Hyperspectral Remote Sensing for Forest Management: Assessing Crown Closure & Classifying Tree Species in Ayubia National Park

This research presentation explores the cutting-edge application of hyperspectral and multispectral remote sensing for sustainable forest management. Focused on Pakistan's ecologically significant Ayubia National Park, the study aims to accurately assess forest crown closure and classify tree species—a critical step towards modernizing forest inventories.

Key Highlights & Objectives:

Advanced Technology: Utilizes high-resolution EO-1 Hyperion (hyperspectral) and Quickbird (multispectral) imagery for precise analysis.

Crown Closure Assessment: Compares feature extraction methods to determine the most accurate technique for estimating this vital bio-physical parameter.

Species Classification: Aims to map the spatial distribution of coniferous and broadleaved tree species in the Western Himalayan moist temperate forest.

Methodology: Integrates rigorous field data collection (GPS, crown measurements) with sophisticated image processing and classification techniques.
Significance: Provides a efficient, cost-effective, and scalable alternative to traditional, labor-intensive forest surveys, especially in inaccessible terrain.

This work is essential for forest ecologists, remote sensing specialists, environmental scientists, conservationists, and policymakers involved in biodiversity conservation and natural resource management.
Keywords: Hyperspectral Remote Sensing, Crown Closure, Tree Species Classification, EO-1 Hyperion, Quickbird, Ayubia National Park, Forest Inventory, Image Classification, Forestry, Conservation, Biodiversity, Pakistan.
Keywords: Hyperspectral Remote Sensing, Crown Closure, Tree Species Classification, EO-1 Hyperion, Quickbird, Ayubia National Park, Forest Inventory, Image Classification, Forestry, Conservation, Biodiversity, Pakistan.

1. Forest Crown Closure Assessment
& Tree Species Classification with
Hyperspectral Imagery
Presented By
Arooj Fatima

2. PRESENTATION OUTLINE
Introduction
Literature Review
Problem Statement
Scope of Study
Objectives
Study Area
Methodology
Anticipated Outcomes
Proposed Timeline
References

3. INTRODUCTION
Crown:
The upper branchy part of a tree above the bole (stem). (BCFT, 1953).

4. INTRODUCTION
Crown Closure:
Crown closure can be defined as percentage or proportion
of ground area covered by the vertical projection of tree
crowns (BCFT, 1953).
The percentage of ground area covered by the vertically
projected tree crown areas. (Canadian Forest Service,
Forest Inventory Terminologies)

5. INTRODUCTION
Crown closure is a bio-physical parameter important for quantifying the energy
and mass exchange characteristics of terrestrial ecosystems such as
photosynthesis, respiration, transpiration and rainfall interception. (Chen &
Cihlar, 1996; Chen et al., 1999; Fassnacht et al., 2000; White et al., 1997).
Crown closure is an approximate indicator of stand density. For this reason, it is
an important variable in the estimation of stand volume from aerial photographs
and in evaluating silvicultural operations and ecological conditions. It has a
significant influence on snow pack accumulation and snow melt.
The main functions of the crown of a tree are:
To display actively photosynthesizing leaves most efficiently to radiant
energy.
To provide for the renewal of the leaves.
The size of a tree crown has a marked effect on, and is strongly correlated
with, the growth of the tree and of its various parts.
Therefore, measurement of crown parameters must obviously concern a forest
manager.

6. INTRODUCTION
Directly measuring of forest canopy crown closure is labor-
intensive and, is thus only practical on experimental plots of
limited size. Consequently, estimating crown closure over
large areas is problematic (Gobron et al., 1997).
Remote sensing techniques, particularly the use of satellite
imagery may offer a practical means to measure crown
closure at the landscape scale or even global scale
(Running et al., 1999).

7. LITERATURE REVIEW
Oladi (2001) performed a study on developing a forest growth
monitoring model using Thematic Mapper Imagery. The author
associated crown closure with pixel values in terms of their surface-
exposure to satellite sensors. As planted tree crown closure is
correlated with their height and diameter at breast height (dbh) in early
stages of plantations, it was expected that a relationship exists between
the canopy closure, height, dbh, and their associated reflectance values.
It was tested in a case study for black spruce (Picea mariana) using
Landsat Thematic Mapper.

8. LITERATURE REVIEW
Pu et al., (2004) conducted a comparison of the
performance of feature extraction methods for mapping
forest crown closure and leaf area index with EO-1 Hyperion
data. The experimental results showed that the energy
features extracted by the WT method was the most effective
for mapping forest crown closure and LAI (mapped accuracy
for CC = 84.90%, LAI = 75.39%).

9. INTRODUCTION
Spectral reflectance of plant species vary with wavelength to different
degrees.
Several authors have studied the spectral difference between plant
species as well as vegetation communities in the laboratory by visually
looking at the shape of vegetation spectra (Knapp et al., 1998; Elvidge,
1990; Vogelmann, 1993).
Researchers have studied the bio-physical and bio-chemical properties
of tree species with the help of leaf spectra (Coops et al., 2001).
Information on leaf optical properties for identifying tropical tree species
using remote sensing has been obtained (Castro-Esau et al., 2006)

10. LITERATURE REVIEW
Groesz and Kastdalen (2007) used hyperspectral imageries and
multispectral imageries for the classification of forest and vegetation in
Stor-Elvdal. The study was carried out to assess and map moose winter
food resources.
Hamada et al., (2007) detected Tamarix species in riparian habitats of
Southern California using high spatial resolution hyperspectral imagery.
The mediterranean invasive specie was studied with the imagery for
weed management.

11. LITERATURE REVIEW
Apan et al., (2004) conducted a study on spectral discrimination and
classification of sugarcane varieties using EO-1 Hyperion hyperspectral
imagery. By using an atmospherically corrected EO-1 Hyperion image
acquired over Mackay, Queensland, Australia, the author analyzed the
apparent reflectance signatures from sample areas of sugarcane
varieties. Discriminant analysis was used to explore spectral separability
and to determine optimum bands and indices.

12. Thus, it can be declared that discrimination of
forest tree species is possible by identifying
spectral differences in tree species.

13. Multispectral sensors may not be effective in distinguishing small
spectral differences of canopies due to similar spectral signatures
between the tree species.
On the other hand, hyperspectral imagery having high spectral
resolution of individual channels i.e. less than10nm is expected to give
better results (Suhaili, 2005).But, the spatial resolution of space-borne
hyperspectral imagery is low i.e. 30m for analyzing individual tree
species.
Hence, this study will employ both multispectral and hyperspectral
imageries for classification of forest tree species.

14. PROBLEM STATEMENT
Forests are a key element of environment. The conservation
and management of these forests is vital for maintaining
environmental stability and ecological biodiversity.
Forest inventories are conducted for the collection of forest
data and thus formulating effective management and
conservation strategies of forests.
These involve intensive field surveys which are quite
cumbersome, time consuming and require man power.
Some forest areas are inaccessible, thus it becomes difficult
to collect data on steep terrains and dense forests.

15. PROBLEM STATEMENT
The parameters that are collected and measured in the
forest include specie type, age, height, stem diameter,
crown diameter, crown closure, volume etc.
Remote sensing is a tool that could provide reliable, up-to-
date data of forests in less time and low cost.
This research poses to study and identify two of the
parameters, i.e. type of species and crown closure with the
help of remotely sensed images taking a step forward in
providing an alternate source of collecting forest data for
Ayubia National Park.

16. SCOPE OF THE STUDY
Species classification is an emerging field with broad
applications to tropical biologists and ecologists, including
tree demographic studies and habitat diversity assessment.
Managers require an understanding of the spatial
distribution of species composition and crown closure to
manage forest resources for particular uses such as
recreation, wildlife production, forestry and watershed
management.
This study will open up new avenues for research in the field
of forest management, forest inventory and remote sensing
in Pakistan.

17. OBJECTIVES
Estimation of forest crown closure of Ayubia National Park
with the help of feature extraction methods employing
Hyperion imagery.
Comparison of the applied methods for crown closure
assessment and deduction of the most suitable method for
Ayubia National Park.
Classification of the tree species and their spatial
distribution in Ayubia National Park using Quickbird &
Hyperion imageries, through image classification
techniques.

18. STUDY AREA
Ayubia National Park will be selected as the study area for this research
as the National Park has a diverse variety of coniferous and
broadleaved tree species in their natural environment.
Ayubia National Park represents one of the best moist temperate forests
of the western Himalayan region.
The altitude of Ayubia National Park ranges from 1600 to 3000m.(5200’
to 9900’)

19. STUDY AREA
Total Area: 33.12 km
2
Area of Interest:
Upper Left Corner:34.063
0
N, 73.38
0
E
Upper Right Corner:34.063
0
N, 73.45
0
E
Lower Left Corner:34.016
0
N, 73.38
0
E
Lower Right Corner:34.016
0
N, 73.45
0
E

Area Center: 34.040
0
N, 73.41
0
E

20. BOUNDARY MAP OF ANP

21. MAP OF STUDY AREA

22. Common Name Botanical Name Frequency
Himalayan Fir Abies pindrow 13%
Spruce Picea smithiana
Chir pine Pinus roxburghii 67%
Blue pine Pinus wallichiana
Deodar Cedrus deodara 12%
Yew/Barmi Taxus baccata
Total Conifers 92%
White Oak Quercus incana 8%
Himalayan Poplar/Pallaachh Popoulus cilliata
Willow/Bins Salix tetrasperma
Maple Acer caesium
Horse chestnut Aesculus indica
Total Broadleaved 8%
Total 100%
Source: Galliat Working Plan, Deptt. of Forests NWFP

23. MATERIALS
Field data:
The field data will be collected by visiting sites in the field, measuring several
parameters (crown closure etc.) and determining coordinates with GPS. 10
sample plots will be taken for a sampling intensity of about 30%. For this
purpose, 3-5 days visit to the study area would be required.
Secondary data:
Some existing secondary data, i.e. topographic sheets (No. 43 F/8) and forest
maps will be used as reference in classification.
Images:
The following image will be used for the study:
Hyperion hyperspectral satellite imagery
Quickbird multispectral imagery

24. COST ESTIMATE
Satellite
Image
Spatial
Resolution
Spectral
Resolution
Spectral
Range
Cost in US $ Cost in Rs.
(If 1$ = 62 Rs.)
25% Student
Discount
Hyperion 30 meter 220 bands 0.4-2.4µm 750 46500 46500
Quickbird 0.6 meter 5 bands 0.4-0.9µm 712 44144 33108
Total 90644 79608

25. FIELD MEASUREMENTS
Sample plot size will be taken around 900-1200 m
2
to ensure inclusion
of a pixel (30-m resolution) of Hyperion image.
Measurements will be taken with a measuring tape.
The length of intercepted parts vertically projected by crowns in
canopies will then be measured and summed.
Formula:
CC (%) = Sum of intercepted crown lengths / Total line length * 100
References:
Avery, T. E. and Burkhart, H. E. (2002). Forest Measurements.
Bonham, C. D. (1989). Measurements for Terrestrial Vegetation.
Pu, R. and Gong, P. (2004). Remote Sensing of Environment. 91: 212-224.

26. GENERAL METHODOLOGY

27. ANTICIPATED OUTCOMES
The anticipated outcomes will be presented as follows:
Forest crown closure of Ayubia National Park.
Deduction of most suitable method for crown closure
assessment of Ayubia National Park.
Forest tree species classification of Ayubia National Park
through image classification methods.

28. PROPOSED TIMELINE

29. REFERENCES
Apan, A., Held, A., Phinn, S. and Markley, J. (2004). Spectral discrimination and classification of
sugarcane varieties using EO-1 Hyperion hyperspectral imagery. www.gisdevelopment.net
Avery, T. E. and Burkhart, H. E. (2002). Forest Measurements. Fifth Edition. McGraw-Hill Series in
Forest Resources.
Bonham, C. D. (1989). Measurements for Terrestrial Vegetation. John Wiley & Sons.
Bork, E., Duffin, E. and Mikati, E. (1995). Spatial analysis of snow cover and topography for predicting
vegetation types in Utah’s Northern Bear Mountain Range. Spring Project.
Buddenbaum, H., Schlerf, M. and Hill, J. (2005) Classification of coniferous tree species and age
classes using hyperspectral data and geostatistical methods. International Journal of Remote Sensing
26(24): 5453-5465.
British Commonwealth Forest Terminologies (1953), Empire Forestry Association. London.
Cudahy, T., Barry, P., Huntington, J. (2001) The mineral mapping performance of Hyperion. CSIRO
Exploration and Mining, Mineral Mapping and Technology Group. Australia.
Castro-Esau, K.L., Rivard, B., Wright, J. and Quesada, M. (2006) Variability in leaf optical properties of
Mesoamerican trees and the potential for species classification. American Journal of Botany 93(4):
517-530.
Chen, J., & Cihlar, J. (1996). Retrieving leaf area index of boreal conifer forests using Landsat TM
images. Remote Sensing of Environment, 55: 153– 162.
Chen, J. M., Leblanc, S. G., Miller, J. R., Freemantle, J., Loechel, S. E., Walthall, C. L., Innanen, K. A.,
& White, H. P. (1999). Compact airborne spectrographic imager (CASI) used for mapping biophysical
parameters of boreal forests. Journal of Geophysical Research, 104 (D22): 27945– 27958.
Chen, Z. S., Hsieh, C. F., Jiang, F. Y., Hsieh, T. H. and Sun, I. F. (1997). Relations of soil properties to
topography and vegetation in a subtropical rain forest in southern Taiwan. Kluwer Academic
Publishers. Germany. Plant Ecology. 132: 229-241.

30. REFERENCES
Coops, N.C., Smith, M.L., Martin, M.E. and Dury, S.J. (2001) Assessing the performance of Hyperion in relation to
Eucalypt biochemistry: Preliminary Project design specifications. CSIRO Forestry and Forest Products. Victoria,
Australia.
Dyer, J. M. (2007). A comparison of moisture scalars and water budget methods to assess vegetation site relationships.
Bellwether publishing Ltd. Journal of Physical Geography. 23: 245-258.
Elvidge, C. D. (1990). Visible and near-infrared reflectance characteristics of dry plant materials. International Journal of
Remote Sensing, 11(10): 1775-1795.
Fassnacht, K. S., Gower, S. T., MacKenzie, M. D., Nordheim, E. V., & Lillesand, T. M. (1997). Estimating the leaf area
index of north central Wisconsin forests using the landsat thematic mapper. Remote Sensing of Environment, 61: 229–
245.
Gong, P., Pu, R., Biging, G. S. and Larrieu, M. R. (2003). Estimation of forest Leaf Area Index using vegetation indices
derived from hyperion hyperspectral data. IEEE transactions on geoscience and remote sensing. 41: 6.
Green, A.A., Berman, M., Switzer, P. and Craig, M.D. (1988). Transformation for ordering multispectral data in terms of
image quality with implications for noise removal. IEEE Transactions on Geosciences and Remote Sensing 26(1): 65-74.
Groesz, F.J. and Kastdalen, L. (2007) Mapping trees and thickets with optical images. Testing the use of high resolution
image data for mapping moose winter food resources. Norwegian Directorate for Nature Management, Norwegian
Space Centre and Hedmark University College.
Hamada, Y., Stow, D.A., Coulter, L.L., Jafolla, J.C. and Hendricks, L.W. (2007) Detecting Tamarix species in riparian
habitats of Southern California using high spatial resolution hyperspectral imagery. Remote Sensing of Environment
109(2007): 237-248.
Hoersch, B., Braun, G. and Schmidt, U. (2002). Relation between landform and vegetation in alpine regions of Wallis,
Switzerland. A multiscale remote sensing and GIS approach. Deparment of Geography. University of Bonn. Germany.
Computers, Environment and Urban Systems. 26: 113-139.
www.elsevier.com/locate/compenvurbsys
Janssen, J.A.M. (2001) Monitoring of salt-marsh vegetation by sequential mapping. PhD thesis. University of
Amsterdam.
Johnson, F. L. (1986). Woody vegetation of Southeastern LeFlore County, Oklahoma, in relation to topography.
Oklahoma Biological Survey. University of Oklahoma. Oklahoma Academic Science. 66: 1-6.

31. REFERENCES
Tsai, F. and Chen, C. F. (2004) Detecting invasive plants using hyperspectral and high resolution
satellite images. Center for Space and Remote Sensing Research. National Central University. Taiwan.
Vapnik, V. N. (1995). The Nature of Statistical Learning Theory. Springer-Verlag, New York.
Vapnik, V.N. (1998). Statistical Learning Theory. John Wiley & Sons, New York.
Vaiphasa, C., Skidmore, A. K., Boer, W. F. and Vaiphasa, T. (2007). A hyperspectral band selector for
plant species discrimination. Science Direct. ISPRS Journal of Photogrammetry and Remote Sensing.
62: 225-235.
Vogelmann, J. E., Huang, C. and Tolk, B. (2002). Factors affecting vegetation cover mapping for
landfire. U.S. Geological Survey. EROS Data Center. Sioux Falls.
Vogelmann, J. E. and Moss, D. M. (1993). Spectral reflectance measurement in the genus Sphagnum.
Remote Sensing of Environment, 45: 273-279.
White, J. D., Running, S. W., Nemani, R., Keane, R. E., & Ryan, K. C. (1997). Measurement and
remote sensing of LAI in rocky mountain montane ecosystems. Canadian Journal of Forest Research,
27: 1714– 1727.

32. THANK YOU!