Differences between Red and Green NDVI, What do they predict and what they don’t predict Shambel Maru
Dec 28, 2015
Differences between Red and Green NDVI, What do they predict and what
they don’t predict
Differences between Red and Green NDVI, What do they predict and what
they don’t predict
Shambel Maru
The science and art of obtaining information about an object, area, or phenomenon through the analysis of data acquired by a device that is not in contact with object, area, or phenomenon under investigation.
The science and art of obtaining information about an object, area, or phenomenon through the analysis of data acquired by a device that is not in contact with object, area, or phenomenon under investigation.
Remote SensingRemote Sensing
What do sensors measure?What do sensors measure?
Measures the amount of variability of the the light source
It filters the light coming to it along the path way
What is actually being measured in sensor systems is spectral radiance or the radiant energy from the target
LightSource
Plant or SoilSurface
Reflected Light
SensingSystem
Light specturmLight specturm
Major components of visible light spectrum are violet, blue, green, yellow, orange and red
Blue and red are used in photosynthesis
Light absorption at different wavelengths
Green leaves have a reflectance of 20 percent or less in the 500 to 700 nm range (green to red) and about 60 percent in the 700 to 1300 nm range (near infra-red).
1=bacteriochlorophyll, 2=chlorphyll a, 3= chlorophyll b, 4=phycoerythrobilin, 5= beta carotene)
VISIBLE Color AbsorbedVISIBLE Color Absorbed
VISIBLE Color TransmittedVISIBLE Color TransmittedVISIBLE Color Transmitted
Violet Blue Green Yellow Orange RedVioletViolet BlueBlue GreenGreen YellowYellow Orange Orange RedRed
Short wavelengthShort wavelengthHigh frequencyHigh frequencyHigh energyHigh energy
Long wavelengthLong wavelengthLow frequencyLow frequencyLow energyLow energy
0.01 10 380 450 495 570 590 620 750 1x106 1x1011
wavelength, nm0.010.01 1010 380380 450450 495495 570570 590590 620620 750750 1x101x1066 1x101x101111
wavelength, nmwavelength, nm
Gam
ma
Ray
sG
amm
a R
ays
Gam
ma
Ray
s
X-R
ays
XX-- R
ays
Ray
s
Ultr
avio
let
Ultr
avio
let
Ultr
avio
let
Infr
ared
Infr
ared
Infr
ared
Mic
row
aves
and
sho
rt r
adio
Mic
row
aves
and
sho
rt r
adio
Mic
row
aves
and
sho
rt r
adio
Rad
io, F
M, T
VR
adio
, FM
, TV
Rad
io, F
M, T
V
Electronic Vibrational Rotationaltransitions transitions transitionsElectronicElectronic VibrationalVibrational RotationalRotationaltransitionstransitions transitionstransitions transitionstransitions
YellowYellow--greengreen YellowYellow VioletViolet BlueBlue GreenGreen--blueblue BlueBlue--greengreen
Characteristics of visible and non-visible portions of the spectra
NDVI, what is it?NDVI, what is it? It is normalized difference vegetation index. Used to measure green biomass (Tucker, 1979) Degree of greenness = chlorophyll concentration Actually measure photo synthetically active radiation absorbed by the canopy (Sellers, 1985) NDVI values vary with absorption of red light by
plant chlorophyll and the reflection of infrared radiation by water-filled leaf cells. It is correlated with Intercepted Photo-synthetically Active Radiation (IPAR).
NDVI, what is it?NDVI, what is it? It is a function of Incident and reflected
light
RNDVI= NIR – Red ,
NIR + Red
GNDVI= NIR – Green ,
NIR + Green
NIR 750-1300 nm Red 600-700 nm Green 550 nm
Where 0< NDVI< 1
Differences between the Green and Red Differences between the Green and Red NDVINDVI
Red band is used to measure green biomass and estimate changes in vegetation state, but it is only sensitive to low chlorophyll-a concentration
(3-5 g/cm2) (Gitelson et al., 1997) Good at early stage
The Green band (520-630nm) is sensitive to a wide range of chlorophyll-a concentration (0.3 – 45 g/cm2) (Gitelson et al., 1997)
May work for late crop stages prediction
What do NDVI predict?What do NDVI predict?
Grain yield from reflectance readings of
NIR and Red (Tucker et al., 1981) Important in variable rate fertilizer
recommendation based on predicted yield Relationship decrease as wheat became ripened. Two late season readings (Feekes 10.5, flowering to
grain filling) may give more stable prediction as compared to that of a single reading (Pinter et al., 1981)
Grain yield from reflectance readings of
NIR and Red (Tucker et al., 1981) Important in variable rate fertilizer
recommendation based on predicted yield Relationship decrease as wheat became ripened. Two late season readings (Feekes 10.5, flowering to
grain filling) may give more stable prediction as compared to that of a single reading (Pinter et al., 1981)
Wheat biomass (NIR) and Nitrogen uptake reliably predicted at Feekes 4 & 5 (Stone et at., 1997, Lukina et al., 2000)
At the same stage percent ground cover and NDVI are co-related with vegetative biomass of wheat.
Wheat biomass (NIR) and Nitrogen uptake reliably predicted at Feekes 4 & 5 (Stone et at., 1997, Lukina et al., 2000)
At the same stage percent ground cover and NDVI are co-related with vegetative biomass of wheat.
What do NDVI predict?What do NDVI predict?
EY
Relationship between Estimated Yield (EY) computed from NDVI at Feekes growth stages 4 and 5, divided by growing degree days and observed grain yield, at six locations, 1998 and 1999.
EY = NDVI1 + NDVI2
GDD
0
1
2
3
4
5
6
0 0.001 0.002 0.003 0.004 0.005 0.006 0.007
INSEY
Gra
in Y
ield
, Mg
ha
-1N*P Perkins, 1998
S*N Perkins, 1998
S*N Tipton, 1998
N*P Perkins, 1999
Experiment 222, 1999
Experiment 301, 1999
Efaw AA, 1999
Experiment 801, 1999
Experiment 502, 1999
N*P Perkins, 2000
Experiment 222, 2000
Experiment 301, 2000
Efaw AA, 2000
Experiment 801, 2000
Experiment 502, 2000
Hennessey, AA, 2000
y = 4E+07x3 - 296260x2 + 970.66x
R2 = 0.64
N o rm alized D iffe re n ce V eg eta tio n In d ex (N D V I)
= N n IR ref – red ref / N n IR ref + red ref
N o rm alized D iffe re n ce V eg eta tio n In d ex (N D V I)
= N n IR ref – red ref / N n IR ref + red ref
(up – dow n)(up – dow n)
exce llen t p red ic to r o f p lan t N up takeexce llen t p red ic to r o f p lan t N up take
0
20
40
60
80
100
120
140
160
180
200
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
NDVI, Feekes 4-6
Ear
ly-s
easo
n pl
ant N
upt
ake,
kg
ha-1
N*P Perkins, 1998
S*N Perkins, 1998
S*N Tipton, 1998
transect Stillw ater, 1999
transect Perkins, 1999
transect Efaw , 2000, Jan
transect Perkins, 2000 Jan
transect Efaw , 2000 Mar
transect Perkins, 2000 Mar
y = 1019.5x3 - 1507.5x2 + 811.5x - 130.32R2 = 0.78
The image above from July 1989 depicting the continent of Africa from the NOAA AVHRR weather satellite illustrates the NDVI concept. Areas in yellow such as the Sahara desert have very low levels of vegetation (low NDVI). Areas in red the such as the tropics along the equator are highly vegetated (high NDVI). The second image was acquired 6 months earlier and illustrate the effect of less rain on the NDVI's (a lowering effect) – notice the shrinking area of reddish areas along the equator.
The Sahel Zone
The image shows senescence in northern hardwoods in some areas, and higher NDVI in conifer dominated areas in the north, and aspen-birch, and mixed conifer and deciduous forests in other parts of the Lake Superior Basin.
Limitations of NDVILimitations of NDVI
It can only measure the surface vegetation biomass at late growth stages
Reading is 2 dimensional not 3D
It can not predict the amount of nitrogen concentration in the plant
Questions?