Lab Exercise-01

Objectives: To study horizontal field demonstration of various sampling techniques - simple, stratified, multi stage, multiphase, non-
random sampling techniques by field exercises.
Theory: Sampling is the process in which enumeration is to be done only in a representative portion of the whole.
Kinds of sampling: There are two main kinds of sampling used in forest inventories:
(A) Random sampling and (B) Non-random-sampling
(A) Random sampling: (Simple, stratified, multistage and multiphase sampling).
Random sampling is that method of sampling in which sampling units comprising a sample are selected in such a manner that all
possible units of the same size have equal chance of being chosen. In this method, sampling units are selected by using a table of
random numbers, by unbiased drawing of card from a well-shuffled pack or by drawing lots. The following variations of this sampling
method are used in forest inventories.
(i) Unrestricted or simple random sampling—The unrestricted or simple
random sampling is one in which sampling units com-posing the samples are
selected by some strictly random process from the whole population or area
without dividing it into homogenous blocks. For example, if sampling unit is
0.1 ha, and the population area is 1000 ha the population is divided into 10,000
sampling units and each of them given a number 0000 to 9999. Now for a 5%
unrestricted random sampling, 500 different four digit reference number are
selected by using a table of random numbers. In this way, every possible set of
500 units has equal chance of being chosen for the sample.
Stratified random sampling – The stratified random sampling is that method of
sampling in which the population is first divided into subpopulations of different
strata and then sampling units are selected from each of them in proportion to
their size. For example, if the forest population of 1000 ha is subdivided on the
basis of site quality into 4 strata with area 300 ha, in 1 quality, 350 ha in II
quality, 200 ha in III quality and 150 ha in IV quality and sampling units are of
0.1 ha, then, for 5% sampling, 150, 175,100 and 75 sampling units will be
selected with the help of the table of random numbers or by drawing lots from
strata of quality I,II,III and IV respectively.
Multi-stage sampling: It is a probability sampling method that can be used in
forest inventory to balance the need for representativeness with logistical
considerations. It divides a population into multiple stages or levels of sampling,
with each stage introducing a new sampling level. It can be more efficient and
cost-effective than sampling the entire population, especially when a
comprehensive list is unavailable. It can be used to estimate forest volume, and
can be based on remote sensing data. For example, a forest population of 100ha
may be divided into 200 sampling units of 0.5 ha. With 5% intensity of
sampling, 10 sampling units will be selected in the primary stage. Then each of
the selected sampling units, may further be broken up into 0.1 ha sampling units.
Thus each of them will have 5 sampling units. Now if one or more of these are
again selected at random in the second stage, it will become two-stage sampling.
The process can be repeated further, if necessary, till the desired stage.

Multiphase sampling – This method of sampling is different from the one described above. In this method some of the same sampling units
are used at the different phases of sampling to collect different information or same information by different methods. Two-phase sampling
is commonly used in forest inventories and because of the two phases, it is often referred to as double sampling. Another example of double
sampling in forest inventories is the procedure using a combination of aerial photographic interpretation and field plots. The first phase of
sampling consists of estimating the volume on aerial photographs of a large number of relatively inexpensive sampling units employing
photo-interpretation and measurement techniques. In the second phase, a subsample of these plots is selected and visited in the field for
direct determination of their volumes. This sub-sample is much smaller since field-plots are more expensive to lie than photo-plots. A
regression is then obtained between filed plot volumes and photo-plot volumes permitting a corrected volume estimate to be made for the
large, less expensive sample of the first phase.
(B) Non-Random Sampling: Non-random sampling is that method of sampling in which samples are selected according to the
subjective judgement of the observer on the basis of certain rules or guidelines indicating what sample should be chosen. As the
selection is based on judgement of the observer, it is likely to be biased. The following variations of this method are used in forest
inventories:
Selective sampling – Selective sampling consists of choosing samples according to the subjective judgement of the observer.
Selective sampling may give good approximations of population parameters if it is properly used by a person with intensive
knowledge of the population. 4 PEA sampling described by Gyde Land and La Bou (picking plots by personal prejudice and /or ease
of access) is one of the forms of selective sampling. For quick and rough estimates, this sampling is often employed.
Systematic sampling – Systematic sampling is that method of non-random sampling in which sampling units are selected according
to a predetermined pattern without recourse to random selection. Most commonly, the pattern envisages regular spacing of units. For
example, sampling units may be located 80 metres apart in rows which may be 200 metres apart. The selection of the first unit may be
either random or according to a fixed arbitrary rule and this will result in distinguishing two varieties of systematic sampling. When
the first unit is selected at random and other according to a fixed pattern, the sampling is referred to as systematic sampling with a
random start.
This method of sampling offers certain advantages as compared to random sampling and is therefore, frequently used in forest
inventories inspite of its shortcoming in estimating the variances of the result. The advantages are even greater in tropical forests
where environmental conditions hamper field work. Among these advantages is simplicity and ease in following it, ease of location of
sampling units and consequent reduction in cost in reaching it, greater certainty of objectivity in the selection of sample and more
uniform distribution of the sample.
Sequential sampling – Sequential sampling is a method of sampling whose characteristic feature is that the number of observations in
the sample is not determined in advance but sampling units are taken successively from a population. Each sample includes all the
sampling units of the former sample. The confidence interval at each stage will go on decreasing at each stage of sampling is stopped
when the desired precision is reached. This type of sampling is usually adopted to test a hypothesis. The hypothesis is tested each time
a new observation is made to determine whether the hypothesis should be accepted or rejected. When the decision is made and the
hypothesis is accepted or rejected, then the sampling procedure is complete and no further sampling unit is selected beyond the
number necessary to make a decision.
 Lab Exercise-02

Objective: To study the demarcation and alignment of plots and strips etc

Theory: Sampling is the process in which enumeration is to be done only in a representative

portion of the whole. In Sampling, the information is obtained only from a part of the population
assuming that it is the representative of the whole. A part is studied and on that basis, the
conclusion is drawn for the entire population.
Sampling unit: The population is divided into suitable units for the purpose of sampling.
Sampling unit’s are of two kinds, viz., those having a fixed area and those having only points.
When the sampling units are only points, it is referred to as point sampling.
SIZE OF FIXED-AREA SAMPLING UNITS
 When the sampling units are of fixed area, it is necessary to describe their size. Small
sampling units are more efficient than larger ones because the larger the number of
independent sampling units observed, the better the precisions.
 On the other hand, the size of the most efficient unit depends on the Variability and
density of stand. Forests of low density are generally - more heterogeneous in stocking.
Selection of small sampling units in such forests may result in a large portion of the
sampling units being empty. In that case, frequency distribution of desired parameters
from units become seriously skewed and does not permit assessment of error using
normal distribution theory. Another factor that influences the size is the relative cost.
 The larger the size for a given intensity, the fewer are- the number of sampling units and
therefore lesser time will be required in travelling to and locating the units. Though the
smaller the units require relatively lesser cost in enumeration, the overall cost is higher
because of higher cost of travelling to and fro and clearing the brush wood etc. therefore
the cost of travel and locating the plot is an important consideration.
Shape of the sampling units: The usual shapes of the sampling units in vogue in India are plots,
strips, topographical units and clusters.
Plots: The term plot is applied to sampling units of small area but of diverse shapes such a
square, rectangular, circular or polygonal. Square and rectangular plots are most commonly used
in forest surveys. They are easy to lay and can be demarcated by four corners. Majority of
sample plots in India are rectangular. Usual size of the plots is 0.1 hectare but bigger plots of 1.0
hectare are often laid in mature crops.
Circular plot is very easy to lie on level ground once its center and radius are fixed. It has the
minimum perimeter for a given area compared to any other geometric shape and thus has the
minimum number of border line trees. In several surveys circular plots of 0.05 hectare (radius
12.62 metres) and 0.1 hectare (radius 17.84 metres) have been used in India. In hilly terrain
radius is determined either by stepping or by correcting for slope. The desired radius is obtained
by multiplying the prescribed radius (12.62 m for 0.05 ha plots) by the slope factor. Slope factors
for some angles and percentage slopes are given below:

The same correction is required for the side of rectangle along slope.

Efficiency of the plot size is determined by the factor (SE) 2 X p,

Where SE is the standard error of a particular plot size and p is the cost of inventory of the plot.
The lower the value, the more efficient is the plot size.
Strips – In certain areas, instead of laying out plots, strips 20 to 40 m wide are laid across the
forest from one end to the other at a particular bearing at regular intervals and inventory of these
strips is made to serve as a sample. For this the survey party decides a base line at one end of the
forest and starts cutting a line through the forest at the fixed bearing from a fixed point on the
base line. This line is the central line of the strip on both sides of which enumeration is done till
the other end of the forest is reached. The party, then, moves to the center line of the next strip
and cutting that line, enumerates the strip back to the base line. Though the entire strip is the
sampling unit, it is often divided into 100 m lengths to serve as recording units. If the volume per
unit area has a tendency to increase in one direction due to fertility gradient or some other cause,
strips are laid parallel to this direction. The width of the strip and the distance between the two
strips determine the intensity of sampling, which can be calculated by the following formula:
Where I is the intensity of sampling,
W is the width of the strip in metres, and
D is the distance in metres between the central lines of the two adjacent strips.
For instance, if strips are laid 100 m apart and the width of the strip is 20m, then the intensity of
sampling is 20%.
Strip method of sampling suffers from certain drawbacks. While it is not possible to use strips as
sampling units in hill, it is difficult to maintain constant bearing and width of the strip even in
plains.
Topographical sampling unit is defined as a sampling unit whose
boundaries are predominantly topographical or natural features such as
nalas, streams, ridges, etc. Topographical units are mostly used as sampling
units in hill forests where other survey methods cannot be conveniently and
accurately carried out.
After the sampling intensity has been decided compact portions of the total
area to be enumerated (after deducting inaccessible or unworkable areas
which are excluded from enumerations) are divided into topographical units
of convenient size varying from 12 to 24 ha first on the map and then on the
ground. The units are then stratified into homogeneous blocks such that the
enumeration of at least two units per block gives the desired intensity of
sampling.
Cluster – The term cluster is used to define a sampling unit which is, in fact,
a group of smaller units. The cluster is the statistical unit where as the
smaller ones are only record units. Information is collected separately in
each record unit and is then merged with information from the other record
units to constitute the information related to the sampling unit (cluster). The
record units are not the statistical units. Inno case should cluster sampling
be understood as being synonymous with two stage sampling.