Secondary forest succession quantification using LIDAR analysis in the southern Appalachians Tucker J. Souther1, Robbie G. Kreza1, Marcus C. Mentzer1, Brian D. Kloeppel1, and Ryan E. Emanuel2 1Department of Geosciences and Natural Resources, Western

1
Secondary forest succession quantification using
LIDAR analysis in the southern AppalachiansTucke
r J. Souther1, Robbie G. Kreza1, Marcus C.
Mentzer1, Brian D. Kloeppel1, and Ryan E.
Emanuel21Department of Geosciences and Natural
Resources, Western Carolina University,
Cullowhee, NC 28723, tjsouther1_at_catamount.wcu.edu
2Department of Forestry and Environmental
Resources, North Carolina State University,
Raleigh, NC 27695
Results
Introduction
Secondary forest succession is an important
ecosystem process that occurs as forested
ecosystems regenerate naturally after
disturbance. As forests undergo succession, the
mass of carbon stored within the ecosystem
changes. The primary goal of the project was to
determine if airplane-based LIDAR (Light
Detection And Ranging) data could be used to
accurately predict tree height, providing insight
into forest structure across a suite of
successional age classes. We used both indirect
and direct forest structure measurements to
determine the structure of the forest during
these successional changes. These measurements
included tree height, forest density, plant area
index, forest basal area, and species
distribution (all collected in the field). To
assess the accuracy of the LIDAR data, four
forest stands of similar species composition
(deciduous montane mixed oak-hickory forest), in
different successional stages (30, 50, 70, and
90 years since disturbance) were utilized for
our measurements. This project was a
collaborative effort between Western Carolina
University (WCU)and four other institutions
across the state of North Carolina.
rho 1.00, P 0.08
rho -1.00, P 0.08
Figure 8. LIDAR derived tree height vs. field
measured tree height in four montane oak-hickory
stands from 30 to 90 years old.
Figure 1. In an effort to provide a statewide
perspective of LIDAR accuracy, four other
institutions were involved with the project,
including UNC-Pembroke, Johnson C. Smith
University, Livingstone College, and North
Carolina State University. Counties are
highlighted in red and study sites are indicated
by green dots with Balsam Mountain Preserve near
Western Carolina University indicated with an
arrow.
Figure 6. (a) Relative importance value of tree
species in four montane oak-hickory sites from 30
to 90 years old. (b) Stand density and stand
basal area vs. stand age across the 30 to 90
year-old sites.
Objectives

• The objectives of this research were to
• Determine if the LIDAR imagery available in
  Jackson County, North Carolina could accurately
  predict tree height as measured in the field
• Determine what impact stand age and the
  concomitant change in forest structure has on our
  ability to accurately predict tree height using
  LIDAR imagery
• Measure plant area index of these stands from
  field-based measurements using the LAI 2000
  instrument
• Compare species composition using forest
  biometrics across several successional stand ages

Legend
Methods
Vegetation Height (m)
High 44.53
Low -1.39
This research was conducted at Balsam Mountain
Preserve (BMP) in Jackson County in western North
Carolina. BMP is an 1800 hectare mountain site
(4400 acres) that was originally owned by
Champion International Paper Company and was
purchased in the late 1990s for an upscale
housing development containing 354 home lots. At
that time, half of the remaining land that was
not being developed was placed into a
conservation easement forming the non-profit BMP
trust. The past land use history (see Figure 2
below) provided us with a range of stand ages in
the montane oak-hickory forest type with dominant
species including chestnut oak (Quercus montana),
northern red oak (Quercus rubra), scarlet oak
(Quercus coccinea), hickory (Carya spp.),
sourwood (Oxydendrum arboreum), red maple (Acer
rubrum), and black locust (Robinia pseudoacacia).
30 Year Plot
Figure 9. Measuring tree stand density and basal
area in the 70-year old stand at Balsam Mountain
Preserve near Western Carolina University.
50 Year Plot
Figure 7. LIDAR derived plot tree height and
stand tree height distributions in four montane
oak-hickory stands from 30 to 90 years old.
median plot and stand heights significantly
different
70 Year Plot
90 Year Plot
Conclusions
Tree Locations

1. Our age chronosequence in montane oak-hickory
   forests resulted in decreasing density and
   increasing basal area from 30 to 90 year-old
   stands and plant area index reached a maximum in
   50 year-old stands.
2. LIDAR derived tree height accurately (linear
   regression slope of 1.01 and R2 0.92) predicted
   tree height when compared to ground based laser
   hypsometer measurements.
3. For all stands, except the 30-year old, the plot
   level LIDAR derived tree height was
   representative of the entire stand (see Figure
   7).
4. Plant area index varied from 3.26 to 4.16 m2/m2
   with a peak in the 50 year-old stand. May 2011
   plant area index values were lower because
   measurements were collected before leaf expansion
   reached a maximum.

Within the Balsam Mountain Preserve study site,
four montane oak-hickory forest stands were
identified across a chronosequence representing
differing successional stages (30, 50, 70 and 90
years since disturbance). Ground-based data
collection for each plot included tree species
and diameter (measured with a DBH tape). In
addition, the three tallest trees in each plot
were identified and height was measured using a
Vertex Laser Hypsometer. Plant area index was
measured using the LAI 2000 leaf area index
meter. The remotely sensed LIDAR data used to
predict tree height were accessed through the
State of North Carolina and developed using
ArcGIS.
Figure 4. The vegetation height layer was
calculated by subtracting the difference in
height of the bare earth data layer from the
first return data layer.
Acknowledgements
We thank the National Science Foundation for
financial support (Award DEB-1110742). We thank
Balsam Mountain Preserve for permission to sample
on their property, especially Michael Skinner,
Ron Lance, and Blair Ogburn. We thank Cody
Amakali at Appalachian State University for LIDAR
data processing.
funding provided by
Figure 2. Forest cutting history at Balsam
Mountain Preserve by the former land owner,
Champion International. Years indicate the most
recent cut for each stand and stands used in this
study are highlighted in yellow.
Figure 3. Each plot in the 30, 50, 70, and 90
year-old stands was arranged similarly with
marked and repeatable locations for the
measurement of each LIDAR pixel. Forest density
and basal area (entire plot), tree height (of
three tallest plot trees), plant area index (on
each of 9 subplots), and GPS readings (on each
plot corner and plot center) were collected.
Figure 5. Plant area index versus stand age as
measured with a LAI 2000 in October 2010 and in
May 2011 in four montane oak-hickory stands from
30 to 90 years old.
Figure 10. Co-authors TJ Souther, Robbie Kreza,
and Marcus Mentzer at Balsam Mountain Preserve.