NOTES AND CORRESPONDENCE Spectral Albedo of Desert ...dust.ess.uci.edu/ppr/ppr_AMY05.pdf · in northern Taklimakan Desert, Ejina in Badain Jaran Desert, and in Tengger Desert. The

Journal of the Meteorological Society of Japan, Vol. 83A, pp. 279--290, 2005 279

NOTES AND CORRESPONDENCE

Spectral Albedo of Desert Surfaces Measured in Western and

Central China

Teruo AOKI, Masao MIKAMI, Akihiro YAMAZAKI

Meteorological Research Institute, Tsukuba, Japan

Sadayo YABUKI, Yutaka YAMADA

The Institute of Physical and Chemical Research (RIKEN), Wako, Japan

Masahide ISHIZUKA

Wakayama University, Wakayama, Japan

Fan-Jiang ZENG, Wei-Dong GAO

Xinjian Institute of Ecology and Geography, Chinese Academy of Sciences, Urumqi, China

and

Jun-Ying SUN, Li-Chao LIU, Mao-Xian ZHOU

Cold and Arid Regions Environmental and Engineering Research Institute, Chinese Academy of Sciences,Lanzhou, China

(Manuscript received 21 September 2004, in final form 7 February 2005)

Abstract

Spectral albedo and nadir reflectance of desert surfaces were measured with spectrometer for spec-tral region from 0.35 to 2.5 mm in western and central China in spring of 2001 and autumn of 2003.Desert surfaces were mainly gobi and/or sand dune around Qira in southern Taklimakan Desert, Aksuin northern Taklimakan Desert, Ejina in Badain Jaran Desert, and in Tengger Desert. The measuredspectral albedos are low (0.05–0.11) at ultraviolet wavelengths but it rapidly increases to 0.2–0.3 at thewavelength about l ¼ 0:6 mm at any site. The wavelength at which the albedo reaches to the maximumis around l ¼ 1.8–2.2 mm, where the albedos range from 0.37 to 0.49 among measured sites. Nadir re-flectance also basically shows a similar spectral distribution with that of the albedo, while the values arelower than albedos at any wavelength. Standard deviations of albedo and reflectance calculated from

Corresponding author: Teruo Aoki, MeterologicalResearch Institute, 1-1 Nagamine, Tsukuba, Iba-raki 305-0052, Japan.E-mail: [email protected]( 2005, Meteorological Society of Japan

measurements for different surfaces at each site indicate that gobi surfaces around Aksu and Ejina havelarge variability compared with other surfaces. These datasets would be useful for the studies of climatesimulation, remote sensing, radiative forcing, and so on.

1. Introduction

Spectral albedo and reflectance of earth sur-faces are important parameters for the earth’sradiation budget and remote sensing studies.However, in situ measured spectral informa-tion of desert surfaces is very limited because ofthe difficulties of access, particularly for west-ern and central part of China. Chinese desertarea is one of the main source region of mineraldust aerosols, whose radiative effect has an im-portant role for the earth climate (Sokolik andToon 1996; Tegen et al. 1996). Spectral albedois one of the key factor to estimate the radiativeforcing by mineral dust (Liao and Seinfeld1998). Especially over desert surface the radia-tive effect by mineral dust in shortwave regionat the top of atmosphere could have both posi-tive and negative values in the possible rangeof refractive index of mineral dust (Aoki et al.2005). Recently, the accuracy of satellite de-rived broadband albedo is much improved byusing such satellite data of multi-channel sen-sor as Moderate Resolution Imaging Spectro-radiometer (MODIS) (e.g., Liang et al. 2002; Jinet al. 2003). However, there are no operationalproducts for spectral albedo. For the studies ofclimate simulation, remote sensing, and radia-tive forcing, it is necessary to compile the data-set of spectral information of desert surfaces.We have made the field measurements forspectral albedo and nadir reflectance of desertsurfaces in Taklimakan Desert, Badain JaranDesert, and Tengger Desert, mainly for thepurpose to compile the datasets, in which thespectral albedos were used in the radiativetransfer model to calculate radiative forcing bymineral dust over desert surface (Aoki et al.2005) and the spectral nadir reflectance couldbe used for satellite remote sensing.

2. Measurement

2.1 Observation sitesObservations of spectral albedo and nadir

reflectance were carried out in four sites shownin Fig. 1. ‘‘Qira’’, ‘‘Aksu’’, and ‘‘Ejina’’ are placenames of oasis (town) near each observationsite. Qira and Aksu are located at the southern

and northern peripheries, respectively, inTaklimakan Desert. Ejina is at the northernpart of Badain Jaran Desert. Site informationand observation condition are shown in Table1. Since there is no oasis near our observationsite in Tengger Desert, we call the site name as‘‘Tengger’’. Surface conditions are classifiedmainly into two types, which are gobi and sanddune. Only in Qira the measurement for farm-land was also done. All observations were madein spring or autumn, in which the atmosphericdust aerosols were richer in spring than in au-tumn. Sky conditions were clear or cloudy withoptically thin cirrus. Landscape and surfacecondition of all observation sites are shown inFig. 2. There were no vegetations at any obser-vation site during our observation periods. Gobisurfaces were generally very flat, consisting ofsoil particles of various sizes and stones with asize of a few centimeters at most. Some part ofgobi surfaces around Aksu consisted of ag-gregated soils as shown in Figs. 2d-2 and 2d-3.Since surface of sand dunes, in general, werenot flat, we selected flat areas for the measure-ments. The surface of farmland was flat, with aroughness of several centimeters in depth, andwas covered by very fine soil particles com-pared with that of sand dune.

2.2 Spectral measurementIn this study we define the albedo a and

the nadir reflectance r are the following non-dimensional parameters, respectively, givenby,

a ¼ F"

F# ; ð1Þ

r ¼ pI"ðyv ¼ 0�; fvÞF# ; ð2Þ

where F# and F" are the downward andupward radiant flux densities (we will simplyrefer to ‘‘flux’’ hereafter), respectively, and I" isthe upward radiance as functions of viewingzenith angle yv and viewing azimuth angle fv.The nadir reflectance is the value of nadir di-rection (yv ¼ 0�) of hemispherical directionalreflectance factor (HDRF).

Journal of the Meteorological Society of Japan280 Vol. 83A

The spectral measurements were made witha grating spectrometer, ‘‘FieldSpec Pro JR’’made by ASD Inc. (USA), which is shown as anexample in Fig. 2b-1. The scanning spectralrange is 0.35–2.5 mm with a spectral resolutionof 3 nm at l ¼ 0:7 mm and 30 nm at l ¼ 1:4 mmand 2.1 mm. The scanning time is one secondwith a sampling interval of 1 nm for thefull spectral range. The detectors are one-dimensional Si photodiode CCDs for l ¼ 0.35–1.0 mm and two different types of InGaAsphotodiodes for l ¼ 1.0–1.8 mm and l ¼ 1.8–2.5 mm.

For albedo measurement, we used the albedoobservation system (Aoki et al. 2000, 2002)(e.g., Fig. 2c-1) with a white reference stand-ard (WRS) ‘‘SRT-99,’’ made by Labsphere Inc.(USA) to measure the downward and upwardradiant flux. As described in above articles,this method has the following advantages: (1)the level of WRS can be maintained accuratelyby means of a spirit level, to be consistentthroughout the measurements; (2) it is not

necessary to know the absolute reflectance ofWRS for albedo measurements since both sidesof WRS are measured and only a relative ratioof reflectance is used; and (3) the cosine prop-erty of reflection is generally better than thatof a cosine collector. Our system, however, un-derestimated the fluxes due to the imperfectcosine property of WRS for reflection and theshading by the optical fiber arm. These effectswere corrected by the method described in Aokiet al. (2000), which requires a spectral diffusefraction in downward solar flux. Since the co-sine property of WRS depends on the zenithangle of incident beam, the correction for thisproperty should be made for the direct and dif-fuse components of illumination on WRS sepa-rately. Aoki et al. (2000) calculated the spectraldiffuse fraction using the radiative transfermodel. In this study, we obtained it from directmeasurements of diffuse and global solarfluxes, because the diffuse fraction is largeunder dust-rich sky conditions. Since the mea-sured solar flux itself is also underestimated,

50°N

30°N

90°E

40°N

100°

E

110°E

120°E80°E

130°E20°N

70°E

Desert

Observation site

< 1500 m

1500 < 3000 m

> 3000 m

Aksu

Qira

Taklimakan Desert

Beijing

Ejina Badain Jaran Desert

Tengger Desert

Fig. 1. Locations where spectral measurements were made in China.

T. AOKI et al. 281March 2005

Ta

ble

1.

Sit

ein

form

ati

ona

nd

obse

rva

tion

con

dit

ion

N1

N2

Des

ert

Sit

en

am

eS

urf

ace

typ

eL

ati

tud

eL

ongit

ud

eA

ltit

ud

e(m

)(*

1)

Da

teL

oca

lti

me

(*2

)y

0(*

3)

(deg

)(*

4)

Sk

yco

nd

itio

n

Ta

kli

ma

ka

nQ

ira

Gob

i3

6� 5

40 0

100 N

80� 4

70 0

800 E

15

26

Ma

r.2

1,

20

01

09

:19

–0

9:5

44

4.7

–4

9.7

50

4–

6C

iQ

ira

Gob

i3

6� 5

50 3

500 N

80� 4

70 2

800 E

13

70

Ma

r.2

2,

20

01

10

:01

–1

0:1

54

1.7

–4

3.3

30

10�

Ci

Qir

aG

obi

36� 5

40 0

100 N

80� 4

70 0

800 E

15

26

Ma

r.2

4,

20

01

08

:45

–1

1:5

23

5.5

–5

4.2

11

00þ

–1

Ci

Qir

aG

obi

36� 5

30 4

900 N

80� 4

60 5

600 E

15

20

Oct

.1

7,

20

03

12

:50

–1

3:3

35

5.3

–5

0.4

54

0þ

–2

Ci

Qir

aG

obi

36� 5

30 2

900 N

80� 4

70 0

100 E

15

31

Oct

.1

8,

20

03

08

:34

–0

9:2

75

3.6

–6

0.8

55

Cle

ar

Qir

aS

an

dd

un

e3

6� 5

50 3

500 N

80� 4

70 2

800 E

13

70

Ma

r.2

3,

20

01

09

:04

–1

1:2

93

6.0

–5

1.5

54

Cle

ar

Qir

aS

an

dd

un

e3

6� 5

40 5

500 N

80� 4

60 4

800 E

14

90

Oct

.1

9,

20

03

09

:37

–0

9:5

75

0.8

–5

2.8

22

9C

iQ

ira

Fa

rmla

nd

37� 0

00 5

600 N

80� 4

30 4

500 E

13

63

Ma

r.2

3,

20

01

12

:41

–1

3:0

14

0.1

–3

8.3

55

Cle

ar

Ta

kli

ma

ka

nA

ksu

Gob

i4

0� 4

10 0

100 N

80� 5

60 3

500 E

10

28

Ma

r.2

7,

20

01

12

:35

–1

2:5

44

1.4

–4

0.0

55

Cle

ar

Ak

suG

obi

41� 2

00 1

600 N

80� 4

10 2

100 E

10

28

Ma

r.2

8,

20

01

08

:41

–0

9:1

15

1.2

–5

5.8

55

Cle

ar

Ak

suG

obi

40� 5

50 4

300 N

80� 0

30 3

100 E

10

75

Ma

r.2

9,

20

01

08

:03

–0

8:3

05

7.6

–6

2.2

55

10�

Ci

Ba

da

inJ

ara

nE

jin

aG

obi

42� 0

10 1

800 N

10

1� 2

30 0

600 E

92

5O

ct.

23

,2

00

30

9:0

2–

10

:10

58

.8–

67

.05

58

–9

Ci

Eji

na

Gob

i4

2� 0

00 5

100 N

10

1� 2

50 0

800 E

89

7O

ct.

24

,2

00

30

9:3

4–

10

:25

57

.8–

63

.15

5C

lea

rE

jin

aG

obi

42� 0

10 3

000 N

10

1� 2

20 4

100 E

93

4O

ct.

25

,2

00

30

9:2

3–

10

:14

59

.0–

64

.86

52

–9

Ci

Eji

na

Sa

nd

du

ne

42� 0

20 0

300 N

10

1� 1

80 4

200 E

92

0O

ct.

23

,2

00

31

0:4

6–

10

:50

55

.5–

55

.81

19

Ci

Eji

na

Sa

nd

du

ne

42� 0

20 0

700 N

10

1� 1

80 4

000 E

91

3O

ct.

24

,2

00

31

1:0

4–

11

:36

53

.9–

55

.13

3C

lea

rT

engger

Ten

gger

Sa

nd

du

ne

37� 5

30 1

400 N

10

5� 0

10 1

000 E

14

03

Oct

.2

7,

20

03

09

:25

–1

0:1

75

4.5

–6

0.1

55

Cle

ar

*1:

Th

ese

alt

itu

des

wer

em

easu

red

wit

hh

an

d-h

eld

GP

Sd

evic

e,so

tha

tth

eh

igh

acc

ura

cyca

nn

otb

eex

pec

ted

.*2

:L

oca

lti

me

isca

lcu

late

dfr

omth

elo

ngit

ud

ea

tea

chsi

te,

bec

au

seB

eiji

ng

Sta

nd

ard

Tim

eis

use

din

all

ofC

hin

a.

*3:

Sol

ar

zen

ith

an

gle

.*4

:N

1a

nd

N2

mea

nth

en

um

ber

sof

mea

sure

dsu

rfa

ces

at

one

site

for

alb

edo

an

dre

flec

tan

ce,

resp

ecti

vel

y.


(a-1) Gobi around Qira (a-2) Surface shown in (a-1) (a-3) Same surface as (a-2)

(b-1) Sand dune around Qira (b-2) Surface shown in (b-1) (b-3) Same surface as (b-2)

(c-1) Farmland in Qira (c-2) Surface shown in (c-1) (c-3) Same surface as (c-2)

(d-1) Gobi around Aksu (d-2) Surface shown in (d-1) (d-3) Same surface as (d-2)

(e-1) Gobi around Ejina (e-2) Surface shown in (e-1) (e-3) Same surface as (e-2)

Fig. 2. Landscape (left column) and surface condition (center and right columns) of all observationsites. Device on the tripod in pictures of left column is an albedo observation system and the box,on which the PC is set, is a spectrometer. Scale in pictures of center and right columns is 17.5 cmin full length.


the correct diffuse fraction was also obtained bycorrecting the measured diffuse and globalcomponents in downward solar flux.

For nadir reflectance measurement, down-ward flux was obtained by measuring the lightreflected from WRS and upward radiance wasmeasured by pointing the optical fiber tip to thenadir desert surface directly. For measureddownward flux, the same correction for cosineproperty of WRS as that for albedo was made.

The spectral albedo (nadir reflectance) of onedesert surface was calculated by averaging fivesamples obtained from measurement pairs fordownward flux and upward flux (upwardradiance). Since the desert surfaces were nothomogeneous, we repeated this procedure sev-eral times (N1 and N2 in Table 1) at each siteon different surfaces of the same surface type tokeep the representativity. The final spectralalbedo (nadir reflectance) at each site wasobtained by averaging these spectra. The stan-dard deviation of these spectra was also calcu-lated to examine the variation of albedos (nadirreflectance) at each site. For three major ab-sorption bands around l ¼ 1.36–1.40, 1.82–1.92, and 2.46–2.48 mm, mainly due to watervapor and carbon dioxide, the spectral albedos(nadir reflectance) are masked because ofvery low signal-to-noise ratio of the measured

spectra. For the usage of other applications,the spectrally-averaged albedo and nadir re-flectance with 20 nm resolution are shown inTables 2 and 3, respectively.

3. Result

Spectral albedo is shown in Fig. 3a and stan-dard deviation in Fig. 3b. Spectral albedo is low(0.05–0.11) at ultraviolet wavelengths but itrapidly increases to 0.2–0.3 at the wavelengthabout l ¼ 0:6 mm at any site. The wavelengthat which the albedo reaches to the maximumis around l ¼ 1.8–2.2 mm, where the albedosrange from 0.37 to 0.49 among measured sites.It is difficult to identify these spectral featuresare common or characteristic because the ex-isting data of comparable spectral albedo mea-sured in situ in other desert areas were verylimited. However, our results are different fromthe spectral feature of dry sand compiled inradiative transfer model Streamer 3.0 (Key2001), which is somewhat high (@0.09) atultraviolet wavelengths and low (0.15) atl ¼ 0:6 mm, and reaches to 0.38 at l ¼ 1:6 mm.According to the MODIS measurements, thebroadband albedos in the Sahara Desert arehigher than those in Chinese deserts (Schaafet al. 2002).

Maximum albedo was observed at farmland

(f-1) Sand dune around Ejina (f-2) Surface shown in (f-1) (f-3) Same surface as (f-2)

(g-1) Sand dune in Tengger desert (g-2) Surface shown in (g-1) (g-3) Same surface as (g-2)

Fig. 2 (continued)


Table 2. Spectral albedo of desert surfaces with spectral resolution of 20 nm

l (mm)Qira Gobi

QiraSand dune

QiraFarmland

AksuGobi

EjinaGobi

EjinaSand dune

TenggerSand dune

0.36 8.503E�02 7.392E�02 9.181E�02 1.050E�01 7.285E�02 6.847E�02 4.708E�020.38 9.667E�02 8.431E�02 1.050E�01 1.164E�01 8.126E�02 7.653E�02 5.309E�020.40 1.124E�01 9.918E�02 1.241E�01 1.322E�01 9.334E�02 8.857E�02 6.238E�020.42 1.271E�01 1.135E�01 1.417E�01 1.462E�01 1.048E�01 9.985E�02 7.219E�020.44 1.452E�01 1.316E�01 1.639E�01 1.633E�01 1.193E�01 1.148E�01 8.591E�020.46 1.601E�01 1.478E�01 1.825E�01 1.784E�01 1.315E�01 1.266E�01 9.872E�020.48 1.710E�01 1.600E�01 1.952E�01 1.896E�01 1.404E�01 1.346E�01 1.090E�010.50 1.852E�01 1.751E�01 2.123E�01 2.029E�01 1.525E�01 1.467E�01 1.235E�010.52 2.019E�01 1.931E�01 2.334E�01 2.184E�01 1.673E�01 1.621E�01 1.421E�010.54 2.197E�01 2.127E�01 2.559E�01 2.366E�01 1.843E�01 1.794E�01 1.655E�010.56 2.387E�01 2.339E�01 2.804E�01 2.615E�01 2.053E�01 1.994E�01 1.974E�010.58 2.552E�01 2.527E�01 3.027E�01 2.887E�01 2.263E�01 2.175E�01 2.317E�010.60 2.656E�01 2.655E�01 3.176E�01 3.078E�01 2.403E�01 2.281E�01 2.561E�010.62 2.711E�01 2.731E�01 3.257E�01 3.179E�01 2.468E�01 2.322E�01 2.693E�010.64 2.753E�01 2.792E�01 3.319E�01 3.251E�01 2.512E�01 2.346E�01 2.787E�010.66 2.790E�01 2.844E�01 3.374E�01 3.315E�01 2.553E�01 2.368E�01 2.872E�010.68 2.826E�01 2.887E�01 3.428E�01 3.377E�01 2.598E�01 2.392E�01 2.952E�010.70 2.863E�01 2.927E�01 3.487E�01 3.442E�01 2.650E�01 2.427E�01 3.037E�010.72 2.903E�01 2.971E�01 3.549E�01 3.510E�01 2.708E�01 2.470E�01 3.128E�010.74 2.945E�01 3.021E�01 3.613E�01 3.579E�01 2.762E�01 2.514E�01 3.215E�010.76 2.981E�01 3.065E�01 3.667E�01 3.635E�01 2.802E�01 2.546E�01 3.283E�010.78 3.013E�01 3.107E�01 3.718E�01 3.687E�01 2.834E�01 2.573E�01 3.334E�010.80 3.028E�01 3.130E�01 3.749E�01 3.720E�01 2.846E�01 2.581E�01 3.358E�010.82 3.029E�01 3.136E�01 3.763E�01 3.734E�01 2.842E�01 2.574E�01 3.362E�010.84 3.023E�01 3.134E�01 3.767E�01 3.743E�01 2.833E�01 2.559E�01 3.355E�010.86 3.015E�01 3.130E�01 3.767E�01 3.747E�01 2.821E�01 2.542E�01 3.346E�010.88 3.012E�01 3.130E�01 3.771E�01 3.754E�01 2.817E�01 2.533E�01 3.348E�010.90 3.015E�01 3.135E�01 3.782E�01 3.766E�01 2.820E�01 2.532E�01 3.361E�010.92 3.029E�01 3.153E�01 3.804E�01 3.787E�01 2.836E�01 2.542E�01 3.387E�010.94 3.044E�01 3.169E�01 3.827E�01 3.805E�01 2.857E�01 2.559E�01 3.420E�010.96 3.073E�01 3.206E�01 3.871E�01 3.847E�01 2.895E�01 2.587E�01 3.464E�010.98 3.115E�01 3.257E�01 3.932E�01 3.904E�01 2.947E�01 2.623E�01 3.524E�011.00 3.133E�01 3.279E�01 3.961E�01 3.929E�01 2.976E�01 2.644E�01 3.558E�011.02 3.148E�01 3.296E�01 3.987E�01 3.948E�01 3.001E�01 2.659E�01 3.590E�011.04 3.161E�01 3.310E�01 4.009E�01 3.963E�01 3.021E�01 2.672E�01 3.615E�011.06 3.175E�01 3.325E�01 4.029E�01 3.974E�01 3.039E�01 2.686E�01 3.641E�011.08 3.191E�01 3.341E�01 4.052E�01 3.985E�01 3.061E�01 2.699E�01 3.668E�011.10 3.208E�01 3.358E�01 4.073E�01 3.993E�01 3.085E�01 2.718E�01 3.697E�011.12 3.228E�01 3.377E�01 4.094E�01 3.997E�01 3.108E�01 2.740E�01 3.730E�011.14 3.259E�01 3.409E�01 4.127E�01 4.015E�01 3.143E�01 2.766E�01 3.769E�011.16 3.293E�01 3.450E�01 4.164E�01 4.040E�01 3.179E�01 2.799E�01 3.802E�011.18 3.328E�01 3.491E�01 4.200E�01 4.064E�01 3.214E�01 2.831E�01 3.839E�011.20 3.368E�01 3.535E�01 4.238E�01 4.086E�01 3.252E�01 2.867E�01 3.881E�011.22 3.411E�01 3.584E�01 4.281E�01 4.112E�01 3.297E�01 2.908E�01 3.925E�011.24 3.455E�01 3.634E�01 4.322E�01 4.136E�01 3.341E�01 2.952E�01 3.970E�011.26 3.497E�01 3.682E�01 4.362E�01 4.157E�01 3.379E�01 2.989E�01 4.011E�011.28 3.540E�01 3.732E�01 4.403E�01 4.181E�01 3.421E�01 3.029E�01 4.053E�011.30 3.579E�01 3.778E�01 4.439E�01 4.201E�01 3.458E�01 3.066E�01 4.089E�011.32 3.617E�01 3.822E�01 4.472E�01 4.216E�01 3.492E�01 3.103E�01 4.123E�011.34 3.648E�01 3.855E�01 4.491E�01 4.222E�01 3.523E�01 3.140E�01 4.157E�011.36 — — — — — — —1.38 — — — — — — —1.40 — — — — — — —1.42 3.635E�01 3.893E�01 4.501E�01 4.164E�01 3.400E�01 3.110E�01 4.024E�01


Table 2 (continued)

l (mm)Qira Gobi

QiraSand dune

QiraFarmland

AksuGobi

EjinaGobi

EjinaSand dune

TenggerSand dune

1.44 3.736E�01 3.999E�01 4.576E�01 4.209E�01 3.539E�01 3.211E�01 4.137E�011.46 3.785E�01 4.057E�01 4.619E�01 4.246E�01 3.612E�01 3.267E�01 4.207E�011.48 3.821E�01 4.103E�01 4.654E�01 4.280E�01 3.662E�01 3.305E�01 4.254E�011.50 3.866E�01 4.148E�01 4.687E�01 4.316E�01 3.709E�01 3.340E�01 4.295E�011.52 3.899E�01 4.185E�01 4.715E�01 4.344E�01 3.748E�01 3.369E�01 4.333E�011.54 3.922E�01 4.217E�01 4.738E�01 4.364E�01 3.780E�01 3.394E�01 4.362E�011.56 3.949E�01 4.248E�01 4.761E�01 4.384E�01 3.814E�01 3.422E�01 4.394E�011.58 3.979E�01 4.280E�01 4.785E�01 4.401E�01 3.849E�01 3.450E�01 4.426E�011.60 4.010E�01 4.312E�01 4.807E�01 4.417E�01 3.884E�01 3.478E�01 4.452E�011.62 4.040E�01 4.341E�01 4.829E�01 4.432E�01 3.915E�01 3.507E�01 4.476E�011.64 4.062E�01 4.364E�01 4.845E�01 4.443E�01 3.943E�01 3.531E�01 4.495E�011.66 4.080E�01 4.384E�01 4.856E�01 4.450E�01 3.966E�01 3.549E�01 4.507E�011.68 4.098E�01 4.405E�01 4.867E�01 4.461E�01 3.987E�01 3.564E�01 4.520E�011.70 4.115E�01 4.424E�01 4.873E�01 4.466E�01 4.004E�01 3.581E�01 4.539E�011.72 4.132E�01 4.445E�01 4.882E�01 4.470E�01 4.017E�01 3.598E�01 4.554E�011.74 4.151E�01 4.468E�01 4.899E�01 4.479E�01 4.037E�01 3.615E�01 4.567E�011.76 4.177E�01 4.494E�01 4.921E�01 4.491E�01 4.085E�01 3.644E�01 4.598E�011.78 4.194E�01 4.508E�01 4.933E�01 4.496E�01 4.112E�01 3.664E�01 4.621E�011.80 4.203E�01 4.513E�01 4.932E�01 4.494E�01 4.140E�01 3.684E�01 4.664E�011.82 — — — — — — —1.84 — — — — — — —1.86 — — — — — — —1.88 — — — — — — —1.90 — — — — — — —1.92 — — — — — — —1.94 3.962E�01 4.398E�01 4.860E�01 4.127E�01 3.729E�01 3.492E�01 4.171E�011.96 4.005E�01 4.432E�01 4.885E�01 4.170E�01 3.776E�01 3.498E�01 4.223E�011.98 4.041E�01 4.456E�01 4.903E�01 4.240E�01 3.842E�01 3.518E�01 4.287E�012.00 4.067E�01 4.466E�01 4.906E�01 4.289E�01 3.884E�01 3.541E�01 4.341E�012.02 4.119E�01 4.500E�01 4.919E�01 4.371E�01 3.977E�01 3.575E�01 4.433E�012.04 4.154E�01 4.530E�01 4.934E�01 4.419E�01 4.023E�01 3.605E�01 4.473E�012.06 4.175E�01 4.547E�01 4.941E�01 4.451E�01 4.072E�01 3.623E�01 4.504E�012.08 4.191E�01 4.556E�01 4.944E�01 4.472E�01 4.107E�01 3.656E�01 4.544E�012.10 4.177E�01 4.551E�01 4.939E�01 4.482E�01 4.100E�01 3.645E�01 4.539E�012.12 4.159E�01 4.535E�01 4.931E�01 4.481E�01 4.089E�01 3.639E�01 4.537E�012.14 4.144E�01 4.527E�01 4.928E�01 4.484E�01 4.076E�01 3.621E�01 4.511E�012.16 4.110E�01 4.501E�01 4.921E�01 4.477E�01 4.028E�01 3.571E�01 4.438E�012.18 4.010E�01 4.426E�01 4.874E�01 4.449E�01 3.882E�01 3.499E�01 4.337E�012.20 3.864E�01 4.293E�01 4.791E�01 4.392E�01 3.657E�01 3.352E�01 4.182E�012.22 3.814E�01 4.251E�01 4.759E�01 4.352E�01 3.575E�01 3.284E�01 4.129E�012.24 3.799E�01 4.244E�01 4.758E�01 4.354E�01 3.603E�01 3.284E�01 4.168E�012.26 3.763E�01 4.211E�01 4.740E�01 4.353E�01 3.622E�01 3.272E�01 4.184E�012.28 3.759E�01 4.207E�01 4.736E�01 4.350E�01 3.634E�01 3.264E�01 4.179E�012.30 3.666E�01 4.113E�01 4.678E�01 4.318E�01 3.527E�01 3.186E�01 4.115E�012.32 3.501E�01 3.955E�01 4.586E�01 4.251E�01 3.371E�01 3.032E�01 3.985E�012.34 3.356E�01 3.813E�01 4.501E�01 4.191E�01 3.200E�01 2.885E�01 3.849E�012.36 3.378E�01 3.848E�01 4.519E�01 4.183E�01 3.176E�01 2.921E�01 3.786E�012.38 3.486E�01 3.975E�01 4.601E�01 4.194E�01 3.248E�01 3.005E�01 3.834E�012.40 3.522E�01 4.005E�01 4.637E�01 4.170E�01 3.280E�01 3.014E�01 3.856E�012.42 3.503E�01 4.001E�01 4.619E�01 4.117E�01 3.243E�01 2.978E�01 3.761E�012.44 3.402E�01 3.929E�01 4.551E�01 4.068E�01 3.158E�01 2.821E�01 3.593E�012.46 — — — — — — —2.48 — — — — — — —

In three major absorption bands around l ¼ 1.36–1.40, 1.82–1.92, and 2.46–2.48 mm, the spectral valuesare masked because of very low signal to noise ratio of measured spectra.


Table 3. Spectral nadir reflectance from desert surfaces with spectral resolution of 20 nm

l (mm) Qira Gobi Qira Sand dune Ejina Gobi Ejina Sand dune Tengger Sand dune

0.36 7.059E�02 5.676E�02 5.967E�02 5.135E�02 3.101E�020.38 8.024E�02 6.422E�02 6.624E�02 5.694E�02 3.469E�020.40 9.340E�02 7.480E�02 7.594E�02 6.557E�02 4.081E�020.42 1.058E�01 8.503E�02 8.522E�02 7.370E�02 4.752E�020.44 1.216E�01 9.837E�02 9.765E�02 8.500E�02 5.754E�020.46 1.345E�01 1.104E�01 1.082E�01 9.405E�02 6.718E�020.48 1.436E�01 1.195E�01 1.158E�01 9.999E�02 7.491E�020.50 1.562E�01 1.310E�01 1.265E�01 1.095E�01 8.640E�020.52 1.713E�01 1.450E�01 1.401E�01 1.220E�01 1.017E�010.54 1.874E�01 1.604E�01 1.558E�01 1.364E�01 1.214E�010.56 2.047E�01 1.769E�01 1.751E�01 1.529E�01 1.487E�010.58 2.197E�01 1.914E�01 1.942E�01 1.681E�01 1.786E�010.60 2.287E�01 2.010E�01 2.067E�01 1.768E�01 2.001E�010.62 2.330E�01 2.063E�01 2.122E�01 1.798E�01 2.115E�010.64 2.361E�01 2.105E�01 2.157E�01 1.814E�01 2.196E�010.66 2.390E�01 2.141E�01 2.190E�01 1.829E�01 2.269E�010.68 2.419E�01 2.170E�01 2.224E�01 1.846E�01 2.337E�010.70 2.448E�01 2.196E�01 2.262E�01 1.869E�01 2.406E�010.72 2.480E�01 2.223E�01 2.302E�01 1.896E�01 2.477E�010.74 2.511E�01 2.253E�01 2.336E�01 1.922E�01 2.541E�010.76 2.539E�01 2.282E�01 2.367E�01 1.945E�01 2.597E�010.78 2.562E�01 2.308E�01 2.389E�01 1.962E�01 2.638E�010.80 2.573E�01 2.320E�01 2.398E�01 1.967E�01 2.658E�010.82 2.573E�01 2.323E�01 2.397E�01 1.961E�01 2.662E�010.84 2.564E�01 2.316E�01 2.385E�01 1.946E�01 2.654E�010.86 2.553E�01 2.307E�01 2.370E�01 1.928E�01 2.644E�010.88 2.547E�01 2.301E�01 2.360E�01 1.916E�01 2.640E�010.90 2.550E�01 2.303E�01 2.361E�01 1.913E�01 2.649E�010.92 2.560E�01 2.313E�01 2.372E�01 1.919E�01 2.669E�010.94 2.575E�01 2.329E�01 2.393E�01 1.935E�01 2.699E�010.96 2.594E�01 2.353E�01 2.420E�01 1.952E�01 2.734E�010.98 2.619E�01 2.387E�01 2.457E�01 1.976E�01 2.786E�011.00 2.634E�01 2.406E�01 2.481E�01 1.992E�01 2.815E�011.02 2.648E�01 2.417E�01 2.500E�01 2.002E�01 2.841E�011.04 2.658E�01 2.425E�01 2.514E�01 2.009E�01 2.861E�011.06 2.670E�01 2.433E�01 2.527E�01 2.019E�01 2.881E�011.08 2.686E�01 2.446E�01 2.547E�01 2.030E�01 2.905E�011.10 2.703E�01 2.461E�01 2.569E�01 2.046E�01 2.933E�011.12 2.724E�01 2.478E�01 2.593E�01 2.067E�01 2.967E�011.14 2.750E�01 2.503E�01 2.622E�01 2.089E�01 2.998E�011.16 2.775E�01 2.531E�01 2.648E�01 2.110E�01 3.025E�011.18 2.800E�01 2.564E�01 2.676E�01 2.135E�01 3.056E�011.20 2.832E�01 2.602E�01 2.711E�01 2.164E�01 3.092E�011.22 2.865E�01 2.639E�01 2.744E�01 2.193E�01 3.127E�011.24 2.902E�01 2.680E�01 2.780E�01 2.226E�01 3.163E�011.26 2.934E�01 2.720E�01 2.812E�01 2.257E�01 3.197E�011.28 2.969E�01 2.768E�01 2.849E�01 2.291E�01 3.234E�011.30 3.003E�01 2.810E�01 2.883E�01 2.322E�01 3.268E�011.32 3.035E�01 2.853E�01 2.914E�01 2.354E�01 3.299E�011.34 3.066E�01 2.893E�01 2.945E�01 2.387E�01 3.331E�011.36 — — — — —1.38 — — — — —1.40 — — — — —1.42 3.039E�01 2.914E�01 2.853E�01 2.371E�01 3.215E�011.44 3.126E�01 3.021E�01 2.976E�01 2.456E�01 3.318E�01


Table 3 (continued)

l (mm) Qira Gobi Qira Sand dune Ejina Gobi Ejina Sand dune Tengger Sand dune

1.46 3.171E�01 3.075E�01 3.033E�01 2.501E�01 3.378E�011.48 3.197E�01 3.114E�01 3.074E�01 2.530E�01 3.416E�011.50 3.233E�01 3.155E�01 3.110E�01 2.557E�01 3.454E�011.52 3.263E�01 3.190E�01 3.142E�01 2.580E�01 3.486E�011.54 3.285E�01 3.219E�01 3.168E�01 2.601E�01 3.512E�011.56 3.308E�01 3.250E�01 3.197E�01 2.626E�01 3.540E�011.58 3.335E�01 3.283E�01 3.231E�01 2.651E�01 3.570E�011.60 3.362E�01 3.314E�01 3.262E�01 2.676E�01 3.596E�011.62 3.389E�01 3.344E�01 3.291E�01 2.701E�01 3.621E�011.64 3.410E�01 3.371E�01 3.316E�01 2.722E�01 3.639E�011.66 3.431E�01 3.398E�01 3.342E�01 2.744E�01 3.658E�011.68 3.455E�01 3.428E�01 3.370E�01 2.765E�01 3.681E�011.70 3.479E�01 3.456E�01 3.396E�01 2.788E�01 3.704E�011.72 3.500E�01 3.483E�01 3.416E�01 2.809E�01 3.724E�011.74 3.514E�01 3.503E�01 3.433E�01 2.824E�01 3.735E�011.76 3.558E�01 3.559E�01 3.494E�01 2.863E�01 3.786E�011.78 3.587E�01 3.586E�01 3.527E�01 2.888E�01 3.814E�011.80 3.609E�01 3.611E�01 3.562E�01 2.910E�01 3.851E�011.82 — — — — —1.84 — — — — —1.86 — — — — —1.88 — — — — —1.90 — — — — —1.92 — — — — —1.94 3.320E�01 3.470E�01 3.205E�01 2.770E�01 3.397E�011.96 3.372E�01 3.509E�01 3.251E�01 2.763E�01 3.452E�011.98 3.425E�01 3.533E�01 3.305E�01 2.781E�01 3.520E�012.00 3.476E�01 3.565E�01 3.353E�01 2.811E�01 3.588E�012.02 3.567E�01 3.642E�01 3.472E�01 2.874E�01 3.697E�012.04 3.590E�01 3.655E�01 3.503E�01 2.894E�01 3.727E�012.06 3.623E�01 3.692E�01 3.562E�01 2.922E�01 3.774E�012.08 3.643E�01 3.696E�01 3.588E�01 2.941E�01 3.804E�012.10 3.637E�01 3.698E�01 3.587E�01 2.939E�01 3.813E�012.12 3.635E�01 3.692E�01 3.586E�01 2.945E�01 3.828E�012.14 3.622E�01 3.687E�01 3.582E�01 2.933E�01 3.810E�012.16 3.566E�01 3.633E�01 3.513E�01 2.874E�01 3.719E�012.18 3.479E�01 3.542E�01 3.374E�01 2.801E�01 3.607E�012.20 3.335E�01 3.389E�01 3.160E�01 2.668E�01 3.448E�012.22 3.254E�01 3.327E�01 3.065E�01 2.583E�01 3.379E�012.24 3.258E�01 3.340E�01 3.102E�01 2.596E�01 3.435E�012.26 3.240E�01 3.310E�01 3.126E�01 2.592E�01 3.463E�012.28 3.234E�01 3.305E�01 3.140E�01 2.582E�01 3.467E�012.30 3.132E�01 3.202E�01 3.029E�01 2.500E�01 3.380E�012.32 3.018E�01 3.083E�01 2.911E�01 2.396E�01 3.302E�012.34 2.876E�01 2.901E�01 2.747E�01 2.270E�01 3.149E�012.36 2.885E�01 2.933E�01 2.735E�01 2.282E�01 3.114E�012.38 2.984E�01 3.044E�01 2.824E�01 2.383E�01 3.166E�012.40 3.007E�01 3.131E�01 2.839E�01 2.410E�01 3.166E�012.42 2.982E�01 3.087E�01 2.806E�01 2.339E�01 3.090E�012.44 2.930E�01 3.089E�01 2.751E�01 2.290E�01 3.000E�012.46 — — — — —2.48 — — — — —

In three major absorption bands in l ¼ 1.36–1.40, 1.82–1.92, and 2.46–2.48 mm, the spectral values aremasked because of very low signal to noise ratio of measured spectra.


in Qira and minimum in sand dune aroundEjina. Aoki et al. (2002) explained the high al-bedo on farmland in Qira is due to fine grainsize of soil particles. Colors of desert surfacesby natural view were brown to yellow for allsurfaces except for sand dune in Tengger Des-ert, which looked red as shown in Figs. 2g-1–2g-3. This difference in color is seen in Fig. 3a,where the spectral albedo in Tengger Desertincreases most rapidly in the visible region.These results suggest that optical properties ofmineral dust, of which ground soil is a source,could be different from place to place even inChinese deserts. From standard deviationshown in Fig. 3b, we see that the spectral albe-dos of gobi surfaces around Aksu and Ejina

vary widely compared with those of other sur-faces, and their spectral variations are large.

Spectral nadir reflectance and its standarddeviation are shown in Figs. 4a and 4b, respec-tively, which basically show the similar spec-tral distributions with those of the albedos.However, the values of nadir reflectance arelower than those of albedo at any wavelength.When the direct solar beam illuminates thesurface, the instrument views both of the sur-faces of soil particles and their shadows exceptfor the case of y0 ¼ 0�. The effect of the shadowsreduces both the albedo and the nadir re-flectance, where the effect depends on y0. In therange of y0 in our observations, this effect ofshadows would be more pronounced for thenadir reflectance than the albedo.

4. Summary

Spectral albedo and nadir reflectance ofdesert surfaces were measured with spectrom-

Fig. 3. (a) Spectral albedo and (b) itsstandard deviation measured atobservation sites shown in Table 1. Inthree major absorption bands aroundl ¼ 1.36–1.40, 1.82–1.92, and 2.46–2.48 mm, the spectral values aremasked because of very low signal-to-noise ratio of measured spectra.

Fig. 4. Same as Fig. 3, but for nadirreflectance.


eter for spectral region from 0.35 to 2.5 mm inwestern and central China in spring of 2001and autumn of 2003. Observed desert surfacesare mainly gobi and/or sand dune around Qirain southern Taklimakan Desert, Aksu innorthern Taklimakan Desert, Ejina in BadainJaran Desert and in Tengger Desert. Spectralmeasurements at each site were made on sev-eral different surfaces of the same surface typeto keep the representativity. The variation ofspectral albedo among measured sites suggeststhat optical properties of mineral dust coulddiffer from place to place, even in Chinese de-serts. Nadir reflectance basically shows a simi-lar spectral distribution with that of the albedo,while the values are lower at any wavelength.These datasets would be useful for the studiesof climate simulation, remote sensing, radiativeforcing, and so on.

Acknowledgments

This work was conducted as part of theAeolian Dust Experiment on Climate Impact(ADEC) project during the period from 2000 to2004 supported by the Ministry of Education,Culture, Sports, Science and Technology ofJapan. English check by Ms. Reiko Watanabewas very helpful to revise the manuscript.

References

Aoki, Te., Ta. Aoki, M. Fukabori, A. Hachikubo, Y.Tachibana, and F. Nishio, 2000: Effects of snowphysical parameters on spectral albedo andbidirectional reflectance of snow surface. J.Geophys. Res., 105, 10219–10236.

———, M. Mikami, and W. Liu, 2002: Spectralalbedos of desert surfaces and size distribu-tions of soil particles measured around Qira

and Aksu in Taklimakan Desert. J. Arid LandStudies, 11, 259–266.

———, T. Tanaka, A. Uchiyama, M. Chiba, M.Mikami, and J.R. Key, 2005: Sensitivity ex-periments of direct radiative forcing caused bymineral dust simulated with a chemical trans-port model. J. Meteorol. Soc. Japan, 83A, 315–331.

Jin, Y., C.B. Schaaf, C.E. Woodcock, F. Gao, X. Li,A.H. Strahler, W. Lucht, and S. Liang, 2003:Consistency of MODIS surface bidirectionalreflectance distribution function and albedoretrievals: 2. Validation. J. Geophys. Res., 108,4159, doi:10.1029/2002JD002804.

Key, J.R., 2001: Streamer User’s Guide, CooperativeInstitute for Meteorological Satellite Studies,University of Wisconsin, 96 pp.

Liang, S., H. Fang, M. Chen, C.J. Shuey, C. Walthall,C. Daughtry, J. Morisette, C. Schaaf, and A.Strahler, 2002: Validating MODIS land surfacereflectance and albedo products: methods andpreliminary results. Remote Sens. Environ., 83,149–162.

Liao, H. and J.H. Seinfeld, 1998: Radiative forcingby mineral dust aerosols: sensitivity to keyvariables. J. Geophys. Res., 103, 31637–31645.

Schaaf, C.B., F. Gao, A.H. Strahler, W. Lucht, X.Li, T. Tsang, N.C. Strugnell, X. Zhang, Y. Jin,J.-P. Muller, P. Lewis, M. Barnsley, P. Hobson,M. Disney, G. Roberts, M. Dunderdale, C. Doll,R.P. d’Entremont, B. Hu, S. Liang, J.L. Priv-ette, and D. Roy, 2002: First operationalBRDF, albedo nadir reflectance products fromMODIS. Remote Sens. Environ., 83, 135–148.

Sokolik, I.N. and O.B. Toon, 1996: Direct radiativeforcing by anthropogenic airborne mineralaerosols. Nature, 381, 681–683.

Tegen, I., A.A. Lacis, and I. Fung, 1996: The influ-ence on climate forcing of mineral aerosolsfrom disturbed soils. Nature, 380, 419–422.


Documents

NOTES AND CORRESPONDENCE Spectral Albedo of Desert ...dust.ess.uci.edu/ppr/ppr_AMY05.pdf · in northern Taklimakan Desert, Ejina in Badain Jaran Desert, and in Tengger Desert. The