Contents...10~50ha 1.73 mil. ha (33%) 1~5ha 1.42 mil. ha (27%) 100haor more 0.85 mil. ha (16%) 100haor larger 3,000 owners (0.4%) 10~50ha 97,000 owners (11%) 50~100ha 0.43 mil. ha
Post on 27-May-2020
0 Views
Preview:
Transcript
Gifu Univ.
Y. Awaya
Air Photo Digital Canopy Height (Airborne LiDAR)
© Mitake town © Nakanihon Air Service
Importance of precise biomass information for forest
managemant: A case study of management planning in Gifu, Japan
1) Gifu University, Gifu, Japan, 2) Gifu Prefectural Research Institute for Forests,
3) Kamo Forest Cooperation
Yoshio AWAYA1) Kuniaki, FURUKAWA2) Tomoyuki, KAWAKATA3)
Applying resource information by RS to forest management.
1
Gifu Univ.
Y. Awaya
1. Background
Necessity of resource information
Forest resource information in Japan and its accuracy.
Small compartment size
2. Resource mapping
Forest type by optical imagery
Stem biomass by LiDAR data
3. Usefulness of resource information by RS
An application for logging planning
4. For space-borne LiDAR
Coarse footprint size
Large area mapping
5. Summary
Contents
2
Gifu Univ.
Y. Awaya
Forestry Agency 2009 Over mature artificial forest
Over mature
35%
Transition 10
years later 67%
March, 2007
Stand age of artificial forest in Japan
15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 100~
Necessity of resource information
Forest resource information:
Field measurement by either statistical or judgment sampling.
Aerial photo interpretation.
→ Accurate wall-to-wall biomass information is rare.
Problems: Over mature & Self sufficient rate of lumber: 27.8% in 2009.
Timber should be used effectively to reduce carbon emission.
→ Forestry agency requests forest entities to make logging plans by
organizing forest owners. → Need of resource information
March, 2017
5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85~ (year) 3
Gifu Univ.
Y. Awaya
Forest base map in 1997
Forest base map & Forest register
Office comp. Area spp. Type density vol. logged slope soil elevation
Sub compartment information
Sub compartment boundaries (red)
on a orthophoto. (Oct. 23, 1997)
Forest register
Examples of
forest
information
in Yonaizawa
National
Forest
4
Gifu Univ.
Y. Awaya
Volume prediction table
1. Trees are thinned regularly.
2. Hatching shows thinned timbers.
3. Seven thinning is scheduled by
age A.
4. C is stand volume at A and CB is
sub-harvest.
5. B is volume of primary trees
which are used for construction.
6. D is the total harvest which is the
sum of stand volume C and
thinned volume at 7 times.
7. Regular thinning is implicitly
designed in the volume
prediction table in the register.
Thinning was executed very
little before 2008.
→ Volume in register doesn’t
show volume in the field. 峰 一三, 1955, 収穫表に関する基礎的研究と信州地方カラマツ林収穫表の調整(収穫表調整業務研究資料 第12号).林野庁林業試験場, 201pp.
材
積
林 齢
D
A
C
B
間伐 1 2 3 4 5 6 7 8
Conceptual diagram of growth prediction
Volume
in register
Volu
me
Thinning
Stand age
5
Gifu Univ.
Y. Awaya
Accuracy of forest information Compartment boundaries are inaccurate in some places. Volume information is inaccurate.
Boundaries over an ortho-airphoto.
Boundaries over a forest type map.
Stem volume map by a register.
Stem volume map by LiDAR data.
It is very difficult to estimate annual logging works and incomes.
→ Tree species are identified and stem volume is estimated by field surveys.
750
500
250
0
(m3/ha)
Evergreen
conifer
Deciduous
broadleaf
6
Gifu Univ.
Y. Awaya
1~5ha681,000 owners
(75%)
5~10ha119,000 owners
(13%)
5~10ha
0.78 mil. ha(15%)
10~50ha1.73 mil. ha
(33%)
1~5ha
1.42 mil. ha(27%)
100ha or more0.85 mil. ha
(16%)
100ha or larger
3,000 owners(0.4%)
10~50ha97,000 owners
(11%)
50~100ha
0.43 mil. ha(8%)
50~100ha7,000 owners
(1%)
(0.91 mil. owners)
(5.21 mil. ha)
Forest owners
Forest area
Difficulty of ground resolution selection
CategoryArea
mil. haArea %
Stockmil. m3
Stock %
National forest 7.7 30.6 1151.8 23.5
Public forest 2.9 11.6 557.7 11.4Private forest 14.5 57.8 3191.0 65.1Total 25.1 100.0 4900.5 100.0
Forestry Agency 2015
Sub-compartments are very small in Japanese private forest. About 17 %
forest area is divided into blocks less than 5 ha.
There are numerous sub-
compartments less than 1 ha.
Therefore ground resolution less
than 30 m is favorable.
Forestry Agency 2015
7
Gifu Univ.
Y. Awaya
Forest resource information by remote sensing
Height measurement by laser beam
Direct measurement
Flight height is estimated by the time
lag between emission and receiving
of laser beam, then surface elevation
is estimated.
Location (X, Y, Z)
by GPS Laser
beam First return
Surface (DSM)
樹冠
高
樹冠
表面
の標
高
DSM DCHM
Last return
Ground (DTM)
DCHM = DSM - DTM
Classification by optical imagery
Indirect estimation
Spectra - color
Color composite
Classification by spectra
Cedar
Broadleaved
Cypress
Mitake, RapidEye 2013/05/03 ©JSI
Based on difference of color
by species. Seasonal change
is also important information.
Breen NIR Red
False color
Canopy
heig
ht E
levatio
n o
f
Canopy s
urfa
ce
8
Gifu Univ.
Y. Awaya
Mapping procedure – forest type & volume
0
100
200
300
400
500
0 5 10 15 20 25
y = -32.0 + 15.2x R2= 0.811
材積
(m
3/h
a)
平均樹冠高 (m)
CH vs. Volume
Deciduous
0
200
400
600
800
1000
1200
0 200 400 600 800 10001200
y = 0.941x
推定
材積
(m
3 ha-1
)
地上調査 材積 (m3 ha-1)
Validation
Evergreen
Accurate DCHM makes precise volume
estimation possible.
Volume estimation models are applied for
cedar, cypress, broadleaves and so on.
GeoEye-1 2010/03/30 ©JSI
Sat. Imagery NDVI Classification
Cedar Cypress Deciduous
30
20
10
0 (m)
DCHM
1000
800
600
400
200
0 (m3/ha)
Stem volume
Forest type map Stem volume map
Average CH (m)
Field V (m3/ha)
Pre
dic
ted
V (
m3/h
a)
Ste
m v
olu
me
(m
3/h
a)
Sat. Image
- Canopy height1m ≧ : Forest
1m < : Non-forest
Evergreen
Veg.Non-veg.
Non-for.For.
Deciduous
Cedar Cypress
- NDVI0.3≧ : Vegetation
0.3< : Non-vegetation
RuleFlow
Supervised
classify
- NDVI0.55≧ : Evergreen
0.55 < : Deciduous
9
Gifu Univ.
Y. Awaya
Forest type map (RS)
Register (Spp.)
Airphoto (© Nakanihon)
Register (Volume)
Cedar
Cypress Other
spp.
Cedar 1
Cypress 3
Red pine 6
Cedar 411 m3/ha
Cypress 224
Red pine 134
(m3/ha)
RS base information vs. forest register
Difference 1. Forest type distribution Cedar in valley Cypress on slope 2. Volume A comparison
Field ≧ RS > Register
Register is cursory. RS information is accurate.
(m3/ha)
0
1500
Volume (RS) 10
Gifu Univ.
Y. Awaya
Natural Artificial Thinning Logging road 2012 2013 2014 2015
Register & base map Information by RS
Comparison of logging plans
The present forest resource information such as forest type distribution is inaccurate
and the logging plan should be replaced using resource information by RS.
Logging plans become reliable. → Accurate estimation of income is possible.
Additional field surveys is required less than before.
Retrieval of plans is required less than before.
Forest owners and foresters can share reliable resource information. 11
Gifu Univ.
Y. Awaya
Mesh size of volume data
Mesh size 10m Mesh size 50m - MOLI footprint
Fifty meter grids cannot show volume well in small sub-compartments.
Foot print size of MOLI is an obstacle in small forest compartments. MOLI
products would be useful in forests larger than 3 to 5 ha. 12
Gifu Univ.
Y. Awaya
0 200 400 600 800 1000 or more
Stem Volume (m3 / ha)
Necessity of large area map
A few prefecture governments
produced large area forest
resource maps by LiDAR data.
Airborne LiDAR data are
important information for
volume mapping, however,
airborne LiDAR data are costly.
Space-borne LiDAR data
would be cost effective and be
useful for deriving and renewal
of resource information in large
areas.
Large area resource maps can
be used for forest planning in
city, prefecture and national
levels.
10 km
Volume map over
Kamo, Gifu, Japan
711 km2
13
Gifu Univ.
Y. Awaya
Artificial forest in the world
There are large artificial forest in the world and management planning
is an important task. MOLI products will be used in forest without
biomass information currently.
Artificial forest area in
the world (FAO 2010)
14
Gifu Univ.
Y. Awaya
Thank you very much for your attention. The aerial photos and a part of LiDAR data were supplied by the Gifu prefecture. Field surveys were supported by staff members and students of Gifu University.
1. Harvest is an urgent matter in Japan, however, resource information is
inaccurate.
2. LiDAR data provides accurate volume distribution maps.
3. A logging plan is designed efficiently using resource information by RS.
4. Producing large area biomass maps is costly by airborne LiDAR,
therefore, space-borne LiDAR would be an important tool.
5. Fine ground resolution information is necessary due to small forest size in
Japan.
6. Biomass information by space-borne LiDAR data can be used for forest
planning by organizations managing large forest.
Summary
15
top related