Swiss Federal Research Institute WSL "Seasonal variability in spectral reflectance of grasslands along a dry-mesic gradient in Switzerland" Achilleas Psomas.

Swiss Federal Research Institute WSL

"Seasonal variability in spectral

reflectance of grasslands along a

dry-mesic gradient in Switzerland"

Achilleas Psomas1,2, Niklaus E. Zimmermann1,

Mathias Kneubühler2, Tobias Kellenberger2, Klaus Itten2

1.Swiss Federal Research Institute WSL,

2. Remote Sensing Laboratories (RSL), University of Zurich

April 29th,2005Warsaw University


Overview

Introduction Objectives Data Processing-Statistical analysis Initial Results Discussion


Introduction Dry meadows and pastures in Switzerland are

species-rich habitats resulting from a traditional agricultural land use.

40% of plant and over 50% of animal species present on dry meadows are classified as endangered

90% of dry grasslands have been transformed to other land cover types

TWW Project "Dry Grassland in Switzerland"(Trockenwiesen und –weiden,1995)

Creation of a federal inventory so ecologically valuable grasslands could be given an increased protection by law.


General ObjectiveObjectives-Field Spectrometry

Examine the potential of using the seasonal variability in spectral

reflectance for discriminating dry meadows and pastures.

Identify the best spectral wavelengths to discriminating grasslands

of different type. Which are the spectral wavelengths with statistical

significant differences?

Identify the optimal time or times during the growing season

for discriminating and classifying different types of grasslands.


Example of grasslands and pastures

Semi-dry [AEMB] Dry [MB]


Data processing-Statistical analysis


Collection-Temporal resolution Field spectroradiometer, Analytical Spectral Devices

FieldSpec Pro 4 grassland types examined along a dry-mesic gradient 12 sample fields at Aargau and Chur 12 repeats (time steps) between March-October 20.000 spectral signatures collected

Structure of dataset


Data processing

Removal of errors mentioned at the field protocol.

Identification of potentially false recordings. Changing weather-moisture conditions. Unforced errors.

Normalization of data : Continuum Removal.


Identification of potential errors


Continuum Removal I It standardizes reflectance spectra to allow comparison of

absorption features.


Continuum Removal II


Statistical Analysis I Statistical significance of spectral response was tested with the

Mann-Whitney U Test (Wilcox test) for a p<0.01 for each wavelength of each field per for recording day.

Analysis was done between individual fields and between each grassland type. (for every individual day)

Continuum removed spectra and the original recordings were tested.

Classification and Regression Tree Analysis (C&RT) on statistically significant wavelength for selection of wavelengths.

Repeated (15x) 10-fold cross validation to optimize the pruning of the tree

Feature space analysis using the Jeffries-Matusita distance.


Statistical Analysis IIIAEMBAEMB MBMB

Wavelengths350nm x 100351nm x 100

..

..2500nm x 100


..

..2500nm x 100


..

..2500nm x 120


..

..2500nm x 120

p-valuep-value

0.0020.038

..

..0.0004

0.0020.038

..

..0.0004

For every day all possible field combination are checked for statistical significance per wavelength.

E.g.: Recording day with 6 fields (AE,AEMB1,AEMB2,MB1,MB2,MB3)Possible combinations : 15Significance tests: 15 combinations x 2000 Wavelengths (variables)

Wilcox test

Wilcox test


Statistical Analysis IV


Preliminary results Details 3 Types

AE: Mesic, nutrient-rich grassland AEMB: Less Mesic, species-rich grassland MB: Semi-dry, species-rich grassland

Aarau 9 time steps

25. Mai 10. Jun 25. Jun 21. Jul28. Jul 15. Aug 23. Aug 02. Sep 18.

Sep


Significant Wavelengths I

Number of statistically significant wavelengths during season.

0

200

400

600

800

1000

1200

1400

1600

1800

2000

Recording Dates

No.

of

ba

nd

s

Continuum Removed

Original Spectra


Significant Wavelengths II

Number of Significant Wavelengths for continuum removed spectra

600

800

1000

1200

1400

1600

1800

2000

25. Mai 10. Jun 25. Jun 21. Jul 28. Jul 15. Aug 23. Aug 02. Sep 18. Sep

Recording Dates

No

. of

ba

nd

s

AEc-AEMBc

AEc-MBc

AEMBc-MBc

Number of Significant Wavelengths for original spectra

600

800

1000

1200

1400

1600

1800

2000


Recording Dates

No

. of

ba

nd

s

AE-AEMB

AE-MB

AEMB-MB

AE AEMB MB -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --

Mesic Dry


C&RT Analysis I

Classification tree:

Variables actually used in tree construction:

b658 b690 b1608 b505 b705 b551 b1441

Number of terminal nodes: 8

Misclassification error rate: 0.07732 = 45 / 582

C&RT for Original spectral recordings - 10th June

2004


C&RT Analysis II: Misclassification error rate

C&RT Misclassification error rate

0

0.05

0.1

0.15

0.2

0.25

Recording Dates

Err

or

rate

Continuum Removed

Original Spectra


C&RT Analysis III: Selected Wavelengths


Feature space distance Jeffries-Matusita Distance

JM distance:Continuum Removed Spectra

1.2

1.3

1.4

1.5

1.6

1.7

1.8

1.9

2

2.1

Recording Dates

JM D

ista

nce AE-MB

AE-AEMB

AEMB-MB

JM distance:Original Spectra

1.2

1.3

1.4

1.5

1.6

1.7

1.8

1.9

2

2.1

Recording dates

JM D

ista

nce

AE-MB

AE-AEMB

AEMB-MB

AE AEMB MB -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --

Mesic Dry


Discussion Increased spectral resolution of hyperspectral recordings provide great

opportunities for discriminating grassland types.

Recordings during the growing season give a better understanding of the spectral differences between grassland types and increase the possibilities for successful discrimination and classification.

Continuum removed spectra gave a smaller number of significant wavelengths but overall better class-separability throughout the season.

C&RT proved to be a powerful statistical approach for optimizing the selection of wavelengths that maximized the class separability .

Processing of the data, statistical analysis ,C&RT analysis and continuum removal was all done with code using the statistical package R, making it

easily reproducible and adjustable.


Thank you for your attention…Thank you for your attention…


Feature space distance

0

5

10

15

20

25

30

35

40

45

50


Bhattacharyya (BH) Distance CR_tree,1,3



0

10

20

30

40

50

60

70

80

90

Bhattacharyya (BH) Distance NO_tree,1,3Bhattacharyya (BH) Distance NO_tree,1,2Bhattacharyya (BH) Distance NO_tree,2,3

Bhattacharyya Distance


25-5-2004

AEMB2 MB2


Preliminary results Seasonal variability of significant wavelengths

600

800

1000

1200

1400

1600

1800

2000

2200

Recording Dates

Fre

qu

ency

AE-AEMBc

AE-MBc

AEMB-MBc

AE-AEMB

AE-MB

AEMB-MB


Preliminary results Number of significant wavelengths for AE-AEMB discrimination

1000

1100

1200

1300

1400

1500

1600

1700

1800

1900

2000


Recording Dates

No

. o

f b

and

s

AEc-AEMBc

AE-AEMB


Preliminary results Number of significant wavelengths for AE-MB discrimination

500

700

900

1100

1300

1500

1700

1900


Recording Dates

No

. o

f b

an

ds

AEc-MBc

AE-MB


Spectral Reflectance - I The total amount of radiation that strikes an

object is referred to as the incident radiationincident radiation = reflected radiation + absorbed radiation

+ transmitted radiation


Scaling-I

Part SensorSpatial

ResolutionSpectral

ResolutionSpatial

CoverageAltitude

A) Field Spectrometry

ASD Field Spectroradiomete

r0.5m 2150 bands 6-8 fields/day 1.5m

B) Imaging Spectrometry HyMap 5m 128 bands 12km x 4km 3km

C) Multitemporal Landsat TM

Landsat TM 30m 7 bands 180km x 180km 700km




Scaling-II


Preliminary results Number of significant wavelengths for AEMB-MB discrimination

1000

1100

1200

1300

1400

1500

1600

1700

1800

1900

2000


Recording Dates

No

. o

f b

and

s

AEMBc-MBc

AEMB-MB


Preliminary results Number of significant wavelengths for sampled fields of 25th June (CR)

1200

1250

1300

1350

1400

1450

1500

1550

1600

1650

Field Combinations

Nu

mb

er o

f W

avel

eng

ths


Additional


Additional


Continuum Removal I

Trees can be used for interactive exploration and for description and prediction of patterns and processes. Advantages of trees include:

(1) the flexibility to handle a broad range of response types, including numeric, categorical, ratings, and survival data; (2) invariance to monotonic transformations of the explanatory variables;

(3) ease and robustness of construction; (4) ease of interpretation; (5) the ability to handle missing values in both response and explanatory variables. Thus, trees complement or represent an alternative to many traditional statistical techniques, including multiple

regression, analysis of variance, logistic regression, log-linear models, linear discriminant analysis, and survival models.


Discussion-Further steps

Separability analysis: Euclidean ,Jeffries-Matusita, Bhattacharyya distance

Perform CART tree analysis using the statistically significant spectral bands.

Upscaling the results of the analysis to HyMap sensor .(5m spatial resolution,128bands spectral resolution).


General ObjectiveTo develop, apply, and test different methods based on remote sensing datasets and techniques for identification and monitoring of dry meadows and pastures in Switzerland

Main project parts: Part A: Field Spectrometry-(Plot to Field)Part B: Imaging Spectrometry-(Field to Region) Part C: Multitemporal Landsat TM approach-(Region to Landscape)


Continuum Removal I It standardizes reflectance spectra to allow comparison of

absorption features.

Spectral absorption-depth method for identifying chlorophyll, water, cellulose, lignin image spectral features

Minimization of factors like atmospheric absorption, soil exposure, other absorbers in the leaf (Kruse et al. 1985; Clark et al. 1987; Kruse et al. 1993a).

A continuum is formed by fitting straight line segments between the maxima of the spectral curve




Classification and Regression Trees (C&RT) Results presented on a tree are easily summarized and

interpreted.

Flexible in handling different response data types and a big number of explanatory variables.

Ease and robustness of construction.

Tree methods are nonparametric and nonlinear

Statistical Analysis II

Swiss Federal Research Institute WSL "Seasonal variability in spectral reflectance of grasslands along a dry-mesic gradient in Switzerland" Achilleas Psomas.

Documents

federal inventory

dry grasslands

warsaw university slide

structure of dataset

drymesic gradient

spectral signatures

different types of grasslands

valuable grasslands