PRINCIPLES AND PRACTICAL

APPLICATIONS OF
POINT SAMPLING
in
FORESTRY
Presented by
SONNEY GEORGE
College of Forestry
Definition
Sampling that uses a critical angle as an expression
of the ratio between the dimensions of the objects
sampled and their distance from the sampling
point.

When the critical angle is exceeded, the object is
included in the sample.
Types of point sampling (Angle Count Sampling )
Horizontal point sampling
Vertical point sampling
Genesis
Bitterlich,1931
o

Diameter of a tree and its distance to a
neighboring tree defined a measurable angle
Comparison of point sampling
with Fixed Plot Method
Fixed plot method
Select sample plots. No and
size of the plots determined
according to Statistical
principles based on the
variability of the stand
Mark boundaries
Measure dimensions of the
plot
Calculate the plot area
Measure the dbh or gbh of
each tree .

Horizontal Point
Sampling
Select sample points. No of
points decided as per the
variability of the stand. No
fixed size
No fixed boundaries
No dimensions to be
measured
No calculation of area
Count the no. Of trees whose
BH cross section exceeds
the critical angle.

Comparison of point sampling with Fixed
Plot Method - Continued
Calculate basal area of each
tree using BA = t d
2
/4 or
g
2
/4t
Find the total for all the
trees. If a 0.1 ha plot is used,
there may be 250 trees at a
spacing of 2m X 2m.
Divide this total basal area
with the total stand area and
then multiply with 10,000 to
get the basal area density in
m
2
/ha.

Multiply this number with
the Basal Area Factor
associated with the fixed
angle device. This gives the
BA per ha.



Procedure
Tally Half Tally Non Tally
Principle : The historical development of the
method and a geometrical explanation is
available in Bitterlich (1984). Here the method
will be explained in terms of the probability
concept of Grosenbaugh (1952b, 1955).
Plan of a
forest area
with the
horizontal
projection
of basal
areas
Land area with tree stems cut at BH
Explanation
Select a number of points N in the plan.
Let n points falls on tree cross sections at BH.
Then n/N gives an estimate of Basal area per ha.

Analogous with selecting balls from a bag.
Let the bag contains n black balls and m white balls.
where total of m+n = N balls.
The proportion of black balls expected in a good
sample will be n/N.
Selecting points Vs balls
Select 1000 points
Let 5 points fell on areas occupied by trees
Then basal area per ha would be 5/1000 ha
(5 /1000) ha per ha = (5/1000)x 10000 m
2
per ha
= 50 m
2
/ ha = .5 %

Ratio of basal area to stand area actually observed in the field is
also about .2% to .7%
The variability is too much
Example of Saw logs
Hence the above scheme is impractical because;
We cant find a practical way of selecting points like that
The variability is too high
The plan of a forest area with a horizontal
projection of the basal areas and
imaginary circular zones around them.
Let d
i
be the
diameter of a tree.
Then diameter of
the circular Zone
around it will be
D
i
.
Where D
i
= f X d
i

and f a constant
factor.
Circular zones - relation to basal areas
Let the diameter of the individual tree be
denoted by d
i
for i = 1,2,...
Their respective basal areas will be equal to
(td
i
2
)/4
Total basal area E g
i
= E (t/4) d
i
2
for i = 1 to n
= (t/4) E d
i
2
for i = 1 to n
Diameter of the corresponding circular zones are
fd
i
for i= 1,2,3...
Total area of the circular zones is
E A
i
= E (t/4)(f d
i
)
2
for i = 1 to n
Circular zones - relation to basal areas - contd.
= E (t/4)f
2
( d
i
)
2
for i = 1 to n
= f
2
(t/4) E ( d
i
)
2
for i = 1 to n
= f
2
X Total basal area
OR Total area of zones/ f
2
= Total Basal area
Or ratio of the total area of zone to the total
basal area will be f
2
: 1
Inherent variation is very very less.
The efficiency of sampling is greatly
increased.
A few sample points are sufficient.
Estimation of zonal area
Count no. Of Zones.
Let the average number of zones counted is n.
Total area of zones will be n m
2
/m
2
of the stand.
or n ha/ha of the stand.
Hence Zonal area corresponding to a count of n
is given by.
Zonal area = n m
2
per m
2
of stand area.
= n X 10,000 m
2
per 10,000 m
2
of stand area.

Estimation of basal area from zonal area
We know Zonal area = n X 10,000 m
2
per ha of
stand area.
Basal area = n X 10,000 X (1/ f
2
) m
2
per ha
of stand area.
Since Total area of zones/ f
2
= Total Basal area
Basal area per ha = average number of zones n
counted X 10,000 X (1/ f
2
) m
2
.
[Where 10,000 X (1/ f
2
) is the Basal Area Factor
referred earlier]
Angle subtended by a tree at the boundary of its
circular zone.
Then in A OPQ
sin u/2 = d/2 = 1/f
fd/2
or u/2 = sin
-1
(1/f) = a
constant
or u = a constant
i.e.. Angle subtended by
the BH cross section of a
tree at the periphery of
its circular zone is a
constant
Q
P u O
f d / 2
d / 2
Angle subtended by the tree inside, outside and
at the periphery of its circular zone.
o is the angle
subtended by the tree
at any point
u is the angle
subtended by the tree
at the periphery of its
circular zone
o = u u = o u o
o o < u
u
o > u
Sides of the fixed angle u when the vertex is
inside, outside and on the periphery of the
circular zone
For a point inside the
periphery the two arms
of the angle passes
through the tree stem
For a point on the
periphery they just
touches the stem
For a point outside the
periphery they are wide
apart from the stem
Wedge prism - principle.
A L
C
N1 u N2

B

A
E P Q
B C R
Procedure
Tally Half Tally Non Tally
Field problems & equipments
Half tally trees
Relascopes:- a general term for Instruments
used for point sampling, incorporating a fixed
angle u.
The simplest Relascope consists of a distance
piece d and a cross piece C as shown below
d c
Calibrations
Let Basal area (in m
2
) per ha G = kz
where z = the no of tallied trees.
Then k = 10,000 X (1/f
2
)
Relationship between BAF and angle u
Earlier we have proved that
Sin u/2 = 1/f
Also BAF = 10,000 X (1/f
2
)
BAF, k = 10,000 X (sin u/2 )
2

= 10,000 X sin
2
(u/2 )
Practical Applications of ACS
Basic Considerations - Inventory layout

Usually systematic sampling on a rectangular grid.
Any sample gives information about the area
immediately surrounding the sample itself.
Results from a circular plot, for example, will be a
good predictor of values in the concentric circle
immediately outside the first, and so on.
In ACS the important results, i.e.. those relating to
the large trees, are gathered over wide areas.
In a close grid these large tree circular areas overlaps.
High accuracy by extrapolation and interpolation
from neighboring sample plots.
Considerations of Bias
Choose points on the map and transfer to the
field.
This kind of a restricted selection is not
biased.
Total sampling area is subdivided into fairly
equal parts.
(Only difference from Stratified Random
Sampling.)
In this respect systematic sampling is a
special kind of stratification.
Density of sampling points.
Dealt with in a voluminous literature (e.g.. Loestsch,
Zohrer and Haller, 1973. )
For a Homogenous stand.
(1) if A < 5 ha, n = 2(\k) A
(2) If A > 5 ha, n = [2(\k) (5
log A
)]/ log A
Where; n is the total number of sampling points.
a, the corresponding distance(in M) on a square grid.
A, the total area (in ha).
So that a =100 \(A/n)
Density of sampling points - Table(K=4).
A ( h a ) l o g A n a ( m )
1 4 5 0
2 8 5 0
3 1 2 5 0
4 1 6 5 0
5 1 8 5 3
6 1 8 5 8
7 1 8 6 2
8 1 9 6 5
9 1 9 6 8
1 0 0 0 3 1 6 7 2 4 5
2 0 0 0 2 4 6 2 8 5
Basal area factor: number of trees per
sample
There is a large literature on this
important item, as for example:
Arvanitis and ORegan (1969); Beers
and Miller (1964); Bitterlich (1973);
Fountain, Hunt and Hassler (1983);
Grosenbaugh (1955); Nyssonen and
Vuokila (1963); Pflugbeil (1964);
Seiber(1973); Spurr (1962); Stohr(1959);
wensel,Leviatan and Barber (1980);
Zohrer (1973).
Basal area factor: number of trees per
sample - continued
When the basal area factor is too small.
Number of sample trees selected will be very high.
Hidden trees may affect the estimates.
When the basal area factor is too large.
Number of trees selected will be too small.
Variability in the number of circular zones becomes high.
Because there wont be any proper overlap of these zones.
Thumb rule
A basal area factor giving 10 - 15 sample trees is considered
ideal by many authors.
Percentage sampled.
We know A
i
/A
s
= g
i
/(GA
s
) where
A
i
= Circular Zone area of tree i in ha , g
i
= basal
area of tree i in m
2
, G = k X 1 is the basal area per
ha for a single tree sampled and A
s
is the area of the
stand in hectares.
[ We know basal area in ha (g
i
) = Zonal area in ha
(Ai) x (1/f
2
) for each tree tallied per ha of land area
or g
i
=A
i
(1/f
2
) 10,000 m
2
per ha = A
i
X

k
or A
i
= g
i
/k in ha per ha of land area for each tallied
tree or A
i
= g
i
/G ]
The estimate of percentage sampled is then given
by A% = 100 g
i
/(GA
s
)
PRECISION
The following table lists some of the related
techniques and compares their accuracy
derived from a simulation study. The actual
volume measured using fixed plot technique
is 619 M
3
.
M e t h o d M e a n
v o l / h a
S D c o n f i d e n c e
l i m i t s ( 9 5 % )
B I T T E R L I C H 6 2 1 1 3 9 6 . 2
K I T A M U R A 6 2 9 1 5 4 6 . 7
U E N O 6 7 9 3 1 5 1 2 . 8
M I N O W A 8 2 5 1 4 3 3 4 7 . 8
Z O H R E R C = 1 . 2 6 2 1 1 2 9 5 . 7
Z O H R E R C = 1 . 4 6 2 4 1 3 3 5 . 9
Applications
Determination of Basal area
Determination of Volume
Estimates derived mainly from basal area
density
Using Volume/basal-area tables
Occularly classifying trees into basal
diameter groups
ACS with diameter obviation
Applications - contd.
Stem number density from diameter
measurements
Other features to be measured on
selected trees
General classifications
Dominant ,suppressed etc.
Elements for volume
DBH, ht, dia at other heights
Elements for quality
Other applications
Determination of thinning needs
Measurement of stand homogeneity
Determination of yield optimum
Crown cover density
Direct angle count
Marshalls method
Determination of dominant height
Determination of Stacked wood density
Determination of Basal area
The most important and primary
application of ACS is in the estimation
of Basal area/ha or the basal area
density. Multiplication of the total
number of tallied trees with the BAF
gives an unbiased estimation of the
Basal area/ha.
Determination of Volume
Estimates derived mainly from basal area density
An ingenious and practical proposal was made by
Ivanyuta (1962) and Strand (1964) making use of
suggestions by Nyssonen (1956) uses stand Form
Height along with Basal Area. We know V = GFH
or V/G = FH
Btterlich (1973a) - average tree height and form factor
for each tree species.
Shanmugasundaram (1989) - felling and actual
measurement of the mean basal area tree. Mean basal
area estimated using ACS, by dividing the total basal
area with the number of trees (both estimated by ACS).
Using Volume / basal area tables
Determination of Volume
Occularly Classifying Trees Into Broad
Diameter Groups
According to Dilworth and Bell (1973), ACS
requires only approximate diameters.
Ns=12732 k (1/d
1
- 1/[d
1
+s])/s (Bitterlich,
1960a)
(provided that an even distribution within the
classes can be assumed)
N
s
= average number of trees per hectare
represented by one in-tree falling in a diameter
class between d
1
and d
1
+s
k= basal-area factor in m
2
/ha
ACS WITH DIAMETER OBVIATION
Beers (1964), following Grosenbaugh (1955).
We know N = k (ft
2
/acre)/g(ft
2
)
where N is the No. Of trees represented by one in-
tree and g its basal area
If d is in inches : N = k /[(td
2
)/( 4 * 12 * 12]= k* 4 *
144/(td
2
)
If the single-tree volume is expressed in cubic feet
by the formula v = b d
2
h, where b is the regression
coefficient, then
V/acre = Nv = [k* 4 * 144 /(td
2
)] * b d
2
h = bhk * 4
* 144/t = Ebh Where E = k * 4* 144/t = 183.35 k
i.e... we have to measure only the heights of in
trees
ADDITIONAL MEASUREMENTS ON ALL
SELECTED TREES - STEM NUMBER DENSITY
FROM DIAMETER MEASUREMENTS
Grosenbaugh (1952b) constructed stand
table using ACS.
We know as per ACS Basal Area per ha.
G = kz where z = no. Of trees tallied.
Each in-tree contributes k basal area/ha
k = N
i
* g
i
or N
i
= k/g
i

z
and N = (k/g
i
)
I= 1
STEM NUMBER DENSITY FROM
DIAMETER MEASUREMENTS -contd.
Extending this to diameter classes.
Each tree counted in class s represents N
s
=
(k/g
s
) where g
s
= basal area of mid diameter
of class and k = basal area factor
If z
s
trees are counted in class s
Then N
s
= (k/g
s
)z
s
and N = E N
s
; where
N = stem number per hectare and N
s
= stem
number per hectare for diameter class s;
OTHER FEATURES TO BE
MEASURED ON SELECTED TREES
ACS gives unbiased estimate of stem
no. Per ha Ni
Hence any quantity/quality measured
on tallied trees can be expanded to per
ha estimates.
This is done by multiplying the
parameter measured, Yi, by Ni.
DETERMINATION OF STACKED-
WOOD DENSITY
Estimated by measurement of the conversion
factor.
In the diagram below the angle gauge has a
ratio of sample area to cross section of 100 : 1.
d d d d d
A o B
d
Principle
d d d d d
A o B
d
Volume of circular zone area = t AB
2

= t (5d)
2
= 25 t (d)
2
= 25 * 4 * [t (d)
2
/4]
= 100 * area of circle A i.e... for a count of n; n
* 1/100 gives the average tree cross sectional
area per unit stacked area which is the
conversion factor.
Procedure
Count cross sections exceeding the arms of the gauge n
which gives unit zonal area in unit stacked area
c s area in unit stacked area = n * (1/100) = Conversion
factor
HIRATAS VERTICAL ANGLE
COUNT - Principle
Total circular zone area per ha = z X 10,000 m
2
where z is the count of tally trees.
Total ht circular area/ha G
h
= z X 10,000 x
(ht circular area/circular zone area)

.
| h
R h
f r o m t h e f i g u r e h = R t a n |
Derivation
= z X 10,000 X (1/f
2
)
( since ratio of ht circular area to circular zone
area is constant = f
2
)
= z X 10,000 (th
2
)/(tR
2
)
= z X 10,000 (h
2
)/(R
2
)
= z X 10,000 tan
2
| -- 2
Hence mean ht circular area g
h
= G
h
/N
--- 3 Where N (= no. of trees) is
determined by horizontal point sampling.
Derivation - contd.
Now mean height circular area can be expressed in
terms of the corresponding ht. As gh = t H
H
2
- 4
Where H
H
is Hiratas height or the height
corresponding to g
h

From 3, 4 and 2 t H
H
2
= G
h
/N= z X 10,000 tan
2
|/N
or H
H
2
= z X 10,000 tan
2
|/(Nt)
or H
H
= [z X 10,000tan
2
|/(Nt)]
1/2

= 100 tan | X\ (z/Nt)
= 100 \ (z/N) because
when | = 60 34 tan (60 34) = \ t
Vertical ACS compared with Horizontal ACS
R o | R
H o r i z o n t a l V e r t i c a l
A n g l e c o u n t s a m p l i n g
Instruments Used in Point sampling
Cross staff
Blade instruments
Thumb as Relascope
Rectangular openings round a disc.
Wedge prism
Spiegel Relascope
Vertical slit angle gauge
Automatic slope adjusting wedge prism
Telerelascope
Wedge prism - principle.
A L
C
N 1 u N 2

B

A
E P Q
B C R