land surface phenology versus traditional observations
DESCRIPTION
Land surface phenology versus traditional observations. Márta Hunkár (1) , Ildikó Szenyán , Enikő Vincze, Zoltán Dunkel (2) Tibor Szerdahelyi (3) (1) Georgikon Faculty University of Pannonia; hunkar@ georgikon.hu - PowerPoint PPT PresentationTRANSCRIPT
1
Land surface phenology versus traditional observations
Márta Hunkár (1) , Ildikó Szenyán , Enikő Vincze, Zoltán Dunkel (2)Tibor Szerdahelyi (3)
(1) Georgikon Faculty University of Pannonia; [email protected](2) Hungarian Meteorological Service; [email protected]; [email protected]; [email protected](3) Szent István University Institute of Botany and Ecophysiolgy; [email protected]
EMS&ECAM 2013 09-13 Sept. Reading UK
EMS&ECAM 2013 09-13 Sept. Reading UK
2
Outline•Aims•Material•Methods•Results•Conclusions
EMS&ECAM 2013 09-13 Sept. Reading UK
3
Aims of our study•Retrieve land surface phenological data
from satellite observations•Comparing traditional phenological
observation data with land surface phenological data
4
EMS&ECAM 2013 09-13 Sept. Reading UK
IntroductionTraditional phenology Land surface phenology
• Dates of phenological phases
• Based on visual observation
• Species-level• Local
• Dates of critical values of vegetation indices
• Based on satellite observation
• Ecosystem level• Areal
EMS&ECAM 2013 09-13 Sept. Reading UK
5
Material
47,6802 o N;19,4217 o E
47,6802 o N;19,4836 o E
47,6385 o N;19,4063 o E
47,6385 o N;19,4681 o E
• Sample area: Iklad (5 km x 5 km) covered by different agricultural and native plants
EMS&ECAM 2013 09-13 Sept. Reading UK
6
Material• Satellite data:
▫ Enhanced Vegetation Index (EVI) from MODIS placed at Terra and Aqua satellites (NASA, LP DAAC, 2011)
from MODIS placed at Terra and Aqua satellites. Data are provided by NASA Land Processes Distributed Active Archive Center (NASA, LP DAAC, 2011). The MODIS Vegetation Index algorithm operates on a per-pixel basis and relies on multiple observations over a 16-day period to generate a composited Vegetation Index both from Terra and Aqua.
We used high resolution (250m x250m pixels) composite data with 8 day frequency for the last ten years (2003-2012) on a sample area with the size of
The MODIS Vegetation Index algorithm operates on a per-pixel basis and relies on multiple observations over a 16-day period to generate a composited Vegetation Index both from Terra and Aqua satellites overlapped with 8 days.
We got high resolution (250m x250m pixels) composite data with 8 day frequency for the last ten years (2003-2012) .The average EVI of that 400 pixels was used to characterise land surface vegetation.
where L is a soil adjustment factor, and C1 and C2 are coefficients used to correct aerosol scattering in the red band by the use of the blue band. The ρblue, ρred, and ρnir represent reflectance at the blue (0.45-0.52μm), red (0.6-0.7μm), and near-infrared (NIR) wavelengths (0.7-1.1μm), respectively. In general, G=2.5, C1=6.0, C2=7.5, and L=1.
where L is a soil adjustment factor, and C1 and C2 are coefficients used to correct aerosol scattering in the red band by the use of the blue band. The ρblue, ρred, and ρnir represent reflectance at the blue (0.45-0.52μm), red (0.6-0.7μm), and near-infrared (NIR) wavelengths (0.7-1.1μm), respectively. In general, G=2.5, C1=6.0, C2=7.5, and L=1.
where L is a soil adjustment factor, and C1 and C2 are coefficients used to correct aerosol scattering in the red band by the use of the blue band. The ρblue, ρred, and ρnir represent reflectance at the blue (0.45-0.52μm), red (0.6-0.7μm), and near-infrared (NIR) wavelengths (0.7-1.1μm), respectively. In general, G=2.5, C1=6.0, C2=7.5, and L=1.
EMS&ECAM 2013 09-13 Sept. Reading UK
7
• where L is a soil adjustment factor, and C1 and C2 are coefficients used to correct aerosol scattering in the red band by the use of the blue band. The ρblue, ρred, and ρnir represent reflectance at the blue (0.45-0.52μm), red (0.6-0.7μm), and near-infrared (NIR) wavelengths (0.7-1.1μm), respectively. In general, G=2.5, C1=6.0, C2=7.5, and L=1.
redred
blue
red
nir
red
nir
LCCGEVI
21
1
Material
EMS&ECAM 2013 09-13 Sept. Reading UK
8
•Traditional observations▫Within the same sample area in 2011 and
2012 Plants:
locust (Robinia pseudacacia); poplar (Populus spp); sessile oak, (Quercus petraea); winter wheat (Triticum aestivum); maize (Zea Mays); sunflower (Helianthus annuus);
Material
native plants
agricultural plants
EMS&ECAM 2013 09-13 Sept. Reading UK
9
•The site of the traditional observations were identified by pixel-level (250m x 250m)
•The EVI values of those pixels were analysed.
Material
EMS&ECAM 2013 09-13 Sept. Reading UK
10
Methods• Spatial average of high resolution EVI data is
used to characterize of the vegetation dynamics. • Temporal variation in EVI data are modelled using
piecewise sigmoid models. Each growth cycle is modelled using two sigmoid functions: one for the growth phase, one for the senescence phase.
EMS&ECAM 2013 09-13 Sept. Reading UK
11
•To identify the land surface phenological transition dates, the rate of change in the curvature of the fitted logistic models is used for each year.
•Specifically, transition dates correspond to the times at which the rate of change in curvature in the EVI data exhibits local minima or maximums.
EMS&ECAM 2013 09-13 Sept. Reading UK
12
Time course of the spatial average EVI values for the sample area
1 15 29 43 57 71 85 99 1131271411551691831972112252392532672812953093233373513650
0.1
0.2
0.3
0.4
0.5
0.6
fitted 2011fitted 2012obs 2011obs 2012
EMS&ECAM 2013 09-13 Sept. Reading UK
13
rate of change in the curvature of the fitted logistic models
1 17 33 49 65 81 97 113129145161177193
20112012
DOY
129143157171185199213227241255269283297311325339353
DOY
F1
F2
F3
F4
Decreasing period of EVIIncreasing period of EVI
EMS&ECAM 2013 09-13 Sept. Reading UK
14
Causes of shifting?
32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 650.0
10.0
20.0
30.0
40.0
50.0
60.0
Above zero temperature sum from 1st of February
20112012
DOY
oC
EMS&ECAM 2013 09-13 Sept. Reading UK
15
•The corresponding phenological transition dates are defined as ▫the onset of greenness increase (F1), ▫the onset of greenness maximum (F2), ▫the onset of greenness decrease (F3), and ▫the onset of greenness minimum (F4).
2002 2004 2006 2008 2010 2012 20140
50
100
150
200
250
300
350
F1F2F3F4D
OY
EMS&ECAM 2013 09-13 Sept. Reading UK
16
Trends of the length of the growing season(F4-F1, F3-F2)
2003 2004 2005 2006 2007 2008 2009 2010 2011 20120
50
100
150
200
250
f(x) = 2.76363636363636 x + 192.2R² = 0.357935180021069
f(x) = 1.89090909090909 x + 99R² = 0.147905043211902
length of growing seasonLinear (length of growing season)duration of max greennessLinear (duration of max greenness)
DAYS
Significant (5%)
EMS&ECAM 2013 09-13 Sept. Reading UK
17
BBCH
10 Sprouting of leaves
19 Beginning of the unfolding of leaves
60 Beginning of flowering
65 Full flowering
69 End of flowering93 Fall of leaves
97 End of falling leaves
Traditional observations of trees
EMS&ECAM 2013 09-13 Sept. Reading UK
18
Traditional observations of trees
1 21 41 61 81 101 121 141 161 181 201 221 241 261 281 301 321 3410
50
100
150
2012
locustpoplaroak
DOY
BBCH
1 21 41 61 81 101 121 141 161 181 201 221 241 261 281 301 321 3410
50
100
150
2011
locustpoplaroak
DOY
BBCH
EMS&ECAM 2013 09-13 Sept. Reading UK
19
EVI values on pixel level
1 17 33 49 65 81 97 113
129
145
161
177
193
209
225
241
257
273
289
305
321
337
353
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
2011
locustpoplaroakFL loc.FL pop.FL oak
DOY
EVI
EMS&ECAM 2013 09-13 Sept. Reading UK
20
EVI values on pixel level
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 450
0.1
0.2
0.3
0.4
0.5
0.6
2012
locustpoplaroakFL locFL popFL oak
DOY
EVI
EMS&ECAM 2013 09-13 Sept. Reading UK
21
Traditional observation of agricultural plantsBBCH wheat
BBCH maize
BBCH sunflower
00 sowing 00 sowing 00 sowing
09 Emergence 09 Emergence 09 Emergence
21Beginning of tillering 13 3 leaves
unfolded 144 leaves unfolded
31First node at least 17 7 leaves
unfolded 53 Inflorescence
55Middle of heading 59 59 End of tassel
emergence 65 Full flowering
65 Full flowering 67 Flowering declining
85 Soft dough 85 85 dough stage 69 End of flowering
92 Over-ripe 89 Fully ripe
99 Harvest99
Harvest99 99 Harvest
EMS&ECAM 2013 09-13 Sept. Reading UK
22
Traditional observation of agricultural plants
-100 -50 0 50 100 150 200 250 300 350 4000
50
100
150
2011
wheatmaizesunflower
DOY
BBCH
-100 -50 0 50 100 150 200 250 300 350 4000
50
100
150
2012
wheatmaizesunflower
DOY
BBCH
EMS&ECAM 2013 09-13 Sept. Reading UK
23
EVI values on pixel level
1 17 33 49 65 81 97 113
129
145
161
177
193
209
225
241
257
273
289
305
321
337
353
-0.1000
0.0000
0.1000
0.2000
0.3000
0.4000
0.5000
0.6000
0.7000
2011
wheatmaizesunflower
DOY
EVI
Heading or flowering
Harvest
EMS&ECAM 2013 09-13 Sept. Reading UK
24
EVI values on pixel level
1 17 33 49 65 81 97 113
129
145
161
177
193
209
225
241
257
273
289
305
321
337
353
0.0000
0.1000
0.2000
0.3000
0.4000
0.5000
0.6000
0.7000
2012
wheatmaizesunflower
DOY
EVI
Heading or floweringHarvest
EMS&ECAM 2013 09-13 Sept. Reading UK
25
Conclusions• Landscape phenology based on EVI is suitable to
evaluate the vegetation dynamic of the season.• Traditional phenology is more sophisticated but it is
difficult to integrate.• EVI values for native plants (trees) can be modeled by
logistic curves. ▫Flowering time relating to EVI is plant specific
• EVI values for agricultural plants show different shape (not logistic).▫The maximum EVI is around flowering and then the
decreasing phase has rather convex curvature.
EMS&ECAM 2013 09-13 Sept. Reading UK
26
Thank you for your attention!