RR 332
~l
Resea·rc:h ·Re.p~ort .. J32.
:
RED AND NEAR-INFRARED
SPECTRAL REFLECTANCE OF SNOW
Harold W. O'Brien and Richard H. Munis
M a r<: h 19 7 5 ·
/-.~
PREPARED FOR
U.S. ARMY MATERIEL COMMAND
U.S. ARMY MOBILITY EQUIPMENT RESEARCH AND DEVELOPMENT CENTER
AND
NATIONAL OCEANIC AND ATMOSPHER1C ADMINISTRATION
BY
CORPS OF ENGINE~RS, U.S. ARMY
COLD REGIONS RESEARCH AND ENGINEERING LABORATORY
HANOVER, NEW HAMPSHIRE
APPROVEO FOR PUSI,..IC REL.EASE; DISTRI8~TION UNL.IMITED.
The findings in this report are not to be
construed as an official Department of
the Army position unless so designated
by other authorized documents.
Unclassified
SECURITY CLASSif=·icATION OF THIS PAGE {When Dsta Entered)
REPORT
DOCU~ENTATION
1.· REPORT. NUMBI;R
GOVT ACCESSION NO. 3.
Research Report 332
4.
r
READ INSTRUCTIONS
BEFORE COMPLETING FORM
PAGE
RECIPIENT'S CATAL9G NUMBER
.
TITLE (and Subtitle)
';·
!5.
TYPE OF REPORT & PERIOD COVERED
6.
PERFORMING ORG. REPORT NUMBER
8.
CONTRACT OR GRANT NUMBER(e)
REb ANDNEAR-INFRARED SPECTRA'L RE.FLECTANCE OF SNOW
7.
AUTHOR(a)
USAMERDC-Order No~ A3137
NOAA-P.O. NA-869-73
Harold W. O'Brien and Richard H. Munis
9. PERFORMING ORGANIZATION NAME AND ADDRESS
·U.S. Army Cold Regions Research and Engineering Laboratory
, Hanover, New Hampshire 03755
11.
DA Project 1T 1611 02B52A
Task 02, Work Unit 007
12. REPORT DATE
CONTROLLING OFFICE NAME AND ADDRESS
USAMC
; USAMERDC
US Dept of Commerce, NOAA
:14.
10. PROGRAM ELEMENT, PROJECT, TASK
AREA & WORK UNIT NUMBERS
March 1975
13. NUMBER OF PAGES
22
MONITORING AGENCY NAME &: ADDRESS(If different from Controlllnf Olllce)
15.
SECURITY CLASS. (of thla report)
Unclassified
I Sa. DECLASSIFICATION/DOWNGRADING
SCHEDULE
16.
DISTRIBUTION STATEMENT (of thla Report)
Approved for public release; distribution unlimited.
17.
DISTRIBUTION STATEMENT (of the •batract entered In Bloclc 20, II different from Report)
:
;
18. SUPPLEMENTARY NOTES
\
;
19. KEY WORDS (Continue on reverse side II necesssry and Identify by bloclc number)
Near-infrared spectral reflectance of snow·
Red spectral reflectance of snow
Snow reflectance
2Q,
ABSTRACT (C'otrtlaue
Gal
rever-
.
a.i.a Ef ~ aad. Identify by bloclc number)
The spectral reflectance of snow in the range of 0.60 to 2.50·pm wavelengths was studied in a cold laboratory using
natural snow and simulated preparations of snow. A white barium sulfate powder w~s used as the standard for
comparison. The high reflectance (usually nearly 100%) of fresh natural snow in the visible wavelength declines
rapidly at wavelengths near and beyond 0.80 pro, as the spectral absorption coefficients of ice increase. Aging snow
becomes only somewhat less reflective than fresh snow in the visible region and usually retains a reflectance greater
than 80%. In the near infrared, aging snow tends to become considerably less reflective than fresh snow. The rate
of decline of near-infrared reflectance due to aging !s strongly affected by the history of the snow during aging.
'
<.
DD
FORM
\ JAN 73
1473
EDrTION OF I NOV 65 IS OBSOLETE
Unclassified
SECURiTY CLASSIFICATION OF THtS PAGE (When Deta Entered)
ii
1Joclassified
SECURITY.CLASSIFICATION OF THIS PAGE(When Data Entered)
Snow aged under c-ertain .conditions may retain 90% or so of its reflectance in th~: visible red, yet may be only about
. 1'0% as refle"ctive as the ·original fresh snow: beyond 2.2 pm . .several environmental factors such as ambient temperature and wind effects which contribute to the variability in snow rCfrectance are discussed-.
...
·;
i
.I
I
Unclassified
.SECURITY CLASSIFICATION OF THIS PAGE(WhenData Entered)
iii
PREFACE·
The investigation covered in this report was conducted by Harold W. O'Brien ~nd Dr. Richard
H. Munis, Res'earch Physicists, Physical Sciences Branch; Research bivisio'n, U.s~ Army Cold
Regions Research and Engineering Laboratory (USA CRREL). Technical monitors for age.ncies
sponsoring this research are R. Navarin, U.S. Army Materiel Command (USAMC), D~R. Wiesnet,
National Oceanic and Atmospheric Administration (NOAA), and D. Gee, U.S. Army Mobility
Equipment Research and Development Center (USAMERDC) ..
The research ~as initiated under USA CRREL In-House Laboratory Independent Research
funds and continued under funding by USAMC{DA Project.lT161102B52A,Mobility and
Environmental Research, Task 02,Military Aspects of Cold Regions-Research,_ Work Unit 007,
Weapon Concealment and Signature Control in ColdRegions), by USAMERDC (Order No.
the U;S. Department of.Commerce, National Oceanic and Atmospheric
A3137),, and
·
·
· ·
Adminis\ration (NOAAY(Purchase Order NA-869-7~).
by
The authors wish to thank Dr. E.P. McClain, Dr. D.F. McGinnis, and D.R. ,Wiesnet of :f\JOAA,
for their helpful comments and Dr. Y.C. Yen· and Capt<:1in J.W. Lan·e of USA CRRELfor their
· ·
critical reviews of the report.
· The contents of this report are not to be used for advertising, publication, or promotionaL
purposes. Citation of trade names does not constitute an official endorsement or· approval of
the use of such commercial products.
/
iv
j
· ~- CONTENTS
•·
\
;
.~
;.
: .j.
··
Abstract .·:... ."........... ~ .-.-.: ~ .............. :................. ·..... .- .. ;·... ;-. :.. .-~. :... :..·........... .-.- .. :... ,.................... ~.
. Preface ...... :.... :;; ........ :........ .-..:.. :..·: ..............................................:........... :.... :: .... ;............ .
Introduction ...... ·.. .- ........ :~; .. :..... :...........;... ; ........................... : .......................... ::.; ......... ;:..
· Spettrophot"omettic ·mettlods ...... .-.- ....-...·... ~· ........-.......... .-.. ~~·········-······.-···············~·~··· ········~
Experimental apparatus·.:.:.: .... :..........'.............. .-: . .- ................ ;; ... ~ ............... :............... ;
Reflectance standard.................................................................................................
Page·
.i .
iii ...
I.
.2
2
2
Sn~o:~~f~2~~~~~~6~~::::::::::::: : : : : : : : : : : : : : : : : : : : : : : : : : : ;: : : : : : : : : : : : : : : : ;~
Preparations ofsimulated snow conditions ... :: .. :.: .. :...... :.: .. :.......... :... :: .......... :; .... :: ... ·..
3·
Dai~::~:~~£::::·~:~·~::::::;·::::: : : : : : : : : ,: ;: : : : : : : : : ·: :·:·:·:;: : : ·: : : ·: :·: ;:': : : .·: : : : · · .·· .:·
Experimental"erro.rs .. :............ :........... :................ .".._.~ .... ~ ...... :........ '.:....; ...... .'.: .. ~.........
D~s·cussion of res.~~t~ .... ·: .... "·\· ..'.. :· .. ·: .... :····:·· ..... :_.: ... ~ ... :..........: ...... ~: ....... :...... :.: ....... ~ ....... ··
4
;.5 5- ·
5_
Parameters influeriQirig the reflectance of snow ............... .' ...... :.. ··~··:............_............
Variation of snow reflectance with wavelength .... ~ .... :: ......... :... ~; .........'.... _.......... :....... ·
Changes .in snow reflectance related to natural aging ........ :..............·..................... :....
6
Effect of melting and refreezing on snow reflectance., .......... ,............ :.... ,............ :····
8
Reflectance of ice-glazed snow ······~·······:,.~ ..............-... ~ ............ ;....:....... .' ............ :... ~.. · I2
.. Effect of drifting and wind-compaction on snow reflectance ..................... .............
13
Snow. refl~ct~nc~ a·t various angles ~ .............. ~· .. :.~ ........... :....... .'.. ::·.: ........... ·...._.... :........ ' 15
Relationshir) o(snow reflect'ance to snow density·......... .'.............. ~<.:
l6
. Dependence of reflectance on microstmctur~ of snow .......... :.'.· ... : .......·..... ~ ............'.. · I6:
Reflections. on reflectance· .... :....• :.:;...........•.. ;............... ~:: ....•.... ,.......·_. ....... _.: ............~ .. :.......
17
Literature·;_cited- ... :•........ :··.-·•·:····;~·.. :........•.... .-.... ,.:; . .-. .- ................. , .................. ,. ····~· .... :......... :::. ·18 ·
...:... : ·....... ;..'.
i
ILLUSTRATIONS
Figure
,
· . ':
· I . · ··
· · ,. i :
·
~
_--_ · • _ _
·
I. InstrumentatiOn for spectral reflectance measurements ......................................
2·
· 2. Typical spectral reflectance c±urve for snow ............. ;...... :·........... :................ :....... · . 6
6
3. Changes in snow reflectance .with natural aging-.(5°- 5°) ......................................
:4. Reflectance of aging snow re.lative to fresher snow ......... ,.. _....._.;..........................
7
5. Changes in snow reflecta~ce ;with nat~ral aging(0°- 30°)............................... ,.:..
}
6. Effects of temperature cond~tions on sift~d snow reflect~nce (5°- 5°) .. ,._............
. 8.
7. Effect~ of temperature con<;lltions~ on sifted snow reflectance (0°- 30°) ..... :........
9
'Effec.ts of temperature condftion·s on old snow reflectance (5°- 5°) .. ::: ........ ~.....
9
9. Effects of temperature conditions on old snow reflectance (0°- 30°) ....... :.........
10
I
.
10. Reflectance of melting and r'efrozen snow relative to original cold sifted snow
. (ra tiQ of curve~ -frotn Fig.l 6) .............. :·. ··:···-~~ .... ,.,,:· .._... ,.,:-:.~·~,-..i;~ ~·:·~· -~ .•••.••••. , ••• . ....
19.
II .. Reflectance . of melting. anq refrozen snow relative to original cold sifted sno~
(ratio of curves from Fig.; 7) ......................................................... :................ .11
12. Reflectance' oftnelting aild refrozen snow relative to original 'old cold snow (ratio
• .· '-· .. · ofctitves frbm Fig.'"8) ... ~ ...... ~.'.. '... :.... ~·:.... ::: ......... :..... : .......,-... ~.'.. ~ ......... :.. :~ .. ::..... _1 1'
.. 13. ~Refle.ttance· o{melting and refrozerl snbw relative to o~riginal old: coid shoW (ratio· ' . ~ ~ of CtH~es from' Fi{ 9) .: ...... :.'.:_.:·.. :...... :...~.:_.·~······-'······: ... :.:.·::·.:.:.': .... :::·........ ~: ... .-:: .... £ 12 .:
14. Reflectance of ice-glazed snow c'rust .. :.:.::·.. :.:..' ..... ~ ..'... :... :~.:~;.::.: ............ :... :.. :.... ::::·· · · '13
15. Effect of snow drifting on snow reflectance........................................................
13
16. Comparison ofrelative reflectance of snow at various source-detector angles ... .-..
14
1
·s.
RED AND NEAR-INFRARED SPECTRAL REFLECTANCE OF SNOW
by
Harold W:O'Brien and Richard H. Munis
INTRODUCTION
.
.
.
-
.
.
Approximately one sixth of the surface of the earth is-pQvered b.Y spo~ arid ice. Many areas of endeavor
in cold regions, both civil and military, require greater knowledge of the spectral characteristics of snow
cover. This project was initiated in response to engineering problems related to the solar energy reflected
from snow covering an underlying structure. Subsequently, interest developed in the application of these
studies to problems in environmental satellite technology and in military surveillance and/or camouflage
activities.
..
Although operational environmental satellites have obtained some snow data remotely, or1e problem
facing the satellite hydrologist is that of understanding the significance of variations in the spectral.refl~ct-.
an~;;eof the snow cover in order to completely evaluate t4e remote observations. The problem becomes
.
more complicat:ed b'ecause'when snow is melting its spec'tral reflectance can change 'drastically on day-to~
day basis.
·.
·
·
·
·
· ·'
·-
a
Military operations 'are limi.ted by the ability to move equipment and personnel without detection. Tre~
quently, the prob.ability o.f detection depends upon the contrast in reflection of solar radiatio~ from objects
compared with that fro~ sri ow covered backgrounds. In~ o-rder to make use of reflective characteristics for
detection or for dmi.ouflage, it is necessary to determine the spectral characteristics of snow~
Very few data have been generated in the past on the spectral reflectance of snow in t}i.e redandnearinfrared ranges. Several studies have been p1ade on the reflectance of snow; howeve.r, generally, optical.
filters have been used to define certain rather broad spectral regions. For example, Hulbert.{l928) investi~
gated approximately 'the same. wavelengths as those investiga'ted in this study, but examined 'the spectr~i
region ih two bands: 0~4 to 0~8 ,urn arid 0.8 to 2.6 ,urn. The studies of Mellor (1965) were ~onducted using ·
narrow-pass interference filters, but extended only to 0.7 ,urn in wavelength. Other studies of interest were
reviewed by the University of Minnesota ( 19 51).
This program was initiated to determine the spectral characteristics of snow in the n~ar infrared {1.00 to
2.50-,um wavelength range) particularly ~ith respect to ·changes in the spectral reflecta~ce bf sno'w cover
during the natural agingproc~ss. It soon became 3:pp.~1·rent that ·more .definitive::parameters thart "aging" had ·
to be considered, Such ~haraeteristics as de~ shy of the snow cover, the history of nat~ral aging· condition's;
and possibly' the snow crystal type app~ared topiay a: significant tole in determining the' spectral reflectance
of the snow. Middleton et al. {1952) considered some of the peculiarities of snow as a··refl'ectin'g·surface ..
mea~ur.e-
During the second winter of the project, the spectrum investigated was expanded to include
ments in. the 0.60 to l.OO~J,Lm wavelength region:
.
Since the authors had no control over naturai agmg conditions due to weather, cert~in co~ditions.of:
aging such as melting·and refreezing were simulated in the cold laboratory.
·
'
.
'
This investigatiori was conduct.~dprincip~lly during the~per~od Janua'ry I972 to Decenibe;r 1973
U.S~ Army Cold Reg,ions R~search arid' Engineering· Labo;atory.
.':
ai the"·
2
REDA!VD NHAR-IJVF.RAREJ)
$PECTR~L 1~EFLt·(:T~NCE
OFS/XOW
Beam· angle and detector adjustabre
w~th resp~ct t.o n_ormal
Broad Spectrum
Monochromator
·Freezer
'SnoVI Sample
· Adjustable
Figure 1. lnstrumelitqt[bn for spectralrcjlectmlce·measurements. ·
SPECTROPHOTOMETRIC METHODS
Experimental apparatus .
The physical arra~get~le.nt of experimental apparatus is depicted in Figure L The optical configuration of .
a Perkin~Elmer Model ·E-13 spectrometer was altered for utilization as a reflectometer.
·
Two monochromator ·gratings were employed: a 1440-line grating for the 0.600 to l.OOO~~m-wavelength
range and a ·576-line grating for the 1.000 to 2.500-pm-wavele.ngth rfin.ge .. In each case, optical. filters were
used to eliminate highe~ order re·flections. A tungsten so.urcefchoppe~ at 13 cycles/sec· and a read sblfide
detector were used for all measiuemerits. Gold first-surface -~1irrors were used to collimate and direct the
:beam einergirig from the monoch.romator. An open~top coii1m·ercial·freezer display case was used as a cold
chamber to mai!1tain snow samples during measurements. The output from the de.tector was r¢corded on
~
a Leeds and Northrup strip chart recorder.
Init'ially the apparatus was set up in a warin laboratory with poorly controlled room -t~mperatures: In
Novermber 1972 the equipme.nt was moved to a cold laboratory with good temperature stabilization·at
10°C.
.
.
.
<t-
..
. .
i.
Reflectance :standard
.
.
. II •.. -.
.
.
.
Eastman White Reflectance Stand~ud (Lot #20 1-2), a white bariutil sulfate powder with excellent reflective.qualities, was used as a standard with which to conipare the snow reflectance measurements. The
·
st'anda.rd ·was pr.epared by filling a milled1Plexigla~ cylihdrical contaii1er (6-in. inside diamete-r and 'l1 in .
.depth) with an amount of the BaS0 4 po~der which, when smoothed and coii1pressed, would conform with
East.man Kodak's ·recommendations on standard density.
1
Procedure
!..
to
A removable platform was inserted at a fixed height, commensurate with"ih:it' heigiH ·which a 'snow .·
sample could.reasonably .be, raised ~ithin the, freezer witl1out SIJbjecti_ng.. the snt?w.~sample to und,ue drafts
. ~nd the,probabil'ity of mel.t,i;1g. The. incident beam was adj.usted to:tl~~ decsired a~gie a,~d ~he;WJ'\ite Reflectai1ce Standard ·placed on the platform, such that the incident beam was centered on the standard powder
suiJace.·:· Tlle-'d'eiectorar,'n:was adjtfs·ted'fdr the desired ;an·gle"c>f:ieflectance;with: the· Whife'Refle~tance
Sta~:1dardas clos~ -'as p'ossib!etto' the ceiiter'ofthe fi~ld. o!f v'iew o,f th•e detector. Aspectra,l scan
·~ruh" was
ot
:.F_,
. _!;.
:
·;: '
; r
1 . . •~ .: ' l
'
i .. ••
3
RED AND NEAR-INFRARED SPECTRAL'REFLECTANCEOFSNOW
then made of. the reflectance in the range of the installed grating. Upon completion of the scan; the other
grating was inserted and a Spec:tralscan was made in its :wavefength rari·ge. Because of some temporal varia. tion of instrument sensitivity, it was fou~d advisable to clwose a"test' wavelength'~ at which to check the-·
recorded signal strength before and after each run.
·
'•
:.:-:
·Similar spectral scans were run on the standard for each combination of angles of incidence and reflection
desired. To eliminate the necessity of repetition of the terms and to simplify future·~ discu'ssion, abbreviated
. designation's signifying an angle ·of incidence-hyphen-angle of reflection will be used in this paper. -For ex'~
ample, "5° -5°" indicates·an-angle of incidence of 5° from normal and an equal angle·.of reflection;.''0°~30°"
indicates that the source beam was at normal.incidence to the reflective surface (staridard' or snow) and that
the reflectance was measured with the detector set at an· angle of30° .to the normal. ·Because of the physfcal
size of the··mirror mounts and detector housing, 5°-5° was the nearest equal-angle reflectance obtainable~
Although more measurements were made at 5°-5° and 0°-30° than at any other angles~ several measurements
were made-at l0°-l0\0°~l5°,0°~20°;and0°-::25\
·
.. -Sitnilar me-asurements were made on each of ~e snow ~amples: Th~ platform holding thesta_ndard ~as_·
removed and the open-top freezer was rolled into position under the mifror system. A screened box con- ·
taining the snow sample was placed on a "lab-jack" in the freezer, leveled, and raised until the snow surface
was in the same position previously occupied by the surface of the BaS0 4 standard. Spectral scans were
run in the same way as on the standard.
Preliminary tests indicated that reflectance measurements were not detectably influenced by the substrate
materials (wood and/or aluminum screening on the bottom of the box) when covered with ab~out 5 or'6 em
of snow. Theref~re, aU measuremeptswere made with a_m~nimum snow ~epth of 10 em.
Natural collection
..
Snow -~amples- ~ere c~lle~ted by placing ·open boxes on 'the grcmnd in a reasonably level area aw.ay: fro111_
the labora-~oiy buildings and allowing show .to fall naturally into. ~he boxes. N~ar the end: of a .fresh. sno~fall,
a precooled, 21-frl home-type deep freeze cabinet was 'rolled, on casters, outside ofthe.laboratory.building
and o~e or 'more of the s~ow boxeswerecarefullyJifted and car~ied to the f~eezer. l'he samples·w.ere then
transport~d in the fr~ezer into the cold laborato-ry.
-·
.
- .· .. - On subsequent days, proVided there had been no precipitation in the meantime, other boxes of naturally
aged snow from the same snowfall were similarly collected. Originally, pine boxes were :used for the::
collections, bu_t it was noted that the sides of the pine boxes caused unnaturally rapid-deterior~tion of the
snow samples. Therefore, light wood frame: boxes were constr~cted and covered on thr~e sideswiJh alumi~
num screen. Upon aging, the snow collected in these boxes showed no signifi~ant differences froJD the .
surrounding snow cover outside the boxes.·
··
Preparations ·of simulated snow conditions>
·.;.
Sufficiently recent fresh snow samples-were not al~ays readily available for measurement. For some
series of tests; older snow was ground in a 'food mill and allowed to fall;loosely into regular snow sampling
boxes. This-milled snow appeared to simulate ri(ltural wind-drifted snow quite closely 'indensity,hardness
and texture. Generally~ the temperature ofthe snow, whether natur·ally fresh or. milled: was held at about
'--19°C.duiing'storage and measurements.
.
. .
.
The aging ofsnow w~s simulated. by taking fresh snow or milled snow and allowing-various degree~ of
warming with the snow _sample still in position for reflectance measurements. Then; sequential spectral ·
reflectimte scans were run While holding.the ambient air temperature as steady as 'possible at each of ~eYeral .
levels: below, m~ar a·nd above the freezing point.
··
.
·
>
•
••
•
4
(:.;;'
Rf.lJ AND NEAR-INFRARED
SPECTRAL
REFLECTAN€E
OF SNOW...
·.....
. . . · -:
.
- ·.
~
~
..
-~-
.,
··~··.:
,·
Snow observations
;
:.:·
.
-: .
. ,•
~
.
~
Early in the project it wa·s intended to do a rather,siinple determimition'.oLthe·generafinfluence.of.a.girig
on the spectral reflectance of snow. As the study progress~d;'iLbecame. obvib.~s thahadditignaLinfornJation
on. the physical characteristics of the snow wo_uld be·;desirable."-Thereafter,()bservations~wer(tgenerally.·made
. of the snow density, hardness and gross particle size:usin·g· usual:meteorologkal.methods.' In addition,. i~- · ·
·occasional Formvar replicas* were made of the snow samples,~ At·the,itime:of..retrieval'Ofsno~·samples from
. the field, observations were made in natural snow cover adjaqent·to the.:colleCtion·boxes:tO av·oid.deshuctive
testing of intended samples. (Early tests indicated no difference between~the:characteristics::qf~~SriQW wfthin"
and those .of "snow outside." the sampl~ boxes.) On othef-·.types:oftsa.inple~;, su'ch ·as the(IlJ'elted~refrozen:· . ,.
samples,' observations were occasionally made on an ~.re~pf,:~he. sap1_ple,\Ve!J~9.!!.! 9f th~,light b~ai_n; qr_~I.1en
possible, observations were made of the sample aft~~:~.9,~r~'':ig n1ps1 w,ere·'c·~~-~kt_ed'.a#.~' ~h~ safri'r,Ie_.·.~.a,(n<?
Jonge_r needed.
· · ' ·.· ~: .·
··
· · · '' '
· ·· · ·
·
:
:~
...
~:;,:.
·._.··.. . ·:.::'-t: ...~.~:-.·~,:.;:.. ,·
._J)
'".:.;-
t •••
:; .
DA. TA ANALYSIS
; i~ ...
~;f.:
,.· ... --- ..
Data reductimi
.
.
:
.
!·f~j-.,. •.;~-- ..
,---- ,;' ··.: ~. :;,· ..
:_~''*'_ .·:'.-:.:·_ ..
~
! . . ,··.· ...:J:,_· . ·•
i.t· .. '.
-·~~-··._.--
:' .. ·;::'.
The strip charts of the spectrophotometer output ·-were digitized ori a Ge·ro·e·r Dim tal' Data Reduction
System .. Corrections for drift in .signal strength were formulated by considering the signal strength at standard
·check points (e.g. BaS0 4 standard.checked at prede,t~~Ip_in~'!: tx.~(~_~velengths and gain) and a curve of signal
strength versus time was plotted for the period of tit11e· covering a"s-equence of runs. From this curve the
.
signal strength before and after each ruwcould be interpolated and corrections applied .... ; '_
.
.. :'~'.:-:.-:··
j:.-r•.
A Digital Equipment Corporation Lab 8/e compl!t~r w.a~_u_tiliz,eq to_peJ(<?.nn t~~ a!1_~lys~~:;:l3qt~ st.~npard
and snow runs were corrected for drift and normaliz~-a-'to:.th:~\ame ga'it.'Each ~now.niK·watill~.n coliipared
with the appropriate standard run by computing and pl~ttirig d~e'"r~·tib 'bt~rio~d~nkct~·D<;e\&h~iidatd ;eflectance point by point throughout' th~ portion of th't'sr.~:dfr~fn\i}ve~r~_il):~ .!~e/~~:ri·;. ·_Y]l.~~-~.iti.ok.:~s()~{'~~bw
runs were compared with other snow rims by the san1J'ill.~:th.o~.(i,._e:.~,me1f111g·ar~'!~.~rB~e-~ ~h9\V_,\\f~r~--S~?'n~1~­
pared with the original cold snow samples).
J ~·~-: · ··'
~-·-· ,_.r...
· · · - · ~-~· ·• ~·-- ···-·
!
I
Experimental errors
.: ;:~;~' ..-
:;. · .-. , .; __ ·'·'
.. .
'· .,_.
· ·rwo experimental discr~pancies, nol promi.nentl~ e~vict~i1~: i~c~~cisl fig ore~',; a~'reat~d li{ ihd ;~nJl)r~e·s :6( . _the form of discontinu~ti~s in the r~_p·~-~{i~9~ ·?:~:~~~~~,~~~t(9q.'ffi~.:~h{i.gai,~ ai6,'~~,4).:f~ :4m.
1
s~me sampl.es in
.
The first of these expe~imental artifacts appeared as a slight mismatch;in,the;magnitude ofthe:relative_,,:
reflectance at the terinin.ation of the 0.60 to j .oo~}lm run compared with that 'at the beginning of the 1.00
to 2.50-J.Lm run. At 1.00 JliTI, each of these runs is near the.limitof,usefulness~of the respective.gniting;:ariCl
there is a tendency to dismiss the discrepancy as· beil}_~,d~~..t9 .~r.cre~s~d ~xpe_riJ!lent_a.l ~~ro~ near_t~ese J,iiJ1itS.
•
,'~~~-._.'~I·.~~·.'
• • .··~·;~:·:~-
.\•":·1.~
•··~-
·. ~
•
..
·.
~',,
1; ,'
'••\>,
The second experin~ental_artifact, when, present, .occu[,red: at 1~7S~Jim ,wavelength ai1d:c6nsist~d of a· ''pip"
in the curve, sometimes accompanied by a very siigh.t< shift.in the trend of the curve. :This~·was·b~e.c~~se·:.theie
was a filter change at L 7 5 Jlm and S<?me runs were made.in two .parts~ii:e·;;,from t;OO.Jo 'L7S•J.Lm wavelength
and from 1. 75 to 2:;50-J.tm wavelength.
This procedure
was follo\Yed,mostly::in
.earlier~r~ns; w~en insthiinent
.
.
..
.
.
";."
:
··*In the replication method employed,-glass miCroscope slide'sa)eccriat'ectlW·iifil'fh~rsi6n',~at'~rribient't~·rrip'e'ra'iuresbe'i~\v
0° C, in a 3% solution of Formvar in ethylene dichloride. Sonfe'slides are· exp9sed to; the \Veather;.a!lowing'a(fe\\'·Silow+ r:. ~'
flakes to fall on the coating of fresh Formvar solution. Oth~r ,f~e.~tt,l_}', ~o~t.~d .sl,i,de~;are ~.~N,lyplace,d: (~~~ ,~p~n -~n ~:~n~'Y;
, surface_~· The ,Formvar solution tends to;en:v;{!lop. snow crys,tal~ touche,? b.Y :t~c~ ~<]a,t~~g. T~e. s}id~~- ar~ .~h~n -~~P~_ ~~ a j , •• 1
. prote,c.te.d_pla~e, belpw 0°C, until the, Fonpvar_coating har~em~. Subsecjuenlremo,vaJ·of the'iC~, preferatily_ by:s_utHimation
'· rather than by· melting, ll~aves a Forrrivar replica of the snow crystal structure which may be stored indefinitely for future
:.:mici-o~c'opic~t~d~ .. -.'" ,, ··
-~ ·
:· · ··· ·~.· · ·· ·
--~ ·
·
·.·· · .,
····;._ · ·. ,·r.·· -:·. · ·
., · ·
,·
:' -
·~ ;
.
.
.
--···-- __.:._..,,~. -. ·· ,.,..,."·..,··-REJJ~NFJ·NEAR-INFRARED SPECTRAL REFLECTANCE OF SNOW
'
. ::::: :' . '....·.:~·;·..
' \_-·,.__,:.··_ · ...... -.~~·:>;! ·, . . ·.·.
:·~.- :'\:.'7,_·:'' :_;•:,·_, ..•..~ -~
;~
.·'-:
~' ..
...~·-'
-..
·.,._
5
. _· .
. sensitivity vaihitionsdue"·to fhictuatiiig room teinp'erattires we're'causing considerable difficulty. in deterIllinJng drift corrections. The .two·partial·.runswere then analyzed as separate runs, _each wit.h its own drift
. ;"correction'>;~Y ~rror.trt deterrlii~ing th:h~e:riotre·ction·s could cause discontinuity·:at. tha(point. In later work,
it was possible to stabilize the room te.mperature.arid minimize this problem~ Stllt an occasional ''pip"
appeared at 1.75 JJ.m because of th~ inconsistency -of the operator in digitizing ·the strip-chart curve across_
the. filter change.
·,
·
When two or more runs were rriade on the same sample under ess~ntially constant conditions;th;e results
we~~ reproducible for the most part within about 1% with occasional deviations of about 3%. The.greatest
po~sibilities of error occurred at the lirriits of each run, as mentioned previously, and in situations where ·
there was a considerable change in signal strength during a run (such as the 'thermal drift previously: mentioned).
off by as
In ~orne instan·ces, hiitial interpolations of the "signal strength versus time" curve appeared to
mu.ch as 10%. Such errors," for the most part, caused rather obvious and nonreproducible irregl.llarifles and/or
discontinuities in the n!flectance curves (in contrast to the rather predictable artifaCts at 1.00-~m and ·1.7-5/J.ffi wavelength) .. The original rec~rd of experimental procedure and resulting data was then reexamined for
procedural or mathematical errors and ·corrected· or rejected as appropriate.
·
· ·
be
It is_d)fficult. toassign very specific qu~ntitative limits to:the experimental error. The au~hors feel that·
the probable error-is apprciximately.•3% throughout most of the spectrum, \vith increasing errot (up to
about 5%) near 0.60 J..UTI and 1.00 JJ.m due to causes previously discussed. The 5% figure would also apply
to the 1.75-JJ.m region in cases·where the runswere:segmented. The probability of error at 1.75 JJ.m may be
subjectively judged by _the_ "smQothness" .of the curve in that region, disreg~rdin.g the sharp "pip," if present,
since the ''pip" indicates only an operator error in a. single data point. Near 2.50 JJ.m, the probable error
becomes increasingly worse due to grating li!Jlitations compounded by the necessity of taking the ratio of
_two sm~l refle.ctances. A r_easonablejudgment would probably be that the error tends to increase from near
3% at 2.30 JJ.m to about 5% near 2.40 JJ.m, becom-ing g'radually worse to a_round 30% at 2.50 JJ.In.
DISCUSSION OF RESULTS
Parameters influencing the reflectance of snow
Anumber of variables: interrelate to determine the spectral reflectance of snow .. Some of tlu!se factors
are basic physical parameters, such as the coefficients of absorptio'n of ice, and the size distributions of the
individual ice crystals. The effects of both of these are wavelength dependent. Other parameters are not as.
easiiy defined and represent environl!lental conditions which influence the structure arid opt-ical character~
istics of a natural snow cover. Middleton et al. (1952) give an excellent account of some of the variables
·
involved in det.ermihing the character of a snow surface.
It is vi~tuallyirhpossible'to separate and examine each individual factor 'which determine{the' spectral
reflectance of snow. An attempt has been made-here 'to discuss some of the significant parameters .and,
where possible, to present res!l_l~s \Vhi~~ rray be ,pre.s~med to illustrat~ tre predominant influences.
Vari~tion of_:~now reflec~ance
with
wavel~~gth
A .typical example of the spectral-reflectance ofsiw~ is shown ·in Figure 2: )n.ge!le.ral,·the ·~r~flectance -is
high and relatively indep.endent of wavelength in the red end of the visible spectrum (0.6to b. 7 JJ.m). ·
Reflectance tends to decline in the near infrared region until,.at about 1.03 JJ.m., a :slight.gain:in reflectance
occurs, resulting in a minor peak at appro~mately 1.09.to l.lO pm .. This coincides with the behavior: pre~
"dieted by Dunkle 'and Beva~·s (I~ 56).'. frorii 1J fo 1.5 phi the reflectan~e de~reas~~ r~pidly exc~pt that at
ab-out 1~2Sjo l.35 :JJ.iri the i_efledance:_ie:v~ls of{artd even ·makes a slight' recovt!.rY. :·T,he ·reflect~hce exhibits
peaks again' atabout 1.83 -~fri:arid ~(about2:24· inTI with a ve-ry 'strong depression' ·of ieflectailce· arouhd 1 ;95
to 2;.05 J.i.m: :These peaks ot refle.ct~n¢e, alth.ough substantial compared with the· iyflectahce in the regions
aroqnd ·J.S p.rrr an.d, ~-~() ~-trn, are small conwared with the reflectance in _the· red spect~um.and in :Uie near"·
, .tinrrared:
.. _
· ·
. . . . · , _1
·· · . ·
-
f
;,.
-
-.
. .
6
.
:
.
.
,.
.
. ··..
.. .
,.
.
.
.
.
NEAR~INFRAREDSPECTRAL-kEFLECTANCE OF SNOW
- . REDAND
SpeCimen no. 730212
Source-de_teet or· 0° - 30°
· ·-- · · -· Sno~-_d>iidition '"Cold, ·sifted, "siigar.,:consistency
. Snow .density 0.357 gfcm 3
·
Snow hardness -~.5 gfcm 2
.1
t
·:.
..
::.-;··
. FigUre
..\
i
2.2
2.4
2 .. 2
2.4
Tyfical spectral rejlectanc·e curve for snow.
...
..
1.8
. --. 2.0 -,
Wovelenoth_, J-Lm
i
Figure3. Chan~es in snow reflectance with natural aging ( 5Cil -5°).
_ Th~ ove~all shape of the reflectance curve appears to be inversely related to 'the .spectral absorption
coefficients of ice (as given by Irvine andPollack 1968).
-.-Changes in-snow.reflectance related
.
:
~'
'
,.
~b;nattiral-~~g.:i.
~
~.
..-
~.f· j
The spectral reflectance of snow iri ·t.he r·ed-' and near~infrared regions· exhibits rather predictable r;:hanges · ·
as t)le sno~ agr.snat:ur;~lly ;: J~h~. l}i~torx . ~f_th~ .~n?_w.,quri~~- ~~:~~ iijfl}J;~J1~e,~;~e ~~gree an.d rate of change
'of reflectance but certain conclusi~~s ~~n P~. drawn <;<?ncerning the general nature of the changes.
Figure J illustrates the changes in reflectance of light snow which fell during the evening of 20-January
1972. A fresh sa,mple (curv.e A) was collected on the morning of the 21st. The weather. was generally overcast and below freezing (about -2°C) during the time between snowfall and. sample collection. The denslty
of the fresh sample was 0.097 g/cm 3 • -
7
RED AND NEAR-INFRARED SPECTRAL REFLECTANCE OF SNOW
· Specimen no.
720 I l l .
A ..
B ..
.C
Source-detector
5° -5°
· Snow condition Natural aging 14 hours 44 hours 70 houn
Snow density
gfcm'
0.097.O.l.f)4
·0.347
Snow hardness
gjcm 2
~4·:5 .
....4
SO+
,-·.-
•c
u
•
.u
0
......
..
~
:>
44 Hour. Natural Agtni
~
·~
....... -···-.
. •":. ~ 70
·.· .. ·
Haur~ Natural Alinl
·· ..·······-··= ........... .
·o.~
WavelenQth, p.m
Figure 4. Reflectance of aging snow relative tf! fresher ( 14-hour) snow (5° - 5°).
Specimen no.
72122
A
B
Source-detector·· . 0°- 30° .
. ...
Snow condition Natural aging Nearly fresh -2 days.
Snow density
gfcm 3 ·
-0.137
-0.21-'6
·Snow hardness
gfcm 2
Negligible.
..... 30
·. -
'~
..
2
·A
~
0
. ~
. g•
.
i
~
u
. 1.0.
'.
WavtltnQth, p.m
Figure 5. Changes in snow r~fled(mcewith natu'ral aging (0
6
: ...
;. 30°).'
s)
was ~etri~~ed ·rrorri !it~ .
Thirty hours later, on the afternoon or22 Janmhy·, a second sample (curve
natural snow cover remaining from the same snowfall. Although the weather was· clear and ~old '(n~ve~
. above freezing and falling·to -20;5°C in the night), the measured density had increased·· slightly to 0;104
.
.
.. ..
g/cm 3 •
A third sample (cu.rve C) was taken onthe afternoon of23 Janu~ry, representing riearl}"th~ee.days:~f'
. ilatudtl aging of the sarrie s'nowfall: This time, however, th~,ambient air temperature had warmed t.b about
+7°C dui'lng the last day ofagin·g an:d the density of thesn~w covet was trteaiv:red asi0.34J:·g/cm·3 ,''! -':.
•
•
••
•
.
•
•
•• '
·••• ; • -~
•
•
t
Point-by-point
ratios of.the reflectance of the. aged.. samples to the fresh sample are; .shown
in'Figure 4.
··-,; ..
. .
.
_... . .. ··:
..Another illustration;ofthe changes in snow reflectance with (lging is shOwn in Figures; .. ~·urve A shows
the. reflectanc~. of freshly fallen s~ow {22 Dec 1972) and. curve.B th,at ofsnow which had. aged naturally. for
two days with ambient temperatures hovering above and below freezing.
.
:
:
~ ~
.
.
.
'
·.
,;
.: '
-',
.l
:}..
.
"'.: 'REDANDNEAR~INFRAREn'SPEcTRALREFLECIY4NCHOF SNOW
8
Spec~men no. 2J~9}08o
Source-dete~tor
5, _- 5
Snow cpridition· ··S~fted
Density Hardness
gfcm 3 ;,_ gfcm2
Curve ..
A .
B .. C. .
D
E .·
F
. 0.8
0.366
o:366·:
0.366
0.525 .
0.525 . -60,000
1.4
1.2
1.0
Origin'al cold sifted snow
Same,. cool, no melting
Same, cold again
Slight surface melting
Sur.face soft and wet
Refrozen
WaveltnQth, p.m
Figure 6. c Effects of temperature conditio1Js on.sifted sriow reflectance (5° -5°).
The significance of these graphs will be more apparent after a discussion of th_e spectral reflectance of
artificially melted and refrozen snoW;.
· ·. · . ·
·
· .
Effect of me~ting and refreezing on snow reflectance
Several ~xperiments were performed to determine the effect 'or inelti.ng and refreezing on the spectral
reflectance_of snow~ Since the· oriJ.y snow readily available at· the time\vas somewhat aged and crusty, snow
samples were prepared by sifting the old snow through a· food milL The resulting snow was very similar
in appearall'ce, texture, and physical properties to naturally windblown snow. The series of runs at 5° -5°
shown in Figure 6 depicts the sequence of events which followed. The first run was m~de with freshly milled
snow continuously cold [about -l9°C (curve A)] ..The· ambient temperature was then raised to near melting
and the snow temperature held around -1.5 to -Q.5°C during the next run (curve B). The sample temperature was then returned to -l9°C·and the reflectance rechecked (curve C). It is obviOl,IS from the proximity
of these curves that temperature changes below melting-have little effect on the spect~al reflectance. During
the fourth run (curve D) a draft of wa~m air was wafted across the surface of the snow causing slight surface
melting, with no visible diminution of;snow volume but-giving the _appearance of a "glistening wet" surface.
· ·-· ·
··
.
·
·
· .' ·
· · :.
· · :·o· .··.· ·-- · o
·.
· · · · · .
The ambient air temperature varied be.fween.+2.5 C and +10 C during the run. With the onset of surface '
melting, the near infrared reflectance _decreased drastically.
By the fifth· run (curve E),the-surfJce·of t.he snow was· soft and wet' toa depth of about
1
/8
in. The
ambient temperature was about +6°C jwithol,lt the warm ajr blower. The snow was then refrozen and
1
anotk~r run (curve F) made with .the ambient tei11perat.ure about -5°C.
The refrozen snow was subsequently milled again, producing a snow still similar to naturally drifted snow.
A series of runs. ~as made at 0°-30° (Fig. 7) _qn the.cold s~o~, Gool snow G!l~~ below .melti~g), surface
_melting sn.~W,,_wet ~eltiJ1g snow'· andjhe same sno~ refro?~n~- Th_e: sno.w condition~ descri~edas "surface
melti~g" and "~et melting" were esse~_ti~lly th,e :same as ,des~rib;e.d, for :t~~ pr~viou,s sampl~: ~xce'pth that, at
the beginning of the run shown by curve C of Figure ,7, the "surface melting~' was very minimal and as the
-"slight'~ 'ineltl~ipiogressed the reflectance declined·.·:.' ; .;
. - . . ',, ,;·.;. ~; .
•
:~ ..
::
~ ;~
.:;,,.
.
•. j
..
~~
_,.Two_mor~.sets <_>f. runs
•
.
•
• ..
~·
••
~ -~ ~.).~~-~-:-~· ~~-;..
•
••
• •
•
... ,•
~~
.;;·,.· \,
1"-
:.•
~·.
'
were,made . on old,;shoyele~_,Joose. giat:I,ular snow: In _these.case&, runswere made
...:.<.;,only on the o.riginal cold snow' (c~rve A),:~m ·~~t:m~·lttng·~IJOW (Gl:lrve B),_a·n~ .o~·Jiie r,eifoze.n.sample (curve
-30~ (f.ig. ,?l·:-.,~t:i~;',~~~~\~~~~n·9~;t wp~s:.~(t9i~q.~p-;se,t:9r~r~,i~;f~t abo,ut · ::
1 :i-.·~c.)at;5:c:?~-,(Fig: ~3}'!~4~~~
9o
..
REDAND_NEAR:JNFRAREDSPECTRALREFLECTANCE OF SNOW
,·.·
1.0
9
_,
Specimen no. 730212
Source~detector 0° ~ .30° ,:
. ,
.
Snowcondition
Sifted,
"sugar:'·c.o'nsistency
. .
.
.:..
__.:'
-"
- ' ' -.
.
•
,~4 ;~·"'.
~
_..,...... -.,
I
·Curve
,:: ..,·
o','l''fl";
Density Hardness
·.'gfcm 3
&/cm 2
Descrip.tion
A
Cold,· ~·sugar" consistency
B
Cool, no melting
C
Surface melting
D 'Wet'melting ··
E . Refrozen
Wave~envth,
11 '
0.357
0.357''
65
65
0.419
0.419
·;20,000
p.m
·Figure 7: EffectsD[temperature conditions on szj'ted snow reflectance (0° · 30°).
Specimen no. 720226
Source-detector 5° - 5°
Snow condition Old snow, stored 1 year
' tie'risitY:
3
6
(/)
Curve
Description
A
Original.old. cold snow
W~t melting
·
Refrozen
If
•>
C
:·•
gfcm
0
a::
~
c:
1
'·
·u0·..
:i•
•i"
... - Fi~ure :8.. ~jJec,ts. oftempera(ure cond,itions on:ok(snow rejlectanc/(5° -_5°).'·:
+3° C in the freezer. This sample had decreased considerably iii volume and Was' Wet to its full depth, the
b.o_ttom layer r~sembling sl~sh. :
:,Additiortaf~Omp'a'risohs:b'f t~e ref1ectances' o'(cold~ mbiting and refrozewsnow'areshown in Figures· I 0- I 3..
In each case, 'the originaf cold ·snow· was used as the· reference ..:The ratios of the 'spectral reflectances of
melting ·~nd ~efrozen snow to cold snow were theJi co~pute<f.
·
..
j
•
;,$.'
,.
t
•l
~
'
_,
•' I •) ....
~:
.•
~
~:
~
t
~
}
~
'
',•
•
,•
. In general, melting lowers the spectral refl~ct~J1ce of.snmy_.:· At ~OIT;l_e.,~av~Jength,s, P;art·of the reflectance.
, _lost by melting may be: regained by refreezing, particularly with source-detector· angle~ of0°- 30°. In the
' .. ,.spectral r.egioir.fiom aB.ouY L9 1to 2.1-Jim wavelength; 'the refleCtance '0( melting and: refroz'en: snows appears·
to Be·.ab6ut the ·sarri·e _as:o.i: greater-· than_ that'ofloose'cohi''snow.·. This·'->c·c~rs_in·ra'regi9.ri-ofhiih·energy
'a6s~rt}ii6litb:Y\v'~ter~·'ai\jd_:ev~n greaier'absbri)tio·n~by.i'c~·:"·sirice this portiofii tne~spectrum rs 1ohe or .·
or
10
RED. AND NEAR-INFRARED SPECTRAL REFLECTANCE OF SNOW
Specimen no. 720226
Source-detector 0° - 30°
.
.
Snow condition Old snow, stored_ I year·
..
0
(/)
·Curve
Description·
Density
gfcm 3
A
Original old cold snow
Wet melting
Refrozen
0.347
0.403
0.403
0
CD
B
•
c
..!
.
.. D
a:
g
0
u
J!
i·
Wavelen9th, Jl-m.
.
Figure 9. ·Effects of temperature conditions on old snow reflectance ( 0°- 30° ).
5.0
Specimen no. 730208
Source-detector so. so
Snow condition Sifted
4.0
•u
c
~
•
a:•
•>
A
Original cold sifted snow
Same, cool, no melting
Same, cold again
Slight surface melting
Surface soft and wet :
Refrozen
c
;:
0
Description
B
·u
Density Hardness
gfcm 3
gfcm 2
Curve
D
.,E
.F
0.366
0.366
0.366
I
I -0.525
0~525
-60,000
!.
waveleri9th; j£m ' ·
2.0
•
a:
. ! · .. :
'
.. ·•
. . ,·,....
:
.
.
. .
'.
·, . . .
; .•
Figure 10. Reflectance of melting and refrozen snow relative to original cold sifted snow(ratio of
. turves from Fig.. 6). .
·.;
,; ..
~
t: .: i
~· • •
•
' ..
·.··
. :. ;,.5.
.
.
..
.
'
·_ REDA]\T[) NE~fl.-IN_FRARED.SJ>EcfRALREFLECFANCF; OFSNOW
ll
Specimen no. 730212 ·
Source-detector 0° " 30° ·
Snow condition ~-Sifted, "sugar" consisten;cy'
.-
I,.'
•uc
Description .
A
Cold, "sugar" consistency
Cool, no melting
Surface melting
Wet melting
Refrozen
f'
u
•
•
G:
B
;
. ·C
D
E
•
-~
..
Density Hardness
gfcin 3 · gfcm 2
Curve
0.357
0.3_57,
r.65
6~
_.
0.419
0.419
29,000
~
G:
O.O_L__.J._....L_..J_.---t_~-....L.-~-L-_.J.-+-~---:-lL.:----L--:"'=-..a..;_-:-'":-__._-'---:-:_~
0:6
0.8
1.0'
. 1.~.-
1.4.
· Wovelenvth, ~m
Figure 11. Reflectance of melting andre[rozensnow relative
: ·curves from Fig; 7).
to original.cold sifted snow (ratio of
Specimen no. 720226
Source-detector·· 5° - 5°
.
. Snow condition Old snow, stored _1 year
Q)
u
Curve
~
A
Description
:·I·
'.·-~-.
·~
Density
gfcm 3
c
u
_Q)
···~·
Original old cold snow
B
. Wet melting
C .:, -Refro'zen. ·
0.347
0.403
0.403
a:
Q)
>
:;:
0
~
a:
1.4
.. j
·. •
,~
...,_ :,_
_wove.le~~t:h'·'~~.
,··•· ··:. . .
,r
:Jligure ~2. J~efledance;ofmeliirig and re/iozen snow relattVe . tb ~riglrialqld'co:la .snliw(ratio bt
=- • · • : ·": -,-,~
eurves' /rorri Ftg: ·s;.·
:. . · ··
. :.~ 1} ..... :
·:·.;
!
i
: r
! ; ...
~-; .
..
; ~
,. .-.
~
·. i .: •.. -~- ~
~
.
Specimen no. 720226
· '''·Source-detector 0° -30° -..... snow conciition ofd sn~w. ~tored '1 year.,·
..
c:
0
A
B_
u
.
0
•; c .
. '·/.·
• D~nsity
:·gfcm3
" ~urve ·
u
. ,j...
Original old cold snow
Wet melting .
..
RefrozeJ;t.
0.347
0:403
0.403
~
:··.:,1.
; •..
0.00'-.6_..__0...,.8'"""'.o.......JL---I.....:O_.....__~I.-2_.._____.L_4,....--Ao.--,'-.6-._.-.---_.-1......8-'-....'-_.~-.-.2......0....;-,_ _.....___._,2~--..-.;.2-'. -.._,_"----_.2.-:-.4___.
W~ve~ehgt_~Jf':m · ·
'· Fi.~re 13.- Rejlectance.ofmeltfng and refrozen snow relative to original old cold snow (ratio"of
·
·
· ·
.. ..c:UrvesJrom Fig. 9 ). ·
·. ·
··
-
.
·:·
~
.
'.
';,.
generally low reflectance for snow, and the_curvesAhe.~efwe.,represent the ratio of two relatively small numbers, caution must be observed in interpreting the significance of specific details. However, there is little
question that increased reflectance of melting and refrozen snow does occur in this region. There is also an
indication that the reflectivity of sriow at wavelengths near 1.5 J.Lm is not greatly reduced by melting and
refreezing, and may in some cases increase. The effects described as occurring at 1.9 to 2;0 J.Lm and around
1.5 J.Lm appear considerably exaggerated in Figures 12 and 13 because the "original old cold snow" had
already deteriorated considerably in reflectance, thereby pe_rmitting a slight increase in reJlectance to
appear as large ratio of increase. one feature of the effect of melting which is--easy .to note, but difficult
to describe quantitatively ,.is the relationship between the reflectance and ~he :degre.e. of melting. Once the
surface of fresh snow becomes wet to a certain slight (but as yet undefined) degree~ the· ~eflectance suddenly
decreases throughout most of the near-infraied portion of th.e spectrum studi~ci: ·when a certain (apparently
very shallow) depth of surface wetness has been attained, softening or wetting ofthe sriow to greater depths
seems to have comparatiyely little additiona~ effect..
..
a
Reflectance of ice-glazed snow:.
.
•
.
I
.
On 8 December 1972, a light snowstorm changed to freezing rain causing an extremely hard smooth ice
glaze on top of the snow. A sal)lple was tak~n at night at just about the time the freezing rain ended, The
1
snow under the crust had a density of 0.148 g/cm 3 • The ice-glazed crust, approximately% in. thick, had a
m~asured density of 0.903 g/cm 3 • The folldwing morning the undisturbed snow cover was subjectively observed t,o ha've ayeiy high specularreflectan~e in the visible range. The spectral reflectance measured in the
nea_ r infrared ·at 5°-5° is shownin Figure 14 In the range of wavelengths from 1.00 to 1.40 J.Lm, the
.
reflectance of the crust is qualitatively sim~~r to that of.oth~!SJ10W, particularly aged and/or refrozen sample_s. l:lowever' beyond 1.4 J.Lm, th((sp'ectr~ jref1ectanc~ of this ic~ cru~ted snow bearslittleres~mblance to
that of other snow samples studied. The ice; glaze refle'cts fairly uniformly from 1.00:to 2.50 J.Lm with
gradual diminution of reflectance and just
s·~gg~stlon of a minor peak in reflectance around 1.9 J.Lm.
Unfortunately, onJy:this on~ run was mad~ i~ the. n~adnfrared on· the natural i.ce: cru~t. A sampl~ ofartificial snow crust was synthesized by spraying i9~ water OJ) ve~y cold snow, however, and similar reflectance
characteristics were observed.
-·
1
•
!
tHe
.
:.\
13
io
:·(/)
, .. c
···;m
··::2
i'
.
. .: ~ .
.
.
l-
..
.·_
..
0.8
Wav.tltnt~th,
p.m
:Figure j 4: Refl~ctance ·of ice-glazed snow-crust:-~· · ·-· · · .................~. ,.
•
~
'
)>-•
"
· ·Specimen no. 721216
0
0
Source-detector ·S --S
.
. ·snow conditio~ A: Near'ly fresh
. Snowderisity
0~12t"-g'icm 3
Snow hardness · -~ 2 gf.cm 2 :.
~ ~
.
0
(/)
·
.
·• .... ,. .
B: Drifted, 2 days aid'
•
0~2S3 gfcm
2
9Q gfcm •
3
·
!.
.
c
:,m
.It
;._.-
:
~
~
· ... ,.c
·····.
a:
u
c
.or.
u
. .!!·
i
0.~.
w.ave_ltnQth, p.m
r. . .
~
'.·EffecLof:drifting'and wind~compaction onsilow:refle.Ctance/;';; ~- :~>r/c.;"
:.;_ ::·~. :,_!
~;-{:-._.- ~~~-_: -.i·;;.{f ~; •. _.. -·;·
.
_,
_·
.
.
.·::·!. . . ·.
·
·,!__ .; ___
,):·.~~
:;;r;}
·::_:·;':;'_-~-- r.~--~-;~:-~_1?·: l~··;::~·r -~)~--~!~:~·- ~~1r-1::
f:;· 1
_:. r~~·: -::.? . ·:::ir!:r·~·;:~~~:-::~
:·only orie ·b.pporturti ty. was .available for st udyi!lg ~}:le;effect,.pf wind_-q I:if!,il)g .on... tJi.~ -~pe,_~_tgll .r~fk~tan~e. pf
... "· :~fio·~~~..-~.~~~
_::_~.;:~~- ·
~..
., ...:~ ..·- _: ~.
_. ____. _.. _{_·- .__ 1;·~;
~-~-:~-~-----~;_:;~-;~
~-~-~"~.:;:!~~-:'1.\;>.t:_i·#·~ '-~<-~ .
':;;Lt::··_,_
;
• i
• ._-..
:d.:__ .. .:.- ......· .....
;-;~
? :·
I
.
•
•
~ -
•
f} __:~:~:;iJ"; ._. . . ,: ;-::
.~r, :·5;-·:~
~
.. ': :~~~'~:~~im~,~~Y.\.~ .. i~:;~r sn_o~.- f~I~ ..quring. th:e p~ri~g. ?f),?;O.;~}\?~~r,s.Rnd,(:<P.~:~:rm~~r,-rn2 9.~:f.B h?Bf~:~n
.tl:'d"·~;
. 16 .:_:.1·:-;·;_~:·
pecember
1972:
·A sample of thts snow.. ·.was
c,olle~t~d_abou~ 2 hq~rs aft~r <;~~~atwn qf, the ~l_l<?W.s.tqrm~~x~d
•· ':- . .
'.::_·. -;:£:.~.~: . ··. -~·; . . . ··
. :· '. .'{:: .::.:n;-:_..
;'_i-- .. ~:.-·~(,i•:~r~P'
.,;;:~~;;-~~-.-==~ .i~£-~.;- ~..:._·-h~:t:~-:![u ?·~-~-!~,_-:-J·.. ~·::~..._..
\- · b'efore:.:sigilifibint"wind ;occurred·. The· ambient temperature during snowfall and up to the fi:me:-of collection
-~ -~i~:~K~~- ~~"~~:e¢?·:~6~~0..-,.~~~:.:.~_;.,, :i;·:~~·~
~
:.:'c,:_~·.;~ ·~·· "':; ::: ,~,· <,, :: -~'-~ ~,::~;:;,~~:I,~f..\~l:~;·-<;;:·.:l:· _,.~--~·::·i-'1 .:;l ~;:.;;~~
1
j-(
..'.: .
'-'':
:·; ,: :-,· .; ·_, . . . :. ,. .·,: ·, :.
·: ,:1 \Th~:;~J.10W: d.ensity- was 0.12-.l g/cm:3 ·and:thehardne·ss;-aJin·ost negligible:(<SY2"g/cn1t:)it:Dtiejo~ e'quipn1e'iit{.~
1
•.. ,
;.; ;·p~f8ti)e~·s;::th~-:satrtpl.e ·V!a~, ~t9rdi' iri 'a':clqs_ed d~ep-free.~:e;.:at:::;=q,~~,c;,_fpc~h-~4t:·~O:lW?'~rc~':h'~fqr.;e[refle.?tat1:ce~.:·.·,
measurements were made. The density and hardness remained essentially unchange.d. The reflectance of this
snow is shown by curve A in Figure 15. On 17 December, at about 1530 hours, another snow sample was
:~
14
RED AND NEAR-INFRARED SPECTRAL REFLECTANCE OF SNOW
'
:. Specimen· no; 740321
Snow condition Freshly fallen
Snow density 0.1 70 gfcm 3
Sriow hardness 8 gfcm 2
.
'0
.Ill
Curve
0
·m
..
..
Source-detector
,.
-.5!
•
~
u
' c
2u
•
i
~avelenQth,
JJ-m
..: .a.
.
'
~Specimen no. 720226 .
Snow condition Freshly fallen,
Curve
"f~athery~'
Source-detector
A
B
c
Figure 16. Comparisonof relative.·r.eflectahce of snow at :various source-detector angles..
retrieved. ·The weather had be~il cold (not above -I °C) and windy (10. to 18kn.ots) on the 16th·. and 17th of
December so that the snow had: drifted considerably and had compacted somewhat. The density and hardness
of the drifted s'now were 0.253 g/cm 3 and 90,-g/cm 2 respectiveiy. The re'flectance or' this drifted snow sample is shown by curve B in Figure 15.
1
Note that the .d-rifted snow e-xhibits a lowe r reflectance than fresh snow.. This rrlight be attributed to the
additiorial aging ~f the sample·. '~However,· snow (partic~larly ·at_' the peaks n~~r 1.8-pm. and 2.2-pm wavelengths)
retains much more ~fits teflectari~e-during t~o days of wind compaction, even with the ·additiondl aging'at
cold temperatures, than snow which has aged naturally and increased commensurately in density without
driftipg. ·This js partis;ularly true -if the .undrifted snow has aged during. periods· of melting temperatures (as
was .Sho~n in Fig. 3 an9 5). Apparently, in these spectral b~nds, crystal metamorphosis <;lue to melting exerts
Spe~irne~ no. 7io:i2o • ··
Snow condition New snow
· ~Snow _d.erisity ·Q.l} 2 gfcJ!l.t. ~- ,
·curve·
Source:detedor··
·A.
B
c
D
wavelenQth, JLm
c.
SpeCimen no . .·1202 20 . _ · .- _;. ·. .
Snow condition Neatly new snow
Snow density 0.144 gfcni 3
•.
·, .·
c5
en
• ;·_ / . !
Curve
0
Source-detector .
.
Ill
.
~
. . .
E .
•
a::
3c:
. ~~
U•
•
i
p.ol_~--~~-L--JL_~~--L2--~--~~.4--_.~=-~~L-~~~~~~~~~--L-~~~
0.6
0.8
LO
L
.,. )·
:;-_·,
WoveltnQth, ·JA-m. ·
't1..
·Figure 16 (\'OI!t'd}
.~
• .l
.'
; '
a greaterinfluen~eupon reflec.tance than. does inech~mfpil restnicturingof.the:snow.by wind-blowing ~nd
comp,action.
.
.
..
.
.
- ..
-::_..k:
::'
'·:··-1!:'
..
-!
.
.::.-.
. ' Mi,ddleton'et ~to9si) giv~ a:good'accmiht 'of g~n.imrie:tric ;11Casuremei1t~·o~ the rcflettance'ofsnow in
.the"VisilSle spet!.rut~. ~ '·. ·::. ·:':· :::> ·" f . ., _-.: ·' .· ·: ~.. ; __ , · :i... ·, ·•· ..-.: ..;::- ::.: ·.. '_; : __. . . . .
;
i
-/~~ ~... :~·;
:- ;·__
·;~_:.;~·-
:1
·
;·~
1
~:..
•
! :·.:~:-(:<·~ri,t
·.
~In ;~he· pr¢seiJ t s·tuqy" t~e.,re fl~cta_qce .curves J~r. a giY:en SJ10~ lare. gene~ally: v~ry,~i.n){lar ,flo~- a variety of ·
,.; ' ' --.angles; indicating_.~greement withthe populady h~ld concept of riear.-perf~~t '_Qiffusivity.ofsno\Y,~ ;rhere is,
however, lack of consistency between samples in the relative magn.itudes of measured reflectances at different
angles, as" shown in-;Figures'-'16a:::d., These figur.es ..were . chosenfor. the.variety. ofangles which they illustrate,
even though more measurements were made at 5° -5° and 0°-30° than at any other angles.
~--~----------------~------~~~----···-------------------·---"--··----~·.···--·-·--·-------···
, ____ , .. _ ... _ _ _ •• ._,.......:... ______ .••.•
-c,<~..-
.RED AND NEAR-INFRARED SPECTRAL,REFLE(;T~CE OF SNOW,
16
Most of the angular comparison runs we~e made early in the project (which accounts for the limited
spectral range shown). From the earlier. data; m'any of which are not shown in the illustrations~ the tentative conclusion w'as'that the' reflectance of snow seems tob((about the: same at 5°-5° and 0°- 30°~'\vith
perhaps slightly greater reflectance at 0°-30° .· On the night of 21 MarcJi 1974, a fairly heavy snowstorm
occurred at Hanover, New Hampshire, presenting an opportunity to. recheck the angular reflectance ~fa
fresh snow. A fresh sample was obtained at 2100 hours. The density was'OJ 70 g/cm 3 and the hardness
was approximately 8 g/cm 2 ; Between 2100 ho~rs and 0419 hours the following morning, spectral scans
were made at several angles, as shown in Figure 16a. Unfortunately, due to an equipment malfunction, the
calibration of the instrument may have been off so that t~e values of reflectance may be in error by a few
percent. However, the error may be considered to have heen the same for all runs, so that the relationship
shown between runs should be essentially valid. Figure 16a shows that for that particular snow-, at least,
the relative reflectance was greatest at 0° - 30°, decreased successively at 0° -25°, 0° - 20°, and 0° - 15°, and
was even less at 5°-5° and 10°-10°.
Relationshjp, ofsnow .. reflectance to snow density
As previously mentione-d, the teflectance of sriow depends a great deal upon its history. Whatever the
subsequent history of a freshly fallen snow, there is a tendency toward densification of that snow. Th_us,
natural aging over a period of time, even at subfreezing temperatures, results in changes in the physical (and
optical) characteristics of the snow caused by settling, sintering, and sublimation. Mechanical working of
the snow, such as-occurs during blowing and drifting snow, causes· densification by settling and wind compaction~ Melting ofsnow, even when slight, not only tends to increase the effective mean grain size and density
by melting the smaller partiCles (such as-dendritic branches) first but also produces fre~ water in the snow.
Densification of snow, by whatever.·mearis, is accompanied by a reduction of the reflectance of the snow less noticeably in the visiple .red region but more and more,significantly in the near-infrared region.
A.s a matter ·ofint(frest, ~ ~urveywas made of snows of several densities, disregarding history, and relating
the reflectance at various wavelengths to the density ofthe snow. Wavde~gths used were l.O,J.~, 1.3, 1.8
and 2.24 pm. In each case, the decrease in reflectance with density showed about a 0.8 correlation at the
0.01 confidence level~
.
Despite the high correlation of decrease of reflectance With._increasing density, it is difficult to assess how
much o( the decrease is due to densification per se and how much must be attributed to the underlying
physical cause of the densification.
Dependence of ~flectance on mi~fostructwle of snow ·
.
- --:.,
.
_;. _ !·.·:
In addition to densificationof the snow 9over, most processes involved in the aging of snow result in an
effective increase in Ice p·artide size.· -Finer' microstructures, such as dendritic branches, are lost by sublimation or melting, while groups of individual crystals become fused to various degrees through sintering
or melting and refreezing.
Dunkle and Bevans (1956) calculated the· theoretical spectral reflectance (0.3 to 1.3 pm) of a snow cover
"for particle sizes of 0.001, 0.01 and 0.1 in ..... " Their theory" .... indicates that, as snow ages and the snow
crystals grow,_ the albedo sh.ould tend to decrease from the initial high values." It would therefore be
anticipated that a reduction in the spectral reflectance would result from the 'combination' of dertsificatioh
.and increased particle size which occurs with aging. This expectation is confirmed by experimental-measurements.
'·
..
Evidence that the spectral reflectance or'snow is preferentially affected by ~a the~ slight changes in 'the·.·
crystal structure of the snow ma{be :inferred from Figures. 16c ·and l6d. A freshl}" fallen snow sarripfe was
collected on the morning of 20 February 1972. The predominant snow crystaltype was spatial dendritic
with plate-like branches. The SJ10W density. was 0.112 g/cm 3 . The reflectance of this snow samp~e.is shown
of measurements,this .snow
was allowed.
to stand. in a closed
in Figure 16c.
After :the initial. series
.
. .
.
.. sam-ple
.
·.
'
.
'.
(
·'
.....
.
~
:,
..
.
17
··
deep~freeze
_. ;.
o
.,
·
~
... .-
.
.-
~·~·
·:
·.--
-~
·. ·
·:~rk.>·_·,..
·-: ·. . . ·
.
·
cabinet at -19. C for two ,days._ Subsequently, thereflect~nce was xemeasured with the r~s_uJts
de:pict~d in- Figur~ _16d. Immedhite'Iy after this ~eries of me'asure~en t~, P1iC~o·s~-opic examinatiOJ1 of the .
_snow-surface revealed ·some loss of structu-ral detail whiCh appeared t:o be dqe fo slight partial melting· of ·
the crystals. Since these measurements Were made in early 1972 Tn a warm:Iabora'toiy' the' cha.nge in crystal
condition was probably caused by penetration of warm convection- currents·from the room into .the open
top of the freezer during measurements. At the time of the second measurement, _the snow density had .
increased to.O.l44.g/cm 3.. From a comparison of the two figures, it.:appears· that, although the relfectance
decreased at all angles, the diffuse reflectance deteriorated considerably more than did the. specular component.
. Although drifted snow observed immediately ·after betng wind-blown has a great~r density' due to 'wind
compaction, than unctrlfted snow; the drifted snow would'be expected to have a smaller;size distribution
and_ more rounded ice particles than snow which has aged. by other processes due to mech(!nical working of
the snow cryshils {tvliddleton et al. 1952). In the limited numbe; of rr{easuren~ents availa~le on winddrifted snow, it was found (Fig. l5)that drifted snow has. a lc{wer spectral ~eflectance th~n before b.elng
wind~blown; in spite of the' resulting smaller crystal size, implying that the spec;tral reflectapce of snow is.
reduced to.some extent by densification per se. Comparing Figure 15 with Figure 3-, it also appears that
drifted snow has slightly higher reflectance, particulady at s'ome wavelengths, than other snow samples .
whichhave otherwise "aged" to approximately the same density,_ thereby indicating the inh~e~ce of
particle·size- on the resultanfreflectarice.
.
.
. .
a
REFLECTIONS ON REFLECTANCE
Many factors influence the red and near-infnired·· spectral refl~ctance of snow. The pri-ncipal parameters
affecting the spectral reflectance, particularly in a natural meteoroiogical e-nvironment, are· so interrelated
as to-defy precise defmition of the role ofindividual factors,
It is to be expected that the spectral absorption <?ficeplays a major role in d~termining -the bvenill shape
of the reflectance curve of any.snow cover. Likewise, wheri. melting occurs, the spectral--absorption ofwater
modifies the reflectance accordingly.
..
- ,,
The crystal size and structur~ of the snow also constitute quantitatively significant and spectrally variable
factors; The'spectral reflecta~ce tends to decreasewithinc~easin'g particle size, aspoi.nted out by Dunkle
and Bevans (1956). The structure of the snow, including crystal orie11tation ~nd corripactipn, play_s a role
less readily defined. ·
·.
· .
· ·.
..··
Because of the poly disperse c~ystal shapes, sizes, and. orientations ~hich occur in a ·na:tural fresh snowfall,
together with the innumerable possible 'sequences or'natunil.aging to which the snow may be subjected, it
is much simpler (and perhaps as valuable from a practical standpoint) to formulate some general conclusions
concerning the spectral reflectance of fresh snow and the reflectance changes that take place during several
·
types of natural aging~
i. Fresh snow has very hi~: r~flectance in the visibl~: r~d, region. The spectral reflectance falls off
sharply in the near-infrared wavelengths and is strongly related, inversely' ~0 the spectrafahsorption coefficients of ice (Fig. 3,curve A).
·
a
'·2: Snow which has aged· under cold conditions, with normal 'settling and ·densification, shows a slight
decrease in red-reflectance and'moderate decreases in infrared reflectance (Fig. 3, curve B).;
. 3.- Sno~ whi~h has drifted uqder c,old conditions, and_.be.com~ densifieq to S()me d~gree by wind-compactiop, also exhibi.ts a slight de~rease in red ;eflectance and ~ode rate' decrea~es itdnfra~ed ·reflectance (Fig. 15).
Howev~r' comp~~~d with cold, aged snow which Jlas rio( been ~ind~bl 0Wf1,. but whi~J:t has achieved' equivalent
densification by natural settling, 'the drifted snow tends 'to have a .siightl~/h~g}le.~reflectance, particularly near
1.~-~m and 2 .2-pm waveiengths (compare Fig. 3 and, 1_5).
__, - .
.'
.. : '
~~
. -·
.•
·.t,
.
\
.
;..
RE.OAI'!D NEAR-iNfRARED SPECTRALREFLECtAtyCE OF SNOW
f8
4, ..Snow which is even slightly m~lt~rig, when .compared with nearly fresh <:_:old snow; has a distinct(y
lower .reflectance in the,red _region: and· ari even mor~ pronounced decrease _in infrared reflectance gene'ially.
There .is an exception;however, inthat the reflectarice,may:stay· nearly· the same as that of fresh snowiorl,;'
even ·incre~se slightly in the. 1.9. to 2.0-J.Lril wavelength region {Fig. 6-13).
5. Refreezing of snow which has .previo·usly 'been exposedto melting temperatur~s has a rehitivelyminor
effect, resulting iri a.reflectance curve gerterally resembling· that of the melting snow. Refr.eezing may, h.owever, cause an increase of reflectance. relative to that of melting snow at wavelengths ·near 1.4 pm ·and 2. i ·J.Lm
{Fig. 6-13).
..
..
.
.
6. The specular reflectance of ice-glazed crust is spectrally similar to that of refrozen snow in the 1'~00
·to 1.40-J.Lm range. The ice glaze ·reflects fairly uniformly from i .00 to 2:50 Jim with gradual diminution of
. reflectan.ce and j.!lSt the suggestion of a mihorpeak in r·eflectance·around 1.9 J.Lm {Fig.14). Unfortul).ately,
'the diffuse component ofreflection from the ice-glazeq crust was not studied ..
7 .. Measurements of snow .reflectance at various angles jndicate that for light fresh snows, with the
so.urce normal·to the snow .surface, the spectral reflectance generally increases slightly with increasing
0
detector a~gles between 15 and 30°·.. The .reflec;tance af5°- S and 10°-10° is slightly les~ (Fig. 16a-d). For
aging snow, results are erratic, pre~umably due. to the variety of microstructural 9hanges involved in the
aging process.
8. The red and infrared spectral reflectance has a strong inverse correlation 'to snow density. Since
densification is the result of many of the natural aging phenomena, it is probable that this correlation is the
result of many concomitant influences,. densification per se being a_ minor contribution .
. LITERATURE CITED
Dunkle, R.V. and J.T. Bevans {1956) An approxiin~te analysis o.f the ~olar reflectance
a snow cover. Journal of Mt;teorology, vol. ~.3, April, p. :i 12-216 ..
~nd transmittance of
Hulbert, E.O .. ( 1? 28) The ultra-viol~t, visible, and infr~r.ed refl~ctivities ofsnow, sand and other substances.
. . . Journal of the Optical SoCiety of America, vol. 17, p. 23-25.
.
Irvine, W.M. and J.B. Pollack (1968) Infrared optical properties of water and tce spheres. Icarus: vol. 8,
p. 324-360.
.
.
.
'
. Mellor, M. (1965) Optical measurement~ of snow. U.S. Army Cold Regions Research and Engineering
·
Laboratory (USA CRREL) Rese.arch, Report _16~. (AD.62i77.5).
.
.
.
. .
.
.
I
.
Middleton, W.E. Knowles and A.G. Mungall {1952) The luminous directional reflectance of snow. Journal
of the OpticalSociety of America, vol. 42, no. 8, p. 572-579. ·
. University of Minnesota (1951) Review of the .properties of snow and ke. U.S. Army Snow, Ice and
Permafrost Establishme!lt (USA SIPRE) Report 4. (AD 696397).: