Spectral Reflectance measurements of geological materials in

Northern Victoria Land, Antarctica Stefano Salvi(1), Francesco Mazzarini(2), Fawzi Doumaz(1), Valerio Lombardo(1), Cristiano

Tolomei(1), (1) INGV – Laboratorio di Geodesia e Telerilevamento,

Via di Vigna Murata, 605 00143 Roma, salvi@ingv.it (2) CNR-CSGSDA, Via S.Maria, 53, 56126 Pisa, ross@dst.unipi.it

Riassunto

Vengono presentati i risultati di tre diverse campagne di misure spettrali eseguite nella Terra

Vittoria Settentrionale, Antartide. Numerosi spettri di riflettanza sono stati acquisiti su diversi tipi

di superfici rocciose, con una risoluzione di 1 nm nell'intervallo spettrale 0.4-2.5 µm. Vengono

descritte la metodologia di misura sul campo e le procedure di analisi preliminare e validazione dei

dati. Gli spettri sono organizzati in una Libreria Spettrale (LILIAN) che contiene anche tutti i dati

ancillari necessari alla loro validazione. Viene descritta l'organizzazione dei dati nella libreria

LILIAN e le modalità di accesso alla stessa via Web.

Abstract

In situ spectral reflectance measurements of geological materials were performed during three

separate surveys in the Northern Victoria Land, Antarctica. We describe the results of the

campaigns, the measurement methodology of spectral reflectance over geological surfaces, the

preliminary data analysis and the structure of the spectral library LILIAN, which contains the

validated spectra. All the data in LILIAN are public domain and can be accessed freely via Internet.

Their main uses will be to constrain the analysis and geological interpretation of remote sensing

data in the 0.4-2.5 µm wavelength range.

Introduction

Lithologic recognition and mapping is one of the primary task of geological oriented remote

sensing. Remote sensing applications are especially useful when extreme environmental conditions

inhibit direct survey such as in Antarctica.

1

Lithologic mapping from remote sensing data implies a typical methodology composed of different

level of analysis: image calibration, correction for atmospheric effects, image analysis (statistical

and/or deterministic), field verification, and final interpretation. The knowledge of the interaction

mechanisms of the electromagnetic radiation with the surface materials is instrumental for the

fullfillment of all these steps. In particular, in the optical range, the most relevant surface property

for geologic remote sensing is the reflectance.

The spectral intensity of the solar radiation reflected by any natural surface is influenced by the

physical and chemical properties of the material. Various electronic and molecular processes result

from the interaction between incident light and rock minerals. In the 0.4-2.5 µm range spectral

features in reflectance spectra of rocks are due to high energy electronic processes at short

wavelengths and to atomic and molecular vibrational processes at longer wavelengths (Hunt and

Salisbury, 1970; Hunt et al., 1973; Hunt, 1980). Spectral features of the most common rock-forming

minerals are stable in the VNIR range and can be used to identify the mineral phases. Since rocks

are formed by minerals, the relative abundance of their mineral constituents is the main factor

controlling the rock spectral signature. Other important factors are structure, texture and alteration

state of the rock surface.

The collection of reflectance spectra of different materials is finalised to the production of a spectral

library. Generally, the main purpose of a spectral library is to feed a remote sensing data analysis

system with the ground truth needed to perform image calibration and reliable geologic

interpretation. Although many spectral libraries have been published, they have limited use in

lithologic mapping, especially with low spectral resolution data, since they are most often

composed of laboratory measurements of mineral samples (usually pulverised). A more practical

approach consists of collecting rock spectra in the field over the actual reflecting surfaces (Salvi et

al., 1997). Such spectra yield a very accurate ground truth for the application of both statistic and

deterministic techniques of image analysis. In the most favourable cases a few well chosen rock

spectra will constrain all the analysis needed for the lithologic mapping of a large area.

This note describes the activities and the results achieved in the frame of the project 3d.1

‘Spectrometric surveys of Antarctic natural surfaces for an integrated study with remote sensing

data’ of the PNRA (Italian Program of Research in Antarctica). The project plan consisted of two

main tasks: the acquisition of spectral measurements of geological surfaces (rocks, regoliths), and

of glaciated surfaces in Northern Victoria Land, Antarctica.

In this paper we describe the procedures for the acquisition and analysis of in situ reflectance

spectra of the main lithologies of the area carried out during three field surveys in the 1994/1995, 2

1995/1996, 1997/1998 austral summers. Several spectra were measured and the rock surfaces

described in terms of their structure, morphology and composition. The entire database was

organised in a spectral library (LILIAN) freely accessible via Internet.

Geographic and Geologic Outline of the Northern Victoria Land

The Transantarctic Mountains (TAM) is a chain running across Antarctica from the Pacific to

Atlantic sides of the continent. The highest percentage of exposed rocks and soils in the continent is

along this mountain belt, especially in its Pacific border: the Northern Victoria Land (NVL); here

the ice free area is more than 5% of the emerged land.

The main regional morphological features are outlet glaciers, neveé and ice sheets. In such

environment most of free-ice areas are strongly affected by periglacial processes producing block-

field, block-sheet, and debris; other consistent deposits are due to glacial processes as glacial drift,

moraines, and raised beaches. There are also wide areas constituted by supraglacial deposits as

supra glacial moraines (Baroni, 1989a, b; Carmignani et al., 1989a, b).

The geology of the study area is characterised by the occurrence of a Proterozoic - Paleozoic

crystalline basement unconformably covered by continental deposits (Beacon Supergroup, Permian

- Triassic) and by volcanic sequences (Ferrar Supergroup, Jurassic - Early Cretaceous); the

Cenozoic - Quaternary alkaline volcanics (McMurdo Volcanics) close the sequence (Ganovex

Team, 1987; Carmignani et al., 1989a, b).

The sequence above described is made up by metamorphic rocks of low to medium high grade

(slate, schist, gneiss, calcsilicate, and amphibolite), intrusive rocks (granites, mafites, and dykes),

volcanic rocks (dolerite, basalt, and pyroclastics), sedimentary rocks (sandstone, siltite), and glacial

and periglacial deposits.

The intrusives are widespread all over the NVL with ages ranging from the early paleozoic (Late

Cambrian - Early Ordovician) to Cenozoic (Eocene - Oligocene). Difference in ages reflects also

different composition of magmatic sequences: the oldest rocks belongs to the crystalline basement

and constitute the Early Ordovician Granite Harbour Intrusives (GHI) of calcalkaline to subalcaline

affinity (Ghezzo et al., 1989); the youngest ones, the Meander Intrusives (MI), are related to the

alkaline Cenozoic magmatism affecting both Ross Sea area and NVL facing coastal area (Muller et

al., 1991). The GHI are constituted by granite, monzogranite, granodiorite, and tonalite rocks,

coupled with mafic products as diorite and gabbro, whereas the MI are made up of alkali granite

and syenogranite. The Cenozoic McMurdoVolcanics (Kyle, 1990) consists of mildly alkaline to

alkaline volcanic products ranging from alkali basalt to trachyte and from basanite and nephelinite

3

to phonolite. The Mount Melbourne Volcano is an expression of this volcanism. All of these rocks

are often covered by eolic and glacial deposits.

Italian geological spectral surveys in the Northern Victoria Land

In this section for each field party the instrument used, the measured type of rocks and the main

results of the campaign will be briefly described.

Austral summer 1994/1995 field survey

In this survey, measurements where planned to be performed by means of a GER IRIS

spectroradiometer with 5 nm of spectral resolution in the 0.4-2.5 µm range. Unfortunately the

severe Antarctic conditions caused the failure of this instrument early in the survey. The field work

was then performed by a spare EXOTEC 100 AX radiometer (property of CNR-IRSS of Milan)

operating in the wavelength range 0.4-1.1 µm. The EXOTEC allows for the calculation of

reflectance values in four channels relative to the spectral windows of Landsat-MSS or TM

satellites, depending on the filters used. With the same instruments Zibordi and Meloni (1991)

made bihemispherical reflectance measurements of some Antarctic surfaces. During the field

survey, MSS landsat satellite filters were used, using optics with field of view (FOV) of 15°; the

methodology of measurement has been discussed in Casacchia et al. (1999). Field data consisted of

spectral measurements, GPS data and field description of the textural and compositional

characteristics of the rock targets acquired in 52 different sites (Cagnati and Mazzarini, 1995;

Bogliolo et al., 1996; Casacchia et al., 1999). During this period, reflectance values of granitoid

surfaces and of their glacial and eolic deposits were measured. A preliminary spectral lithologic

classification was carried out analysing the collected data, relative to a deglaciated area south of the

Italian Terra Nova base (Casacchia et al., 1999); four classes were identified: a) granites

(granodiorite to monzogranite); b) mafic intrusives (gabbro, diorite and tonalite); c) glacial drift; d)

glacial and eolic deposits related to granites (feldspatic sand).

Austral summer 1995/1996 field survey

During the second campaign, the EXOTEC radiometer was used only in five sites, while most of

the reflectance spectra were acquired by means of a newly acquired Field Spec FR

spectroradiometer. This instrument measures the radiance reflected by a surface, or the solar

irradiance, in 2151 contiguous spectral bands (bandwidth 10 nm) between 350 and 2500 nm.

Various optics can be attached to the instrument via a fiber optic and the system is calibrated in the 4

laboratory with a standard illumination source to obtain absolute radiance measurements. The

reflectance spectra are obtained ratioing the radiance spectrum of the target surface to the radiance

spectrum of a standard reference panel measured in the same illumination geometry. The

instrument was delivered shortly before the beginning of the campaign, therefore the

spectroradiometer had to be tested directly during the Antarctic expedition (Bogliolo et al., 1996).

Several reflectance spectra of the main geological formations were measured for each of 14

different sites (Salvi, 1996). The instrument performed quite well in almost all conditions but the

quality of the spectral measurements suffered from a low S/N in the short wave infrared range and

from a damaged fiber optic cable.

Several types of rocks were spectrally sampled during this expedition. Particularly, granites

(monzogranites, granodiorites and tonalites), high grade metamorphic rocks (schists, gneisses and

amphibolites), metasedimentary rocks (meta-sandstones, meta-siltsones, meta-limestones), mafic

intrusive rocks (gabbros and diorites), pyroclastic rocks and basaltic and doleritic lavas, glacial

deposits (moraines), regolithic covers. GPS coordinates and field description of the textural and

compositional characteristics of the measured rock surfaces were recorded.

Austral summer 1997/1998 field survey

To increase the number of spectral measurements and to improve the quality of the spectra collected

in the previous campaign, a new survey was carried out in November/December 1997. The Field

Spec FR spectroradiometer was re-calibrated and a new fiber optic mounted to improve the overall

S/N. Rock surfaces were spectrally sampled in 17 sites; considering the improved instrument

performance, part of the measurements from the previous campaign were repeated and better

quality spectra were acquired. In 10 sites, spectral measurements from helicopter were also

performed to improve the surface characterization at the pixel scale of remote sensing images.

During this campaign reflectance spectra of granites (monzogranites and granodiorites), high grade

metamorphic rocks (schists, gneisses and amphibolites), regolites from alkaline lavas, pyroclastic

rocks, basaltic and doleritic lavas and Quaternary-Holocene beach deposits were acquired (Salvi,

98). Geological samples were collected for all the measured surfaces, to be later analysed and yield

the necessary information for the ancillary data files (see library description).

A further objective of the campaign was to identify a few sites to be used for the calibration and

atmospheric correction of remote sensing images, and measure their spectral reflectance. For this

purpose the site had to match the following requirements: high spectral homogeneity over a

minimum area of 60 x 60 m, very small local relief, gentle slope (0-10 degrees), easy identification

5

on remote sensing images. Availability of ground sites with a wide range of reflectances, some of

them possibly located at different elevations to account for the difference in atmospheric paths, was

also required. A few of the measured sites had these characteristics and were marked as possible

calibration sites.

Measurement methodology

Most of the surfaces measured with the 4-bands EXOTEC radiometer during the 94/95 and 95/96

campaigns were subsequently re-measured with the FieldSpec spectroradiometer. We verified the

consistency of the different measurements (see Figure 5 later in the text), thus we will describe in

the sequel methodology and results for the FieldSpec campaigns only.

The typical measurement procedure in the field consisted of using a tripod to hold the fiber optics

perpendicular to the surface at a distance generally between 0.3 and 0.6 meters; a FOV of 25° was

used, thereby sampling the surface over a circular area with diameter between 13 and 24 cm. Some

stop-and-go measurements were also done (i.e. holding the optics in hand and moving over the

surface after each measurement) to obtain an average spectral response over heterogeneous

materials as for instance poligenic regoliths and glacial drifts.

To minimize the effects of variable solar zenith angle on the reference panel response (see later in

this chapter) a particular measurement setting was also used (coded TI in LILIAN, see next

chapter). In this setting the reference panel was placed perpendicular to the solar illumination

direction (solar zenith angle = 0) and the optics was also aligned with this direction (view angle =

solar zenith angle). The target surface was chosen as close to the same geometry as possible.

During the 1995/1996 campaign some test measurements were executed from the helicopter at 50 to

100 m heights, in order to sample areas comparable to the pixel dimensions of satellite images (20 -

40 m) and obtain an average of the spectral contributions for non uniform surfaces. The spectra

measured during the aerial tests of the 1995 campaign did not compare well with the ground

measured spectra, mainly because the measurement time required to obtain a good signal was about

2-3 minutes and it was difficult to maintain a stable illumination geometry and to keep the same

surface area within the instrument FOV. These problems were partially solved during the

1997/1998 campaign, when the higher S/N of the improved instrument allowed for faster

measurements. Moreover for most sites the measurement distance was reduced to 10-15 m, which

corresponds to a 5-7 m-wide circular area, ensuring a better target uniformity during the

measurement. This measurement setting resulted in a good correspondance between ground and air

measurements (Figure 1).

6

As an additional requirement for the execution of good measurements the cloud cover had to be

close to zero; a minimal cover was tolerated when located far from the sun disc and was in any case

reported in the ancillary data file. The instrument was used most often in radiance mode, i.e. the

radiance reflected from the target surface and the reference panel were measured and stored as a

independent spectral files. Notwithstanding the requirement of clear sky and stable overall

illumination, the radiances reflected from target and panel were measured in short succession to

minimize the changes in illumination conditions. For all measurements, the reference panel used

was a 30 x 30 cm Spectralon board whose absolute directional/hemispherical reflectance factor was

certified.

For each target surface a large number of spectra (up to several tens for the less homogeneous

materials) was measured and averaged to increase the S/N and account for the compositional

variability.

For each spectral channel the absolute reflectance was then calculated as:

Rabs = (L target / L reference) * R reference [1]

Where:

L target is the radiance of the target,

L reference is the radiance of the reference panel (measured under the same illumination conditions),

R reference is the directional/hemispherical reflectance factor of the reference panel.

An important issue to be addressed before using field-measured reflectances concerns the

correction for the effect of the illumination geometry on the reference panel response (Jackson et

al., 1992; Alberotanza et al., 1993).

The absolute reflectance of the Spectralon material is determined at the factory in an integrating

sphere with isotropic irradiance, thus measuring the directional/hemispherical reflectance factor

R(0°/h) (normally the view angle = 0°, which is also common for field measurements).

In the field the typical measurement with natural illumination occurs in a directional/directional

geometry (the skylight contribution is very small and can usually be neglected) where the solar

zenith angle (θ) can vary over a range of several tens of degrees. For spectral measurements in the

period November – January in the Northern Victoria Land, θ ranges between 50° and 63° (the

actual value for each spectrum can be calculated using the ancillary data Lat, Lon, Date, Local time 7

and a commercial program for ephemerides calculation). Bidirectional calibration factors for

Spectralon panels have been obtained from laboratory experiments (Jackson et al., 1992;

Alberotanza et al., 1993) and can be used to calculate absolute directional/directional reflectance

spectra from the uncalibrated field spectral data.

We decided to leave the user the choice of which calibration factors to use, and did not apply any

directional/directional reflectance calibration to the LILIAN data. Since the calibration consists in

the multiplication for a number usually between 0.8 and 0.98, it would change only to a minor

extent the shape of the spectrum and the depth of absorption features. The latter should therefore be

cleary distinguishable even in the uncalibrated spectral data.

Moreover we executed at several sites measurements in TI mode (see previous section) which, for

smooth and quasi-lambertian targets give a good approximation of the directional/directional

reflectance.

LILIAN, the reflectance spectral library of antarctic lithologies

Introduction

The absolute reflectance spectra calculated according the [1] for all measured surfaces were then

examined for noise evaluation and for a first validation. Some spectra contained considerable

noise or spectral artifacts due to erroneous measurement setting or change in the illumination

conditions between the target and reference radiance measurements, and were discarded.

The remaining spectra form our reflectance spectral library LILIAN. The acronym LILIAN

indicates that most of its content regards antarctic lithologies, although a minor number of non-

geologic antarctic surfaces are also included.

A fundamental step in the creation of a spectral library is the data validation, i.e. the association

of each spectrum to a set of ancillary data pertaining to:

the compositional characteristics of the measured surface;

the geographical and geological characteristics of the site;

the instrument characteristics and all environmental parameters which could have influenced

the measurement.

While information on the site, instrument, and environmental conditions were noted at the

measurement time, the composition of the measured surface was in general only visually

described in the field and representative samples were collected to proceed later to a more

detailed petrographical analysis.

8

Different analysis techniques were used to characterize the mineral composition of the samples.

Igneous rocks characterized by holocrystalline texture allowed for a modal analysis. For these

rocks one thin section (occasionally two when a textural anisotropy was evident) was made,

trying when needed to cut the section through the alteration band. The mineralogical

composition was established using a polarized-light microscope to resolve the optical features of

the different crystals within the rock. Petrographic microscope provided also useful information

of the texture and structure of the rocks. The fraction of each mineralogical species was

calculated using a point-counter to retrieve the modal distribution of the paragenesis. This

method allows also for the estimation of the glass content for the hypocrystalline lavas or the

fraction of vacuoles for pumices and scoria.

For siltstones and claystones a qualitative mineralogical composition was obtained by X-ray

diffrattometry, which allows to discriminate among the various clay minerals.

In the digital library LILIAN, the analysis data have been integrated when possible with pictures

of the thin sections and of the measured surfaces (in a WEB-compliant format) and with

diffractograms (as spreadsheet files). LILIAN spectra and accessory files can be searched,

browsed, or downloaded from the following Web address:

http://www.ingv.it/labtel2/LibrerieSpettral/LILIAN.htm

The LILIAN structure and file formats

We used a rather simple structure for LILIAN, preferring a plain file organization to the use of

data base software packages and complex data formatting.

Three main types of data files are used: spectral data files, wavelength data files and ancillary data

files. Additional information can be stored in image files, diffractometric data files, etc.

The spectral data file contains the absolute reflectance spectrum in the following format (Fig. 3):

Rec. 1 : Name of the spectral data file itself

Rec. 2: Definitions of the numerical fields in the following records, typically: Wavelength, Name of

lithology measured

Rec. 3 to the EOF: central wavelength (in nm) of the spectral band, absolute reflectance value (in

the range 0-1)

A wavelength data file exists for each instrument (or instrument configuration), containing:

Rec. 1: Header record: definitions of the 4 numerical fields in the following records

9

Rec. 2 to EOF: Band number, Central wavelength, Minimum wavelength, Maximum wavelength

(all in µm)

The ancillary data file is named after the target surface and contains its complete description

(Figure 2); it also contains the pointers to the other files: spectral file, wavelength file, image file.

In particular the first 24 records of the ancillary data file contain the site and surface descriptions,

the record # 25 contains the number N of the spectral measurements available for that surface, while

the following N blocks of 26 records contain the instrument and environmental data relative to each

of the N measurements. Each block contains also the pointer to the actual spectral data file, i.e. the

name of the file itself. The file is in ASCII format and each record contains one or more keywords

descriptive of the data fields. For character fields the no data is the slash character (/), for numeric

fields is the number 999. For some character fields the data are coded (Tables 1, 2, 3, 4).

A detailed description of the records of the ancillary data file follows, see sample file in figure 2.

01 FILE_NAME : Contains the name of the ancillary data file itself. The file name (see Figure 2) is composed by

: two characters defining the material type according to Table 1 + first six letters of the

lithology (or else) name + the underscore character + a two-figure number to distinguish

amongst similar lithologies. A three-letter file extension is also attached to identify the project

(e.g. .ant for Antarctic program).

Table 1: Codes of the various materials of the spectral library

Elements, Alloys, Carbides, Nitrides,

Phosphides

01

Sulfides, Selenides, Tellurides, Arsenides,

Antimonides, Bismuthides

02

Halides (Chlorides Fluorides etc.) 03

Oxides 04

Hyroxides 05

Carbonates 06

MINERALS Borates, Nitrates 07

Sulfates, Chromates, Molybdates,

Wolframates

08

Phosphates, Arsenates, Vanadates 09

Nesosilicates 10

Sorosilicates 11

10

Cyclosilicates 12

Inosilicates 13

Phyllosilicates 14

Tectosilicates 15

Intrusive 21

Volcanic 22

ROCKS Metamorphic 23

Sedimentary clastic 24

Sedimentary organic 25

Sedimentary chemical 26

SOILS and PALEOSOILS 30

PLANTS 40

OTHER 50

02 ROCK_NAME : Name of rock, mineral or geologic material (e.g. Limestone,

Hematite, etc.).

03 CLASS : Group of mineral or (only for sedimentary rocks) rock

classification.

04 LOCALITY : Geographical reference of measurement site from large to small

scale. 05 LAT_LON_EL : Latitude and longitude of measurement

site, in decimal degrees, elevation in metres above sea level.

06 GEOL_FORM : Geological formation of measured surface (if

applicable).

07 SAMP_PROVE : Sample provenance if any.

08 REFERENCES : Cartographic and bibliographic source materials.

09 REFERENCES : Continuation of previous record.

10 STRIKE_DIP : Local strike and dip of measured surface.

11 ROCK_COMPO : Mineral composition of the rock ascertained by visual

examination (not present if sample is not a rock or if detailed

petrographical analysis is present at Records 15-22).

12 MIN_FORMUL : Standard mineral formula (for minerals).

13 SAMP_CODE : Code of sample collected from the measured surface and name of

depository institution.

14 ANALYS_WHO : Operator and/or organization who performed the laboratory

analysis.

15 ANALYS_TYP : Type of laboratory analysis (coded as from Table 2).

Table 2: Codes for laboratory analysis

11

XRD X-ray diffraction

EMP Electronic microscope

XRF X-ray fluorescence

PAT Petrographic analysis on thin section

E1 Major elements analysis

E2 Minor elements analysis

16 ANALYS_RES : Results of laboratory analysis.

17 ANALYS_RES : "

18 ANALYS_RES : "

19 ANALYS_RES : "

20 ANALYS_RES : "

21 ANALYS_OUT : Name of files containing additional analysis information.

22 ANALYS_OUT : Same as previous record.

23 RESERVED : Reserved for next versions.

24 RESERVED : Reserved for next versions.

25 NUM_MEAS : Number of spectral measurements associated to this particular

sample/surface. This number is equal to the number of blocks of

26 record which follow from record 26 to the EOF.

26 DATA_TYPE : Type of data in the spectral file (e.g. radiance, reflectance,

emissivity, etc.). If reflectance, its referral is indicated

(e.g. Absolute, Spectralon, etc.).

27 MEAS_WHO : Operator (organization) who executed the measurement.

28 MEAS_TYPE : Type of measurement: FIELD, LAB, HELI, etc., and geometry of

sensor-target configuration (e.g. nadiral viewing: TP, viewing

with minimum phase angle: TI).

29 INSTRUMENT : Make of instrument and model.

30 FOV_&_DIST : Sensor Field-Of-View and distance from target during this

measurement.

31 SPECTR_INT : Effective spectral interval in microns (min-max), and nominal

spectral interval. The first may be narrower due to noise

increase at the ends of the nominal interval.

32 SPECTR_RES : Spectral resolution in microns.

33 SPECTR_BND : Number of spectral channels of instrument. Used to read spectral

and wavelength data files.

34 WHITE_REF : Type of reference panel for reflectance measurements (e.g.

Spectralon).

35 ILL_SOURCE : Code of illumination source for this measurement, see Table 3

Table 3

Code Description

12

SUN Sunlight and skylight

SKY Only skylight

LMP Lamp

36 MEAS_DATE : Date of this measurement dd/mm/yyyy.

37 LOCAL_TIME : Local time of this measurement.

38 METEO_COND : Description of local meteorological conditions.

39 METEO_COND : Continuation of previous record.

40 SAMP_STATE : Code for state of sample for laboratory measurements, see Table

4.

Table 4

Code Description

WHOLE Solid surface of intact sample

POWDER Pulverized sample

41 GRAIN_SIZE : If previous record is POWDER, here is found the min and max

grain size.

42 SURF_DESCR : For WHOLE samples and for in situ measurements, this record

reports a concise description of the measured surface (texture,

roughness, alterations, lichens, etc.).

43 SURF_DESCR : Continuation of previous record.

44 TEMP : Sample temperature (for spectral measurements in the SWIR and TIR) or

air temperature (for radiance/reflectance measurements in the

VNIR)in degrees Celsius

45 DATA_ANALY : Notes regarding the data analysis of the original spectral data

(e.g. averaging, filtering, resampling of data, etc.).

46 DATA_ANALY : Continuation of previous record.

47 WVL_FILE : Name of wavelength data file (file extension .WVL).

48 SPEC_FILE : Name of spectral data file. The file name is identical to the

name of the ancillary data file but the file extension is the

progressive number of the spectrum.

49 NO_REC_INI : Number of initial text records of the spectral data file. In

general 2 such records are present.

50 RD_FORMAT : Format of numerical records in the spectral data file.

51 MEAS_UNIT : Dimensions of quantities for each column of the spectral data

file (e.g. µm, nm, W/m2, mW/cm2, Dimensionless (DMSL), etc.).

Use of reflectance spectra from LILIAN

13

The principal use of LILIAN will be in the interpretation of optical remote sensed images of the

Antarctic continent and in particular of the Northern Victoria Land. Since most of the spectra

pertain to rock surfaces, they could hardly be used for geologic mapping in different

petrographic provinces. Moreover the peculiar surface alteration processes acting in Antarctica

make a comparison with spectra of similar rock types coming from different environments

difficult.

Although commercial satellite data to date available do not attain a high spectral resolution, new

instruments are being tested which dramatically increase the possibilities for lithologic mapping

from space data. The Earth Observing – 1 spacecraft, launched by NASA’s GSFC at the end of

2000, demonstrated the use of an imaging spectrometer with a 10 nm spectral resolution in the

VNIR-SWIR range and a 30-m pixel resolution (http://eo1.gsfc.nasa.gov/Technology

/Hyperion.html). The Hyperion instrument will hopefully be selected for following, more

operational, missions, and its high spectral resolution will certainly favour a full exploitation of

deterministic lithologic mapping, fostering the use of ground-measured spectral libraries.

The high resolution spectra in LILIAN can be used in conjunction with multispectral data, as the

6-band Landsat data, resampling each spectrum to the TM/ETM bandwidths, but the resolving

power for lithologic discrimination will be limited.

We present in Figure 4 an example of a LILIAN spectrum for the sample described in Figure 2.

Figure 5 shows instead the same rock surface sampled by the Exotech radiometer in the MSS

bandwidths and a comparison with the reflectance values in the same bands convolved from the

high resolution spectrum.

Conclusions

We described three spectral measurement surveys carried out in the Northern Victoria Land,

Antarctica, in the framework of the Italian Antarctic Research Program, Project

3d.1‘Spectrometric surveys of Antarctic natural surfaces for an integrated study with remote

sensing data’. We presented the methodology for the measurement of in situ spectral reflectance

over geological surfaces, the preliminary data analysis and the structure of the spectral library

LILIAN, which contains the validated spectra. All the data in LILIAN are public domain and can

be accessed freely via Internet. Their main uses will be to constrain the analysis and geological

interpretation of remote sensing data in the 0.4-2.5 µm wavelength range.

14

Acknowledgments

We thank E. Zilioli for lending the EXOTECH radiometer during the 1994/1995 and 1995/1996

surveys. We are very grateful to all the logistic personnel of the Italian base at Terranova Bay for

the help given in many occasions, and especially to the scouts and helicopter pilots who were very

comprehensive and collaborative during the field measurements. References

Alberotanza L., Canossi I., Pavanati M., Ramasco C., Zibordi G., (1993). Spettri di riflettanza di

superfici naturali: metodologia di misura ed applicazione ad acque lagunari. Rivista Italiana

di Telerilevamento, 2, 17-20.

Baroni C. (1989a). Geomorphological map of the Northern Foothills near the Italian Station (Terra

Nova Bay, Antarctica), 1:20000 scale. Mem. Soc. Geol. It., 33, 195-211.

Bogliolo M., Buongiorno M.F., Cagnati A., Casacchia R., Mazzarini F., Salvatori R., Salvi S.,

Salvini F., Tomasoni R., Valt M., Zilioli E., (1996). Radiometria di superfici glaciologiche e

geologiche in Antartide. ING publication n. 580, 54 pp.

Cagnati A., Mazzarini F., (1995). Report Campagna Antartica 1994-95, Progetto 3b-3b.1

Telerilevamento. Rapporto Campagna Antartica Estate Australe 1994-95, Decima Spedizione,

Progetto Antartide ANT 95/02, 163-167.

Carmignani L., Ghezzo C., Gosso G., Lelli A., Lombardo B., Meccheri M., Montrasio A., Pertusati

P.C., Salvini F. (1989a). Geological map of the area between David and Mariner Glaciers,

(Victoria Land-Antarctica), 1:500000 scale. Mem. Soc. Geol. It., 33.

Carmignani L., Ghezzo C., Gosso G., Lelli A., Lombardo B., Meccheri M., Montrasio A., Pertusati

P.C., Salvini F. (1989b). Geology of the Wilson Terrane in the area between David and

Mariner Glaciers, (Victoria Land-Antarctica). Mem. Soc. Geol. It., 33, 79-99.

Casacchia R., Mazzarini F., Salvatori R., Salvini F., (1999). Rock-type discrimination by field, TM

and SPOT data, Tarn Flat, Antarctica. International Journal of Remote Sensing 20, 2, 403-

420.

GANOVEX Team, (1987). Geological Map of the North Victoria Land, Antarctica (1:500,000).

Explanation Notes. Geol. Jb.., Vol. B66, 7-79.

Ghezzo C., Baldelli C., Biagini R., Carmignani L., Di Vincenzo G., Gosso G., Lelli A., Lombardo

B., Montrasio A., Pertusati P.C., Salvini F. (1989). Granitoids from the David Glacier-Aviator

Glacier segment of the Transantarctic Mountains, North Victoria Land, Antarctica. Mem.

Soc. Geol. It., 33, 143-159.

15

Jackson R.D., Clarke T.R., Moran M.S., (1992). Bidirectional calibration results for 11 spectralon

and BaSO4 reference reflectance panels. Remote Sensing of Environment, 40, 231-239.

Hunt G.R., (1980). Electromagnetic radiation: the communication link in remote sensing. In Siegal

B.S., Gillespie A.R., Remote Sensing in Geology. John Wiley & Sons Inc., Usa.

Hunt G.R., Salisbury J.W., (1970). Visible and near-infrared spectra of minerals and rocks: I

Silicate minerals. Modern geology, 1, 283-300.

Hunt G.R., Salisbury J.W., Lenhoff C.J., (1973). Visible and near-infrared spectra of minerals and

rocks: VII Acidic igneous rocks. Modern geology, 4, 217-224.

Kyle P.R., (1990). McMurdo Volcanic Group western Ross Embayment. In: LeMasurier &

Thompson (eds.), Volcanoes of the Antarctic Plate & Southern Oceans, AGU, Washington,

Antarctic Research Series, 48, 19-145.

Muller P., Schmidt-Thome M., Kreuzer H., Tessenshon F. & Vetter U., (1991). Cenozoic

Peralkaline Magmatism at the Western Margin of the Ross Sea, Antarctica. Mem. Soc. Geol.

It., 46, 315-336.

Salvi S., (1996). Report Campagna Antartica 1995-96, Progetto 3b-3b.1 Telerilevamento. Rapporto

Campagna Antartica Estate Australe 1995-96, Undicesima Spedizione, Progetto Antartide

ANT 96/07, 123-124.

Salvi S., (1998). Report Campagna Antartica 1997-98, Progetto 3b-3d.1 Rilievi spettroradiometrici

di superfici naturali in Antartide per uno studio integrato con dati telerilevati. Rapporto

Campagna Antartica Estate Australe 1997-98, Undicesima Spedizione, Progetto Antartide

ANT 98/02, 139-140.

Salvi S., Bogliolo P.M., Buongiorno M.F., (1997). A reflectance spectral library for Antarctic rock

types in the range 0.4-2.5 µm. In: The Antarctic Region: Geological Evolution and Processes

(C.A.Ricci ed.) Terra Pub, Siena, 1111-1116.

Zibordi G., Meloni G.P. (1991). Correction of bihemispherical reflectance measurements for

noncosine response of 2π steradian optics: a methodology and its application to Antarctic

surfaces. Rem. Sens. Environ., 37, 55-62.

16

0.05

0.07

0.09

0.11

0.13

0.15

0.17

0.190.21

0.23

0.25

400

500

600

700

800

900

1000

1100

1200

1300

1400

1500

1600

1700

1800

1900

2000

2100

2200

2300

2400

2500

Wavelength, nm

Abso

lute

refle

ctan

ce

Helicopter measurementGround measurement

Granodiorite, Mt. Crummer

Figure 1. Comparison between reflectance spectra of the Mt. Crummer granodiorite measured at the ground over a 25

cm circular area (dashed line), and measured from the helicopter over a ~10 m circular area (solid line).

17

01,FILE_NAME :,"21GRANIT_01.ANT" 02,ROCK_NAME :,"Granite" 03,CLASS :,"/" 04,LOCALITY :,"Campo Oasi, Terranova Bay, Northern Victoria Land, Antarctica " 05,LAT_LON_EL :,"- 74.693, 164.101, 50" 06,GEOL_FORM :,"Granite Harbour Intrusives" 07,SAMP_PROVE:,"XIII Italian Antarctic Expedition" 08,REFERENCES :,"/" 09,REFERENCES:,"/" 10,STRIKE_DIP :,"999,99" 11,ROCK_COMPO :,"Leucogranite with centimeter-sized feldspars" 12,MIN_FORMUL :,"/" 13,SAMP_CODE :,"/,/" 14,ANALYS_WHO :,"/,/" 15,ANALYS_TYP :,"/" 16,ANALYS_RES :,"/" 17,ANALYS_RES :,"/" 18,ANALYS_RES :,"/" 19,ANALYS_RES :,"/" 20,ANALYS_RES :,"/" 21,ANALYS_OUT :,"/" 22,ANALYS_OUT :,"/" 23,RESERVED :,"/" 24,RESERVED :,"/" 25,NUM_MEAS :,"1" 26,DATA_TYPE :,"Reflectance, Absolute" 27,MEAS_WHO :,"S. Salvi, Geodesy and Remote Sensing Lab.-Ist. Naz. Geofisica e Vulcanologia" 28,MEAS_TYPE :,"Field,TP" 29,INSTRUMENT :,"Analytical Spectral Device, Fieldspec FR 634 2" 30,FOV_&_DIST :,"25,0.6" 31,SPECTR_INT :,"0.35, 2.50, 0.35, 2.50" 32,SPECTR_RES :,"0.01" 33,SPECTR_BND :,"2151" 34,WHITE_REF :,"Spectralon" 35,ILL_SOURCE :,"Sun" 36,MEAS_DATE :,"12/12/1997" 37,LOCAL_TIME :,"14:35" 38,METEO_COND :,"100 % CLEAR SKY" 39,METEO_COND :,"/" 40,SAMP_STATE :, "/" 41,GRAIN_SIZE :,"999, 999" 42,SURF_DESCR :,"Flat surface with centimeter-sized irregularities due to protruding feldspar crystals" 43,SURF_DESCR :,"with specular facets, limited and sparse yellowish alteration" 44,TEMP :,"7" 45,DATA_ANALY :,"Average of 3 spectra" 46,DATA_ANALY :,"/" 47,WVL_FILE :,"FieldSpec_FR.wvl" 48,SPEC_FILE :,"21GRANIT_01.ant" 49,NO_REC_INI :,"2" 50,RD_FORMAT :,"ASCII 2 floating numbers per record" 51,MEAS_UNIT :,"nm","DMSL"

18 Figure 2. Example of Ancillary data file (see text for explanation

19

21GRANIT_01.1 Wavelength Granite - Campo Oasi 0.35 0.1317 0.351 0.1324 0.352 0.1333 0.353 0.1344 0.354 0.1357 0.355 0.1367 …….. ………

Figure 3: Example of spectral data file.

20

XIII Italian Antarctic ExpeditionSample 21GRANIT_01.ANT, Campo Oasi

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

400

500

600

700

800

900

1000

1100

1200

1300

1400

1500

1600

1700

1800

1900

2000

2100

2200

2300

2400

2500

Wavelength, µm

Abso

lute

Figure 4. Example of a LILIAN reflectance spectrum. Sample 21GRANIT_01.ANT: granite surface at Campo Oasi.

21

Campo Oasi granite

0

0,05

0,1

0,15

0,2

0,25

0,3

0,35

0,4

0,45

0,5

A B C DLandsat MSS channel

Abso

lute

refle

ctan

ce

FieldSpec FR

EXOTEC AX100

Figure 5. Comparison of spectral data sampled by the Exotech radiometer in the MSS bandwidths with the

reflectances convolved in the same bands from the spectrum in Figure 4.

22