LAND USE CLASSIFICATION AND CHANGE DETECTION

OF RANGAMATI SADAR UPAZILA, CHITTAGONG BY

USING MULTI-TEMPORAL REMOTELY SENSED
IMAGERY

SUBJECT: PROJECT PAPER

COURSE NO: FOR

THIS PAPER HAS BEEN PREPARED AND SUBMITTED AS A PARTIAL FOUR YEAR
PROFESSIONL DEGREE B.Sc. (Hons.) IN FORESTRY

SUBMITTED BY
EXAMINATION ROLL NO: 16208048

REGISTRATION NO: 16208048

SESSION: 2015-2016

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

INSTITUTE OF FORESTRY AND ENVIRONMENTAL SCIENCES

UNIVERSITY OF CHITTAGONG

CHITTAGONG-4331

SEPTEMBER, 2021
Page | 1
LAND USE CLASSIFICATION AND CHANGE DETECTION
OF RANGAMATI SADAR UPAZILA, CHITTAGONG BY
USING MULTI-TEMPORAL REMOTELY SENSED
IMAGERY

SUBJECT: PROJECT PAPER

COURSE NO: FOR

THIS PAPER HAS BEEN PREPARED AND SUBMITTED AS A PARTIAL FOUR YEAR

PROFESSIONL DEGREE B.Sc. (Hons.) IN FORESTRY

Supervised by Submitted by
Dr. Mohammad Mozaffar Hossain Ainun Nahar
Professor Examination Roll No: 16208048
Institute of Forestry and Registration No: 16208048
Environmental Sciences, Session: 2015-16
University of Chittagong University of Chittagong

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
INSTITUTE OF FORESTRY AND ENVIRONMENTAL SCIENCES

UNIVERSITY OF CHITTAGONG

CHITTAGONG-4331

SEPTEMBER, 2021
Page | 2
DECLARATION
The author hereby declares that the work presented in this project paper has been composed
by me under the Institute of Forestry and Environmental Sciences, university of Chittagong
(IFESCU) and it has not been presented in any previous application for a degree. It is neither
been submitted nor accepted for any other degree or course by any University or Institute.

...............................................................

Ainun Nahar

Examination Roll No: 16208048

Registration No: 16208048

Session: 2015-2016

Institute of Forestry and Environmental Sciences,

University of Chittagong,

Bangladesh.

Page | 3
CERTIFICATION
This is to certify that Ainun Nahar, Examination Roll No: 16208048, Session: 2015-2016 has
prepared this paper entitled “Land Use Classification and Change Detection of Rangamati
Sadar Upazila, Chittagong by Using Multi-Temporal Remotely Sensed Imagery” under
my supervision. I do hereby approve the style and continents of this project paper. This
project paper has been prepared as a requirement for the partial fulfillment of the degree of
four year B.Sc. (Hons.) in Forestry from the Institute of Forestry and Environmental
Sciences, University of Chittagong, Bangladesh.

…………………………………………..

Dr. Md. Mozaffar Hossain

Professor

Institute of Forestry and Environmental Sciences

University of Chittagong

Bangladesh

Page | 4
Page | 5
DEDICATED
TO
MY BELOVED PARENTS

Page | vi
ACKNOWLEDGEMENT
All praise and gratitude goes to almighty God, he most benevolent, gracious and merciful,
who has made me able to accomplish the tedious job and write up this paper.

I express my deepest gratitude to my respective supervisor Dr. M. Mozaffar Hossain,

Professor, Institute of Forestry and Environmental Sciences, University of Chittagong, for his
indispensable and expert guidance, direct concern, constructive criticism encouragement and
inspiration throughout this project work.

I am highly grateful to Assistant Professor Tanjina Hasnat Tuli for her inspiration and
continuous guidance throughout this research work. She taught me the basic knowledge of
the research and patiently corrected my mistakes and errors during my work. She always
gave me new idea and skills. This study definitely would not have been possible unless her
cordial co-operation.

I am also grateful to my friends Mehera, Hoimanti, Raha, Setu, Afroza, Tasnim, Sabrina who
have been with me throughout my studies and have encouraged me to continue my study.

At last, I would like to express my heartiest gratitude to my beloved parents, brother and
sisters for their continuous inspiration during entire period of my study.

Author

Page | vii
TABLE OF CONTENTS

DECLARATION.......................................................................................................................................iii
CERTIFICATION......................................................................................................................................iv
TABLE OF CONTENTS............................................................................................................................vii
ABSTRACT............................................................................................................................................xii
CHAPTER ONE........................................................................................................................................1
INTRODUCTION.....................................................................................................................................1
1.1 General overview.....................................................................................................................1
1.2 Objectives................................................................................................................................2
1.3 Justification of the study..........................................................................................................2
CHAPTER TWO.......................................................................................................................................4
LITERATURE REVIEW..............................................................................................................................4
2.1 Land.............................................................................................................................................4
2.2 Land use/ Land cover...................................................................................................................4
2.3 Land use/land cover change.........................................................................................................5
2.4 Drivers of Land use/Land cover change......................................................................................5
2.5 History of land use/land cover changes in different countries......................................................6
2.6 Land use in Bangladesh...............................................................................................................7
2.7 Land use change in Bangladesh...................................................................................................7
2.8 General insights on sectorial causes of land use change..............................................................8
2.8.1 Multiple causes.....................................................................................................................8
2.8.2 Natural variability.................................................................................................................8
2.8.3 Economic and Technological Factors...................................................................................8
2.8.4 Demographic factors.............................................................................................................9
2.8.5 Cultural Factors....................................................................................................................9
2.8.6 Globalization........................................................................................................................9
2.9 Land use/land cover classification...............................................................................................9
2.10 Geographic Information System and Remote Sensing.............................................................10
2.11 Application of Remote Sensing and GIS to study the long term changes in land use and land
cover................................................................................................................................................11
2.12.1 Mapping Procedure...........................................................................................................13
2.12.1.1 Direct mapping..............................................................................................................13

Page | viii
2.12.1.2 Indirect Mapping............................................................................................................13
2.13 Mapping software....................................................................................................................14
2.13.1 ArcMap 10.3.....................................................................................................................14
2.13.2 Google Earth.....................................................................................................................15
CHAPTER THREE...................................................................................................................................16
STUDY AREA........................................................................................................................................16
3.1 Study Area.................................................................................................................................16
CHAPTER FOUR....................................................................................................................................18
MATERIALS AND METHODS.................................................................................................................18
4.1 Materials....................................................................................................................................18
4.1.1 Collection of literature and relevant information................................................................18
4.1.2 Software used.....................................................................................................................18
4.2 Method......................................................................................................................................18
4.2.1 Data Used...........................................................................................................................18
4.2.2 Shape files of the area.........................................................................................................19
4.2.3 Imagery of the area.............................................................................................................19
4.2.4 Image processing and classification....................................................................................19
4.2.5 Construction of map............................................................................................................20
4.3: Methodology of image classification........................................................................................22
4.3.1 Band Combinations using ArcGIS......................................................................................22
4.3.2 Image sub setting using ArcGIS.........................................................................................22
4.3.3 Supervised Classification....................................................................................................23
4.3.4 Final map preparation.........................................................................................................24
4.3.5 Methodology for land use and land cover change (LULC) detection..................................24
4.4 Accuracy Assessment................................................................................................................25
CHAPTER FIVE......................................................................................................................................26
RESULTS AND DISCUSSIONS................................................................................................................26
5.1 Land use category identification and change assessment using multi-temporal satellite data from
1990-2020........................................................................................................................................26
5.2 Land use status of the study area from 1990-2020.....................................................................26
5.3 Land use and land cover change pattern of Rangamati Sadar Upazila in 1990..........................27
5.4 Land use and land cover change pattern of Rangamati Sadar Upazila in 2000.........................29
5.5 Land use and land cover change pattern of Rangamati Sadar Upazila in 2010..........................31
5.6 Land use and land cover change pattern of Rangamati Sadar Upazila in 2020..........................33

Page | ix
5.7 Relative changes in land use in Rangamati Sadar Upazila.........................................................35
5.7 Overall accuracy and Kappa statistics ( K ¿) for 1990, 2000, 2010 and 2020 supervised
classification....................................................................................................................................37
CHAPTER SIX........................................................................................................................................38
CONCLUSIONS AND RECOMMENDTAIONS..........................................................................................38
6.1 Conclusions...............................................................................................................................38
6.2 Recommendations.....................................................................................................................38
6.3 Limitations of the study.............................................................................................................39
REFERENCES........................................................................................................................................40
APPENDICES.........................................................................................................................................46
APPENDIX 1: Error Matrix showing accuracy of 1990 supervised classification of land uses of
Rangamati Sadar Upazila.....................................................................................................................46
APPENDIX 2: Error Matrix showing accuracy of 2000 supervised classification of land uses of
Rangamati Sadar Upazila.....................................................................................................................46
APPENDIX 3: Error Matrix showing accuracy of 2010 supervised classification of land uses of
Rangamati Sadar Upazila.....................................................................................................................47
APPENDIX 4: Error Matrix showing accuracy of 2020 supervised classification of land uses of
Rangamati Sadar Upazila.....................................................................................................................47

Page | x
LIST OF TABLES

Table 2.1 Land use and land cover classification system for use with the remote sensor data
(Anderson, 1976)......................................................................................................................10

Table 4.1 The list of software used..........................................................................................18

Table4. 2 Detailed information about satellite image..............................................................19

Table4. 3 Strength of agreement for Kappa Statistics..............................................................25

Table 5. 1 Land use pattern in 1990.........................................................................................28

Table 5. 2 Land use pattern in 2000.........................................................................................30

Table 5. 3 Land use pattern in 2010.........................................................................................32

Table 5. 4 Land use pattern in 2020.........................................................................................34

Table 5. 5 Category wise land use distribution of Rangamati Sadar Upazila from 1990 to
2020..........................................................................................................................................35

Table 5. 6 Land use change assessment of Rangamati Sadar Upazila on time frame data
(1990-2020)..............................................................................................................................36

Table 5. 7 Overall Classification accuracy and overall kappa statistics of supervised

classification of 1990, 2000, 2010 and 2020 imageries of Rangamati Sadar Upazila.............37

Page | xi
LIST OF FIGURES

Figure 3.1 Map of the Study Area............................................................................................17

Figure 4.1 Steps for Downloading image................................................................................20

Figure 4. 2 Procedure to generate base map in ArcGIS...........................................................21

Figure 4. 3 Procedure of Satellite image classification............................................................21

Figure 4. 4 Procedure of stacking image using ArcGIS..........................................................22

Figure 4. 5 Procedure of clipping image in ArcGIS................................................................22

Figure 4. 6 Procedure of creating Area of Interest (AOI) shape file........................................23

Figure 4. 7 Procedure of supervised classification...................................................................23

Figure 4. 8 Process of final map preparation in ArcGIS..........................................................24

Figure 5. 1 Land use map of Rangamati Sadar Upazila 1990..................................................27

Figure 5. 2 Distribution of land use categories (%) of Rangamati Sadar Upazila...................28

Figure 5. 3 Land use map of Rangamati Sadar Upazila 2000.................................................29

Figure 5. 4 Distribution of land use categories (%) of Rangamati Sadar Upazila 2000..........30

Figure 5. 5 Land use map of Rangamati Sadar Upazila 2010.................................................31

Figure 5. 6 Distribution of land use categories (%) of Rangamati Sadar Upazila 2010..........32

Figure 5. 7 Land use map of Rangamati Sadar Upazila 2020..................................................33

Figure 5. 8 Distribution of land use categories (%) of Rangamati Sadar Upazila 2020..........34

Figure 5. 9 Relative changes in land use of Rangamati Sadar Upazila....................................36

Page | xii
ABSTRACT

Land is one of the most essential input resources in a densely populated country like
Bangladesh. When different factors (such as geophysical, proximity, socioeconomic, and
climatic) alter, the geographical distribution of land uses changes dramatically. The major
goal of this research is to determine land cover changes, particularly deforestation and
degradation, in the Rangamati Sadar Upazila's evergreen and semi-evergreen rainforests.
Using satellite pictures, the study assesses land use/land cover changes and degradation of
natural forest in Chittagong's Rangamati Sadar Upazila between 1990 and 2020. Landsat TM
and Landsat 8 OLI-TIRS images were analyzed using ArcGIS to determine the spatial and
temporal dynamics of land use/land cover changes. For classification, maximum likelihood
supervision was utilized. Accuracy was assessed using ArcGIS and Google Earth Pro, with
kappa statistics used to determine the percentage of errors avoided.The overall supervised
classification accuracy of the Landsat-derived land use/land cover maps ranged from 81% to
90%. The analysis revealed that substantial growth of agricultural land (23.90%), settlement
(192.1%) and degraded forest (84.10%) in Rangamat Sadar Upazila, Chittagong over the
study period (1990-2020) resulted significant decrease in the area of forest (15.67%) and
barren area (28.8%). The study quantified the patterns of land use/land cover change for the
last 30 years for Rangamati Sadar Upazila, Chittagong and found that forest area decreased
rapidly due to high rate of agricultural expansion and degrading forest through various
artificial land use techniques such as (shifting cultivation, joom cultivation and burning forest
area etc). According to this study, effective policies should be implemented as soon as
possible to protect residual forest coverage and restore it to its former state.

Page | xiii
CHAPTER ONE

INTRODUCTION
1.1 General overview
Land use and land cover change is the key driver of environmental change. Only a few
landscapes remained on the earth's surface that had not been affected or are being
transformed in some way by humans. Influences of human on the planet and landscape
development have had a significant impact on the natural environment. Land use patterns,
which are a key component of many current environmental concerns, are driven by
anthropogenic forces (Riebsame et al., 1944).

Land cover information can be reliably obtained by using remote sensing technology and
geographic information system. Remotely sensed data is used to examine the spatial extent
and temporal changes of land cover. When the spectral features of vegetation or other cover
types in a given location vary over time, this is known as change detection. It happens as a
result of temporal effects, as spectral response variation involves contexts (Hoffer, 1978).
Change detection is a method of perceiving and tracking the changes in an object or
phenomenon over time (Singh, 1989).

According to Turner and Meyer (1994), “Land use changes are likely to cause some land
cover change but land cover may change even if the land use remains unchanged”. Economic
progress, population increase, technological advancements, and environmental changes all
influence land use. The rates of land use shift frequently in lockstep with population
expansion, but they normally decline locally as acceleration of economic development
(Houghton, 1994). Land use change, particularly in the previous few centuries and especially
in the last few decades, has contributed to global climate change via increasing GHG
emissions. Land use activities are estimated to account for 20-75 percent of all relevant GHG
emissions in the atmosphere (Penner, 1994).

Since land use and land cover change is a dynamic and ongoing process, substantial research
on LULCC patterns, as well as their social and environmental ramifications at various
geographical and temporal scales, is necessary (lopez et al., 2016). LULC has emerged as a

Page | 1
key study subject in recent years, as it has the potential to produce large-scale environmental
changes. Information about LULC can be quite useful in natural resource management
(Kantakumar and Neelamsetti, 2015).

Bangladesh is a small South Asian country with 10.96% forest coverage in 2016, with almost
evergreen and semi-evergreen tropical rain forests mainly elongated over Chittagong, Cox's
Bazar, Chittagong hill tracts, and Sylhet, and accounting for 4.54% of the total country's
landmass, 44 percent of which is national forest land (Mamnun and Hossen, 2020). Primary
forests of Bangladesh have been destroyed due to rapid population expansion, poverty, and a
significant shortage of fuelwood, feed, and lumber (Islam et al., 2001).

1.2 Objectives
The general objectives of the study was using remote sensing and geographic information
systems to analyze the transformation of forest land into degraded forest, agricultural land, or
other types of land use in the study area from 1990 to 2020. These precise objectives are
listed below in order to achieve the stated goal:

1. To classify land cover of the study area in different period using supervised
classification.
2. To analyze periodical changes of land cover of the study area.
3. To prepare land use maps of the study area.

1.3 Justification of the study

In the Chittagong Hill Tracts of Bangladesh, tropical wet evergreen/ semi-evergreen and
deciduous forests predominate; these forests, which cover 43 % forest land, have been
severely degraded and destroyed as a result of shifting cultivation, excessive wood extraction,
encroachment, settlement, urbanization, and poor management practices (Hossain et al.,
2008; Ahammad and Stacey, 2016).

Using Landsat time-series imagery from 1990 to 2020, this research attempts to examine,
analyze, and depict the characteristics of LULCC in Rangamati Sadar Upazila, Rangamati,
which is currently suffering major environmental issues due to the rapid loss of its natural
resources, posing a threat to existing forest cover (Nath et al., 2014). During the nineteenth
century, the southeastern tropical hill forest lost 50% of its natural coverage due to shifting
cultivation, incendiary fires, population pressure from both indigenous and low-land settlers,
the encroachment of reserve forests, accelerating deforestation, soil erosion, agroforestry

Page | 2
practices, horticultural and commercially teak plantation destruction, and the encroachment
of reserve forests (Rasul et al., 2004; Ahammad and Stacey, 2016). The almost damaged
tropical forest region of Rangamati Sadar Upazila was chosen as the study area, and utilizing
multi-temporal remote sensing and GIS techniques, it was attempted to determine its LULCC
image from 1990 to 2000, in order to limit degradation and protect forest cover.

Page | 3
CHAPTER TWO
LITERATURE REVIEW

2.1 Land
Land is considered as the solid surface of earth that is not permanently submerged in water.
The concept 'land use' refers to the human activity or economic purpose that a piece of land
serves. The term "land" is defined by the United Nations as a delineable portion of the earth's
terrestrial surface that includes all aspects of the biosphere immediately above or below this
surface, such as near-surface climate, soil and terrain formations, surface hydrology
(including shallow lakes, rivers, marshes, and swamps), near-surface sedimentary layers,
associated groundwater reserve, the plant and animal populations, the human settlement
patterns and physical results of past and present human activities (terracing, water storage or
drainage structures, roads, buildings etc) (FAO, 1995).

2.2 Land use/ Land cover

Land use/ land cover change is a hybrid category. Land use denotes the human employment
in the land and it elaborately studied by the social scientists. Land cover denotes the physical
and biotic character of the land surface and is studied largely by natural scientists.
Connecting this two are proximate sources of change. Human activities directly alter the
physical environment. Proximate source changes the land cover with further environmental
consequences that ultimately feed back to affect land use (Meyer and Turner, 1992).

Land cover refers to the various types of earth’s surface, including the distribution of
vegetation, water, soil, and other physical features of the land. While land use refers to the
way in which land has been used by humans and their habitat, usually with an emphasis on
the functional role of land for economic activities (Liping et al., 2018).

About 90% of the land surface of the earth is covered by different types of vegetation. To
determine the types of vegetation light, temperature, energy, available water, natural or
induced soil fertility are act as the realizing factors. It’s needed to realize the present and
future minimum requirements of land quantity and quality for the increasing number of
people and for sustainable life. Such observation and research are required for global,
regional and national levels (Turner et al., 1994).

Page | 4
2.3 Land use/land cover change
Land use patterns were influenced by a range of societal factors, resulting in changes in land
cover that influenced biodiversity, water and radiation budgets, trace gas emissions, and other
processes that, in turn, influenced global climate and biosphere. Land usage and land cover
changes are currently major components of the wider challenge of global environmental
change (Riebsame et al., 1994).

Modification and conversion are two types of changes in land cover caused by land use. A
change in condition within a cover type is referred to as modification (such as, unmanaged
forest to a forest managed by selective cutting). Conversion is the process of switching from
one type of cover to another (e.g., deforestation to create cropland or grassland). However
(Riebsame et al., 1994) prudently portrayed examples of conversion in their paper.

2.4 Drivers of Land use/Land cover change

In recent years, land use and land cover changes (LUCC) have emerged as a key topic in
study on global environmental change (Fan et al., 2017).

Initial efforts aimed at modeling land-use change have focused primarily on biophysical
attributes (e.g. altitude, slope or soil type), given the good availability of such data.
Incorporation of data on a wide range of socio-economic drivers of change is however
required (Turner et al., 1995).

There are two biophysical and eight socio-economic drivers out of the ten most significant
drivers which act to convert agricultural land to waterbodies. From these climate dynamics
and extreme events and changes in urban and rural settlement area were driving the changes
from forest to shrub land in the southeast region of Bangladesh. In the southwest region, the
conversion from agricultural land to standing waterbodies was mainly driven by urban
household, distance to cities, precipitation dynamics, population growth and major roads (Xu
et al., 2020).

There are few studies of land-use/land-cover change which provides integrated assessment of
the driving forces and consequences of that change, particularly in Africa. Reid et al., (2000)
conducted research about these land use changes and their objectives were to determine how
driving forces at different scales change over time, how these forces affect the dynamics and
patterns of land use/land cover, and how land-use/land-cover change affects ecological
properties at the landscape scale.

Page | 5
In Modhupur Sal Forest, Bangladesh a study was conducted to assess the land use/cover
changes from 1972 to 2015. The study presented land use/land cover of the area which
enumerated that the main driver of natural forest loss was human activities like settlement,
commercial plantation, agriculture and agroforestry practices (Abdullah et al., 2015).

2.5 History of land use/land cover changes in different countries

In last 300 years the human sphere of influence on world’s natural environment has
increased at an alarming rate. The deforestation of the orient region (currently Lebanon) is a
good example on how human activities of different kinds have affected the land cover in this
area. Along with European settlement, emphasized land use activities rapidly spread through
the Americas. During the nineteenth century and early twentieth century, nations such as the
United States, Canada, Argentina, the Former Soviet Union, and Australia developed at the
expense of their natural vegetation cover. In these countries land use activities slowed down
over the last 50 years. In the developing countries, land use activities increased mainly during
the twentieth century, and continue today. These developing countries are depleting their
natural resource base in order to expand their agricultural production. Deforestation and
agricultural expansion has been the major pattern of land cover change over the last three
centuries. However, changes in land use practices such as agricultural land management, fire
suppression, and urbanization have also been momentous drivers of global change
(Goldewijk and Ramankutty, 2009).

Comprehensive research on land use changes in Asia is available during the period 1880 –
1980 (Richards and Flint, 1994). This involves 13 countries (India, Sri Lanka, Bangladesh,
Myanmar, Thailand, Laos, Cambodia, Vietnam, Malaysia, Brunei, Singapore, Indonesia and
the Philippines). In this area as a whole, forest/woodland and wetlands declined over a
hundred years by 131 million ha (-47%). At the same time, the cultivated area increased by
106 million ha. Therefore, 81% of the lost forest and wetland vegetation seems to have been
converted into expanding agricultural land. Intensified timber extraction for domestic and
export markets and the exploitation of firewood, fodder and forest products all contributed to
deforestation in this part of the world.

Bangladesh has passed through dramatic land use and land cover changes in past years, but
quantitative analysis of LULCC drivers at the national scale does not exists so far. (Xu et al.,
2020) conducted research about the dynamic of land use change in Bangladesh. The results
suggest that major LULCC occurs in Bangladesh between agricultural land and waterbodies

Page | 6
and between forest and shrubland. Because of extraordinary anthropogenic pressure and
climate variability and change, coastal land use across the world has experienced remarkable
rapid change over the recent decades. The area under shrimp cultivation greatly increased at
the rate of 2·05% per annum in the Ganges Tidal Floodplain. The majority of the shrimp area
gained from conversion of single cropland. The area of settlement considerably increased
over time. The dynamics of land use change have been attributed to low and unstable food
production in the coastal region (Islam et al., 2014).

The land use history of Indonesia shows two different faces. The islands of Java and Bali
accommodate a huge population supported by intensively managed paddy rice cultivation,
upland field crops and countless home gardens. Not surprisingly, Before 1880 Java had been
largely deforested. In contrast, the islands of Sumatra, Kalimantan (Borneo) and Sulawesi
support relatively low populations and were scarcely exploited before 1920 (Richards and
Flint, 1994).

Richards and Flint (1994) calculated a forest loss of 40% for this region during the last
century. Important driving forces were a livestock increase of 108% (217 million head), a rise
in the cultivated area by more than 40% of 44 million ha and a staggering population increase
by well over half a billion people (210%).

2.6 Land use in Bangladesh

Bangladesh's three principal land uses are agriculture, forestry, and human settlements.
Human habitation has always been driven to the periphery and hazard-prone places as growth
of population been accelerated. Due to the rapid rise of industrialization and urbanization
over the last few decades, the demand for land for non-agricultural activities and urban use
has expanded dramatically (Ali et al., 1987).

2.7 Land use change in Bangladesh

In Bangladesh, human population demands on land resources have escalated. Islam and Weil
(2000) investigated the effects of land use change on soil quality, In Bangladesh. They
discovered that soil quality has deteriorated as a result of the destruction of natural forests
and subsequent soil farming. Soil quality deteriorated greatly (44%) under farming, although
this can be restored by replanting extremely adaptable and fast-growing Acacia (Acacia sp.)
and grass species.

Page | 7
2.8 General insights on sectorial causes of land use change
Land use and land cover change (LULC) is the most apparent indicator of human intervention
and the leading source of biodiversity loss and other forms of land degradation.

2.8.1 Multiple causes

Land use change is rarely adequately understood when numerous driving variables such as
population increase, poverty, and infrastructure are combined. Individual and social
responses, on the other hand, are influenced by changing economic situations, which are
mediated by institutional considerations. Markets and legislation, which are increasingly
influenced by global influences, generate opportunities and limits for new land uses. Extreme
biophysical occurrences might also cause further alterations. Different human-environment
situations react to and distort the influence of drivers in diverse ways, resulting in different
land-use change paths. The set of abiotic and biotic factors that determine the natural capacity
to land use change varies between localities and region (Lambin et al., 2001).

2.8.2 Natural variability

Dryland ecosystems are frequently affected by synergetic effects of large-scale climate and
anthropogenic forcing variables. When one of this changes, the ecosystem becomes more
vulnerable to changes in the others. Transition triggers are generally preceded by the
alternation of humid and dry periods. The former causes an increase in the human demand on
resources, whereas the latter causes irreversible damage. Rainfall variability, as well as
herbivore or fire-induced depletion of above-ground biomass, have acted as forcing factors in
the evolution of dryland ecosystems. High demand on land resources is amplified by highly
changeable ecosystem circumstances caused by climate fluctuations, especially in dry to sub
humid climatic conditions (Puigdefabregas, 1998).

2.8.3 Economic and Technological Factors

Many experts feel that technological improvement that increases yield is the only way to
meet the agricultural needs of a growing population and economy without intruding on
primary forests. The impact of agricultural technical development on deforestation is
determined by the types of technologies and farmers engaged, as well as the circumstances in
which the change occurs. Unless it replaces labor with capital and expels displaced labor to
the agricultural frontier, technological advancement in the intensive sector is generally
helpful for forest conservation (Angelsen and Kaimowitz, 2001).

Page | 8
2.8.4 Demographic factors
Livelihood diversification is “the process by which rural households construct an increasingly
diverse portfolio of activities and assets in order to survive and improve their standard of
living” (Ellis, 2000). Demographic changes are related with the formation of families and
elements of their life cycle, as well as the shift from high to low rates of fertility and
mortality (as suggested by demographic transition) (Trincsi and Turner, 2014). Other non-
demographic factors such as government policy, changes in purchasing patterns, economic
integration, and globalization also have a role in migration (Nath et al., 2000).

2.8.5 Cultural Factors

A variety of cultural factors can influence land use decisions. Culture and other
environmental changes are frequently related (Singh, 1989). According to Macleod and
Congalton (1998) four characteristics of change detection are essential while monitoring
natural resources.

 Observing that changes have happened

 Determining the nature of the change
 Measuring the aerial extent of the change
 Evaluating the spatial pattern of the change

2.8.6 Globalization
By reducing regional borders, diminishing national bonds, and increasing interdependence
among individuals and nations, globalization amplifies or drives our driving force. The
impact of these drivers on land use is accelerated or buffered by globalization (Barbier EB,
2000).

2.9 Land use/land cover classification

Several categorization systems were devised for use with remote sensing techniques and used
as the foundation for debate at a land use information and classification conference held in
Washington, D.C. in 1971. Land use and land cover differ based on topographical
characteristics on the Earth's surface. In order to build a category for remote sensing
techniques that would give a framework to meet the needs of the majority of users, specific
criteria for evaluation must be developed first (Anderson, 1976).

Page | 9
Table 2.1 Land use and land cover classification system for use with the remote sensor
data (Anderson, 1976)

1. Built up land 2. Agricultural land

 Urban  Cropland
 Rural  Pasture
 Industrial  Orchards, groves, nurseries and
 Transportation ornamental horticultural areas
 Mixed urban/built up  Other Agricultural land
land

3. Rangeland 4. Forest land

 Herbaceous rangeland  Deciduous forest land
 Shrub and brush  Evergreen forest land
rangeland  Mixed forest land
 Mixed rangeland
5. Water 6. Barren land
 Stream water  Dry salt flats
 Lakes  Beaches
 Reservoirs  Sandy areas other than beaches
 Bays and Estuaries  Bare exposed rock
 Transitional areas
7. Tundra 8. Perennial snow or ice
 Shrub and bush tundra  Perennial snowfields
 Herbaceous tundra  Glaciers
 Bare ground tundra
 Wet tundra
 Mixed tundra

2.10 Geographic Information System and Remote Sensing

A Geographic Information System (GIS) is a computer-aided system for collecting, storing,
analyzing, and visualizing geographic data. A number of software solutions are now available
to aid this process. They can, however, differ greatly from one another, in part due to the way
they display and operate with geographic data, but also because to the relative importance
they place on these distinct tasks (Eastman, 2001).
Page | 10
By making it simple to organize and store, access and retrieve, manipulate and synthesize,
and apply knowledge to the solving of issues, geographic information systems assist us in
managing what we know. GIS stands for Geographic Information System, which is a
computerized system capable of recording, storing, analyzing, and displaying spatially
referenced data (Longley et al., 2005).

Over the years, a lot of effort has gone into combating ambiguity in geographic information
in its various forms. Until far, the majority of research on dealing with inaccuracy, fuzziness,
error, and related difficulties has concentrated on problems with spatial data and their
immediate products, which are often representations of geographical objects or fields
(Couclelis, 2003).

Any procedure that gathers information about an object, place, or phenomenon without
coming into direct contact with it is known as remote sensing. Because of this broad
definition, the word "remote sensing" has come to be connected with the measurement of
interactions. A good example of a remote sensing device is our eyes. We can obtain
information about our surroundings by measuring the amount and quality of visible light
energy that is reflected back to us. It reflects off objects in our area of view from some
external source (such as the sun or a light bulb) (Eastman, 2001).

Remote sensing relies on the detection of spectral variations in the energy reflected or emitted
by a target feature. The identification of changes in the brightness of objects and features is at
the heart of any image acquired by remote sensing. Every sensor has a size restriction for the
smallest area that can be individually captured as an image entity (Campbell and Wynne,
2011).

2.11 Application of Remote Sensing and GIS to study the long term changes in
land use and land cover
A research lead by Dewan and Yamaguchi (2009) was done for the time frame 1975 to 2003
used by satellite image and socio-economic data in greater Dhaka, Bangladesh, to evaluate
land use/land cover changes and urban expansion. The geographical and temporal dynamics
of land use/cover changes were quantified using three Landsat images, a supervised
classification system, and a post-classification change detection technique in GIS. The land
use/cover maps provide about 85 to 90 percent accuracy. The study found that across the
study period, significant increase of built-up regions in Greater Dhaka resulted in large

Page | 11
reductions in the extent of water bodies, agricultural land, vegetation, and wetlands. The
study found that across the study period, significant increase of built-up regions in Greater
Dhaka resulted in large reductions in the extent of water bodies, agricultural land, vegetation,
and wetlands. Because Bangladesh lacks trustworthy and current data, the land use maps
created in this study will aid in the creation of sustainable urban land use planning decisions
as well as the forecasting of potential future changes in growth patterns.

Abdullah et al., (2015) conducted study in the Modhupur Sal (Shorea robusta) forest in
Bangladesh's central region, which is a vulnerable natural resource. Human activities have
had a significant impact on this forest. The current study studied and assessed the loss in
aerial expanse of natural Sal forest using Landsat satellite data from 1972 to 2015. In
comparison to 1972, the amount of original natural Sal woodland was reduced by 75, 64, and
31% in 2015, 2010, and 1989, respectively. The vast degradation of natural Sal forest was
shown in this investigation. The study effectively used Landsat Multi Spectral Scanner
(MSS), Thematic Mapper (TM), and Operational Land Imager (OLI) data to monitor forest
cover changes, which might be used for low-cost but effective forest resource monitoring in
Bangladesh and other regions of the world.

Khan et al. (2015) conducted research in Bangladesh's southwest coastal districts. In this
study, satellite photos were evaluated for 13 years (1999–2012) to detect changes in land
cover. The results of image analysis revealed a 30 percent growth in shrimp culture ponds in
the last 13 years. Agricultural land and vegetation fell by 48 and 3%, respectively, while
barren and built-up regions grew by 73 percent.

Roy and Saha (2016) conducted research in Chittagong district, a rapidly rising port city in
Bangladesh's south-eastern region. The Chittagong district's temporal land use pattern
changes were studied using Google Earth satellite pictures and ArcGIS. Google earth maps
from 2002 to 2014 were utilized to create maps for every two years in the Chittagong district
in this study. According to the findings, the urban area and barren land are rising day by day
with the passage of time from 2002 to 2014, while forestland is dwindling at an alarming rate.
The water body is also shrinking day by day, though at a slower rate.

On the Eleyele Forest Reserve in Ibadan, Oyo State, Akingbogun et al. (2012) investigated
the integration of Remote Sensing and Geographic Information System (RS/GIS) for use in
urban growth effects. Eleyele Forest Reserve was identified and classified using Landsat TM
satellite Remote Sensing data from 1972 to 2000. With the help of GIS analytical capabilities,

Page | 12
a GIS database of land use categories and their location during a 28-year period (1972-2000)
was created and evaluated. The findings revealed that population increase in the vicinity of
the forest puts a lot of strain on the forest plantation.For this reason forest reserve has
suffered seriously and if it continues the whole reserve would have been converted to a bare
ground. Remote Sensing Technology in combination with Geographic Information System
can give the reliable information on land use dynamics.

2.12 Maps and mapping for natural resource management

Maps are useful tools for not only communicating global information to the public, but also
for identifying problems and answers for policymakers. They are a great tool for
geographically communicating information and can thus be used to target policy
interventions (Maes et al., 2012). Maps can help with spatially explicit prioritization and
problem identification, particularly when it comes to synergies and trade-offs between
different ecosystem services and biodiversity (Maes et al., 2012). Mapping ecosystem
services is critical for quantifying and valuing ecosystem services and ensuring effective
ecosystem management. There is considerable evidence that the spatial mapping of
ecosystem services at regional and global scales is a fast increasing study area, as noted in
(Naidoo et al., 2008) and recent studies (Troy and Wilson, 2006).

2.12.1 Mapping Procedure

There are two methods for mapping key resources: direct and indirect mapping. Aerial
pictures and remote sensing data can be used directly to map observable and measurable
features, or databases with ground data can be used indirectly to map these.

2.12.1.1 Direct mapping

For direct mapping, we must collect or acquire satellite imageries from AOI (area of interest)
from various available data sources, and then process and analyze those data using various
tools.

2.12.1.2 Indirect Mapping

Indirect mapping is primarily based on field data that has been collected and processed. We
will need to use a GPS device and collect various field-based primary and secondary data for
these purposes. The coordinate location and tracking of numerous natural and manmade
elements are captured using GPS devices. Primary data is gathered in a variety of methods

Page | 13
through surveys. Secondary information is gathered from a variety of sources. Expert advice
is also taken into account.

2.13 Mapping software

2.13.1 ArcMap 10.3

ArcMap is the main application of ArcGIS, and it's used to do everything from basic GIS
operations to specialized, user-specific jobs. Here is a list of some of the most typical
workflows that ArcMap can handle:

 Work with maps— To explore information, browse around your map documents,
toggle layers on and off, query features to get the rich attribute data behind the map,
and visualize geographic information, open and use ArcMap documents.
 Print maps— ArcMap can handle everything from simple to complex mapping.
 Compile and edit GIS datasets— ArcMap is one of the most used tools for
automating geodatabase datasets. The new and updated features are recorded in the
layer's dataset after selecting layers in the map document to change.
 Use geoprocessing to automate work and perform analysis—Geographic
information system (GIS) is both visual and analytical. Any geoprocessing model or
script can be run in ArcMap, and the results can be viewed and worked with using
map visualization. Geoprocessing can be used for analysis as well as to automate
various monotonous operations such as map book generation, data link repair in a
collection of map documents, and GIS data processing.
 Organize and manage your geodatabases and ArcGIS documents— The catalog
pane in ArcMap allows users to arrange all of their GIS datasets and geodatabases, as
well as map documents and other ArcGIS files, geoprocessing tools, and a variety of
other GIS data sets.
 Publish map documents as map services using ArcGIS for Server— Geographic
information is published as a set of map services to bring ArcGIS content to life on
the web. For publishing a map document as a map service, ArcMap provides a simple
user interface.
 Share maps, layers, geoprocessing models, and geodatabases with other users—
ArcMap has capabilities that make packaging and sharing GIS datasets with other
users a breeze.

Page | 14
 Document your geographic information— One of the main goals of GIS
communities is to describe geographic information sets to projects in order to
facilitate data exchange and search. Organizations that employ standard-base
metadata can also use the ArcGIS metadata editor to document datasets.
 Customize the user experience— ArcMap contains customization features, such as
the ability to build software add-ins to add new functionality, to simplify and optimize
the user interface, and to automate tasks using geoprocessing (“What is ArcMap?-
Help ‫ ׀‬ArcGIS for Desktop,” n.d.).

2.13.2 Google Earth

Google Earth Engine is a cloud-based platform for planetary-scale geospatial analysis that
applies Google's massive computing power to a variety of high-impact societal issues
such as deforestation, drought, disaster, disease, food security, water management,
climate monitoring, and environmental protection. It is a one-of-a-kind integrated
platform in the field, designed to empower not only traditional remote sensing scientists,
but also a far broader audience lacking the technical capacity to use typical
supercomputers or large-scale commodity cloud computing resources. Google Earth now
has four versions: the desktop client Google Earth Pro, open-source Google Earth
Enterprise, browser-based Google Earth 9, and Google Earth VR. Earth Engine is made
up of a multi-petabyte data catalog that is ready for analysis, as well as a high-
performance, intrinsically parallel compute service. It's managed and accessed via a web-
based interactive development environment (IDE) and an Internet-accessible application
programming interface (API), which allows for rapid prototyping and viewing of
outcomes (Gorelick et al., 2017

CHAPTER THREE

STUDY AREA
Page | 15
3.1 Study Area
The study was carried out in the Rangamati Sadar Upazila situated in Rangamati district
under Chittagong division. It is a significant biodiversity hotspot with a tropical semi-
evergreen forest ecosystem that is currently being degraded due to continual deforestation
and degradation. Rangamati Sadar Upazila is located in between 22°30' and 22°49' north
latitudes and in between 92°04' and 92°22' east longitudes. The area is bounded
by Naniarchar and Langadu Upazilas on the north, Kaptai and Belaichari Upazilas on the
south, Barkal and Juraichari Upazilas on the east, Kawkhali Upazila on the west
(Banglapedia, 2012).

The actual total area of the Rangamati Sadar Upazila is 54649 ha with a population of
124,728 among which 66,212 are male and 58,516 are female and population growth rate is
2.3%. Population density of this area is relatively low. Population density of this area is 227.9
per square kilometer (BBS, 2011). Various indigenous communities such as Chakma,
Marma, Tanchangya, Tripura, Khumi, Chak, Lushei, Pankho belongs to this upazila. In1983,
Rangamati Sadar Upazila was established as an upazila (Banglapedia, 2012). It is comprised
of 6 unions, 1 municipality and 178 villages. Average literacy rate is 64.5%, male is more
literate (69.7%) than female (58.6%) (BBS, 2011).

Main water bodies of Rangamati Sadar Upazila are Karnaphuli River and Kaptai Lake which
took the one-third area of Upazila. Main sources of income are agriculture (32.5%), service
(56.4%) and industry (11.1%) (BBS, 2011).

Main crops Paddy, sugarcane, maize, pulse, cotton, tobacco, potato. Main fruits Mango,
jackfruit, litchi, banana, pineapple, papaya, lemon. The amount of pacca road is 91.25 km,
semi pacca road is 60 km, kacha road is 210 km and water way is 100 km round the year
(BBS, 2011). There is no railway line in Rangamati sadar Upazila.

Page | 16
Figure 3.1 Map of the Study Area

CHAPTER FOUR
Page | 17
MATERIALS AND METHODS
4.1 Materials

4.1.1 Collection of literature and relevant information

The information related to the project paper was collected from different books, journals,
seminars proceedings and published papers from the Institute of Forestry and Environmental
Sciences. Very recent information was collected from different websites such as Google
Scholar, Bing, Springer, Nature etc.

4.1.2 Software used

Several GIS tools namely ArcGIS, USGS Earth Explorer and Google Earth Pro were used for
processing and mapping of images.

Table 4.1 The list of software used

SN Software Purpose
1 ArcGIS version 10.3 GIS analysis and mapping
2 Google Earth Pro For the matching of digitize boundary

4.2 Method

4.2.1 Data Used

The latest high resolution satellite imagery provided by USGS (United States Geological
Survey) for Landsat 8 OLI-TIRS (Operational Land imager/Thermal Infrared Sensor), and
Landsat 5 (Thematic Mapper) for the time period of 2020,2010,2000 and 1990 were used for
visual image interpretation, land use identification and classification of land use. The spatial
resolution of all Landsat image was 30 m. While image selection, cloud and unwanted shade
free imagery were set as criteria. Imagery which has cloud could substantially reduce the
accuracy of the classification. For this reason pictures were taken in January to March.
Because in Bangladesh November to February is winter season and March is the transitional
period of winter to summer. On the other hand Monsoon period in Bangladesh is cloudy and
days in winter are mostly cloud free. Due to this reason imageries from January to March
were the main sources for this study. A total of 50907 ha of the land area estimated for the
whole Rangamati Sadar Upazila after supervised image classification by using Arc GIS 10.3.
But the actual total area of the Rangamati Sadar Upazila is 54649 ha (BBS, 2011), which is

Page | 18
very close to estimated value of this study. So this value is considered as accurate value of the
study.

4.2.2 Shape files of the area

Vector file of Bangladesh consisting Rangamati Sadar Upazila area has been collected from
DIVA GIS. It is a free platform for collecting country map vector file.

4.2.3 Imagery of the area

USGS Earth EXPLORER is a simple, intuitive web map interface for collecting Landsat
satellite imagery developed by USGS. All of the Landsat satellite images of the required
periods (1990, 2000, 2010, and 2020) were explored and downloaded from the USGS website
by opening an account.

Table4. 2 Detailed information about satellite image

Satellite Sensor Path/ Acquisition Spatial Spectral Bands (µm) Data Sources
Row Date Resoluti
on (m)s

Landsat 8 OLI- 136/44 04/02/2020 30 B2 (Blue): 0.45-0.51

TRS B3 (Green): 0.53-0.59
B4 (Red): 0.64-0.67
B5 (NIR): 0.85-0.88

U.S Geological Survey

Landsat 5 TM 136/44 16/02/2010 30 B1 (Blue):0.45-0.50
B2 (Green): 0.52-0.60
B3 (Red): 0.63-0.69
B4 (NIR): 0.77-0.90

Landsat 5 TM 136/44 28/01/2000 30 B1 (Blue):0.45-0.50

B2 (Green): 0.52-0.60
B3 (Red): 0.63-0.69
B4 (NIR): 0.77-0.90

Landsat 5 TM 136/44 16/01/1990 30 B1 (Blue):0.45-0.50

B2 (Green): 0.52-0.60
B3 (Red): 0.63-0.69
B4 (NIR): 0.77-0.90

4.2.4 Image processing and classification

Image processing and performing supervised image classification helps to extract information
from imageries. .Arc GIS 10.3 software was used for image processing. With the Arc GIS
Spatial Analyst extension, the Multivariate toolset provides tools for both supervised and
unsupervised classification.
Page | 19
C
M
A
p
O
r
f
R
o
G
e
b
l
g
n
i
c
t
a
d
Steps related to download imagery

Figure 4.1 Steps for Downloading image

4.2.5 Construction of map

Mapping is based on mainly the collected and analyzed data. ArcGIS is used for preparing
the map. A base map of the study area was prepared in ArcGIs.

Page | 20
p
T
M
w
D
8
d
B
m
I
-
b
tS
V
(
s
)
L
u
lc
C
a
e
r
A
ig
y
n
o
5
f Figure 4.2 Procedure to generate base map in ArcGIS

Figure 4.3 Procedure of Satellite image classification

4.3: Methodology of image classification

Satellite image classification consists of a set of work using ArcGIS for analyzing and
classification of image. Google earth Pro facilitates identifying object during classification of
imagery.

4.3.1 Band Combinations using ArcGIS

Different images of different bands are stacked in order to analyze the satellite image. The flowchart
shows the procedure of stacking images

Page | 21
A
W
g
w
y
u
I
S
p
m
s
d
n
)
7
(
a
b
e
t
c
i
lR
L
j
P
T
O
x
E
r
V
M
'C
k
o
d
b
T
A
y
L
r
V
M
D
p
n
m
e
t
s
a
R
c
i
l
C
k
o
'
x
g Figure 4.4 Procedure of stacking image using ArcGIS

4.3.2 Image sub setting using ArcGIS

Sub setting of image involves clipping and extracting mask layer using Arc GIS. Area of
interest is clipped by using a vector layer (shape) file.

Figure 4.5 Procedure of clipping image in ArcGIS

Figure 4. 6 Procedure of creating Area of Interest (AOI) shape file

Page | 22
G
T
v
f
R
g
u
N
F
t
m
r
p
s
h
d
n
a
e
S
I
O
A
c
i
l
C
k
'o
4.3.3 Supervised Classification
Supervised classification is the technique which is used for quantitative analysis of remote
sensing image data. As its core is concept of segmenting the spectral domain into regions that
can be associated with the ground cover classes of interest to a particular application
(Richards, 2013).

Figure 4. 7 Procedure of supervised classification

Page | 23
O
u
U
b
T
y
j
h
C
N
t
c
w
,A
G
k
M
g
L
l
o
f
r
s
d
n
i
e
v
S
p
a
m
4.3.4 Final map preparation

Figure 4. 8 Process of final map preparation in ArcGIS

4.3.5 Methodology for land use and land cover change (LULC) detection
Maximum likelihood classification (MLC) approach is being widely used for land use change
assessment (Islam et al., 2018). ML is mostly used and convenient to apply with satisfactory
accuracy. Or calculating the area of the processed image following formula used in ArcGIS

Area= ¿ ¿ ------------------- (1)

The magnitude change for each land use class was calculated by subtracting the area
coverage from the 2nd year and initial year as shown in equation (2) (Islam et al., 2018).

Magnitude = magnitude of the new year – magnitude of the previous year ------------- (2)

Percentage change for each land use type was calculated by dividing magnitude change by
the base year (the initial year) and multiplied by 100 as shown in equation (3) (Islam et al.,
2018)

Percentage Change =
magnitude of change ×100
Base year
----------------- (3)

To obtain annual rate of change for each land use type, the difference between final tear to
initial year which represents magnitude of change between corresponding year was divided
by the number of study year i.e. 1990-2000 (10 years), 2000-2010 (10 years), 2010-2020 (10
years), and 1990-2020 (30 years) respectively as appropriate using equation (4) (Islam et al.,
2018).

Page | 24
Final Year−Initial Year
Annual Rate of Change = ----------------------------------- (4)
No. of Years

4.4 Accuracy Assessment

Accuracy assessment for the supervised land use classification was done for the 1990, 2000,
2010, 2020 image by using Arc GIS and Google earth pro. From the classifier 36, 37, 35, 33
points were generated randomly for 1990, 2000, 2010 and 2020 supervised images
respectively. Each and every point consisted specific color tone and the pixel value was
recognized by the software itself when the data sets were considered as reference values. All
the randomly generated points then identified by user and assigned in different classes. This
process was done for the 4 supervised classification images. The correctly identified points
were considered as classified values. An Error Matrix and Kappa Statistics were also
generated from this reference and classified data.

In the Error Matrix the rows denotes the categories as derived from the reference values. The
diagonal of matrix shows the agreement of the form-to categories which indicate the error
(omission and commission errors) that remains between the classified and reference data
(Afify, 2011).

P 0−P
K ¿= ………………………….. (5)
e

1−Pe
r
Po = ∑ Pii ……………………………. (6)
i=1

r
Pe = ∑ (Pi1 ¿ × Pi 2) ¿ ……………………….. (7)
i=1

Here,

r = Number of rows in the error matrix

Pii = The proportion of pixels in row i and column i

Pi1 = The proportion of the marginal total row of i

Pi2 = The proportion of the marginal total column of i

Table4. 3 Strength of agreement for Kappa Statistics

Kappa Statistics Strength of Agreement
<0 Poor
0.00-0.20 Slight
0.21-0.40 Fair
0.41-0.60 Moderate
0.61-0.80 Substantial
0.81-1.00 Almost perfect/perfect

Page | 25
CHAPTER FIVE

RESULTS AND DISCUSSIONS

5.1 Land use category identification and change assessment using multi-
temporal satellite data from 1990-2020

It was expected that results of land use mapping of Rangamati Sadar Upazila would provide
information about aerial distribution of land use categories and identification and estimation
of land use changes over past 30 years. The land use map realized from different landsat
imageries which shown in figures 5.1, 5.3, 5.5 and 5.7. The aerial distribution of various land
use classes for the year 1990, 2000, 2010, 2020 and their change scenarios in between
different time frames are shown in table 5.1 and 5.2.

5.2 Land use status of the study area from 1990-2020

Land uses of the study areas were categorized into following 6 groups:
I. Forest
II. Water Body
III. Agricultural Land
IV. Settlement
V. Degraded forest
VI. Barren Area

Page | 26
5.3 Land use and land cover change pattern of Rangamati Sadar Upazila
in 1990

Figure 5. 1 Land use map of Rangamati Sadar Upazila 1990

Page | 27
Table 5. 1 Land use pattern in 1990

Class Area (Ha) Area Percentage

Forest 30638.26784 60.18
Water Body 10912.68921 21.435
Agricultural Land 5269.914206 10.351
Settlement 311.921235 0.6127
Degraded Forest 2942.479833 5.779
Barren Area 833.780259 1.637

In order to know the past land use pattern of the study area, I tried to focus on landsat
imagery for the year 1990. Different land use categories had been identified and used as past
reference for the image 1990. The identified land use pattern was verified in cross-
examination at present context, whether or not the category exists or converted into another
pattern. The land use pattern which was identified (6 categories) for the year 1990 are listed
in table 5.1 and shown in figure 5.1. From the identified land use categories, the highest
categories was forest area (30638.26 ha, 60.18%), Water body (10912.68 ha, 21.4%),
Agriculture (5269.9 ha, 10.35%), Settlement (311.92 ha, 1%), Degraded forest (2942.47 ha,
5.77%), Barren area (833.78 ha, 1.63%).Figure 5.1 and 5.2 shows the land use pattern of the
study area in 1990. In these figures, green color represents Forest area which was more
prominent in Rangamati Sadar Upazila, red color represents settlement which was observed
in middle of the Rangamati Sadar Upazila, blue color represents water body and yellow color
represents agricultural land area, pink color represents degraded forest and golden color
represents barren area.

Area
2%
1% 6%
Forest
10% Water Body
Agricultural Land
Settlement
Degraded Forest
21% Barren Area
60%

Page | 28
Figure 5.2 Distribution of land use categories (%) of Rangamati Sadar Upazila

5.4 Land use and land cover change pattern of Rangamati Sadar Upazila
in 2000

Figure 5.3 Land use map of Rangamati Sadar Upazila 2000

Page | 29
Table 5.2 Land use pattern in 2000

Class Area (ha) Area percentage

Forest 31997.27942 62.854
Water Body 12377.42919 24.313
Agricultural Land 2109.697576 4.144
Settlement 395.462544 0.776
Degraded Forest 3545.801413 6.965
Barren Area 485.342246 0.953

After 1990 land use classification, land use pattern of 2000 was visually interpreted through
ArcGIS 10.3 and it was also used as the base for identification of changes in land use
categories. Based on 2000 image classification results, the highest category was forest area
(31997.27 ha, 62.85%) of total land area. Other land uses were water body (12377.42 ha,
24.31%), agricultural land (2109.69 ha, 4.14%), settlement (395.46 ha, 1%), degraded forest
(3545.80 ha, 6.96%) and barren area (485.34 ha, 1%).

Area
1%
1% 7%
4% Forest
Water Body
Agricultural Land
Settlement
24% Degraded Forest
Barren Area
63%

Figure 5.4 Distribution of land use categories (%) of Rangamati Sadar Upazila 2000

Page | 30
5.5 Land use and land cover change pattern of Rangamati Sadar Upazila
in 2010

Figure 5.5 Land use map of Rangamati Sadar Upazila 2010

Page | 31
Table 5.3 Land use pattern in 2010

Class Area (ha) Area Percentage

Forest 31166.56221 61.221
Water Body 9970.700547 19.585
Agricultural Land 3194.937653 6.275
Settlement 898.661379 1.765
Degraded Forest 3953.883094 7.766
Barren area 1723.250157 3.385

In the year 2010, area of forest (31166.56 ha, 61.22%) was dominant but the forest area
reduced than the year 2000 and settlement area (898.66 ha, 1.76%) also increased as well as
degraded forest (3953.88 ha, 7.76%) and barren area (1723.25 ha, 3.38%) also increased
compared to year 2000.

Area
3%
8%
2% Forest
6% Water Body
Agricultural Land
Settlement
Degraded Forest
20% Barren area
61%

Figure 5.6 Distribution of land use categories (%) of Rangamati Sadar Upazila 2010

Page | 32
5.6 Land use and land cover change pattern of Rangamati Sadar Upazila in
2020

Figure 5.7 Land use map of Rangamati Sadar Upazila 2020

Page | 33
Table 5.4 Land use pattern in 2020

Class Area (ha) Area Percentage

Forest 25834.81458 50.748
Water Body 11594.60761 22.7752
Agricultural Land 6529.789375 12.826
Settlement 911.221023 1.789
Degraded Forest 5443.508332 10.693
Barren Area 593.21111 1.165

Due to shifting cultivation, joom cultivation and increasing of agricultural land area the forest
area of Rangamati Sadar Upazila significantly change from the year 2010 to 2020. For this
image 6 categories were considered similar to the year 1990, 2000 and 2020 image
classification. It is noticed that the forest area (25384.8 ha, 50.74%) is still dominant in the
Rangamati Sadar Upazila. Other land uses are Waterbody (11594.6 ha, 22.77%), which
increased slightly compared to 2010, Agricultural land (6529.78 ha, 12.82%), which
increased prominently than 2010 and also degraded forest (5443.50 ha, 10.69%) area
increased comparatively. From the image analysis, it was observed that barren area (593.21
ha, 1.16%) decreased.

Area
1%
11%
2% Forest
Water Body
Agricultural Land
13% Settlement
51% Degraded Forest
Barren Area

23%

Figure 5.8 Distribution of land use categories (%) of Rangamati Sadar Upazila 2020

Page | 34
5.7 Relative changes in land use in Rangamati Sadar Upazila
Relative changes in land uses of Ranagamati Sadar Upazila were assessed based on data

presented in Table 5.5 and 5.6. Trends of land use changes from 1990-2010 showed some

negative changes; but land use changes pattern from 2010-2020 showed comparatively better

changes than 1990-2010 time period. An area of 5331.75 ha Forest decreased from 1990-

2010 and 3334.85 ha agricultural land increased due to shifting and other cultivation. But

overall land use changes from 1990-2020 showed positive changes in agricultural land,

degraded forest, waterbody and settlement area. While forest and barren area decreased

relatively (Table 5.6; Figure 5.9).

Table 5.5 Category wise land use distribution of Rangamati Sadar Upazila from 1990
to 2020

Land Use (A) in Land Use (B) in Land Use (C) in Land Use (D) in
Land Use 1990 2000 2010 2020
Category Area (ha) % of Area (ha) % of Area (ha) % of Area (ha) % of
1990 land 2000 land 2010 land 2020 land
Forest 30638.26 60.18 31997.27 62.854 31166.56 61.221 25834.81 50.748

Water Body 10912.68 21.435 12377.42 24.313 9970.7 19.585 11594.6 22.7752
Agricultura
5269.91 2109.69 3194.93 6529.78
l Land 10.351 4.144 6.275 12.826
Settlement 311.92 0.6127 395.46 0.776 898.66 1.765 911.22 1.789
Degraded
2942.47 3545.8 3953.88 5443.5
Forest 5.779 6.965 7.766 10.693
Barren
833.78 485.34 1723.25 593.21
Area 1.637 0.953 3.385 1.165
50907 100 50907 100 50907 100 50907 100
Total

Page | 35
Table 5. 6 Land use change assessment of Rangamati Sadar Upazila on time frame data
(1990-2020)

Land Use Change (B-A): Land Use Change (C-B): Land Use Change (D-C): Land Use Change (D-A):
1990-2000 2000-2010 2010-2020 1990-2020
Land Use
Category Change Annu Change Change Annua Change Annua
d Area al rate d Area Annual d Area l rate d Area l rate
1990- % of 2000- % rate of 2010- % of 1990- % of
2000 chang chang 2010 chang change 2020 chang chang 2020 chang chang
(ha) e e (ha) (ha) e (ha) (ha) e e (ha) (ha) e e (ha)
- - -
Forest +1359. +4.43 +135. -2.59 5331.7 -17.10 533.1 4803.4 -15.67 -
01 90 -830.71 -83.07 5 7 5 160.11
- +16.2
Water +1464. +13.4 +146. 2406.7 -19.44 +1623. 8 +162. +681.9 +6.24 +22.7
Body 74 47 2 -240.67 9 39 2 3
- - +51.4 +104. +23.9
Agricultur 3160.2 -59.9 316.0 +1085. 4 +108.5 +3334. 3 +333. +1259. 0 +41.9
e Land 2 2 24 2 85 485 87 9
+127. +192.
Settlemen +26.7 +8.35 2 +1.39 +1.25 1 +19.9
t +83.54 4 +503.2 +50.32 +12.56 6 +599.3 7
+11.5 +84.9
Degraded +603.3 +20.5 +60.3 +408.0 0 +1489. +37.6 +148. +2501. 9 +83.3
Forest 3 33 8 +40.80 62 962 03 6
+255. - -
Barren -41.7 +1237. 06 +123.7 1130.0 -65.5 113.0 -28.8
Area -348.44 -34.84 91 9 4 04 -240.57 -8.019

Note: (+) sign denotes increase and (-) sign denotes decrease of magnitude of change of
land use category in different time frame

4000
3000
2000
1000
0
Forest Water Body Agricultural Land Settlement Degraded Forest Barren Area
-1000
-2000
-3000
-4000
-5000
-6000
Changed Area 1990-2000 (ha) Changed Area 2000-2010 (ha)
Changed Area 2010-2020 (ha) Changed Area 1990-2020 (ha)

Figure 5.9 Relative changes in land use of Rangamati Sadar Upazila

Page | 36
5.7 Overall accuracy and Kappa statistics ( K ¿) for 1990, 2000, 2010 and
2020 supervised classification
Overall Classification Overall Kappa Statistics
Accuracy
1990 2000 2010 2020 1990 2000 2010 2020
Accuracy
Assessment
88% 89% 81% 90% 0.682 0.708 0.642 0.8820
4 5 7
Table 5.7 Overall Classification accuracy and overall kappa statistics of supervised
classification of 1990, 2000, 2010 and 2020 imageries of Rangamati Sadar Upazila
Summary of supervised classification accuracy and kappa statistics is shown in above 5.7
table and appendices 1, 2, 3 and 4.The highest accuracy was found for 2020 supervised
classification (90%) and the lowest for 2010 (81%).Kappa statistics is a measurement
mechanism between reference data and user identified classified data. Kappa value is also
used to check accuracy in classification having a kappa value (0.81-1.00) which denotes
almost perfect/perfect match between the classified and referenced data in the classification
system (Landis and Koch, 1977; J. van Vliet et al., 2011; Yang and Lo, 2002). Kappa
statistics for 2020 image classification shows a value of .8820 denoting almost perfect/perfect
match between the classified and referenced data in classification system.

Page | 37
CHAPTER SIX

CONCLUSIONS AND RECOMMENDTAIONS

6.1 Conclusions
Land cover mapping using remote sensing is a highly efficient mapping approach (Jayakumar
and Arockiasamy, 2003). The study used topographic maps and remote sensing to classify
multi-temporal land use change in the Rangamati Sadar Upazila in the Chittagong district.
The dynamics of land use change were examined using a post-classification comparison.  The
result revealed that Rangamati sadar Upazila experiencing rapid agricultural growth and
degraded forest area is also increasing day by day, resulting in the swift loss of evergreen
forest area. Overall settlement area mainly increased in the time frame of (2000-2010) and
slightly increased in 2020. On the other hand, Forest area decreased rapidly and barren area
slightly decreased due to agricultural land use expansion.

According to the land use policy 2001, permanent conversion of land use which related to
environmental degradation is strictly prohibited (MoL Governments of the people’s Republic
of Bangladesh, 2001). But this study found that, in Rangamati Sadar Upazila permanent
conversion of land use is still continuing more frequently.

 Land use classification accuracy was satisfactory. The Landsat OLI-TRS (2020) data had the
best precision, while the TM image had the lowest accuracy (2010). The land cover change in
Rangamati Sadar Upazila was quantified using GIS and remote sensing techniques in this
study.

6.2 Recommendations
According to the study, remote sensing techniques should be used to monitor and analyze a
vast area with natural resources, according to the study. This could be a better approach to
show policymakers the scenarios over time and operate as a decision support system.

Forest area decreased rapidly due to high rate of agricultural expansion and degrading forest
through various artificial land use techniques such as (shifting cultivation, joom cultivation
and burning forest area etc). Government should take proper initiative to restore degraded
lands specially degraded soil and forests in order to prevent further degradation. Conversion
of good agricultural land to non-agricultural purposes and forest to other land uses should be
strictly prohibited as per the directives of the National land Use Policy 2001.

Page | 38
6.3 Limitations of the study
 Landsat image are freely available but the problem is lower resolution (30) where
Quickbird, IKOPNIS, SPOT have high resolution which are not freely available.
 Images could not be selected in the same month for selection of cloud free images.
 Actual images of 1990, 2010 and 2020 could not be cross-matched properly due to the
lack of reference material.

Page | 39
REFERENCES
Abdullah, A. Y. M., Masrur, A., Adnan, M. S. G., Baky, M., Al, A., Hassan, Q. K., &

Dewan, A. (2019). Spatio-temporal patterns of land use/land cover change in the

heterogeneous coastal region of Bangladesh between 1990 and 2017. Remote

Sensing, 11(7), 790.

Abdullah, H. M., Mahboob, M. G., Rahman, M. M., & Ahmed, T. (2015). Monitoring natural

sal forest cover in Modhupur, Bangladesh using temporal landsat imagery during

1972–2015. Int. J. Environ, 5(1), 1-7.

Afify, H.A., 2011. Evaluation of change detection techniques for monitoring landcover

changes: A case study in new Burg El-Arab area. Alexandria Eng. Journal, 50, 187–

195.

Ahammad, R. and Stacey, N. (2016). Forest and agrarian change in the Chittagong Hill

Tracts region of Bangladesh. Center for International Forestry Research (CIFOR).

190-233.

Akingbogun, A. A., Kosoko, O. S., & Aborisade, D. K. (2012). Remote sensing and GIS

application for forest reserve degradation prediction and monitoring. In  First FIG

Young Surveyors Conference Knowing to Create the Future. 1-27.

Anderson, J. R. (1976).  A land use and land cover classification system for use with remote

sensor data, 964. US Govt. Printing Office.

Angelsen, A. & Kaimowitz, D. (2001). When does technological change in agriculture

promote deforestation. Tradeoffs or synergies, 89-114.

Barbier, E. B. (2000). The economic linkages between rural poverty and land degradation:

some evidence from Africa. Agriculture, Ecosystems & Environment, 82(1-3), 355-

370.

Page | 40
Campbell, J. B., & Wynne, R. H. (2011). Introduction to remote sensing. Guilford Press.

Cegielska, K., Noszczyk, T., Kukulska, A., Szylar, M., Hernik, J., Dixon-Gough, R.,

Jombach, S., Valánszki, I., Kovács, K. F. (2018). Land use and land cover changes in

post-socialist countries: Some observations from Hungary and Poland. Land Use

Policy, 78, 1-18. doi:10.1016/j.landusepol.2018.06.017

Couclelis, H. (2003). The certainty of uncertainty: GIS and the limits of geographic.

Dewan, A. M., & Yamaguchi, Y. (2009). Land use and land cover change in greater Dhaka,

Bangladesh: Using remote sensing to promote sustainable urbanization. Applied

geography, 29(3), 390-401. doi:10.1016/j.apgeog.2008.12.005

Dewan, A. M., & Yamaguchi, Y. (2009). Using remote sensing and GIS to detect and

monitor land use and land cover change in Dhaka Metropolitan of Bangladesh during

1960–2005. Environmental monitoring and assessment, 150(1), 237-249.

Eastman, J. R. (2001). Guide to GIS and image processing Volume. Clark Labs, 2, 1-144.

Ellis, F. (2000). Rural livelihoods and diversity in developing countries. Oxford university

press.

Fan, Y., Yu, G., He, Z., Yu, H., Bai, R., Yang, L., Wu, D., (2017). Entropies of the Chinese

Land Use/Cover change from 1990 to 2010 at a County level. Entropy 19 (2), 51.

Goldewijk, K. K., & Ramankutty, N. (2009). Land use changes during the past 300

years. Land-Use, Land Cover and Soil Sciences, 1, 147-168.

Gorelick, N., Hancher, M., Dixon, M., Ilyushchenko, S., Thau, D., & Moore, R. (2017).

Google Earth Engine: Planetary-scale geospatial analysis for everyone. Remote

sensing of Environment, 202, 18-27. doi:10.1016/j.rse.2017.06.031

Hoffer, R. M. (1978). Biological and physical considerations in applying computer-aided

analysis techniques to remote sensor data. Remote sensing: The quantitative

approach, 227-289.

Page | 41
Hossain, M. K., Alam, M. K. and Miah, M. D. (2008). Forest restoration and rehabilitation in

Bangladesh. IUFRO World Series, 20(3), 21-66.

Houghton, R. A. (1994). The worldwide extent of land-use change. BioScience, 44(5), 305-

313.

Islam K, Jashimuddin M, Nath B. and Nath T. K. (2018) Land use classification and change

detection by using multi-temporal remotely sensed imagery: The case of Chunati

wildlife sanctuary, The Egyptian Journal of Remote Sensing and Space Science, 21,

37-47.

Islam, K. R., & Weil, R. R. (2000). Land use effects on soil quality in a tropical forest

ecosystem of Bangladesh. Agriculture, Ecosystems & Environment, 79(1), 9-16.

Islam, K. R., Ahmed, M. R., Bhuiyan, M. K. and Badruddin, A. (2001). Deforestation effects

on vegetative regeneration and soil quality in tropical semi‐evergreen degraded and

protected forests of Bangladesh. Land Degradation & Development, 12(1), 45-56.

Islam, M. R., Miah, M. G., Inoue, Y., 2014. Analysis of Land use and Land Cover Changes

in the Coastal Area of Bangladesh using Landsat Imagery. LDD. Vol. 27, 899-909.

Jayakumar, S., & Arockiasamy, D. I. (2003). Land use/land cover mapping and change

detection in part of Eastern Ghats of Tamil Nadu using remote sensing and

GIS. Journal of the Indian Society of Remote Sensing, 31(4), 251-260.

Khan, M. M. H., Bryceson, I., Kolivras, K. N., Faruque, F., Rahman, M. M., & Haque, U.

(2015). Natural disasters and land-use/land-cover change in the southwest coastal

areas of Bangladesh. Regional Environmental Change, 15(2), 241-250.

Page | 42
Lambin, E. F., Turner, B. L., Geist, H. J., Agbola, S. B., Angelsen, A., Bruce, J. W., ... & Xu,

J. (2001). The causes of land-use and land-cover change: Moving beyond the

myths. Global environmental change, 11(4), 261-269.

Landis, J. R., & Koch, G. G. (1977). An application of hierarchical kappa-type statistics in

the assessment of majority agreement among multiple observers. Biometrics, 363-374.

 Liping C, Yujun S, Saeed S (2018) Monitoring and predicting land use and land cover changes

using remote sensing and GIS techniques—A case study of a hilly area, Jiangle, China. PLoS

ONE 13(7): e0200493.

Longley, P. A., Goodchild, M. F., Maguire, D. J., & Rhind, D. W. (2005). Geographic

information systems and science. Ed. 2. John Wiley & Sons.

Macleod, R. D., & Congalton, R. G. (1998). A quantitative comparison of change-detection

algorithms for monitoring eelgrass from remotely sensed data. Photogrammetric

engineering and remote sensing, 64(3), 207-216.

Maes, J., Crossman, N. D., & Burkhard, B. (2016). Mapping ecosystem services.

In Routledge handbook of ecosystem services, 188-204. Routledge.

Mamnun, M. and Hossen, S. (2020). Spatio-temporal analysis of land cover changes in the

evergreen and semievergreen rainforests: A case study in Chittagong Hill Tracts,

Bangladesh. International Journal of Forestry, Ecology and Environment, 02(02), 87-

99.

Nath, S. S., Bolte, J. P., Ross, L. G., & Aguilar-Manjarrez, J. (2000). Applications of

geographical information systems (GIS) for spatial decision support in

aquaculture. Aquacultural Engineering, 23(1-3), 233-278.

Page | 43
Penner, J.E. (1994). Atmospheric chemistry and air quality. In Changes in Land use and land

cover: A global perspective. 175–210. Edited by W.B. Meyer and B.L. Turner.

Cambridge University Press, Cambridge

Puigdefábregas, J. (1998). Ecological impacts of global change on drylands and their

implications for desertification. Land degradation & development, 9(5), 393-406.

Rai, R., Zhang, Y., Paudel, B., Li, S., & Khanal, N. R. (2017). A synthesis of studies on land

use and land cover dynamics during 1930–2015 in Bangladesh. Sustainability, 9(10),

1866.

Rasul, G., Thapa, G. B. and Zoebisch, M. A. (2004). Determinants of land-use changes in the

Chittagong Hill Tracts of Bangladesh. Applied geography, 24(3), 217-240.

Reid, R. S., Krushka, R. L., Muthuni, N. et al. (2000). Land-use and land-cover dynamics in

response to changes in climatic, biological and socio-political forces: The case of

southwestern Ethiopia landscape ecology, 15, 339-355.

Richards, J. A. (2013). Supervised classification techniques. In Remote Sensing Digital

Image Analysis, 247-318. Springer, Berlin, Heidelberg.

Richards, J. F., Flint, E. P., & Daniels, R. C. (1994). Historic Land Use and Carbon Estimates

for South and Southeast Asia: 1880-1980 (NDP-046) (No. ORNL/CDIAC-61; NDP-

46). Environmental System Science Data Infrastructure for a Virtual Ecosystem;

(CDIAC), Oak Ridge National Laboratory (ORNL), Oak Ridge, TN (USA).

Riebsame, W.E., Meyer, W.B. & Turner, B.L. (1994). Modeling land use and cover as part of

global environmental change. Climatic Change, 28, 45–64. doi:10.1007/BF01094100

Roy, B., & Saha, P. (2016). Temporal analysis of land use pattern changes in Chittagong

district of Bangladesh using google earth and ArcGIS. In Proceedings of the Annual

Page | 44
Int’l Conference on Chemical Processes, Ecology & Environmental Engineering

(ICCPEE’16), Pattaya, Thailand. 28-29.

Singh, A. (1989). Review article digital change detection techniques using remotely-sensed

data. International journal of remote sensing, 10(6), 989-1003.

Trincsi, K., & Turner, S. (2014). Mapping mountain diversity: Ethnic minorities and land use

land cover change in Vietnam's borderlands. Land Use Policy, 41, 484-497.

Troy, A., & Wilson, M. A. (2006). Mapping ecosystem services: practical challenges and

opportunities in linking GIS and value transfer. Ecological economics, 60(2), 435-449

van Vliet, J., Bregt, A. K., Hagen-Zanker, A. (2011). Revisiting kappa to account for change

in the accuracy assessment of land-use change models. Ecol. Model. 222, 1367-1375.

Xu, X., Shrestha, S., Gilani, H., Gamma, K. M., Siddiqui, B. N., Jain, A. K. (2020).

Dynamics and drivers of land use and land cover changes in Bangladesh. Regional

Environmental Change 20, 54. doi:10.1007/s10113-020-01650-5

Yang, X., Lo, C. P. (2002). Using a time series of satellite imagery to detect land use and

land cover changes in the Atlanta, Georgia metropolitan area. Int. J. Remote Sens. 23,

1775-1798.

Page | 45
APPENDICES
APPENDIX 1: Error Matrix showing accuracy of 1990 supervised classification of land
uses of Rangamati Sadar Upazila

Reference Data
Error Matrix Settlement Barre Agricultura Degraded Water Fores Row
n area l land forest body t Total
Settlement 5 0 0 0 0 0 5
Barren area 0 3 0 0 0 2 5
Classified Data

Agricultural land 0 0 5 0 0 0 5
Degraded forest 0 0 0 3 0 1 4

Water body 0 0 0 0 5 0 5
Forest 0 0 0 1 0 10 11
Column Total 5 3 5 4 5 13 35

APPENDIX 2: Error Matrix showing accuracy of 2000 supervised classification of land

uses of Rangamati Sadar Upazila

Reference Data
Error Matrix Settlement Barre Agricultura Degraded Water Fores Row
n area l land forest body t Total
Settlement 5 0 0 0 0 0 5
Barren area 1 5 0 0 0 0 6
Classified Data

Agricultural land 0 0 7 0 0 0 7
Degraded forest 1 0 1 2 0 0 4
Water body 0 0 0 0 5 0 5
Forest 0 0 1 0 0 9 10
Column Total 7 5 9 2 5 9 37

Page | 46
APPENDIX 3: Error Matrix showing accuracy of 2010 supervised classification of land
uses of Rangamati Sadar Upazila

Reference Data
Error Matrix Settlement Barre Agricultura Degraded Water Fores Row
n area l land forest body t Total
Settlement 5 0 0 0 0 0 5
Barren area 0 3 0 0 1 0 4
Classified Data

Agricultural land 0 0 4 0 2 0 6
Degraded forest 0 0 0 3 0 2 5
Water body 0 0 0 0 6 0 6
Forest 0 0 0 1 0 6 7
Column Total 5 3 4 4 9 8 33

APPENDIX 4: Error Matrix showing accuracy of 2020 supervised classification of land

uses of Rangamati Sadar Upazila

Reference Data
Error Matrix Settlement Barre Agricultura Degraded Water Fores Row
n area l land forest body t Total
Settlement 6 0 0 0 0 0 6
Barren area 0 6 0 0 0 0 6
Classified Data

Agricultural land 0 0 6 0 3 0 9
Degraded forest 0 0 0 7 0 0 7
Water body 0 0 0 0 8 0 8
Forest 0 0 0 1 0 4 5
Column Total 6 6 6 8 11 4 41

Page | 47