Vegetation Dynamics and SeasonalResponses of North and South Americafrom EOS-MODIS Vegetation Indices

Alfredo Huete1, Kamel Didan1, Tomoaki Miura1,Edna P. Rodriguez1, Xiang Gao1,

and Laerte G. Ferreira2

1Department of Soil, Water, and Environmental Science,

University of Arizona, Tucson, AZ2Universidade Federal de Goiás - UFG

Goiânia, GO, 74.001-970 Brazil

MODIS (MOD 13)Vegetation Index Products

K The MODIS Products include 2 Vegetation Indices(NDVI, EVI) and QA produced at 16-day and monthlyintervals at 250m/ 500m, 1km, and 25km resolutions

K The narrower ‘red’ MODIS band provides increasedchlorophyll sensitivity (band 1),

K The narrower ‘NIR’ MODIS band avoids highly variablewater vapor absorption (band 2),

K Use of the blue channel in the EVI provides aerosolresistance.

Global EVI composite (Sept. 30 - Oct. 15, 2000)(Spatial and temporal intercomparisons of vegetation

activity)

Hammer Aitoff Projection

1000 km 5000 km

TBRS, University of ArizonaSCF Group, http://tbrs.arizona.edu

water 0.2 0.4 0.6 0.8 1.0

EVI

TBRS, University of ArizonaSCF Group, http://tbrs.arizona.edu

NDVI

water 0.2 0.4 0.6 0.8 1.0

Colorado River DeltaVegetation Spatial and Temporal Dynamics with MODIS NDVI 2000-

2001

June 9May 25

Apr 7

Mar 22

Mar 6

Feb 18

Feb 2

Jan 17

Jan 1Dec 18 Dec 2

Spring Summer

FallWinter

Sept 13

Sept 29

Oct 15

Oct 31Nov 16

June 25

July 11

Aug 12

Aug 28

MODIS Vegetation Indices

ARVIAtmosphereResistance

)Re(Re*)1(

LdNIR

dNIRL+ρ+ρ

ρ−ρ+

SAVI

Canopy BackgroundCorrection

Soil-adjusted Vegetation Index

NDVINormalized DifferenceVegetation Index

dNIR

dNIR

Re

Re

ρρ

ρρ

+

−

EVI

Enhanced Vegetation Index

)Re(Re*

21 blueCdCNIRLdNIRG

ρρρ

ρρ

−++

−

)(

,

ReRe dbluedrb

rbNIR

rbNIR

ρ−ργ−ρ=ρρ+ρρ−ρ

Geographic location of validation activitiesand time series plots.

Cascades

Jornada &White Sands

KonzaPrairieKonzaPrairie

SALSA

Sevilleta

HarvardForestHarvardForest

MODIS MNR (RGB) DOY 225-240

MODIS MNR (RGB) DOY 225-240

Harvard Forest (mixedhardwood and conifer forests)

Cascades (mixedconifer forest)

Konza TallgrassPrairie

SALSA (semiaridshrub/ grass)

White SandsSevilleta LTER(semiarid shrub/ grass)

Jornada LTER (semiaridshrub/ grass)

North American sites withdiverse vegetationconditions

Seasonal Patterns of Major Biomesin North America

Effective LAI0 2 4 6 8

0.0

0.2

0.4

0.6

ZeroDarkBright

25% 50% 95%

Needleleaf Broadleaf

Effective LAI0 2 4 6 8

0.0

0.2

0.4

0.6

0.8

1.0

Broadleaf

Effective LAI0 2 4 6 8

0.0

0.2

0.4

0.6

0.8

1.0

Needleleaf

25% 50% 95%ZeroDarkBright

EVI

NDVI

-30

-20

-10

0

10

20

30

49 70 91 112

133

154

175

196

217

238

259

280

301

322

343

364 19 40 61 82 103

124

145

166

187

208

229

250

271

Tem

pera

ture

, C

Semi-arid Vegetation DynamicsApril 8 May 26 June 26 July 13

July 29Aug 14Sep 15Oct 1

WGEW WGEW –– ExotechExotech EVI EVI

Airplane Data

Ground Data

MODIS 250m EVI

EVI from MODIS & EVI from MODIS & ExotechExotech

300m AGL 150m AGL

NDVI

EVI

Dry Season(129 ~ 144)

Wet Season(200 ~ 208)

NDVI EVI

JER

###############################################################################################################################################

##################################################################################

#############################################################################################

#########################################

############################

################################################################################

######################################################

###################################################################

#############################################################

##########################

##################################################################################################################################################################################################################

#

Jornada Jornada –– ExotechExotech EVI EVI

Green = high EVIOrange = low EVI

ETM+ NDVI

Scaling:ground > mquals > ETM+ > MODIS

Ground ASD MQUALS ASDScaling up

MQUALS UTM ETM+ UTM Scaling up

Band convolution

ASDBand

convolutionETM+

MODIS

ETM+ UTM MODIS ISIN

Scaling up

Reprojection

MQUALS vs. ETM+ (JER)(May 2000/2001)

MQUALS reflectance/ VIs0.0 0.1 0.2 0.3 0.4

ET

M+

refle

ctan

ce/ V

Is

0.0

0.1

0.2

0.3

0.4

MMMMMMMMMMMMMM

MM MM M

MMMMM MMMM

MMMMMM

MMM M

MMMMM MMMM

MMM

MMMM

MM

M

MMMM

MMMM MM

MMMMMMMMM

MMMMMMMMMMMM

MMMMMMMCCC

CCCCCCCCC CC CCC

CC

CCC

CC CCC

C

CCC

CC

CCC

CCC

CCC

CCCCCC

CCCC

CC

C

CC

CC

C

C

CC

CCC

CCC

C

CCCCC

CCCCC C

CCCCCCCCCC CCCCCC

CCCCCCC

CCCC

CCCCCCCCCCCCCCCC

CCCCCCC

CCCC

M

M

M

MM

grassland transition mesquiteCreosoteC

C

C

CC

BlueRedNIRNDVIEVI

MQUALS reflectance/ VIs0.0 0.1 0.2 0.3 0.4

ET

M+

reflectance/ VIs

0.0

0.1

0.2

0.3

0.4

M

M

M

M

MC

C

C

C

C

ET

M+

Ref

lect

ance

s/ V

I’s

MQUALS vs. MODIS (JER)(May 2000/2001)

MQUALS reflectances/ VIs0.0 0.1 0.2 0.3 0.4

MO

DIS

reflectances / VIs (250m

)

0.0

0.1

0.2

0.3

0.4

CCCCCCCCCCCC CC

CCCC

C CC

CCC

C

C

CCCC

C CC

CC

C

C

C

CCCCCCCCCCC CC C

CCCCCCCCCCCCCCMMMMMMMMMMMMMMMM

MMMMMMMMMMMMM

M

MM

MMMMMMMM

MM

M

M

MMMMMMMMMMMMMM

MMMMMMMMMMMMMM

MQUALS reflectances/ VIs0.0 0.1 0.2 0.3 0.4

MO

DIS

reflectances / VIs (250m

)

0.0

0.1

0.2

0.3

0.4

C

C

C

C

CM

M

M

M

M

BlueRedNIRNDVIEVI

M

M

M

MM

C

C

C

CC

grassland transition creosotemesquite

MO

DIS

Ref

lect

ance

s/ V

I’s

Multitemporal VI Comparisons(MODIS, ETM+, MQUALS, Ground)

DOY120 140 160 180 200 220 240 260 280 300

Vegetation Indices

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

E E

E

E

E EE

E

Mesquite

rain

DOY120 140 160 180 200 220 240 260 280 300

Vegetation Indices

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

E E

E

E

E EE

E

composited MODIS (500m)nadir single day (500m)ETM+EMQUALS

Creosote

rain

E

G Ground G

NDVI EVI

DOY120 140 160 180 200 220 240 260 280 300

Vegetation Indices

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

E E

E

E

E E

E

E

G

G G

G

G G

G

G

Transition

rain

DOY120 140 160 180 200 220 240 260 280 300

Vegetation Indices

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

EE

E

E

E E

E

E

G

G GG

GG

GG

GrasslandDOY

240 250 260 270 280 290

Biophysical

0

10

20

30

40

50% Green coverGreen LAI *100

rain

Seasonal Patterns of Major Biomesin South America

ETM+DOY 211,2001

(Tapajos)

Seasonality observed by MODISTapajos National Forest, Brazil

Dry Season

Brasilia National ParkBrasilia National Park

• Blue = ASD

•Yellow = DigitalImages

Brazil Olympus Digital ImagesBrazil Olympus Digital Images

ASD from Brasilia NPASD from Brasilia NP

Araguaia Araguaia National ParkNational Park

• Red = ASD

•Yellow = DigitalImages

MODIS VI Long Term Stability MonitoringMODIS VI Long Term Stability Monitoring

Saharan Desert Sites (h18v05)

MOD13A2 (V3)Mar 6 – Mar 21, 2001False Color Composite (1, 2, 3)

TzTzTz2Tz2

MODIS VI Long Term Stability MonitoringMODIS VI Long Term Stability Monitoring

Atacama Desert Sites (h11v11)

MOD13A1 (V3)Sep 30 – Oct 15, 2001False Color Composite (1, 2, 3)

AtBAtB

AtDAtD

AtMAtM

AtPAtP

AtAt

MODIS VI Long Term Stability MonitoringMODIS VI Long Term Stability Monitoring

VI vs. Sun Zenith Angle PlotsNDVI (Daily)NDVI (Daily) EVI (Daily)EVI (Daily)

MODIS VI Long Term Stability MonitoringMODIS VI Long Term Stability Monitoring

VI Comparisons with Time-LagsNDVI (Daily)NDVI (Daily) EVI (Daily)EVI (Daily)

*α – 1 month ~ 1 year time lags with similar geometries.

MODIS Vegetation Index performanceMODIS Vegetation Index performance

EVI EVI_QA NDVI NDVI_QA

Nov

embe

r 16

-30

1-1

5

NDVI

water 0.2 0.4 0.6 0.8 1.0water 0.2 0.4 0.6 0.8 1.0

EVI

MODIS Vegetation Index performance

Transect 1

Transect 4

Transect 3

Transect 2

Transect 1: This transect spans different vegetation types under different atmosphere conditions. It also show the response of both VIs to aerosol, clouds, snow and vegetation activities)

Transect 1 (pixel 200 to 400):Medium to high aerosol is prevalent is this section of the Transect. EVI response is more stable and shows more resistance to aerosol. NDVI on the other hand is very unstable.

VI Validation

• VI products are provisionally validated,however,

• Vegetation Indices are widely used inregional to global applications,

• It is these intended uses or product outputs,that need to be validated,

• Product inputs, however, propagate into theoverall uncertainty of the VI (atmospherecorrection, residual clouds, calibration)

Next Steps in Validation

• Summer 2001 campaigns in Brazil andArgentina need to be analyzed,

• BRDF and cloud effects remain in thecompositing routine,

• Snow/ice problems persist in both VI’s,

• Assessment of feasibility of using snowproduct and BRDF products.