometer supporting imaging spectrometer validation and hyperspectral sensor modelling Remote Sensing Laboratories Department of Geography University of Zurich Zurich Switzerland Schaepman Strub G Schaepman M Dangel S Painter T and Martonchik J 2005 About the use of reflectance terminology in imaging spectroscopy EARSeL eProceedings 4 2 2005 191 202 Available online http las physik uni oldenburg de eProceedings vol04 2 04 2 schaepman strubl pdf Accessed 06 01 2009 Schmidt H amp Glaesser C 1998 Multitemporal analysis of satellite data and their use in the monitoring of the environmental impacts of open cast lignite mining areas in Eastern Germany International Journal of Remote Sensing 19 12 2245 2260 Shafii B Price WJ Prather TS Lass LW amp Thill DC 2004 Using landscape characteristi
151. r taken from nadir with the camera held at shoulder height moving across the site from west to east ng n5 and n6 taken from nadir with the camera held at a 1 meter height or as the vegetation height will allow with camera on full zoom moving across the site from western edge to centre and then to eastern side _esl and _es2 east sky taken of the eastern sky at horizon and at 45 degrees respectively wsl and ws2 west sky taken of the western sky at horizon and at 45 degrees respectively If the east and west sky are obscured photographs of the north and south sky are taken instead labelled as ns1 ss1 etc ZI zoom 1 taken towards zenith angle with the camera held vertically with no zoom and provides a record of the atmosphere around the Sun hl height 1 taken of the height of plant with measuring ruler in view if species is clumped Note the number of the photographs according to the camera name convention Wherever possible the measuring pole is included in the photographic images of the ground setup A 1 24 Moving to the next plot Lower the horizontal bar ensuring that the fibre optic cable is not bent kinked or pinched Fold up stabilizing mechanism and rest onto wooden panel and move on to next plot A 1 25 Setup for the next plot Once the equipment is setup on the side of the target point opposite the sun and the viewing geometry has been checked t
152. r and target and accurate calibration of both the sensor and data are factors that have prevented the transfer of knowledge from one application to another and also limited the commercialisation of field and imaging spectroscopy applications The conceptual diagram in Figure 11 highlights not only the factors that need to be considered within the experimental design to maximise the accuracy of spectra but also highlights the need to document these components as spectral metadata including the capture of photographs of the sky conditions and target It is only once consideration is given to the experimental design of spectral collection and that accurate metadata including photographs are captured that we can begin to populate spectral libraries representing reference spectra and use these spectra for separability and similarity assessment studies across applications For the full capability of spectral sensing technology to be exploitable it is essential that a well populated spectral library exists and is accessible in a user friendly way by the user of this technology Gomez 2001 This necessitates a consistent and repeatable spectral collection method with standards adhered to and the inclusion of metadata The advantages of collecting spectra with the future view of data transfer are that data quality improves systematic bias is reduced variability associated with data collection is minimised extraneous factors can be accounted for and
153. r even ground as the spectrometer is sensitive to vibrations The spectrometer should always be transported carefully It is common practice at SSD to transport the spectrometer within a vehicle It is not appropriate to transport the spectrometer in a tray back ute for risk of temperature vibrational and or dust damage For spectral campaigns that are located by sealed roads the spectrometer can be secured with a seatbelt and transported in the back passenger seat The spectrometer can be warming up when transported in this mode For longer trips the spectrometer is transported while secured in the black pelican case usually securely positioned in the back of a station wagon or back seat of a sedan The pelican case must be out of direct sunlight to prevent high temperatures affecting the spectrometer Preferably the vehicle air conditioning is constantly running during transport On smooth surfaces the spectrometer can be wheeled to and from the vehicle via the use of the case handle Figure A 1 The spectrometer buggy setup The spectrometer is seated securely in the seat of the buggy and shaded from direct sunlight The spectrometer is weather resistant but definitely not waterproof or airtight and therefore should not be operated or transported under rainfall or dusty conditions Ideally the spectrometer and reference panels are kept in a dry and dust and salt free environment 97 During transport and when the spectrometer is not in
154. ra Implications for data sharing In Proceedings Workshop on hyperspectral remote sensing and field spectroscopy of agricultural crops and forest vegetation 10 February 2006 University of Southern Queensland Toowoomba Queensland 21 22 Pfitzner K Esparon A amp Bollh fer A 2008 SSD s Spectral Library Database In Proceedings of the 14 Australasian Remote Sensing and Photogrammetry Conference Darwin 29 September 3 October 2008 Phinn S Scarth P Gill T Roelfsema C amp Stanford M 2008 Field spectrometer and radiometer guide Centre for Remote Sensing amp Spatial Information Science School of Geography Planning and Architecture The University of Queensland Version 8 Phinn S Hill GJE Johnston S amp Kirwood C 1991 Landsat Thematic Mapper imagery for mapping disturbed and rehabilitated vegetation North Stradbroke Island Queensland Queensland Geographical Journal vol 47 51 Pinter PJ Jackson RD amp Moran SM 1990 Bidirectional reflectance factors of agricultural targets A comparison of ground aircraft and satellite based observations Remote Sensing Environment 32 215 228 Portigal F Holasek R Mooradian G Owensby P Dicksion M Fene M Elliot M Hall E amp Driggett D 1997 Vegetation classification using red edge first derivative and green peak statistical moment indices with the Advanced Airborne Hyperspectral Imaging System AAHIS In Third International Airborne Remote Sensing Conference and
155. rev4 web pdf Analytical Spectral Devices ASD Inc 2000 FieldSpec Pro User s Guide Analytical Spectral Devices Inc Boulder CO USA Analytical Spectral Devices ASD Inc 2001 Field spectrometry Techniques and instruments Available online at http www asdi com Field o20Spectroscopy screen pdf Web site accessed May 18 2005 Analytical Spectral Devices ASD Inc 2002 Field Spec Pro User s Guide January 2002 USA Anderson K Milton EJ amp Rollin EM 2003 Sources of uncertainty in vicarious calibration Understanding calibration target reflectance In Proceedings of the IEEE International Geoscience and Remote Sensing Symposium IGARSS 2003 Toulouse France IEEE New Jersey USA CD Arnold GW Ozanne PG Galbraith KA amp Dandrigde F 1985 The capweed content of pastures in south west Western Australia Australian Journal of Experimental Agriculture 25 117 123 Barret EC amp Curtis LF 1992 Introduction to environmental remote sensing 3d edn Chapman amp Hall London Baumgardner MF Silva LF Biehl LL amp Stoner ER 1985 Reflectance properties of soils Advances in Agronomy 38 1 44 Bayliss P Bellairs S Pfitzner K amp Vink S 2004a Revegetation of Nabarlek minesite Preliminary characterisation of vegetation on the minesite and on adjacent natural landscapes in September 2003 Internal Report 488 October Supervising Scientist Darwin Unpublished paper Bayliss P Pfitzner K amp Bellairs S 2004b
156. rgeted opportunistically Native species will be more thoroughly addressed during the Ranger Trial Landform research 24 Table 2c Summary of target native grass species important for Ranger and Nabarlek Native grasses Genus Species Ranger Nabarlek Alloteropsis semialata v Aristida holathera v v Aristida hygrometrica v Aristida ingrate v Bothriochloa bladhii v Brachiaria holosericea v Chrysopogon fallax v Dicanthium fecundum v Digitaria bicornis v Digitaria gibbosa v v Dimeria ornithopoda v Ectrosia agrostoides v Eragrostis potamophila v Eragrostis spartinoides v Eragrostis sp v Eriachne agrostidea v Eriachne burkittii v Eriachne ciliata v Eriachne major v Eriachne shultziana v Heteropogon contortus v Heteropogon triticeus v v Imperata cylindrical v Mnesithea formosa v Pseudopogonatherum contortum v Pseudopogonatherum irritans v v Pseudoraphis spinescnes v Rottbeollia cochinchinensis v Schizachyrium fragile v v Sehima nervosum v Sorghum brachypodum v Sorghum intrans v Sorghum plumosum v Sorghum stipodeum Spermacoce sp v Sporobolus pulchellus v Thaumastochloa major v Yakirra nulla v v identified by Brennan 2005 Hollingsworth and Meek 2003 EWLS 2005 and Daws et al 2008 identified by Bayliss 2004a amp b 25 Four pastoral grasses with potenti
157. rnum Alysicarpus schomburgkii Alternanthera augustifolia Bergia pusilla Blumea axillaris probably Blumea sp1 Blumea tenella Bonamia pannosa Buchnera asperata Buchnera sp Cartonema parviflorum Cartonema spicatum Cartonema trigonospermum Cyanthillium cinereum Euphorbia muelleri Euphorbia schizolepis Euphorbia schultzii Fabacea sp Galactia tenuiflora Gomphrena flaccida Goodenia armstrongiana Goodenia pilosa Goodenia porphyrea Haemodorum sp Hybanthus enneaspermus Hydrolea zeylanica Jacquemontia browniana Ludwigia octovalvis Ludwigia perenis Malachra fasciata Marsdenia viridiflora Minuria macrorhiza Mitrasacme connata Mollugo pentaphylla Murdannia graminea Pachynema junceum Pachynema sphenandrum Phyllanthus eutaxioides Physalis minima Polycarpaea holtzei Polygala longifolia Polygala triflora Ptilotus corymbosus Pycnospora lutescens Sauropus ditissoides Scoparia dulcis Sebastiana chamaelea Sowerbaea alliacea Spermacoce stenophylla Stylidium semipartitim Stylidium turbinatum Thysanotis banksii Utricularia chrysantha Aeschynomene americana Alysicarpus vaginalis Euphorbia heterophylla Euphorbia hirta Hyptis suaveolens Macroptilium atropurpureum Macroptilium lathyroides Sida acuta Sida rhombifolia Stylosanthes hamata Stylosanthes viscosa Tridax procumbens 22 N Ipomea abrupta Ipomea diversifolia Ipomea eriocarpa Ipomea sp1 Merremia quinata Tephrosia remotiflora Xenostegia tridentata Passiflora foet
158. rocessing and analysis An IT based approach Journal of Spatial Science 51 2 93 Hueni A 2007 SPECCHIO User Guide Remote Sensing Laboratories University of Zurich 1 1 71 Hueni A amp Kneub hler M 2007 SPECCHIO a system for storing and sharing spectroradiometer data SPIE Newsroom December 2007 DOI 10 1117 2 1200711 0956 Online at http spieorg x18220 xml Hueni A Nieke J Schopfer J Kneub hler M amp Itten K 2009 The spectral database SPECCHIO for improved long term usability and data sharing Computers amp Geosciences 35 3 557 565 Hueni A Nieke J Schopfer J Kneubuhler M amp Itten KI 2007 Metadata of spectral data collections Proceedings 5th EARSeL Workshop on Imaging Spectroscopy Bruges Belgium April 23 25 2007 Hunt GR 1977 Spectral signatures of particulate minerals in the visible and near infrared Geophysics 42 502 513 Hunt GR 1979 Near infrared 1 3 2 4um spectra of alteration minerals Potential for use in remote sensing Geophysics 44 1974 1986 Hunt GR amp Ashley RP 1979 Spectra of altered rocks in the visible and near infrared Economic Geology 74 1613 1629 Hunt GR amp Salisbury JW 1971 Visible and near infrared spectra of minerals and rocks II Carbonates Modern Geology 2 23 30 Hunt GR Salisbury JW amp Lenhoff CJ 1971a Visible and near infrared spectra of minerals and rocks III Oxides and hydroxides Modern Geology 2 195 205 Hunt GR Salisbury JW a
159. rrimah Farm or CSIRO CP_20 _mm_dd eg C Data 2007 Field Data 2007 Croc Park CP_2007_05_17 A 1 12 Start High Contrast RS3 instrument software Start RS3 to obtain an interface like that illustrated below RS 64664 Display Control GPS Help 3S DO de ADDIS Dark Current None Taken Current 0 25 m White Reference None Taken Spectrum Save Y lab 003 0 1 m j Optimize Parms 5 a iana A eco Vnir IT 17 ms Swirl G 500 0 2048 Swir2G 500 0 2048 y MS SEM m 1250 1500 1750 2000 2250 2500 Wavelength nm amp Latitude A 1 13 Connect GPS via USB to the laptop The GPS should be set up with NMEA output Connect the GPS via USB to the laptop This must be done once the laptop is running otherwise the computer recognises the USB connection as a mouse and the actual mouse will be disabled Under RS3s GPS menu enable the GPS The coordinates will be recorded in the spectral header file and coordinates can be seen displayed in the lower left corner of the screen A 1 14 Spectral measurement setup saving data Go to Menu Control Spectrum Save or press Alt S 81 RS 64664 Display Contro GPS Help ooo Dark Current Take Dark Current measurement F3 390 Initialize Radiometric measurement F9 Take White Reference measurement F4 Adjust Configuration None Taken Optimize instrument settings culo Abort Spectrum Collection
160. rs A critical assumption in spectral measurements using single FOV instruments is that the BRF can be accounted for The essential field calibration procedure consists of the measurement of the response Vs of the instrument viewing the subject and measurement of the response V of the instrument viewing a level reference surface to produce an approximation to the BRF of the subject Robinson amp Biehl 1979 Duggin amp Philipson 1982 Deering 1989 Milton 1987 Rs 6i Qi Os P 7 X R 9 Qi 9 P X K r where R 0 0 is the bidirectional reflectance factor of the reference surface R is required to correct for its non ideal reflectance properties including non ideal reflectivity and non Lambertial behaviour and K measured reflectance of standard reflectance in band pass rS The amount of reflected EMR from the surface is expressed as a proportion of that which fell on the surface thereby compensating for the intensity and spectral distribution of the light source Milton 2001 Assumptions are that the incident radiation is dominated by its 36 directional component clear sky the instrument responds linearly to entrant flux the reference surface is viewed in the same manner as the subject and the conditions of illumination are the same the entrance aperture is sufficiently distant from the subject and the angular FOV is small with respect to the hemisphere of reflected beams limit of 20 angular FOV
161. s Council of Australia 2002 that can be assessed over the mining lease with remotely sensed data One research component of SSD is to evaluate remotely sensed data for land cover condition assessment and monitoring in the mine environment including post mining revegetation assessment Remote sensing techniques are routinely applied for vegetation applications at landscape scales In contrast to the landscape scale minesite applications often require large scale mapping discriminating covers at a high resolving power of highly variable surface covers The disturbed mining environment often composed of mixtures of plant species soils and rocks covers only relatively small areal extents Figure 1 The Alligator Rivers Region ARR m Escarpment Lowlands KNP boundary The identification and discrimination of vegetation cover at minesites are critical considering the role vegetation plays in preventing soil erosion and sedimentation by stabilising landforms The spatial arrangement of vegetation type and condition is an important component of studies of bio geochemical cycles and land use change Dungan 1998 and defining an ecosystem in terms of diversity and abundance may aid in revegetation management plans The health of vegetation provides an indication of other processes that may be hidden under vegetation cover Geological and chemical conditions may induce discrete patterns in revegetated areas such as senescence because certain p
162. s such as fire development or mowing In addition the sites needed to be in close proximity to each other and the eriss laboratory to reduce travel times Replicate plots have been established around the greater Darwin region with support from Commonwealth and Territory Government Departments and private industry via access to land from CSIRO Berrimah Farm and Crocodylus Park respectively Figure 5 East Point el Figure 5 Proximity map of Darwin area sites Table 3 summarises the species sampled for the spectral database during 2006 07 Figures 6 8 show the location of the species sampled in 2006 Observations over the late wet season of 2007 have shown that not all plots remain compositionally pure and in some cases the dominant cover present in 2006 has been out competed by another cover type or grazed In addition new species have become readily 26 identifiable after the wet season flush All species have been identified by the Northern Territory Herbarium Where a plot became overgrown with another species of pure composition the site was given a new identifying code for the following sampling season Table 3 Species sampled for the database during 2006 07 Genus Species CP CSIRO BF Status Aeschynomene americana CP26 Weed shrub Andropogon gayanus CP21 Weed grass Brachiaria humidicola CS10 BF02 03 Pasture grass p
163. so that the standard panel measurements are consistent The pistol grip mounted to the tripod is fitted with a laser pointer to ensure the focus point is centred Samples including the standard panels are positioned with the focus point centred in the middle of the sample and this position is checked before each measurement Figure 15 illustrates the laboratory set up Note that the white surroundings of the laboratory would have adjacency effects The laboratory walls and bench appear bright as the photograph has been taken with the fluorescent lights switched on Black matt walls would be ideal and we are in the process of updating all laboratory surfaces to matt black Figure 15 Spectrometer and laboratory white panel setup Note that the laboratory is a dark room rather than the white walls illustrated for this setup photograph 42 4 5 3 VNIR and SWIR spectrometer detector condition monitoring in the laboratory It is recommended to use a known discrete emission light source for verifying calibration in the VNIR and periodic examination of the absorption features in the spectra of materials with known characteristics for the SWIR detectors Beal 1999 Prior to an ASD spectrometer being dispatched or after the return of a spectrometer to ASD Inc wavelength calibrations on the spectrometer instrument are undertaken and the calibration relationship between wavelength and channel number in the controlling computer s asd ini file is installe
164. son amp Biehl 1979 by measuring radiation reflected from a surface accompanied by a near simultaneous measurement of radiation reflected from a reference panel in order to calculate a BRF for the surface Jackson et al 1987 Intelligent use of the BRF technique is an accurate and practical means to obtain the spectral optical properties of targets needed for advances in remote sensing Robinson amp Biehl 1979 Further there are mechanisms to check the BRF of the sequential measurements In most field measurements it is the reflectance factor RF that is estimated Robinson amp Biehl 1979 Reflectance factor is defined as the ratio of the radiant flux actually reflected by a sample surface to that which would be reflected into the same reflectance beam geometry by an ideal perfectly diffuse Lambertian standard surface irradiated in exactly the same way as the sample Nicodemus et al 1977 in Deering 1989 Robinson amp Biehl 1979 Rollin et al 2000 4 3 2 Standard panels Field reference panels are used to standardise measurements of target radiant flux in order to derive the RF on the assumption that the flux reflected from the panel can be used as a surrogate of the incident global irradiance Kimes amp Kirshner 1982 in Rollin et al 1995 1997 1998 2000 This assumes that the viewing and illumination geometries are exactly the same for the target and the reference panel The requirements of the standard reference are that the panel
165. specially near water band locations Salisbury 1998 Field measurements are therefore commonly restricted to a period around solar noon when the solar geometry is changing least and when the errors due to the angular response of the reflectance panel are at a minimum Gu et al 1992 in Milton et al 1995 Salisbury 1998 Rollin et al 2000 At SSD in situ measurements are made positioned on the side of the target point opposite the sun around the wings of solar noon When measuring spectra in even slightly varying or limiting conditions optimisation is performed frequently radiance mode is viewed occasionally to verify that signal saturation is not occurring ASD 2002 and a new solar irradiance and repeat WR sequence for every target sequence is recorded An accurate record of geographic location time sun azimuth and altitude and localised environmental conditions accompany spectral data The centre point of each sampling plot site is measured and documented with a dGPS The exact sampling position relative to the target can change over the fortnightly temporal scales as measurements are made positioned on the side of the target point opposite the sun The location is measured with each spectral reading using a USB GPS and recorded in the spectral header file although there is a generalised offset of 1 metre between the buggy position and the target sample site Section 4 7 2 In addition a written record of the location with respect to the
166. standing of both the species present and their distribution across the minesite as a result of poor spectral discrimination in a small spectral space 2 5 2 Hyperspectral remotely sensed data for revegetation applications Hyperspectral platforms typically record data over the VNIR SWIR in up to hundreds of narrow channels Examples include Airborne Visible Infrared Imaging Spectrometer AVIRIS Compact Airborne Spectrographic Imager CASI Airborne Multispectral Scanner AMS and Hyperspectral Mapper HyMap The pixel size of airborne data is subject to the capability of both the sensor and flying height of the platform Hyperion is a research based sensor onboard the Earth Observation EO 1 satellite that co orbits and has the same pixel size as that for Landsat TM with 220 bands across the VNIR SWIR The CHRIS sensor operates in two modes Mode 1 works with 62 spectral bands at a spatial resolution of 34 m while Mode 2 used for studies of waterbodies presents 18 bands at 17 m Guanter et al 2005 Many examples of mapping and monitoring vegetation using hyperspectral remotely sensed data exist eg Chewings et al 2000 Lewis 2000 Goel et al 2003 McDougal et al 1999 used AVIRIS data to group vegetation into three general groups high and moderate chlorophyll content dry and green vegetation and dry vegetation There are isolated examples of 13 discriminating weedy and native vegetation using hyperspectral data For example Underwoo
167. sting in adjacent areas of Kakadu National Park to form an ecosystem of long term viability which would not require a maintenance regime significantly different from that appropriate to adjacent areas of the Park Further to the RPA requirements the Environmental Requirements of the Commonwealth of Australia for the Operation of Ranger 1 Northern Territory of Australia Mining Management Act 2006 authorisation number 0108 04 variation of authorisation number 0108 03 19 Uranium Mine 1999 state that operations should not result in change to biodiversity or impairment of ecosystem health outside of the RPA and that the operations at Ranger will not result in any adverse impact on Kakadu National Park through the introduction of exotic fauna or flora Hollingsworth and Meek 2003 describe six vegetation communities comprising Eucalypt savanna woodlands and a Melaleuca sedge grassland as analogue descriptions for the ecosystem reconstruction for the RPA They recommend a list of 60 candidate species including overstorey midstorey and understorey species for restoration of the landform based on their commonness dominance and similarity to community structure in similar adjacent areas in Kakadu National Park The ground covers or understory described included 40 species and these species are ranked with importance values across habitats In order of importance for ground covers they list the following grasses Sorghum intrans Aristi
168. such as the wavelength position depth and width of an absorption feature may be diagnostic descriptions as can be reflectance magnitude and slope Traditional spectral research related spectral observations with geological materials for example see Hunt amp Salisbury 1971 Hunt et al 1971a b Rowan et al 1977 Hunt amp Ashley 1979 Hunt 1977 1979 Clark et al 1990 and biophysical measurements Collins 1978 Horler et al 1983 Boochs et al 1990 Elvidge 1990 Spectrometry has been extended to novel applications such as the urban environment eg Herold et al 2004 and coral reefs eg Joyce amp Phinn 2003 2 3 Spectral remote sensing Coupled with recent advances in remote sensing systems and expectations of future developments in satellite technology have been the increasing need to measure in situ reflectance spectra Spectral signatures are fundamental means of data representation and analysis in all forms of passive reflected sunlight remote sensing Differences in the spectral response from remotely sensed data are a function of the target and environmental background the illumination and viewing geometries and the spectral spatial and radiometric response of the remote sensor The relationships between spectral signatures and the biological chemical physical and atomic structure of gases water vegetation and soils has been explored using remote sensing techniques in areas of atmospheric chemistry plant physiology geological
169. supervising scientist report Standards for reflectance spectral measurement of temporal vegetation plots a OO a OOS OT K Pfitzner R Bartolo G Carr A Esparon amp A Bollh fer y Australian Government Sd a Department of Sustainability Environment Water Population and Communities Supervising Scientist It is Supervising Scientist Division policy for reports in the SSR series to be reviewed as part of the publications process This Supervising Scientist Report has been formally refereed by two external independent experts Authors Dr Kirrilly Pfitzner Spatial Sciences and Data Integration Group Environmental Research Institute of the Supervising Scientist GPO Box 461 Darwin NT 0801 Australia Dr Ren e Bartolo Spatial Sciences and Data Integration Group Environmental Research Institute of the Supervising Scientist GPO Box 461 Darwin NT 0801 Australia Geoff Carr Northern Land Council GPO Box 1222 Darwin NT 0801 Australia Environmental Research Institute of the Supervising Scientist at time of writing Andrew Esparon Spatial Sciences and Data Integration Group Environmental Research Institute of the Supervising Scientist GPO Box 461 Darwin NT 0801 Australia Dr Andreas Bollh fer Environmental Radioactivity Group Environmental Research Institute of the Supervising Scientist GPO Box 461 Darwin NT 0801 Australia This report should be cited as follows Pfitzner K Bart
170. t To attach the fibre optic cable to the pistol grip on the measurement pole remove the fibre optic cap and place the cap in a secure place so that the fibre optic can be recapped at the end of measurement collection Unscrew the crimp on the pistol grip and insert the end of the cable through the crimp and all the way into the pistol grip until the tip of the fibre optic can be seen protruding through Tighten the crimp so that the cable is held in place but be careful to not over tighten as this will damage the fibre optic cable Screw on the 8 FOV attachment Ensure that the laser pointers are attached on either side of the pistol grip and check the light source of these using the remote control Connect the spectrometer to the laptop by securing the parallel cable into the spectrometer and laptop ports Always turn on the spectrometer before the laptop irrespective of the power option mains or battery powered to avoid damage to the spectrometer electronics Never 99 turn on the PC first as the current this generates may damage the spectrometer Ensure that the time and date displayed on the PC is correct The GPS should be set up with NMEA output From the laptop start the RS3 program and then plug in the GPS cable to the laptop Go to RS3 s GPS settings and enable the GPS It should find the GPS and show the coordinates in the lower left corner of the screen A log file of the coordinates where measurements are taken will be stored in
171. t are detailed This document provides a detailed description of the scientific and operational requirements needed to successfully conduct the SSD Spectral Database Project with the focus on field standards and laboratory calibrations The standards are transferable to other researchers and applications The only difference required may be that of the fore optic height and target field of view The timing of this research is appropriate given the decommissioning timeline for the Ranger uranium mine and the associated requirement for a robust rehabilitation monitoring method combined with expectations of advances in technology such as hyperspectral and newer generation VHR satellite platforms The project is applicable to any application requiring information on the spectral separability of ground cover species 1 2 Hypothesis and research objectives The hypothesis is that with a well designed approach to collecting field spectral measurements and metadata extraneous factors can be accounted for accurate post processing of spectra can be performed and the first database of northern Australian spectra relevant to the mine environment can be developed These data can be assessed to gain knowledge on the usefulness of remotely sensed data for vegetation applications Spectra can be assessed for similarity and separability within and between species and analysed over time Spectra can also be resampled to band numbers and widths of existing
172. t be used to detect small weed infestations or weeds mixed with other vegetation Lass et al 2005 Light and scattered weed infestations represent the highest priority for control but these forms are most difficult to detect from remotely sensed data Moody amp Mack 1988 in McGowen et al 2001 Quantitative information about vegetation often requires high spectral resolution data because vegetation types are chemically similar and most healthy plants show absorption bands that are almost identical with broadband remotely sensed data Clark et al 1995 Fitzpatrick et al 1990 Price 1994 Recent spatial spectral and radiometric advancements of remote sensors provide applicable test data for minesite applications including revegetation assessment and monitoring The challenge in monitoring minesite environments using remotely sensed data is to differentiate cover types with wide spectral variation across an inherently variable land surface and over different capture times Differentiation of introduced weeds native ground and tree canopy cover exposed soil and some mineral assemblages is required over both the minesite and surrounding country as these index local environmental conditions in addition to contributing to an overall rehabilitation assessment Frequent landscape scale cover data are required to adequately assess the pervasive effects of ecological disturbances such as those caused by fire weeds and high winds Remotely sensed data that
173. ta specifications particularly spatial scale spectral resolution and geometric instability have restricted the use of such data for routine monitoring and minesite applications For vegetation applications aerial photography has not been widely used because of the absence of quantitative data high cost variable interpretation and the requirement for manual scanning or digitising Arnold et al 1985 in Lass et al 2005 although is no longer the case with digital photography Broadband satellite data include Landsat Thematic Mapper TM which records data over seven visible near infrared VNIR to shortwave infrared SWIR bands at 30m pixel sizes and the Advanced Very High resolution Radiometer AVHRR which records data over six visible thermal infrared TIR bands at 1 km pixel sizes The Moderate Resolution Imaging Spectroradiometer MODIS instrument that provides high radiometric sensitivity 12 bit in 36 spectral bands ranging from the VNIR to TIR with two bands imaged at a nominal resolution of 250 m at nadir five bands at 500 m and the remaining 29 bands at 1 km The Advanced Spaceborne Thermal Emission and Reflectance Radiometer ASTER is characterised by pixel sizes between 15 90 m with 14 spectral bands covering the VNIR to TIR regions Broadband satellite data can be used to detect vegetation patterns including weed infestations only after they become dense and widespread Carlson et al 1995 in Underwood et al 2003 but canno
174. tance 1500 2000 500 TODO Wavelength nm 1000 ER Um 1000 ER ona Wavelength nanometers Wavelength nanometers 2007 10 17 2007 11 25 2008 01 24 Figure 10 Selected photographic and spectral examples for one plot of Digitaria swynnertonii Arnhem Grass over a period of time 30 25 3 4 Project limitations There are limitations in the project design with respect to the range of species sampled the number of replicate plots of a given species sampled and differences in soil conditions and localised atmospheric conditions These limitations have been acknowledged and thought given to minimise the influence on spectral results 3 4 1 Species range and replication The fortnightly plots were established to enable all spectra to be gathered efficiently For the required frequency of spectral measurements high travel times were not feasible and site selection was therefore restricted to the greater Darwin region and limited to a few locations The species available for sampling were dependent on these criteria although species of potential spectral sampling are continually being sourced particularly during flowering the post wet season Priority species are those weeds occurring in the wet dry tropics Smith 1995 amp 2002 that are of concern to the revegetation success at minesites by increased threat of disturbance such as fire or those species that potentially threaten rehabilitation success particularly with
175. tandard setup is marked permanently on the laboratory bench The laboratory set up is similar to that recommended by ASD 2002 The laboratory is fitted with two 200 500 Watt quartz halogen cycle tungsten filament lamps in housings with aluminium reflectors The illumination lamps are warmed up for 30 minutes prior to any spectral measurements to ensure they are stabilised both in current and thermally G Fager 2006 pers comm The two standard lamps are each positioned on a tripod The lamps on the tripods are fixed 1 m from the surface to minimise interference fringes at an angle of 30 degrees from the surface and at a horizontal distance of 50 cm ASD 2001 The tripod positions are marked in place on the laboratory bench defining a constant illumination distance and angle orientation so that the flux density remains the same The steady electrical power supply is used and whenever a lamp bulb needs to be changed both bulbs are replaced at the same time to ensure a similar output The spectrometer fore optics are mounted on a tripod at a height of 51 cm with the collecting optics of the spectrometer nadir to the sample This provides an instantaneous field of view IFOV of approximately 0 9 cm 7 0 cm and 22 2 cm for 1 8 and 25 degree FOV lenses respectively see Table 5 The 8 degree FOV lens is used unless otherwise stated The location of the fibre optic focus is marked on the bench The standard panel dimensions are also marked on the bench
176. temperatures and vibrations and should not be transported under direct sunlight or left in the car without the air conditioner running It is not appropriate to transport the spectrometer in the back of a ute For travel on sealed roads only the spectrometer can be secured with a seatbelt in the back of a sedan This way the spectrometer can be left switched on during transport as part of the warm up time Once at the field location ensure the transport vehicle is in a safe and secure location then unpack the field equipment A 1 4 Set up buggy Unfold buggy so that it is stable on three wheels and place the warmed up spectrometer into the buggy seat without the top panel Run the adjustable velcro straps on buggy though the spectrometer handle and comfortably tighten so that the spectrometer is vertical and secure Place the smaller wooden brace and large horizontal wooden panel onto the buggy and secure This will keep the spectrometer shaded and form a shelf to place the laptop and WR box Place the laptop and WR panel box on the buggy s horizontal wooden bench top panel Connect the laptop to the spectrometer via the serial cable 79 Load the auxiliary equipment into the buggy basket laser plumbline level counterweight pen field sheets camera portable weather station A 1 5 Set up measuring equipment Set up the wooden tripod stabilisation structure at the desired plot The setup will be positioned on the side
177. the plot to capture in site variability Soil inter space and leaf litter are also recorded if visualised during the growing season Operators and assistants dress in low reflective dark coloured clothing Deering 1989 and maintain a distance from the target with the stabilising pole to minimise any interference As shade eg under a tree is illuminated principally by skylight and background radiance some identified sites that are dense and homogenous have been found unsuitable for spectral sampling due to their proximity to other vegetation 4 6 7 Standardised photographic recording Photographic recording of the sky conditions and the state of the ground target at the time of spectral measurements can be helpful in interpreting and determining the data quality Deering 1989 In addition to scaled setup and nadir photographs photos of the eastern and western sky if these views are obscured then north and south views to be obtained as well as the hemisphere are documented to support quantitative and qualitative measurements of the hemispheric component Photos of sampling sites and sky conditions are best taken from the same location enabling the viewer to compare the target with similar backgrounds Different backdrops can distract the viewer Sky condition photos also contain pieces of familiar backgrounds eg horizon features to serve as reference points enabling the viewer to visualise the scale of clouds from one point in time compared
178. ting field measurements Table A1 Required field equipment Item Description Category Item Description Category 1 Buggy Field Equipment 18 8 Fore optic ASD Equipment 2 Buggy Wooden Brace Field Equipment 19 Trigger Fore optics Holder ASD Equipment 3 Buggy Wooden Flat Panel Field Equipment 20 Laptop ASD Equipment 4 Camera Field Equipment 21 Spare Laptop Batteries ASD Equipment 5 Spare Batteries AA Batteries Field Equipment 22 Serial Connector Cable ASD Equipment 6 Level Field Equipment 23 White Reference Panel in Box ASD Equipment 7 Laser Plumb Line Field Equipment 24 Field Notes Folder Stationery 8 Weather Station Field Equipment 25 Pens Stationery 9 Probe Holder Tripod Field Equipment 26 Field Data Sheets Stationery 10 GPS Field Equipment 27 Permanent Marking Pen Stationery 11 2 metre ruler Field Equipment 28 Weed Books Reference 12 Tripod Chain Brace Field Equipment 29 Brock s Plant Book Reference 13 Sample Bags Field Equipment 30 Fencing wire Maintenance 14 Access Key Croc Park only Field Equipment 31 Pliers Maintenance 15 Spectrometer ASD Equipment 32 Wire Cutters Maintenance 16 Charged Spectrometer Batteries ASD Equipment 33 Flagging Tape Maintenance 17 Pelican Case ASD Equipment 34 Mash Hammer Maintenance 78 The field measurement standards can be followed for measurements other than the temporal sampling of the vegetation plots with the only difference being the height of FOV and WR panel relative to the grou
179. tion were characterised on sample transects located on rehabilitated areas of the minesite and on adjacent natural reference sites and compared The results of the surveys conducted during a dry and wet season Bayliss et al 2004b are summarised in Table 1 and below Table 1 Plant species found on transects in the late wet season May 2004 Source Bayliss et al 2004b crasses ues O VINES sepoes Native Weed Genus species Native Weed Genus species Native Weed Genus species N N N N N N N N N N N N N N N N N N N N N N WwW Ww Ww Ww Ww Ww Ww Ww Ww Ww Ww Ww w Ww Aristida holathera Aristida ingrata Bothriochloa bladhii Chrysopogon fallax Digitaria bicornis Digitaria gibbosa Dimeria ornithopoda Eragrostis potamophila Eragrostis spartinoides Eriachne burkittii Eriachne major Heteropogon contortus Heteropogon triticeus Imperata cylindrica Pseudopogonatherum contortum Pseudopogonatherum irritans Pseudoraphis spinescens Rottbeollia cochinchinensis Schizachyrium fragile Sorghum plumosum Yakirra nulla Andropogon gayanus Chloris inflata Chloris gayana Chloris virgata Cynodon dactylon Echinochloa colona Melinis repens Paspalum plicatulum Pennisetum pedicellatum Pennisetum polystachion Setaria sp Sporobolus sp Urochloa maxima Urochloa mutica e e aa e a a a a e a a A E A E A e a a a E a E E E A r a a E A E A A r A Ar E A a a a A e A E A A e e a a A a A G Allium ce
180. tmospheric conditions measured with a cosine receptor or calibration targets development of spectral attributes surface water vegetation soil minerals and rocks goniometric measurements to develop and test models describing the relationships between the directional spectral reflectance of surfaces and their biophysical attributes and for feasibility and cost benefit analyses prior to remotely sensed data acquisition The types of questions that may be addressed in a minesite vegetation rehabilitation feasibility study include e Isa land cover type separable e What spectral and spatial scale is required for separation 14 e What is the best time of year for maximum separability of a land cover type Curtiss amp Goetz 2001 Detailed examples of ground based reflectance spectrometry for remote sensing feasibility studies converting data from radiance to reflectance the development of spectral libraries and the role of spectral libraries in multispectral data analysis can be found in Pfitzner 2005a Many remote sensing applications will remain in the research realm without a knowledge base to define expectations of species separability likeliness over time The spectral separability of vegetation provides special difficulties because the spectral behaviour is described by a small number of independent variables Price 1992 correlated due to their chemical composition Portigal et al 1997 Uncertainties in the physiological
181. tp fsf nerc ac uk resources guides pdf guides asd guide v2 dn pdf MacArthur A 2007c Field Guide for the ASD FieldSpec Pro Radiance Irradiance Measurements in Raw DN Mode Natural Environment Research Council Field Spectroscopy Facility NERC FSF Version 2 Available online http fsf nerc ac uk resources guides pdf guides asd guide v2 Radlrrad pdf MacArthur AA MacLellan C J amp Malthus TJ 2006 What does a spectrometer see Natural Environment Research Council Field Spectroscopy Facility NERC FSFO Available online http fsf nerc ac uk news RSPSocPoster pdf McCall J Gunn J amp Struik H 1995 Photo interpretive study of recovery of damaged lands near the metal smelters of Sudbury Canada Water Air amp Soil Pollution 85 2 847 852 McDougal RR Clark RN Livo KE Koklay RF Rockwell BW amp Vance JS 1999 Preliminary materials mapping in the Oquirrh Mountains region for the Utah EPA Project using AVIRIS data Summaries of the 8th Annual JPL Airborne Earth Science Workshop ed RO Green NASA JPL AVIRIS Workshop 8 11 February 1999 JPL Publication 99 17 291 298 McGowen IJ Frazier P amp Orchard P 2001 Remote sensing for broadscale weed mapping is it possible In Geospatial Information and Agriculture conference Information for Better Production and Environmental Management Incorporating Precision Agriculture in Australasia 5th Annual Symposium 16 19 July 2001 Australian Technology Park Everleigh Sydne
182. tral measurements in the field Resources are economised by a modified buggy This set up is designed for the temporal vegetation plot sampling but is appropriate for any field campaign where the buggy is to be utilised The only difference that may be required for other applications is that of the height at which the fore optics are mounted and the resultant GFOV The buggy houses the equipment required for spectral and metadata recording The spectrometer is housed in the seat of the buggy and secured into position using Velcro straps A modified platform shades the spectrometer and houses the controlling laptop and WR panel at a height of one metre from the ground surface The WR panel is protected in a wooden box The lid of the box is opened only when WR measurements are being made to minimise contamination of the surface There is a tradeoff between minimising contamination of the surface of the WR panel and the influence of the wooden box lid on spectral on adjacency effects A stabilising pole and measurement pole are clipped to the buggy during transport This setup requires one person only to record all spectral and metadata see Figure A 2 SOS Figure A 2 Scaled set up standard photograph A USB GPS is connected to the laptop recording the position of the laptop in the spectrum header file A weather station is carried on the buggy shaded by the top wooden panel Data sheets and digital camera are housed in the mesh carry baske
183. ude the spectrometer buggy and fore optic setup The photos can also be used as a quality control mechanism to confirm the correlation between documented and visual environmental conditions All photos and spectra are linked to a metadata record so environmental conditions when the data were collected can easily be referenced Figure 30 displays the spectrum list for the same metadata and photo page illustrated in Figure 29 The structure allows the user to easily query information Selected spectra can be viewed and overlaid to give a visual comparison Figure 30 provides an example of a solar radiance target and white reference spectra Each spectrum has certain characteristics that can be used to categorise them through an iterative process As part of the quality control procedure all spectra are processed through an algorithm that categorises the spectrum into a specified group depending on the defined boundary conditions Spectra that are not found to fit into a known category are marked as undefined not defined by the Classify algorithm These 61 spectra require further examination and may indicate problematic conditions For example the Not Defined entry circled in Figure 30 shows a stepped appearance with values between 0 8 0 9 in the VNIR and around 0 9 in the SWIR as well as a drop off in reflectance beyond 2200 nm The spectrum represents an invalid white reference measurement and any target measurements made in associ
184. uds are described by the percentage of sky covered by clouds and according to altitude high mid low by the standard BOM method http www bom gov au info clouds Cloud cover is measured by dividing the sky up into eights known as oktas and estimating how much sky is covered by cloud see Table A 1 If there are patches of individual cloud an estimate is made of how much of the sky they would cover if they were all put together Photographic recording of the sky conditions see Section 4 7 7 accompany the description of cloud cover in oktas in the metadata It is essential that no sampling be undertaken when clouds are passing overhead and typically sampling is not undertaken when the cover is greater than 4 oktas Table A3 Cloud cover 0 oktas Clear skies 1 okta Almost clear skies just the odd cloud 2 oktas Mostly clear skies only a quarter of the sky covered by cloud 3 oktas Partly cloudy just over half the sky is cloudless 4 oktas Partly cloudy half of the sky covered by cloud 5 oktas More than half the sky covered by cloud 6 oktas Mostly cloudy only a quarter of the sky showing 7 oktas Almost overcast just a small amount of sky showing 8 oktas Overcast no sky showing 9 oktas Sky obscured by fog Cloud is divided up into ten different types which are identified by their height and form see Figure A 3 and if the operator is confident in cloud classification then these descriptions are useful additions to the meta
185. umber of replicates under different conditions would be spectrally sampled To minimise and account for external variation species are sampled from maintained plots that maximise a homogenous response ie non target species are continually removed from the plot The soil inter space is spectrally measured wherever a lt 100 cover is obtained and detailed metadata is used to define any change in localised conditions both within the target plant and external condition 31 4 Factors affecting spectral reflectance measurements 4 1 Introduction Spectral measurements need to be accurate and precise representations of the target material but there are a variety of factors that affect the quality of spectral measurements Careful consideration must be given to the methods adopted to undertake spectral measurements and to the variety of factors including the optical propagation and those environmental and experimental issues that can affect the quality of resultant spectral data Critical issues for making in situ spectral measurements have been reported eg Nicodemus et al 1977 Duggin amp Philipson 1982 Milton 1987 Curtiss amp Goetz 2001 Milton et al 1995 Jupp 1997 Salisbury 1998 Schaepman 1998 Milton 2001 and these include the properties of the atmosphere timing of measurement height of measurement orientation of measurement FOV spectral averaging and calibration of the spectral data Milton 1987 Deering 1989 Rollin et al 2000
186. und Description 3 Cloud Cover Solar Spectrum Target spectrum Target spectrum Target spectrum Target spectrum White Reference get spectrum White Reference 11 04 2007 11 04 2007 11 04 2007 0372007 White Reference White Reference View Details vi White Reference e Points over 1000 Decision Points 0 9 1 0 Not Defined Figure 30 The spectral data associated with the photographs illustrated in Figure 29 5 1 2 Analysis of spectral data Once all suspect spectra have been filtered out analysis can commence on the high integrity data There are a number of spectral analysis management systems available online including SAMS Rueda amp Wrona 2003 SPECCHIO Hueni 2007 Hueni amp Kneub hler 2007 SPECtrum Processing Routines SPECPR Clark 1993 Kokaly 2005 and SpectraProc Hueni amp Tuohy 2006 SSD also has expertise in computing language and interactive environments for algorithm development data visualisation data analysis and numeric computation There are a number of toolboxes available including project specific signal processing techniques These may be tested in parallel to benchmark any custom methods produced by this project A future report will document the post processing procedure as specialised processing techniques are required for high dimensional feature space data Both feature and spectral space methods will be used such as quantifying similarity and dissimilarity variation in red edge
187. ures that provide the best separation These data can be resampled to indicate whether or not current multispectral systems can resolve important features for vegetation land cover mapping and condition monitoring in the mine environment The standards described here were developed to provide a consistent and repeatable method for recording spectra that minimises the influence of extraneous factors in spectral reflectance radiance and irradiance measurements The standards should be used to routinely obtain accurate and precise spectral measurements A literature review of the factors affecting in situ spectral measurements was undertaken to define what equipment needed to be calibrated what features needed to be characterised how the equipment should be calibrated how the features should be characterised and how the required measurement accuracy could be obtained The report identifies the key parameters that determine the accuracy and uncertainty of spectral measurements systems and the resultant measured data from them The method considers the factors affecting spectral data outlined in Pfitzner et al 2005 and provides standards to collect time series spectra of vegetation that maximise the spectral response of the end member itself Pfitzner amp Carr 2006 Pfitzner et al 2006 A detailed description of the measurement process developed to collect reference spectra and ancillary metadata is then given This report details the scientific and op
188. urge detection with hyperspectral remotely sensed data Journal of Range Management 56 1 106 112 Pfitzner K Bartolo RE Ryan B amp Bollh fer A 2005 Issues to consider when designing a spectral library database In Spatial Sciences Institute Conference Proceedings 2005 Spatial Sciences Institute Melbourne ISBN 0 9581366 Pfitzner K 2005a Ground based spectroscopy do we need it In Applications in Tropical Spatial Science Proceedings of the North Australian Remote Sensing and GIS Conference 4 7 July 2005 Darwin NT CD Pfitzner K 2005b Remote sensing for minesite assessment examples from eriss In Applications in Tropical Spatial Science Proceedings of the North Australian Remote Sensing and GIS Conference 4 7 July 2005 Darwin NT CD Pfitzner K amp Bayliss P 2006 Revegetation monitoring 60 cm pixels and an object oriented approach In Proceedings of the 13th Australasian Remote Sensing and Photogrammetry Conference Canberra November 2006 Pfitzner K amp Bollh fer A 2008 Status of the vegetation plots for the spectral library project Internal Report 546 December Supervising Scientist Darwin Unpublished paper 74 Pfitzner K Bollh fer A amp Carr G 2006 A standard design for collecting vegetation reference spectra Implementation and implications for data sharing Journal of Spatial Sciences 54 2 79 92 Pfitzner K amp Carr G 2006 Design and implementation of vegetation reference spect
189. useful record of the type and condition of the target measured the way in which the target was measured and the environmental conditions at the time of measurement Whilst spectral data can be acquired quickly in the field the acquisition and recording of spectral metadata does increase the time required for the field campaign However the increase in usefulness of fully described spectral data far outweighs the small additional investment in time required for metadata descriptions of associated spectra This report focuses on the standards for reflectance spectral measurement developed by the Supervising Scientist Division SSD The standards described here relate specifically to the Spectral Database Project and in particular standards for measuring terrestrial vegetative ground covers The Spectral Database Project aims to provide a database of reference spectral signatures over the 400 2500 nm range pertinent to the study of cover and condition of minesites and surrounding country Vegetative ground covers shrubs and trees soils and minerals mine related features and built up features will be incorporated into the database The ground cover component aims to investigate the use of remotely sensed data to discriminate ground cover plant species using spectral data acquired by in situ spectrometry To do this dense and homogenous plots of key ground cover species pertinent to the success of minesite rehabilitation including native and
190. vasion of weedy grass species and fire Bayliss et al 2004a amp b Remotely sensed data have shown promise at Nabarlek in both assessing revegetation covers and mapping and monitoring the area of disturbance and recovery of threats such as fire Pfitzner 2005b Experience has shown that compositionally different vegetative species show spectrally similar responses at the Quickbird resolution and the Nabarlek example showed that vegetative species are spectrally confounded within the Quickbird 4 dimensional spectral space Pfitzner 2005b Pfitzner et al 2006 Despite poor spectral resolving power very high spatial resolution of the pan sharpened product aids in the differentiation and identification of different vegetative species However an extensive knowledge base of the distribution of species cover combined with an object orientated rather than pixel based mapping approach is required Pfitzner amp Bayliss 2006 to map the complex vegetation cover WorldView 2 satellite data have now been acquired over the Ranger uranium mine and surrounding country With suitable spatial resolutions from satellite image data now available the advantages of using satellite over airborne data are that costs are known image captures can be planned in advance and new data captures can be tasked when disturbances such as fire are realised The disadvantage of the VHR approach is the extensive fieldwork component that is initially required to gain an under
191. ve the surface multiplied by the FOV of the solid angle that admits light see Section 4 3 Tan 0 5 FOV x height m x 2 x 100 GFOV cm SSD acquires in situ target measurements positioned on the side of the target point opposite the sun as suggested by Deering 1989 ie measuring setup in the solar principal plane A bubble leveller attached to a stabilising pole at a horizontal distance of 1m is utilised Figure 21 to ensure nadir viewing Two remote controlled laser pointers are attached to either side of the bubble leveller to provide the centre point Mounting the pistol grip on a tripod and immobilising both optical cable and FOV is recommended for reflectance measurements requiring high repeatability and accuracy Salisbury 1998 and our experience has shown the stabilising pole is required to reduce the variations in spectral measurements seen whenever wind is a limiting factor see Section 4 7 5 Measurements are made at a sensor zenith angle of 0 nadir with an 8 FOV so that the angle of acceptance is less than 20 full angle Baumgardner et al 1985 Deering 1989 Milton 1987 According to ASD pers comm it is better to move the sensor during data takes to minimise FOV problems It is a trade off as moving the instrument might give a better representation of the target but the pointing direction will be harder to maintain At nadir the only significant geometric concerns are the IFOV or GFOV and its relationship to the
192. veloped to account for adequate spectrometer warm up time laboratory monitoring of the spectrometer and reference panels image documentation of the target and environmental conditions photographs of the target at nadir scaled set up horizon photographs and hemispherical photographs subject information at the time of sampling classification condition appearance physical state measurement information instrument mode date local time data collector s fore optics number of integrations reference material height of measurement from target and ground viewing and illumination geometry environmental conditions general site description specific site location geophysical location sun azimuth and altitude ambient temperature relative humidity wind speed and direction weather instrument and sky conditions and of course reflectance spectra Section Al outlines the standards implemented for spectral field measurements over vegetation plots The equipment required in the field is outlined and this can be used as a checklist when packing for a fieldtrip Section A2 outlines the standards implemented for spectral measurements made in the laboratory Section A3 provides details on care and transport of the spectrometer Section A4 outlines the steps involved in set up of the spectrometer in the field Section A5 details cloud descriptions Section A6 provides instructions for standardised photographic recording A 1 Standards for collec
193. vering the Nabarlek minesite 6 Figure 3 Subset of the Nabarlek minesite covering the rehabilitated plant run off pond area 7 Figure 4 Illustration of green senescing and drying spectra of Digitaria milanjiana Jarra digit grass taken in the months of April May and October respectively in the Top End of Australia 9 Figure 5 Proximity map of Darwin area sites 26 Figure 6 Location of CSIRO Sites 28 Figure 7 Location of Crocodylus Park Sites 28 Figure 8 Location of Berrimah Farm sites April 2007 29 Figure 9 Examples of vegetation plots used to record the spectral reflectance of selected species over time 29 Figure 10 Selected photographic and spectral examples for one plot of Digitaria swynnertonii Arnhem Grass over a period of time 30 Figure 11 Conceptual diagram of the factors affecting spectral measurement Figure 12 Attenuation versus length of permanent FR fibre Figure 13 Typical 8 Hemispherical reflectance of a 99 calibrated Spectralon reflectance panel Figure 14 Obtaining the GFOV Figure 15 Spectrometer and laboratory white panel setup Figure 16 Mercury Argon Emission Spectrum Figure 17 Mylar transmission Spectrum Figure 18 Solar radiance spectrum measured in the field Figure 19 Standard Spectralon panel measurements are essential metadata for reflectance spectra Figure 22 Direction position and FOV Figure 23a amp b Weighted plumb line ensures sampling is obtained from central position of white pan
194. weedy grasses herbs vines and sedges were established The spectra of these species were measured over time at fortnightly intervals The spectral data were accompanied by metadata descriptions and photographic records using the methods described in this report This work was undertaken because management of both operating and rehabilitated minesites requires comprehensive information on species distribution and composition Traditional ground based surveys for floristic mapping involve time consuming fieldwork that is often very stressful for workers in the tropical environment Remote sensing has the potential to greatly reduce the requirement of ground based surveys for floristic mapping Broad band remote sensing sensors that have historically been used extensively for mapping of plant communities vi are however not sufficiently sensitive to allow discrimination of individual plant species Relatively recent advances particularly with respect to hyperspectral and very high spatial resolution sensors offer the potential for application to the mine environment The data obtained with the spectral database project will show whether or not there is potential for fine spectral resolution remote sensing products to map vegetation cover and condition based on spectral signatures at scales appropriate to the mine environment An evaluation of the most suitable wavelengths for spectral separation of cover species may identify specific spectral feat
195. wing season and used spectral changes resulting from seasonal variability to identify wavelengths most suitable for quantification of temporal changes using CASI data Research has also assessed the potential for detailed remote sensing measurements of vegetation chemistry eg Dury et al 2000 Datt 2000a and reflectance spectrometry has been used to correlate remotely sensed responses with biophysical changes including leaf area and leaf area Index LAI Birch et al 1998 Pu et al 2003 canopy species Cochrane 2000 Datt 20002 water content Datt 1999b Hunt amp Rock 1989 Hunt et al 1987 leaf biochemistry Dawson 2000 Kokaly amp Clark 1999 Kokaly 2001 Lamb et al 2002 Buschmann et al 1994 Curran amp Milton 1983 Belanger et al 1995 characterisation of leaves and flower bracts Hunt et al 2004 and stress Clark et al 1995 Narrow spectral bands such as those from hyperspectral sensors are required to resolve features such as the red edge chlorophyll and water absorption in vegetation spectra that may be indicative of the vegetation composition and vigour including dieback stress or morbidity and the success of results depend on many factors including the localised conditions the timing of data capture the characteristics of the sensor and the methodology of mapping used 2 6 The need for the collection of in situ spectra In situ reflectance data are collected for the calibration and validation of hyperspectral data a
196. y CD ROM Mifsud J 1996 Vegetation establishment in tropical environments In Mine Rehabilitation in Tropical Environments Short course Notes 18 20 December 1996 Darwin NT organised by Australian Centre for Minesite Rehabilitation Research Kenmore Qld Miller JR Wu J Boyer MG Belanger M amp Hare EW 1991 Season patterns in leaf reflected red edge characteristics International Journal of Remote Sensing 12 7 1509 1523 Milton EJ 1987 Principles of field spectrometry International Journal of Remote Sensing 8 1807 1827 Milton EJ 2001 Methods in Field Spectroscopy NERC EPFS http www soton ac uk epfs methods spectroscopy shtml Web site accessed May 18 2005 Milton EJ amp Goetz AFH 1997 Atmospheric influences on field spectrometry Observed relationships between spectral irradiance and the variance in spectral reflectance In Seventh International Symposium on Physical Measurements and Signatures in Remote Sensing ISPRS Courchevel France 109 114 73 Milton NM amp Mouat DA 1989 Remote sensing of vegetation responses to natural and cultural environmental conditions Photogrammetric Engineering and Remote Sensing 55 8 1167 1173 Milton EJ Rollin EM amp Emery DR 1995 Advances in field spectroscopy In Advances in environmental remote sensing eds Danson FM amp Plummer SE Wiley UK Milton EJ Schaepman ME Anderson K Kneubuhler M amp Fox N 2009 Progress in field spectroscopy Remote Sensin
197. ynthesis and higher reflectance at green wavelength regions Absorption in the red region at 680 nm and a rapid increase in reflectance from 680 to 780 nm are known as the red edge Milton amp Mouat 1989 Slaton et al 2001 which often forms an extreme slope Cell structure controls reflectance in the near infrared NIR Campbell 1996 primarily from the internal structure of plant leaves as a function of the number and configuration of the air spaces that form the internal leaf structure Danson 1995 Slaton et al 2001 The near infrared spectra of leaves result from a complex combination of scattering processes and overlapping absorptions arising from water and biochemical components Kokaly 2001 The reflectance of healthy green vegetation increases dramatically in the NIR where from about 700 1300 nm a plant leaf typically reflects 40 50 of the energy incident upon it Lillesand amp Kiefer 1994 Water content controls reflectance in the mid infrared Campbell 1996 Hunt et al 1987 Hunt amp Rock 1989 Healthy vegetation beyond 1300 nm typically absorbs or reflects incident energy with reflectance peaks at about 1600 and 2200 nm Absorptions occur as a result of water absorption around 1400 and 1900 nm with the exact position of water absorption bands varying For example Murphy and Wadge 1994 found that short blade grass showed strong leaf water absorption bands at 1450 and 1930 nm Using spectroscopy Kokaly and Clark 19
198. ype and distribution of ground cover even or clumped The cover may be described as even cover or uniformly covering the ground or clumped into distinct clumps across the plot e height of ground cover including maximum height and mean height density or the height of most biomass cover e apparent phenology vegetation health and growth stage using such terms as green flush flowering seeding senescing drying dead e any disturbances visualised such as the plot having been flattened by rain trampled by animals etc e pattern of distribution between species or age classes Where a correlation is being established other than the interaction of the target with EMR other measurements will be required leaf area index or cover moisture canopy height chemical analysis of compounds biomass height and leaf angle distribution For soil characterisation specifically colour pH moisture sample and field description roughness texture moisture are required For such variables a description and photograph are the minimum requirement for metadata records Where samples are taken for further analysis eg x ray diffraction chlorophyll concentration sample numbers should be associated with their reflectance and metadata records A photograph detailed description and sample s are referenced in the metadata record that is linked to the spectra Tables 7 and 8 show example metadata recorded for one morning of sampli

SSR195 - Standards for reflectance spectral measurement of

Contents

Download Pdf Manuals

Related Search

Related Contents