Updating Sappi growing stock information
at stand level using aerial LiDAR data
Precision Forestry Symposium 2017
One
Sappi
Intentional
evolution
2020
vision
André Wise
Mensurationist
Research, Planning and Nurseries
Forest measurement evolution
1.
Stand level: Sampling (3%)
a)
b)
2.
Stand level: Full census – Transition
a)
b)
3.
Contemporary technology, traditional concepts
Semi spatial
Sub stand level: Full census
a)
b)
Intentional
2
Traditional technology
Non-spatial
evolution
Contemporary technology, contemporary and traditional concepts
Fully spatial
Ground-based enumeration
Traditional forest measurement
Intentional
3
evolution
Ground-based enumeration
Stand level: Sampling (3%)
a)
b)
Intentional
4
evolution
Traditional technology
Non-spatial
Empirical growth modelling
Height/SI
40
30
20
10
0
0
5
10
15
Utilizable Tons
BA
400
50
40
DBH
300
30
200
20
100
10
0
TPH
0
1500
1000
500
0
0
Intentional
5
evolution
5
10
15
5
10
15
0
0
5
10
15
Stand level: Sampling
Intentional
6
evolution
Source: Sappi LiDAR updates 2016
Stand level: Enumeration sampling
Age of enumeration data
Intentional
7
evolution
Source: Sappi enumeration data 2017
Stand level LiDAR measurement
Intentional
8
evolution
Stand level LiDAR measurement
Stand level: Full census – Transition
a)
b)
Intentional
9
evolution
Contemporary technology, traditional concepts
Semi spatial
Stand level LiDAR measurement
Height percentile extraction
Intentional
10
evolution
Percentile selection
Example percentile fitting E.dunnii
80
25
15
20
25
15
20
Percentile value
25
25
95
25
90
20
Enum HM
Edun
evolution
20
Enum HM
Edun
20
15
Intentional
15
Enum HM
Edun
15
Percentile value
15
11
20
Percentile value
20
15
Percentile value
25
85
25
15
20
Enum HM
Edun
25
Stand level LiDAR measurement
Canopy height model
Intentional
12
evolution
Stand level LiDAR measurement
Canopy height model
Intentional
13
evolution
Stand level LiDAR measurement
Pixel percentile value
Intentional
14
evolution
Stand level LiDAR measurement
Pixel percentile value
Intentional
15
evolution
Stand level LiDAR measurement
Mean stand height
Intentional
16
evolution
LiDAR Height/SI
Update classes:
1. Young (Eucalyptus < 3, pine < 7 years)
Site Index update from updated SI map
2. Previously enumerated
Update of height from LiDAR
3. Older (Eucalyptus > 3, pine > 7 years), non-enumerated compartments
Update of Site Index from LiDAR
LiDAR mean height
(HM2)
Intentional
17
evolution
Conversion factor per
species
LiDAR dominant height
(HD2)
Site Index function
SI = f(HD2, Age2, AgeSI)
LiDAR SI
Example of SI differences before and after update
Relative change in SI. Red shows decrease, green shows increase.
Intentional
evolution
Source: Hlelo plantation. Sappi LiDAR updates 2016
Stand level SI comparison
All species, enumeration <5 years before LIDAR
Intentional
19
evolution
Source: Sappi LiDAR updates 2016
Basal area update (Previously enumerated)
Stand Enumeration data
TPH 1
Basal Area/ha 1
(BA1)
Dominant height (HD1)
Enumeration Age (Age1)
Intentional
20
evolution
Updated LiDAR Enumeration
TPH survival model
TPH2 = f(Age1,Age2,TPH1)
Basal area projection model
BA2 = f(BA1, TPH1, TPH2, HD1,
HD2, Age1 ,Age2)
TPH2
LiDAR dominant height
(HD2)
LiDAR Age (Age2)
Basal Area/ha
(BA2)
Stand level LiDAR coverage
Data measured and updated in 2016
Intentional
21
evolution
Source: Sappi LiDAR updates 2016
Sub stand level LiDAR measurement
Intentional
22
evolution
Sub stand level LiDAR measurement
Sub stand level: Full census
a)
b)
Contemporary technology, contemporary and traditional concepts
Fully spatial
Intentional
23
evolution
Source: Venus SI coverage. Sappi LiDAR data 2016.
LiDAR TPH
Intentional
24
evolution
LiDAR stem count adjustment
CHM Underestimate
Intentional
25
evolution
LiDAR stem count Adjustment
Intentional
26
evolution
Source: E. dunnii sliver plots. <1 ha. 1 to 7 yr old stands.
LiDAR DBH estimate
Output based on PSP data inputs (Not LiDAR measurements)

Fitting of beta functions per species/region based on PSP
and other enumeration data (include site variables).

DBH = f(LiDAR height, LiDAR TPH, Beta)
Intentional
27
evolution
Source: Sappi PSP data. E. dunnii.
LiDAR measurement outputs
Pixel level forest measurement
TPH
HM
800
15
150
100
50
600
50
1000
Frequency
20
150 200 250
1200
TPH
100
25
Frequency
-3354000 -3353800 -3353600 -3353400
HM
10
0
0
400
10
15
20
25
400
600
v
v
Dq
Vol
1000
1200
15
14
50
50
16
100 150 200 250
18
Frequency
20
20
100
25
22
Frequency
-3354000 -3353800 -3353600 -3353400
24
150
Vol
Dq
10
0
0
12
12
14
16
18
20
22
24
10
v
-84100
-83900
-83700
-83500
-84100
-83900
-83700
Intentional
evolution
15
20
v
-83500
LID
1097440
28
800
HM
17.08
TPH
894
Dq
16.59
Vol
9084.89
25
LiDAR measurement outputs
Pixel-level forest measurement
Intentional
29
evolution
Conclusion
Intentional
30
evolution
Plantation growing stock variables
Full coverages
Variable
LiDAR
Enumeration
Plantation level
Coverage
100% (max 2 year old data)
30% (1-8 year old data)
Compartment level
Sample % Area Data source
Sample % Area Data source
Height/SI
100%
Direct
<3%
Direct
TPH
100%
Adjusted
<3%
Direct
Diameter
100%
Derived
3%
Direct
Volume
100%
Derived
3%
Derived
Intentional
31
evolution
Acknowledgements
Dr. Horst Kassier
Dr. Riyad Ismail
Giovanni Sale
Nico Hattingh
Sappi Research Team
Sappi GIS Team
Intentional
32
evolution