Presentation

Report
Elizabeth McGarrigle
Ph.D. Candidate
University of New Brunswick
Dr. John A. Kershaw
University of New Brunswick
1
Elizabeth McGarrigle
Ph.D. Candidate
University of New Brunswick
Dr. John A. Kershaw
University of New Brunswick
2
Sampling in Riparian Zones
 Typically higher species richness
 Slope, drainage, soils - heterogeneity
 To capture full range of species, need to capture full
range of growing conditions
 Sampling high perimeter to area ratio
 Traditional sampling schemes
 Boundary overlap/slop over
 Requires boundary overlap correction
 Alternatives?
3
What is sector sampling?
 Developed for use in small patches left after harvest
(Iles and Smith 2006)
 Eliminates bias caused from edge effect
 No correction required for boundary overlap – plot ends
at boundary
 Plot placed anywhere in patch, random azimuth
determined and predetermined angle used to project
plot to stand boundary
 Single or balanced sector
4
What is sector sampling?
 Developed for use in small patches left after harvest
(Iles and Smith 2006)
 Eliminates bias caused from edge effect
 No correction required for boundary overlap – plot ends
at boundary
 Plot placed anywhere in patch, random azimuth
determined and predetermined angle used to project
plot to stand boundary
 Single or balanced sector
5
Is Sector Sampling Applicable?
 Advantages in riparian zones
 No boundary overlap correction required
 No predetermined sampling location required
 Possible disadvantages
 Azimuth down middle of stand = High number of trees
 Is it efficient?
6
Overall Objective
 How does sector sampling compare with traditional
sampling methods in riparian zones?
 Ability to quantify:
 Density
 Basal area
 Species composition
 Options:
 Implement each method in the field
 Tree map riparian zone and simulate sampling methods
7
The Riparian Zone
 373 meter stretch of stream
 1050 trees stem mapped,
diameter measured
8
The Riparian Zone
 12 species total (balsam fir,
largetooth aspen, red maple)
 1400 trees/ha
 26 m^2/ha
 15 cm quadratic mean diameter
9
The Riparian Zone
 Lots of corners
 Some dropped tapes
 4 hornet/bee nests
10
Sampling Types
 Circular fixed area plots corrected using walkthrough




method
Angle count sampling (ACS) boundary overlap
corrected using walkthrough method
Strip plots perpendicular to stream
Sector Sampling
Riparian zone divided into 6 sections and sampling
types simulated
11
Strip Plots
 First strip randomly placed
in first 20 meters
12
Strip Plots
 2nd and 3rd strips each 20
meters downstream
 Two size strips sampled:
 2m
 4m
13
Point Locations
On Strip
 On each strip one point
location on each side of
stream
 6 points total
 Each point:
 Fixed area, ACS and
Sector plots
14
Circular Plots
 Two sizes sampled:
 3.01 meter radius
(1/350th of hectare)
 4.61 meter radius
(1/150th of hectare)
15
Sector Plots
 Two sizes sampled:
 10 ° angle
 20 ° angle
16
Centered Balanced
Sector Plots
 Two sizes sampled:
 10 ° angle/4
= 2.5 ° per sector
 20 ° angle/4
= 5 ° per sector
17
Sampling strategy
Type
Small Plot
Large Plot
Section
Sample per
Section
Strip Plot
2 meter full
width
4 meter full
width
All
3
Fixed Area
3.01 m. Radius 4.61 m radius
All
6
Angle Count
Sample (ACS)
Metric basal
area factor - 3
Metric basal
area factor – 2
All
6
Sector
10 ° angle
20 ° angle
All
6
Centered
Sector
Four 2.5°
angles
Four 5° angles
All
1 with 4
balanced
sectors
18
Looking at the Data
 Observed and predicted graphs of density and species
composition
 By section
 Overall
 Overall basal area
6
2
 Average error over 6 sections


o

p
1 i i
AverageError 
 Observed density
n
 Observed species comp
 Average error versus number of trees measured
19
Density
 Black – larger plots
 Red – smaller plots
 Generally all types
clustered around
observed
 Plot sizes not
significantly different
in majority of sections
 3/6 sections sector
plots are above all
other types
20
Density
 Black – larger plots
 Red – smaller plots
 Balanced sector
underestimates
overall density
21
Species Composition












ASH – ash
BE – beech
BF – balsam fir
GB – grey birch
LA – largetooth aspen
RM – red maple
SM – sugar maple
SP – spruce
STM – striped maple
WB – white birch
WP – white pine
YB – yellow birch
22
Species Composition by Section
23
Species Composition Overall
Small Plots
Large Plots
24
Basal area
 Black – larger plots
 Red – smaller plots
 ACS and Strip
sampling closest to
observed
 Centralized sector is
underestimating
basal area
25
Average Error of Density
 Black – larger plots
 Red – smaller plots
 Centered sector has
lower error and
number of trees
measured than
single sectors
 Strip plots have
lowest error
26
red maple
Average Error in
Species Composition
 Black – larger plots
 Red – smaller plots
largetooth aspen
balsam fir
27
Conclusions
 After 1050 trees, 4 nests(lots of stings), some
equipment fishing and some simulation…
 Variability in sector plots in density and species
composition predictions
 Centralized sector plots with balanced sectors
performing better than single sectors
 Strip plots more accurate in predicting overall density
and require fewer trees measured than other types
28
What’s next?
 Are the results consistent?
 Run simulation of sampling scheme again
 Different size sector angles
 Different configuration for the balanced sectors


3 or 4 sectors per point?
Optimal angle to use?
29
Acknowledgements
 John Kershaw – Fisherman, nest marker, dog bringer
and simulation coder
Questions?
30

similar documents