INTEGRATION OF REMOTE SENSING IN WATERSHED STUDIES:

A CASE STUDY OF CHAWIA & FURURU TAITA TAVETA COUNTY

KENYA.

BY:

ROBERT MOKUA NYANG’AU

MG 20/PU/36039/20

MSC STUDENT

A THESIS SUBMITTED TO PWANI UNIVERSITY, IN PARTIAL FULFILLMENT OF

THE REQUIREMENT FOR THE AWARD OF A MASTERS DEGREE IN
ENVIRONMENTAL SCIENCE.

SEPTEMBER 2023

i
 DECLARATION

I, Robert Mokua hereby declare that this thesis, “Integration of remote sensing in watershed
springs studies: A case study of Chawia & Fururu, Taita Taveta County Kenya.” Except for
permissible references from other sources, which have been duly acknowledged, the proposal
consists entirely of my work that will be produced from research to be undertaken under the
supervision of Drs. Melckzedeck Osore, Stanley Nadir, and Mike Teucher, and that no part of it
has been published or presented for another degree elsewhere.

Date……………………….

Signed…………………….

ROBERT MOKUA NYANG’AU

MSC. STUDENT

DEPARTMENT OF ENVIRONMENTAL AND EARTH SCIENCES.

ii
Approval

We confirm that this thesis by the candidate was under our supervision. Without the author's
and/or Pwani University's explicit written consent, no part of this proposal may be duplicated.

DR. MELCKZEDECK OSORE Senior Research Scientist

Kenya Marine & Fisheries Research Institute (KMFRI)

PO Box 81651, Mombasa 80100,

Kenya.

Date………………… Signed………………..

DR. STANLEY NADIR

Deputy Regional Director /Senior Research Scientist

Kenya Forestry Research Institute, Coast Eco-Region Research Program
P.O Box 1087 – 80200, Malindi
Supervisor

Date………………… Signed………………..

DR. MIKE TEUCHER

Supervisor
Martin-Luther-University Halle-Wittenberg, Germany.
Institute for Geosciences and Geography
Department of Geoecology.

Von-Seckendorff-Platz 4, D-06120 Halle.

Date 08/09/2023…

iii
 DEDICATION

Pacifica Ntabo, Denis Nyang'au, Sammy Moya, and Mary Moya, you are the individuals that
inspire me, and this proposal is dedicated to you. I would like to appreciate Mr. Sammy Moya
(Uncle), (Mary Moya), and Penninah Rosanna (Auntie) for the unwavering support you have
given me throughout my study period. In addition, I must be thankful to the Almighty God who
has been faithful and gracious to me.

iv
 ACKNOWLEDGMENT

My greatest gratitude goes to God almighty for his protection and guidance during my entire
studies. I want to thank Dr. Melckzedeck Osore and Dr. Stanley Nadir and all the staff at Pwani
University especially the Department of environmental and Earth sciences and express my
sincere gratitude to my supervisors for their leadership, support, attention to detail, and hard
work have set an example I hope to match someday.

I would also like to express my sincere appreciation to DAAD –German Academic Exchange
Services for their Research Grant to support my research expenses and fieldwork through
financial and professional expert support throughout the research period.

Finally, I want to thank all my colleagues for providing a challenging and wonderful experience
throughout the entire proposal development period.

v
 LIST OF ABBREVIATIONS

GPS- Global Positioning System

GCP-Ground Control Points

GIS -Geographic Information System

NGO- Non-Governmental Organization

USGS- United States Geological Survey

UN- United Nations

UNEP- United Nations Environmental Programmes

LU-LC- Land Use and Land Cover

DEM- Digital Elevation Model

ASTER- Advanced Spaceborne Thermal Emission and Reflection Radiometer

GDEM- Global Digital Elevation Model

GLCF-Global Land Cover Facility

UAV-Unmanned Aerial Vehicle

AOI- Area of Interest

MCDM- Multi-Criteria Decision Analysis Methodologies

vi
 LIST OF FIGURES

Figure 1: Schematic representation of a watershed: Source: https://estuarymetal.com/watershed-

hydrological-function/...................................................................................................................25
Figure 2: Chawia Forest patch and degraded Fururu forest fragments. Source: Petri K. E. 200730
Figure 3: Site map showing Chawia and Fururu fragmented forests within the Taita Hills.........31
Figure 4: Taita Taveta County elevation map Taita Taveta County Integrated Development Plan
2018-2022 2018)............................................................................................................................32
Figure 5: View of the Taita Hills (Source: Camera image by Author, 4 September 2021............34
Figure 6:Showing physical location of the mapped springs in Chawia forest & Figure 7:Showing
physical locatikon of the mapped springs in Fururu Forest...........................................................50
Figure 8:showing delineated watersheds in Chawia & Figure 9: Showing delineated watersheds
in Fururu........................................................................................................................................51
Figure 10: Showing the Phenological Land cover changes in Chawia from 1987-2022..............54
Figure 11:Showing the Phenological Lnad cover Changes in Furur Forest from 1987-2022.......54
Figure 12: Showing difference in vegetation Index (NVDI) in percentages.................................55
Figure 13: Showing Land use status in chawia and Fururu forested ecosystems..........................56
Figure 14: Map showing Chawia-Fururu Land use i 2022............................................................57
Figure 15:Showing Chawia Land cover in 2022 & Figure 16:Showing Land cover in 2017.......58
Figure 17:Showing land cover in 2019 & Figure 18:Showing land cover in 2009.......................58
Figure 19:Showing Land cover in Fururu in 1987........................................................................59
Figure 20:Showing Fururu Land cover in 2009 & Figure 21:Showing Fururu Land cover in 2017
.......................................................................................................................................................59
Figure 22:Showing Fururu land cover in 2019 & Figure 23:Showing Fururu land cover in 2022
.......................................................................................................................................................60

vii
 LIST OF TABLES

Table 1: Characteristics of Landsat images and used in monitoring land cover change...............37

Table 2:Landsat 8 images has 11 bands........................................................................................38

Table 3: Landsat 7 image has 9 bands...........................................................................................38

Table 4: Land use image combinations.........................................................................................39

Table 5:Land cover change (Ha) from 1987 to 2022....................................................................54

Table 6: Accuracy summary..........................................................................................................59

Table 7: Error matrix.....................................................................................................................60

viii
 TABLE OF CONTENTS

DECLARATION...........................................................................................................................ii

DEDICATION..............................................................................................................................iv

ACKNOWLEDGMENT...............................................................................................................v

LIST OF ABBREVIATIONS......................................................................................................vi

LIST OF FIGURES.....................................................................................................................vii

ABSTRACT................................................................................................................................xiii

CHAPTER ONE..........................................................................................................................14

INTRODUCTION.......................................................................................................................14

1.0 Background information......................................................................................................14

1.1 Statement of the Problem.....................................................................................................16

1.2 Justification of the study......................................................................................................17

1.3 Scope and Limitations of the study......................................................................................18

1.4 Overall objective..................................................................................................................18

1.4.1 Specific objectives:...........................................................................................................19

1.4.2 Research Questions...........................................................................................................19

1.5 Hypotheses...........................................................................................................................19

1.6 Significance of the study......................................................................................................20

CHAPTER TWO.........................................................................................................................22

LITERATURE REVIEW...........................................................................................................22

2.1 The concept of a watershed..................................................................................................22

2.2 Watershed Management Approach......................................................................................23

2.3 Intergrated watershed management (IWM).........................................................................24

2.4 Watershed degradation and rehabilitation............................................................................24

ix
 2.5 Remote sensing and Geographic Information Systems.......................................................26

2.6 Use of remote sensing data to analyze Land Uses and Land Cover....................................27

2.7 Case studies in which remote sensing and GIS were used to investigate watershed
hydrology and changes in land use and land cover....................................................................28

2.7.1 A Case study of Integration of Remote Sensing and GIS for Jammu and Kashmir, India.
....................................................................................................................................................28

2.7.2 A Case study of Gangajalghati Block, West Bengal, India..............................................28

2.7.3 A Case study: Exploration of groundwater potential using GIS and remote sensing in
Embu County, Kenya.................................................................................................................29

2.8 Land cover changes during the last decades........................................................................29

CHAPTER THREE.....................................................................................................................31

MATERIALS AND METHODS................................................................................................31

Introduction................................................................................................................................31

3.0 Description of the study area................................................................................................31

3.1 Study sites and Land Cover situation in Taita Hills.............................................................31

3.1.1 Study site...........................................................................................................................32

3.1.2 Physiographic and Natural Conditions..............................................................................33

3.1.3 Physical and Topographic features...................................................................................33

3.1.4 Agro-Ecological Zones.....................................................................................................34

3.1.5 Climate conditions............................................................................................................35

3.2 Assumptions of the study.....................................................................................................36

3.3 Methodology........................................................................................................................37

3.3.1 Mapping the physical location of springs in the fragmented Chawia and Fururu
watersheds (Physical mapping using remote sensing of satellite images).................................37

Mapping of springs and wells within the watersheds................................................................37

Introduction................................................................................................................................37

x
3.3.2 Watershed Delineattion and Hydrological modelling.......................................................37

3.3.4 Land use, land cover change classes within and around the forested watersheds period
1987-2022..................................................................................................................................40

3.3.5 Landuse classification using sentinel 2a satellite image...................................................40

3.4 Data Analysis.......................................................................................................................41

3.4.1 Satellite Image processing and analysis............................................................................41

3.4.2 Confusion matrix/Accuracy Assessment..........................................................................41

3.4.3 Time Series Analysis of the Study area satellite images..................................................42

3.4.5 To identify and delineate the original watersheds in fragmented Chawia, Fururu forested
watershed using Geographical information systems (GIS) and remote sensing........................42

Introduction................................................................................................................................42

3.4.6 DEM Preparation and Generation.....................................................................................43

3.4.7 Filling the DEM................................................................................................................43

3.4.8 Creating a Flow Direction and Accumulation..................................................................43

3.4.9 Creating Watershed Pour Points.......................................................................................44

3.5 Watershed Delineation.........................................................................................................44

Introduction................................................................................................................................44

3.6 Data analysis;.......................................................................................................................45

3.6.1 Photography......................................................................................................................45

3.6.2 Observations and Ground truthing....................................................................................45

3.7 To determine the categories in land use and land cover LU-LC through spatial Time series
comparison of the LU-LC maps and overall catchment characteristic......................................45

Introduction................................................................................................................................45

3.8 Data sources........................................................................................................................45

3.8.1 Satellite Images.................................................................................................................45

3.8.2 Image Preparation, Restoration, Rectification, and Enhancement....................................46

xi
 3.8.3 Image Processing and classification..................................................................................47

3.8.4 GIS image processing tool................................................................................................48

CHAPTER FOUR.......................................................................................................................50

RESULTS & DISCUSSION.......................................................................................................50

3.9 To map the physical location of springs and current status springs in the fragmented
Chawia and Fururu watersheds..................................................................................................50

4.0...............................................................................................................................................51

To identify and delineate the original watersheds in the fragmented Chawia, Fururu forested
watersheds using GIS and RS....................................................................................................51

4.1 To determine the Land uses and Land cover classes through spatial Time series
comparison of the overall catchment characteristics.................................................................52

4.1.1 Land Use Land Cover Analysis from 1987 to 2022.........................................................52

4.1.2 Percentage of change detection in Land use Land cover from 1987-2022.......................55

4.1.3 Land Cover Analysis from 1987 to 2022..........................................................................59

4.2 Accuracy summary...............................................................................................................61

4.2.1 Error matrix.......................................................................................................................62

4.3 Discussion............................................................................................................................62

CHAPTER FIVE.........................................................................................................................69

SUMMARY, CONCLUSION AND RECOMMENDATIONS...............................................69

4.4 Conclusion............................................................................................................................69

4.5 References............................................................................................................................73

APPENDICES..............................................................................................................................81

Appendix 1: Work plan..............................................................................................................81

Appendix 2: Budget...................................................................................................................82

Appendix 3: Equipment used.....................................................................................................83

xii
 ABSTRACT

Increased water use needs due to population growth and the current environmental problems due
to climate change have resulted in the drying up of springs, and a reduction in their discharge
volumes in Taita Hills. These problems called for the study and exploration of groundwater
resources in Taita Hills. The main objective of the study was to evaluate the potential of
integrating both Geographical Information Systems (GIS) and Remote Sensing (RS) in studying
the status and spatial distribution of recent springs in the Chawia and Fururu forested watersheds.
In this study, physical mapping of the springs was done by demarcating the location using
Geographic Positioning System (GPS). The study identified and delineated the original
watersheds in the Chawia and Fururu Forest fragments using GIS and RS using satellite data for
watershed mapping and determined the categories for Land Use Land Cover (LU-LC) through
spatial comparison of their (LU-LC) maps and catchment characteristics by performing sentinel
data analysis.

Mapping of springs was part of the data collection process and RS data from Landsat 7, 8 and
Sentinel 2 were acquired, as well as processed and analyzed. For watershed delineation, the data
was examined using the Environmental Systems Research Institute, Inc (ESRI) ArcMap GIS
Hydrology tool, and image classification was done using the Maximum Likelihood algorithm,
with classification accuracy verified using the confusion matrix approach. GIS image-processing
program was used to ground truth data and process Digital Elevation Model (DEM) data. From
the results: Ninety one (91) springs were mapped and visualized in Chawia and Fururu forested
Watersheds and it was found out that Fifty three (53) springs were in Chawia whereby thirteen
(13) were found to be inactive and thirty (30) were still active in Chawia Compared to Thirty
eight (38) springs in Fururu which were found to be all active. The study provided information
on the LU-LC classes in the study area, and how they influence the functions of the springs in
providing water as well as informing environment managers on watershed management. Six land
use classes were identified, Built-up (66.3%), Forestland (26.1%), Agriculture (7.13%), Water
bodies (0.23%), Grassland (0.04%), and Bare land (0.01%). The study also showed that the
Chawia forest cover has reduced from the year 1987 (70%) of Vegetation Index to 2022 (60%)
and Fururu had increased by 10% from 70% in 1987 to 80% in 2022.

xiii
Keywords: Springs, Land use; Cover Change; ; watershed; remote sensing ; Sentinel-2; Chawia;
Fururu

xiv
 CHAPTER ONE

INTRODUCTION

1.0 Background information

Water is regarded as life and the core of sustainable development and it is indeed very vital for
socio-economic development, healthy ecosystems, and human survival. The demand for water
use around the world has been growing at more than twice the rate of population increase in the
last century, and, although there is no global water scarcity as such, there is an increasing
number of regions are having chronic issues of water (Changwony et al., 2017). According to
Changwony et al. (2017), today's water scarcity is exacerbated by droughts, which damage both
surface water resources and generate long-term imbalances, reducing water supply and
disrupting land cover and biodiversity. In general, spring occurs where the ground surface and
the impermeable rocks cross with the ground water table. This is the point at which water flows
from an aquifer to the earth's surface. The majority of the time, the occurrence is determined by
the features of the rocks that recharge, such as lithology, the topsoil's porosity and permeability,
hydrogeomorphology, and precipitation (Taloor et al., 2020). Since groundwater can consistently
provide the necessary amount of high-quality water, it is a good resource; however, effective
water conservation plans are crucial for the long-term usage of groundwater. Excessive
groundwater extraction and improper aquifer recharging are common causes of groundwater
depletion in many places. In order to promote effective use and systematic management of
groundwater resources, precise estimation and prediction of groundwater recharge should be
made. (Lee et al., 2020).

There are more than 10 Cubic kilometers of water on earth all over the world. Although most of
the planet is covered by water, the greater quantity of water is available in forms unavailable to
terrestrial and freshwater ecosystems (Jackson et al., 2001). The rate and intensity at which the
global freshwater crisis and the environmental problems associated with the water crisis has been
underrated over decades with over 1 billion people on earth being without reliable access to
water which is very critical in achieving the United Nations Millennium Development Goals

15
(United Nations University & Institute for Water, 2012). The majority of people on earth live in
cities. Currently, it is estimated that 5 billion people live in cities. Within a decade, 6.5 billion
people are anticipated to live in metropolitan areas. Urbanization will have an impact on
terrestrial ecosystems and raise demand for the services they provide. (Abad-Segura et al., 2020).
The effectiveness of the world's environmental systems has been significantly impacted by the
growing change of the land surface caused by anthropogenic activity. Natural vegetation is being
converted to agricultural land at a rapid rate in the developing globe, which has led to a decrease
in the availability of water, soil, and vegetation resources. It alters surface and subsurface
vegetation in the watershed by affecting hydrological processes as evapotranspiration,
interception, and infiltration.(Ruttoh et al., 2022).

According to a report by the United Nations Inter-Action Council (2012), the water crisis'
environmental effects are damaging to our natural water supplies, and this is due to population
increase, which creates competition with nature. According to the United Nations report, it
shows that a great number of springs and rivers do not make it to sea and there is widespread
surface and groundwater contamination and human activities that make water unfit for use and
even drying up springs (United Nations University & Institute for Water, 2012).

In Kenya, with a population of 50 million, 32 percent of Kenyans rely on unimproved water

sources, such as springs, ponds, shallow wells, and rivers, while 48 percent of Kenyans lack
access to basic water solutions (Kanoti et al., 2019). The natural surface water sources in the
country are not enough to satisfy the vast population and the groundwater is a more important
feature of the natural environment (Muthoni, 2005).

Mountain springs are the forms in which groundwater ensures water security for the rural
population. Springs are the primary and reliable source of water in the hilly and mountainous
areas of Taita Taveta County. The livelihood of the community depends on it. Providing clean
water for agriculture and households and beverages (Niraula et al., 2021). Springs are indeed the
terminal points of the water tower flow systems, and knowledge of their spatial distribution is
essential for getting a better understanding of groundwater flow, vulnerable and water
availability (Iván et al., 2020). Taita Hills forest fragments (Chawia and Fururu) are known to be
among the many of the Taita Hills forest fragments covering and 86.0 ha and 5.0 ha respectively
(Hohenthal et al., 2015). Unequal and uneven allocation of water distribution and drying of

16
springs spatially and temporarily present a major problem for the sustainable development of
these regions. Many people in Taita Hills live in rural areas, the main socio-economic activity is
farming, agricultural labor, and therefore availability of water is essential and a backbone to
these people’s lives (Hohenthal et al., 2018).

Therefore, this study aimed to provide the much-needed hydrological information of springs in
Chawia and Fururu forested watersheds using RS and GIS techniques. The study provided
information on the physical location, status, discharge volumes of the springs within the
watershed. Relevant parties were utilized in this data to forecast water availability situations for
planning and decision-making. Furthermore, the data will contribute to a scientific understanding
that will be beneficial in the management and conservation of the Chawia and Fururu
watersheds.

1.1 Statement of the Problem

The forested Taita Hills are the headwater areas for many springs, streams, and rivers including
Iriba, Kisagala, and Sasenyi in Fururu and Laghonyi, Kengwa, and Mwakuhola in Chawia,
which provide water for agriculture and domestic use to the lowland communities. But due to
overexploitation and overuse of natural resources and environmental degradation is putting a lot
of pressure on water resources (Hohenthal et al., 2015). According to the study by Hohenthal et
al., (2015) from the 1960s, the indigenous forest cover has been reducing drastically due to
population increase, expansion of intensive agriculture, plantation of exotic tree species, and
road construction.
The slopes of Taita Hills were densely vegetated and covered with forest in a dense cloud forest,
but the current state of the patch of the back wood was destroyed and native to the place where
the forest was located since the 1960s during the arbitration. Some of the patches remain. Nine in
the mosaic of human settlements, small country plots, and extraordinary tree houses, leaving the
first 400 hectares of forest areas of Chawia (86 hectares), Fururu (5 hectares), and Ngangao (46
hectares). Land use in Chawia and Fururu has been swamped by increased horticulture and
brushing as a result of the rapid growth of the population, and this has forced the neighboring
local area to climb the slopes, creating actual erosion, soil disintegration, and lowering water
tables in the slopes. The increase in the human population. (1969:110,142 ….1979:147,597

17
(+33.3%), 1989:207,273 (+40.4%), 1999:246,671 (+19.0%), 2009:284,671 (+15.4%), and 2019:
340,671 (+19.7%) KNBS Census Report (2019).

Consequently, the incorporation of GIS and RS techniques advances watershed depiction and the
land cover characterization from DEM information from Landsat. The use of GIS and RS in
concentrating on the Chawia and Fururu Watersheds helped in information securing on the
hydrological foundation of the two watersheds, group the land uses, and land cover changes in
the catchments. Rapid industrialization, urbanization, intensive agriculture, and changing
lifestyles have all generated new needs that are increasing yearly. Environmental deterioration
and severe water scarcity are results of the indiscriminate exploitation of surface and
underground waterways. The issue has been made worse by the rainfall's spatiotemporal
fluctuation.

1.2 Justification of the study

Remote sensing plays a vital role in the conservation and watershed management and achieving
optimum planning and operation of water resources (Masud & Bastiaanssen, 2017). Satellite data
is an alternative attractive for data required by hydrological models and provide spatial
information for decision-makers (Masud & Bastiaanssen, 2017). Remote sensing exhibits its
presentation quality on worldwide dreary estimation ability like scene investigation, land use
grouping, scene environmental change recognition, and landscape demonstrating. Tragically,
66% of the earth's surface is constantly covered by mists all through the year (Wang et al., 2009)
leading to significant issues in optical frequency remote sensing. Therefore the geographical
coverage of remote sensing data is an advantage of using RS instead of field data and the
geometrical resolution and development during the last decades. Thus it is possible today to
investigate small-scale phenomena with satellite-based RS data.

Esche et al., (2002) also stated that over half of the world is covered in the cloud at any given
time, making the main test of making repeatable and strong instructions. This can be achieved by
comprehension and tending to cloud tainting since mists/shadows cast obscure optical
symbolism influencing the precision of vegetation gauges, environment framework throughout
an expansive scope of existence scales (Wang et al., 2009). This influenced the satellite
gathering to the GPS gadget during hands-on work information assortment. Financial

18
imperatives: Because the neighborhood springs are diffused across Taita slopes, and the
complete review interaction demanded numerous assets, this review was a pricey endeavor.

The GIS information helped with evaluating watershed conditions through displaying effects of
human exercises to imagining effects of elective situations in understanding the danger to water
sources to the dynamic cycles; Land use arranging and backwoods restoration plots in Taita
slopes. If the review was not attempted, the local area and public would not have known about
the meaning of water sources as catchment regions and begin preserving them as water pinnacles
and in this way, more springs will keep on evaporating, as more land cover is lost. Additionally,
gave information to strategy definition and dynamic interaction. Therefore the geographical
coverage of remote sensing data was an advantage of using RS instead of field data and the
geometrical resolution and development during the last decades. Thus it was possible today to
investigate small-scale phenomena with satellite-based RS data.

1.3 Scope and Limitations of the study

The study focused on the use of GIS and RS to study springs in Chawia and Fururu watersheds
in a section of Taita Hills forest around the Ngerenyi Sub County. Furthermore, inside Taita
Taveta County, the study was limited to Chawia and Fururu forests near Ngerenyi campus area
water sources. When carrying out this study, it was believed that the lack of understanding of
both farmers and workers from local communities would be a barrier, which would limit the
amount of data available for the study. Financial constraints this study was such an expensive
venture, as the local springs are scattered across Taita Hills and the whole study process required
many funds. Because nearly half of the planet is covered in the cloud at any given moment,
cloud cover in the research area may influence satellite availability by GPS devices during
fieldwork data collecting.

Therefore, bringing the most significant challenge in creating repeatable and robust
classifications is by understanding and addressing cloud contamination since clouds/shadows
cast blur optical imagery affecting the accuracy of vegetation estimates, climate systems over a
broad range of time and spatial scales this therefore affected the satellite receiver to the GPS
device during fieldwork data collection. Remote sensing is Useful for generating environmental

19
indicators that can be integrated with collateral data and social indicators, Synoptic view, Multi-
resolution, multi-spectral, repetitive offers appropriate method for quick, unbiased mapping and
monitoring of natural resources both in space and time domain and also Timely and accurate
information on spatial distribution landuse, soil, vegetation density, forest, geology, water
resources etc.

1.4 Overall objective

The main objective of the study was to evaluate the potential of integrating Geographical
Information systems and Remote Sensing in studying the spatial distribution of springs in
Chawia and Fururu forested watersheds

1.4.1 Specific objectives:

i. To map the physical location of springs and current status springs in the fragmented
Chawia and Fururu watersheds

ii. To identify and delineate the original watersheds in the fragmented Chawia, Fururu
forested watersheds using GIS and RS.

To determine the Land uses and Land cover classes through spatial time series
comparison of the overall catchment characteristics.
1.4.2 Research Questions

i. What is the current physical location and status (dried up vs. active) of the
springs in the fragmented Chawia and Fururu watersheds?

ii. What is the delineation extent of the watersheds of the springs/wells in the
Chawia and Fururu forest fragments?
What are the Land uses and Land cover classes through spatial time series comparison of
the overall catchment characteristics?
1.5 Hypotheses

The study tested the Non-directional Alternative Hypotheses;

20
 i. The locations of springs and distribution in Chawia and Fururu forests watersheds have
changed compared to historical location and some historical springs have dried up
perhaps due to change in Land uses.
ii. The size of the watershed/catchment has changed

iii. There are changes in the land use classes and land cover categories within the
watershed/catchment areas

1.6 Significance of the study

The information generated from the study was imperative to successful planning and decision-
making to strengthen the national environmental management systems by providing data on
Integrated Watershed Management. The information can assist Kenya’s process towards
achieving the targets associated with the sustainable development goals and its efforts to achieve
vision 2030. The study was also to model water availability scenarios for planning and decision
making by relevant parties as well as county spatial planners. In addition, the information
contributed to scientific knowledge useful in dealing with Chawia and Fururu watershed
management and conservation for Integrated Watershed Management.

By identifying the many types of pollution existing in the watershed, how those pollutants are
carried, and making recommendations for ways to decrease or remove those pollution sources,
watershed management aids in the control of pollution of the water and other natural resources
within the watershed. A watershed's natural resources and water quality will be impacted in some
way by every activity that takes place within it. The quality of the resources in a watershed can
be impacted by new land development, runoff from already developed regions, agricultural
operations, domestic activities including gardening and lawn care, septic system use and
maintenance, water diversion, and vehicle maintenance. In order to adequately address these
activities and lessen the negative effects of pollution, watershed management planning
thoroughly identifies those that have an impact on the health of the watershed.

The planning process results in a cooperation between all parties who are impacted by the
watershed, which is another reason why watershed management is significant. Since all partners
have a stake in the health of the watershed, that cooperation is crucial to the effective
management of the land and water resources in the watershed. In situations when resources may

21
be scarce, it is also a useful technique to organize the implementation of watershed management
plans according to priority. The development of computer-based Geographic Information
Systems (GIS) and satellite remote sensing technology has made it possible for environmental
planners to have accurate overall perspectives with the least amount of time and money spent.

In order to enhance and maximize the use of water resources and watershed management
programs, the current study uses GIS in conjunction with remote sensing technologies. By doing
so, the study area will experience overall socioeconomic development and poverty reduction. It
is understood that increasing water supplies to the area to best meet the basic needs of drinking
and sustaining the agriculture, on which the majority of the population depends, is the most
effective way to end drought and famine and to reverse the process of desertification in an area.

In order to evaluate groundwater potential zones, remote sensing (RS) and geographic
information systems (GIS) have proven to be a viable and affordable tool. Researchers can
efficiently study, analyze, and manage groundwater resources thanks to RS's ability to collect,
process, and manage enormous amounts of data at once. (Guduru & Jilo, 2022). As a result,
during the past few decades, groundwater potential zone delineation using remote sensing and
GIS-guided multi-criteria decision analysis methodologies (MCDM) has become an increasingly
useful tool for managing water resources. In this study, ecosystems were assessed using satellite
photos, and the health and integrity of these ecosystems were assessed using spectral indices.
The parameterization of the yearly curve of spectral vegetation indices enables the derivation of
indices for productivity, seasonality, and phenology—three aspects essential to the functioning
of ecosystems—to define the reference conditions and assess the changes that take place.
Therefore, it was feasible to discover directional trends of change, important from the standpoint
of global change, through the analysis of a lengthy time series.

The goal of this study was to describe the growth in the body of knowledge regarding the use of
remote sensing in forest management to enhance ecosystem services with the hope of advancing
research and identifying new areas of inquiry. As a result, in this context, satellites enable
consistent, integrated, synoptic, and numerical observations of the planet and the evolution of its
surface across time. As a result, the study's suggested approaches for managing the ecological,
economic, and sociocultural balance offer benefits that are felt globally, including the

22
preservation of biodiversity and ecosystems, a decline in rural poverty, and a lessening of some
of the effects of climate change in the study area.

CHAPTER TWO

LITERATURE REVIEW

2.1 The concept of a watershed

A watershed is an area that drains into a stream and can also be referred to as a small catchment
from which all precipitation flows into a single stream forming a natural way to dispose of the
runoff as efficiently as possible (Swami, 2011). A watershed remains to be a crucial part of
environmental geography, which is capable of indicating and defining areas for water resources
management. Generally, the physical characteristics of the watershed, including topography,
geomorphology, lithology, bedrock, and geological structures, are not changing, but they
influence the set up of the drainage system of the watershed (Waiyasusri & Chotpantarat, 2020).
There is at least one watershed, also known as a catchment, drainage basin, or basin, for every
single point on the terrestrial surface. There are six levels for watersheds based on size for data
management and water resource planning: region, subregion, basin, sub-basin, watershed, and
sub-watershed. (Asgari, 2021).

Human activities have a direct influence on the quality and quantity of groundwater and related
natural resources which are affected by human activities and that the direct connection with the
upstream activity of the spring. Therefore, to observe the watershed conditions of the drainage
basin, remote sensing, and GIS techniques are used to study. Hereby, remote sensing provides
synoptic coverage, data from inaccessible areas/regions and former time steps. Thus, the
construction of a detailed base map. GIS provides an easy way to update and analyze spatial data
and the investigation gives a quantitative description of the drainage system, which is an
important aspect of watershed characterization (Farjana Birajdar, 2015).

23
Both biotic and abiotic elements are present in the watershed, including confined human
populations. Te watershed offers crucial functions required for society's well. Despite the
benefits of watersheds, the quantity and quality of services have been dropping over time, and
this is due to a combination of factors including climate change, land use and cover, and the
demand for these services from the world's population, which is expanding at an alarming rate
(Ruttoh et al., 2022).

2.2 Watershed Management Approach

An iterative process of integrated decision-making about the uses and alterations of lands and streams
within a watershed is known as watershed management. Stakeholders have the option to weigh various
objectives and uses for natural resources during this process, as well as to think about how their collective
actions may affect the long-term sustainability of these resources. (Srinivas et al., n.d.). The quantification
of the effects of possible and planned measures on pollution loadings and water quality is made possible
by watershed models. To address the functions and conflicts in a watershed, they are crucial tools.
(Erturk et al., 2006). As a result, they help us comprehend the procedures involved in pollution
from diffuse sources in particular. They can produce a variety of options based on the
watershed's criteria, which might also lead to planning strategies, and it would be simple to
evaluate the effects of human-induced activities. (Fadil et al., 2011). With the development of
simulation interfaces utilizing tools like Geographic Information System (GIS) and Remote
Sensing (RS), watershed modeling began with the estimation of water amount for engineering
applications. This led to the assessment of both the quantity and behavior of point and diffuse
source pollutants. (Erturk et al., 2006).

The watershed management approaches essentially envisage three main things; to enhance water
availability, to prevent erosion, and to generate biomass and GIS techniques have been a growing
trend for watershed management because it helps in rapid analysis and classification of land use
(Umrikar, 2015). Watershed Management and GIS works with the administration and checking
of water assets all the more productively and viably to empower spatial information or data about
the accessibility and nature of water assets in a watershed can be surveyed and observed over the
long run and inside a predefined time scale (Mirrah & Kusratmoko, 2017).

According to Bera and Bandyopadhyay (2012), water and land are likely nature's best
endowments in today's world, and these precious regular assets should be protected and

24
maintained with caution for ecological assurance and biodiversity balance. The greatest danger
to the world's freshwater assets is land debasement, which decreases spring stream rates and
brings down groundwater levels, which prompts silting up of estuaries and accordingly adversely
influences human prosperities (Bera & Bandyopadhyay, 2012). The creators contend that the
main answer for this issue is watershed the board utilizing space innovation and GIS.

These are the key tools that may be used to extract watershed borders for hydrological
assessment and Land use Land cover change detection. Bera & Bandyopadhyay, (2012), decided
the best land the board practice that limits water shortage for human and rural watersheds and in
this way tend to land debasement, deforestation, urbanization, and escalated impromptu agrarian
exercises (Bera & Bandyopadhyay, 2012).

2.3 Intergrated watershed management (IWM).

Integrated watershed management IWM is the process of managing human activities and natural
resources on a watershed basis while taking into account local community interests, social,
economic, and environmental challenges, as well as things like the effects of growth and climate
change (Integrated Watershed Management - NVCA, 2020).

There are social and political issues at the local and national levels all over the world as a result
of the effects of shifting land cover patterns on water resources. For instance, the move to
agriculture in the Al-Baha region of Saudi Arabia has led to a variety of changes in the
ecosystem's structure and function, which has led to a general degradation of the ecological
services offered by the region's natural system. (Mahmoud & Alazba, 2015).

The most significant agricultural activity, particularly in dryland agriculture, is watershed

management. It contributes more to the conservation of runoff water from various sources. The
primary source of water in the watershed is rainwater collection (Gurjar et al., 2017). In dry
locations, the captured water is utilized to irrigate crops under water stress. Additionally, it is

25
applied to the crops as supplemental irrigation or as irrigation that saves the lives of various
agricultural crops. (Gurjar et al., 2017).

2.4 Watershed degradation and rehabilitation

One of the promising scientific methods for evaluating and managing water resources is surface
hydrological indicators. The selection of a water recharge location, watershed modeling, runoff
modeling, watershed delineation, groundwater prospect mapping, and geotechnical research are
all dependent on drainage morphometric analysis. (Singh et al., 2014). A common LULC shift that
is of great concern all across the world is the conversion of forests, which make up about 30% of
the world's land, to agricultural fields.

Over one million km2 of tropical rainforests and moist deciduous forests on Earth were
reportedly destroyed between 1981 and 1990, according to the UN Food and Agricultural
Organization. This represents an average annual deforestation rate of 0.8% for these forests over
the course of the decade. (Rajaei et al., 2021). Forests offer a variety of advantages, including
flood protection and ecosystem services, therefore their removal has a significant impact on
water supplies. While the loss of forests is a global concern, it is more of a concern in developing
nations because there is less cooperation between the government and stakeholders, a reliance on
conventional management techniques for pollution and flood mitigation, and fewer resources
available to research and manage these impacts [2]. Projecting the effects of forest loss (due to
conversion to alternative LULCs) on water availability and quality in these nations is therefore
crucial. (Rajaei et al., 2021).

The use of a multi-stakeholder approach when managing natural resources is essential for
decision-making, improves the quality of policy decisions, and increases the acceptance of
policies throughout implementation. (Mwamuye et al., 2021). Multi-stakeholder involvement
aids in the coordination of policies across sectors for the management and maintenance of
mangrove forest ecosystems. In addition to providing a focal point for the development of shared
perspectives, accepted solutions, and interactions to establish consensus regarding the specified
objectives, public participation boosts confidence and excitement in mangrove forest

26
conservation. Local populations have extensive indigenous knowledge that enhances and
supplements scientific knowledge and methods of managing mangrove forest resources locally
(Mwamuye et al., 2021).

Figure 1: Schematic representation of a watershed: Source:

https://estuarymetal.com/watershed-hydrological-function/

2.5 Remote sensing and Geographic Information Systems

Remote sensing is the science of acquiring, processing, and interpreting images that record the
interaction between electromagnetic energy and matter (Vora, 2011). The science and technique
of obtaining information about an object or an area through the analysis of data acquired by a
device that is not in contact with the object and the result is usually stored in image data (aerial
photographs are also considered as image data) (Vora, 2011).

In addition, Change Detection is the process of identifying differences in the state of an object or
phenomenon by observing it at different times where the time and accuracy of change detection
on the earth’s surface can give us a better understanding of the relationships and interactions
27
between human activities and the natural environment. Therefore, guides decision-making and
management of the use of resources (Alqurashi & Kumar, 2013). The effectiveness of spatial
digital approaches for locating, evaluating, and monitoring the LULC in the Chawia-Fururu
watershed has been shown through the application of RS and GIS (Ruttoh et al., 2022).

GIS is an effective tool for creating digital databases that include a variety of thematic maps
(such as information on wells, crops, and socioeconomic data) as well as non-spatial data (such
as information on slope, rock kinds, landforms, etc.). Using various analysis/modeling
techniques, thematic and tabular data are further integrated in a GIS environment. By fusing
socioeconomic and other auxiliary data (such as population, depth of weathering, well and crop
inventory, etc.) with thematic information (land use/cover, slope, drainage, soil, etc.), GIS aids in
arriving at alternative land use practices, selection of sites for water harvesting structures, etc. for
a watershed (Pande et al., 2021).
2.6 Use of remote sensing data to analyze Land Uses and Land Cover

Land cover is the physical and natural front of the outer layer of the land, while land use is the
condition of human exercises like farming, ranger service, and building development that modify
land surface cycles. In other words, land use refers to the patterns of human activity for its
exploitation, whereas land cover refers to the physical and biological cover of the surface. Both
natural and manmade factors contribute to the change in the land's cover and use. (Sewnet,
2016). Land uses Land Cover (LU-LC) and their elements are essential in affecting the
environment conditions and capacities and in this manner, the progressions in LU-LC speeds up
soil disintegration rates and effects water bodies in a watershed since there are numerous
associations between land surface attributes and the hydrological cycle (Sewnet, 2016). Thusly,
the utilization of GIS and RS information along with a high goal from satellite symbolism causes
these advances to seem ready and can have an enormous effect ashore asset the board drives
associated with the checking of Land use and Land cover planning and chance discovery (Singh
et al., 2014). According to Pande et al., (2021), a review on the evaluation of land use and land
cover change identification planning in the Pangari watershed in India. GIS, and remote sensing
advances were incorporated for investigation and changes in LU-LC classes, utilizing a directed
arrangement strategy to recognize Land use classes with the assistance of field date, where the
review showed a variety of agrarian land and woods cover in Region 1. (Pande et al., 2021).

28
Remote sensing data provides detailed information on the land use/cover (crop land, forest,
wastelands, grasslands etc.) of the watersheds, which is very useful in prioritizing the watersheds
(Pande et al., 2021).

2.7 Case studies in which remote sensing and GIS were used to investigate watershed
hydrology and changes in land use and land cover.

CASE STUDY 1:
2.7.1 A Case study of Integration of Remote Sensing and GIS for Jammu and Kashmir,
India.

A study was done by Jasrotia et al., (2016) on the integrated remote sensing and GIS approach
for delineation of groundwater potential zones in Devak and Rui watersheds of Jammu and
Kashmir, India. Remote sensing, GIS, and Fieldwork techniques were combined and used to
delineate the groundwater sources map of Devak and Rui watersheds. Hydro geomorphology,
drainage density, and Land use Land cover maps were all created using remotely sensed data.
Advanced Space borne Thermal emission and Reflection (ASTER) digital elevation model
(DEM) data was used for preparing slope maps, physical tests, which generate a radial flow of
water towards the pumping well, were used to prepare discharge maps. All of these maps were
converted to raster data and given a theme weight and class score based on the criteria used. The
groundwater potential map was created by combining these maps in a GIS context using the
Weighted Index Overlay Analysis (WIOA) approach, and the groundwater prospects zones were
displayed in five categories: excellent, good, moderate, low, and run-off zone. The results from
the study revealed that the excellent zone covers 13.5%, good zone 22.7%, moderate zone
15.8%, low zone 18.5% and run off zone 29.5% of the total study area. The results depicted the
groundwater potential zone found to be helpful in better planning and management of
groundwater resources in the study area (Jasrotia et al., 2016).

CASE STUDY 2:
2.7.2 A Case study of Gangajalghati Block, West Bengal, India.

In another study conducted by Das (2017), employed the integration of remote sensing and GIS
techniques to delineate groundwater potential zones for groundwater management of

29
Gangajalghati Block, Bankura District, West Bengal. Survey of India topographical maps and
satellite imageries along with other data and field observation data were utilized to prepare
various thematic layers For example, drainage, slope, lineament, geology, hydro geomorphic
units, Land use Land cover, soil, and depth to water level to the hard rock (Das, 2017). The
weighted index overlay method was utilized to identify groundwater potential zones, and all
thematic used that influence groundwater occurrences were assessed and integrated through GIS
analysis based on weights assigned. According to the study's findings, the study area can be
divided into four distinct water potential zones. Bore well yield revealed that medium yielding
wells (50-100 L per min) are in and around the moderate to high lineament density zone and
buried pediment medium region and validation revealed that a good correlation to the observed
groundwater potential zonation. This study was useful for better planning and management of
groundwater resources as well as providing quick perspective guides for groundwater
exploitation in aquifer systems (Das, 2017).

CASE STUDY 3:
2.7.3 A Case study: Exploration of groundwater potential using GIS and remote sensing in
Embu County, Kenya.

In a study conducted by Chepchumba (2019) remote sensing and GIS geospatial approaches
were used to identify groundwater potential zones in Embu County, Kenya (Chepchumba, 2019).
The lineament layer was obtained by processing Landsat 8 ETM + images with ENVI 4.7
Principal Component Analysis (PCA) and automatically extracting from the Principal
Component Images (PCI) with the LINE module of Geomartica software (Chepchumba, 2019).
All thematic layers in Embu County were converted to raster format before utilizing weighted
overlay analysis in GIS to map groundwater zones. The study was viewed as a way of modeling
geospatial data for mapping of groundwater potential zones for planning by local authorities,
with about 78 percent of the total area falling under 'high' to ‘very high' zones, indicating that
over three-quarters of the study area falls under good groundwater potential zones.
(Chepchumba, 2019). Therefore, remote sensing is the ideal method for observing the effects of
implementation, particularly changes in water distribution, water bodies, and land use/cover.

2.8 Land cover changes during the last decades

30
Agricultural intensification, deforestation, and urbanization are the mainland use that has
contributed immensely to tremendous changes in land cover on global scale (Ndalilo et al.,
2021). These activities are largely driven by the need to meet immediate human needs at the
expense of environmental conservation resulting in the decline in biodiversity through the loss,
modification, and fragmentation of habitats, degradation of soil and water quality, and
overexploitation of native species (Ndalilo et al., 2021).

Agrarian development is the primary driver of land cover change in the Taita Slopes with the latest
things endured, it is normal that rural regions will possess 60% of the review region by 2030. Model
reenactments demonstrate that agrarian development is probably going to occur transcendently in
marshes and lower regions. The massive advancements in land inclusion result in modifications in
the Taita Slopes' water towers arrangement, where the advances occur here up in the Taita
Slopes, but the effects are also visible and known in lower locations around Mwatate or Voi.

A few spots stay undisturbed or went through insignificant advances, e.g holy places/woodland
altars (fighis), subsequently, identification and solution of consecrated locales and current
wells/springs could assist with understanding pathways of results (Maeda, 2011). With some of
the fastest population growth rates in the world, Sub-Saharan East Africa is under increasing
pressure to change its land cover. Over the past 30 years, East Africa has experienced fast
economic expansion and environmental change, leading to significant gains in livelihood,
occasionally at the expense of natural ecosystems (Bullock et al., 2021). The vegetation cover
enhances infiltration, lowers runoff, and promotes water retention. Additionally, the type of
vegetation cover can be modified into appropriate configurations to lessen erosion. (Yuksel et al.,
2008). By using the Chawia/Fururu watershed as an assessment region, the current study aimed
to better understand changes in land cover and usage. This is due to the fact that the dispersed
watersheds are important sources of drinking water for the neighborhood as well as for irrigation.
The two watersheds, Chawia and Fururu, which are part of the Ngerenyi watersheds, serve as the
primary water sources for the nearby community. The research is crucial since the watersheds
contain a lot of springs and wetlands. For the purpose of creating a solid environmental planning
and management strategy, precise information on land cover changes and the causes and
processes underlying them is crucial. Consequently, the goal of this study was to identify
historical and current trends in land cover/use changes. (Yuksel et al., 2008).

31
 CHAPTER THREE

MATERIALS AND METHODS

Introduction

This chapter describes the study area; it shows the profile of Chawia and Fururu fragments
within Taita Taveta County. It explains the research design that will be utilized to collect data,
analyze it, and report the results. In this study an integrated use of multispectral satellite data, the
digital elevation model (DEM) and Kenyan topographical sheets were utilized for database
generation and extraction of various drainage parameters. The details of data type, software and
sources used are discussed below:

3.0 Description of the study area

3.1 Study sites and Land Cover situation in Taita Hills

Chawia, Fururu, Mbololo, Ngangao, Vuria, and Susu forest patches are major portions of the
environmental mosaic. Small and degraded forest patches (Fururu) the forest edges and
plantations (Eucalyptus) and agricultural land.

32
Figure 2: Chawia Forest patch and degraded Fururu forest fragments. Source: Petri K. E. 2007

3.1.1 Study site

The Chawia and Fururu forest fragments are among the 12 forest fragments of Taita Hills at
around 2000 m above sea level. The fragments are in Wundanyi around the Ngerenyi area in
Taita Taveta County covering Chawia (86.0 ha) and Fururu (5.0) ha (Mkuzi, 2020). Ngerenyi is
some 181 miles or (291 km) (Chemuku, 2018) southeast of Nairobi, and lies on
38° 20' 62E", 3° 29' 00" S in the verdant and humid Taita Hills, located on the dry Tsavo plains
between the Tsavo East and West National Parks (Wekesa et al., 2020). The hills cover an area
of 1000 km2 and are surrounded by semi-arid Acacia/Commiphora shrubland and dry savanna
(Erdogan et al., 2011). While the surrounding plains are at an elevation of 600–900 masl, the
Taita Hills rise abruptly in a series of ridges with the highest peak of Vuria at 2208 m, although
the average elevation of the hills is 1500 masl (Wekesa et al., 2020).

33
Figure 3: Site map showing Chawia and Fururu fragmented forests within the Taita Hills

(By Author).

3.1.2 Physiographic and Natural Conditions

3.1.3 Physical and Topographic features

Taita Taveta County is classified into three major topographical zones, as follows; Upper zone –
which comprises Mwambirwa, Taita, and Sagalla hills regions with altitudes ranging from 600
meters to two, 2200 meters above sea level (Taita Taveta County Integrated Development Plan
2018-2022 - Devolution Hub, 2018). The zone is suitable for horticultural farming. Lower zone –
which includes plains where the national parks, mines, and ranches are found (County, 2018).
Volcanic foothills zone – which covers the Taveta region with underground water and springs
sourcing from Mt. Kilimanjaro (Taita Taveta County Integrated Development Plan 2018-2022 -
Devolution Hub, 2018.)

34
Figure 4: Taita Taveta County elevation map Taita Taveta County Integrated Development Plan
2018-2022 2018).

3.1.4 Agro-Ecological Zones

The high, medium, and low potential areas of agro ecological zones (AEZ) are the three main
areas that can be recognized. The highlands fall into the high potential category, while the
transitional zone falls into the medium potential category, and the lowlands go into the low
potential category.(Jaeztold et al., n.d.). The Taita hills' lower highland zone (LH2) and upper
midland zone (UM3, UM4) are followed by the lower midland zone (LM4, LM5, LM6) and
lowland zone (LM4, LM5, LM6) (L5, L6) High potential (AEZ 2 and 3), medium potential
(AEZ 4) and low potential (AEZ 5 and 6) can be further classified into high rainfall and low
rainfall sub-zones:

35
The Taita Hills' greatest altitudes are part of a very tiny lower highland zone (LH2) The Taita
Hills are located in the upper midlands (UM3). The valley bottoms have more agricultural
potential and are better suited to growing vegetables. Zones UM4 and LM4 are dominated by
steep slopes. It is too dry to produce maize in the foothill zone, however some early maturity
new sorghum and millet types can do well. The majority of the lowlands are classified as LM5,
LM6, L5, and L6 and are dry for the most of the year(TAITA TAVETA County Integrated
Development Plan 2018-2022 - Devolution Hub, n.d.). Climate and relief determine the
ecological zones in Taita Taveta County. The Taita Hills, which form the highlands, covers
approximately 1,000 km2 and rises to a maximum elevation of 2,208 meters above sea level
(Vuria Peak)(Taita Taveta County Integrated Development Plan 2018-2022 - Devolution Hub,
2018).

3.1.5 Climate conditions

Taita Taveta County is mainly dry, except Taita Hills, which are considerably wet. The
southeasterly winds influence climate in the area, whereby hilly areas have ideal conditions for
moisture condensation, which then results in relief rainfall (Wekesa et al., 2020). The Taita Hills
are the ones who get the most rain. The Taita Hills' high potential zones (LH2 and UM3) receive
more than 900 mm of rainfall per year. (For example, Wundanyi 1300 mm and Wesu 1400 mm).
Temperatures range from 15 to 200 degrees Celsius. With greater temperatures and evaporation,
the medium potential areas receive 700 to 900 mm of rain. There is a net water surplus in the
Taita hills(Taita Taveta CADP, 2018 County, 2018). Rainfall distribution is usually uneven, with
higher rainfall amounts being recorded in highland areas as compared to the lowlands (Wekesa et
al., 2020). Annually, mean rainfall is 650 mm. The average temperature in Taita Taveta County
is 230C, with lows of 180C in hilly areas (Sagalla, Taita, and Mwambirwa) and rising to about
250C in lower zones (Taita Taveta County Integrated Development Plan 2018-2022 - Devolution
Hub, 2018.)

36
Figure 5: View of the Taita Hills (Source: Camera image by Author, 4 September 2021

3.2 Assumptions of the study

The springs to be sampled shall be indicative of the entire number of springs/wells and farmers
in Taita Taveta County, Ngerenyi region, Kidaya Ngerenyi sub-county, as well as their activities.
The respondents will be presumed to be reasonable, and the comments from local guides and
experts will be assumed accurate representations of their experiences. Weather patterns will also
be assumed to be favorable for fieldwork activities during the research.

37
3.3 Methodology

3.3.1 Mapping the physical location of springs in the fragmented Chawia and Fururu
watersheds (Physical mapping using remote sensing of satellite images).

Mapping of springs and wells within the watersheds

Introduction

Data collection involved mapping of springs and wells by talking to the village elders, and
people who might know historic springs and localization with the help of a GPS device. This
helped to produce a precise map of recent and historic springs and wells, which was important
for analyzing spatial distribution to identify former watersheds and to identify pathways of
consequences due to drying up of springs and wells. The prescription of springs, if accessible
with pictures and remarks on floristic elements and hints to former land usage in the direct
neighborhood (special trees, like fig trees).

3.3.2 Watershed Delineattion and Hydrological modelling

The study purposed to use Landsat images with a ten-year sequence between 1987 and 2019. In
this study, the physical mapping of the springs was done with the help of a GPS to localize the
active and inactive springs in the fragmented Taita hills watersheds. The study used remotely
sensed data from Landsat and sentinel-2 in Digital elevetation model (DEM) format to delineate
the Chawia and Fururu watersheds. To analyze vegetaion change, the use of satellite-based
remote sensor data has been widely applied to provide a cost-effective means to develop LC
coverages over large geographic regions. Satellite images such as Sentinel-2A, Landsat 8, 4-
5mss &1-5mss were acquired from U.S. Geological Survey.The Normalized Difference in
Vegetation Index (NDVI) quantifies vegetation by measuring the difference between Near-
Infrared (NIR) which the vegetation strongly reflects and red light (which the vegetation
absorbs/has a low reflectance). NDVI values range from -1 to +1. This index accounts for the
converse reflectance properties of green vegetation in the near infrared (NIR) and red (R) parts
of the solar spectrum and is defined as NDVI =(NIR-RED)/(NIR+RED).Negative values of
NDVI (Normalized Difference Vegetation Index)-1 corresponds to water. Values close to zero (-
0.1 to 0.1) generally corresponds to barren areas of rock, sand, or snow.

38
Table 1: Characteristics of Landsat images and used in monitoring land cover change
dynamics.

Low positive values represent shrubs and grasslands (approximately 0.2 to 0.4), while high
values indicates dense natural/indigineous forest.(0.5 to 1). NDVI calculation formula For
Landsat 8 image,cloud cover 10% =NIR-RED/NIR+RED:Where RED is B4, 636-673nm & NIR
is B5, 851-879nm.This applies for landsat 8 image from 2013 to 2022.using band (B5) and (B4)
For computation of NDVI. B5 is NIR & B4 RED. Landsat 1-5
(NIRB04-REDB02)/(NIRB04+REDB02) NDVI calculation formula For Landsat 5 image,cloud
cover 10% =NIR-RED/NIR+RED. This applies for landsat 7, 5 image from 1970 to 2013. using
band (B4) and (B3) For computation of NDVI. B4 is NIR & B3 RED.

ID Scenes Satellit Altitude Spatial Number Cloud Date

e sensor Resolution of cover
bands level

LM05_L1TP_166062_19870218 MSS 705km 60m 4 Low 1987/02/18

LT04_L1TP_166062_20091113 MSS 705km 60m 4 Low 2009/11/13

S2A_MSIL1C_20170126T073251 MSI 786km 30m 13 Low 2017/01/26

S2A_MSIL1C_20190327T073251 MSI 786km 30m 13 Low 2019/03/27

S2A_MSIL1C_20220209T072619 MSI 786km 30M 13 Low 2022/02/09

3.3.3 Sentinel-2 A

39
By the combination of different optical wavelength of satellite data provides increased
opportunities for more more frequent cloud-free surface observations due to variable cloud cover
as different satellites overpass time and dates(J. Li & Roy, 2017). In this study the Landsat-8
images, sentinel-2A and sentinel-2B together provided a summary of revisit intervals and the
number of observations that are useful in terrestrial monitoring and Assessment.The images
came from sentinel 2, NDVI =(NIR-RED)/(NIR+RED) where RED is B4 664.5nm & NIR is B8
835.1nm.

Table 2:Landsat 8 images has 11 bands

Band Name Bandwith (µm) Resolution (m)

Band 1 coastal 0.43-0.45 30
Band 2 blue 0.45-0.51 30
Band 3 green 0.53-0.59 30
Band 4 red 0.64-0.67 30
Band 5 NIR 0.85-0.88 30
Band 6 SWIR 1 1.57-1.65 30
Band 7 SWIR 2 2.11-2.29 30
Band 8 pan 0.50-0.68 15
Band 9 Cirrus 1.36-1.38 30
Band 10 TIRS 1 10.6-11.19 100
Band 11 TIRS 2 11.5-12.51 100

Table 3: Landsat 7 image has 9 bands

Band Name Bandwith (µm) Resolution (m)

Band 1blue 0.45-0.52 30
Band 2 green 0.52-0.60 30
Band 3 red 0.63-0.69 30
Band 4 NIR 0.77-0.90 30
Band 5 SWIR1 1.55-1.75 30
Band 7 SWIR2 2.09-2.35 30

40
 Band 8 pan 0.52-0.90 15
Band 6 TIR 10.40-12.50 30

3.3.4 Land use, land cover change classes within and around the forested watersheds for
the period 1987-2022

3.3.5 Landuse Classification using Sentinel 2a Satellite image

Table 4: Land use image combinations

Combinations R G B

Natural Colors 4 3 2

False color Infrared 8 4 3

False color Urban 12 11 4

Agriculture 11 8 2

Atmospheric penetration 12 11 8a

Healthy vegetation 8 11 2

Land/Water 8 11 4

Natural Colors with Atmospheric Removal 12 8 3

Shortwave Infrared 12 8 4

Vegetation Analysis 11 8 4

41
3.4 Data Analysis

The physical map of localized springs and their distribution over the fragmented watersheds in
Chawia and Fururu will be generated using ESRI ArcGIS software. To verify the ground truth
and to locate the springs present in the watersheds field visits will be made by the use of GPS to
the location of the historical and recent springs. Analysis and visualization of the status of found
springs.

3.4.1 Satellite Image processing and analysis

In this study, Environmental Systems Research Institute, Inc.(ESRI) Arc GIS Desktop 10.8 was
used to perform the data analysis where supervised classification from Landsat /sentinel_2 was
done. The Image classification and processing, watershed delineation using DEM (SRTM)
prepared Tiff format dataset.

3.4.2 Confusion matrix/Accuracy Assessment

The accuracy of the LULC maps created from the pixel- and object-based classifications was
evaluated. Thus, other more suitable statistical tests were chosen and applied, both for the maps
and the reference samples, in addition to their overall accuracy being assessed based on quantity
disagreement and allocation disagreement indices. Because they can take into account the
association between the appropriateness of test samples, as well as their categories and location,
these indexes were thought to be more appropriate. According to the less-than-ideal match in the
proportions of the classes, quantity disagreement is the value of alteration between reference
samples and a produced map. However, allocation disagreement is the degree of spatial variance
seen among pixels belonging to categorized classes derived from validation samples.The ESRI
accuracy evaluation technique was used to quantify how well pixels were sampled into the
correct land cover classes. The ESRI ArcGIS software were performed this, followed by the
classification tab and the accuracy assessment tool. The ground truth reference data was
collected using a stratified random sampling method. The ground truth points was superimposed
on a map of land usage and land cover, and the value was calculated (Sarkar, 2018). A confusion
matrix was generated and placed such that class membership determined by image classification
is along the y- axis. In this manner, correct values fall along the major diagonal of the matrix
(Sarkar, 2018).One of the pioneers of time series analysis is Richard Denis in forecasting

42
algorithms and the extent of information arranged in chronological order (Ayele et al., 2018).
This is a statistical methodology appropriate for longitudinal research designs that involve single
subjects that are measured repeatedly at regular intervals over time (“Time-Series Study,” 2010).

3.4.3 Time Series Analysis of the Study area satellite images

This refers to the processes of removing distortions and anomalies within the images that might
hinder image classification hence affecting the analysis of the results. In the first stage, satellite
images from Landsat7&8 between 1987 to 2022 were used to analyze trends of deforestation in
the area. The date and month of data acquisition were all during the dry season allowing cloud-
free images (Kaburi & Odera, 2014). As previously said satellite, data is divided into bands and
covers various sceneries on the earth's surface. The image enhancement function were used to
improve the image's visual quality to aid image interpretation by increasing visual distinctions
between the image's features. Algorithms were used to adjust the contrast brightness, picture
sharpness, and color renderings of features in an image. Two methods of enhancements were
applied in this study – brightness, and contrast by using histogram equalization to increase
contrast over the whole image and edge enhancement to bring in contrast in linear features such
as river channels (Kaburi & Odera, 2014).

3.4.5 To identify and delineate the original watersheds in fragmented Chawia, Fururu
forested watershed using Geographical information systems (GIS) and remote sensing.

Introduction

The slope map for the region of the wooded watershed was created using a 30-m DEM that was
collected from Sentinel 2. To keep the flow going to the catchment outlets, sinks and flat sections
were removed using DEM analysis. This objective involves drawing a timeline outlining the
most important events that have had an impact on the water resources or livelihoods of people in
the Taita Hills within a period from 1987 to the present. Furthermore, the good and negative
consequences of these events will be discussed. This activity will aim to understand local
people's perceptions of the changes that have occurred in land use and water resources. People
remember things differently and this exercise will be a way of creating the big picture of the
events that might have led to certain issues. Data collection will use the mapped springs using

43
GIS. Acquisition of remote sensing data from Landsat where the images will be processed and
analyzed.

3.4.6 DEM Preparation and Generation.

The DEM serves as the foundation for defining a watershed, and it can be accomplished using
the ESRI ArcGIS platform's hydrological extension. Many geospatial and hydrological modeling
tasks, including watershed delineation, can be accomplished with the use of a Geographic
Information System (GIS) and a Digital Elevation Model (DEM). This study will explore the use
of contour data and features extracted from topography maps to derive DEM from watershed
delineation in Taita Hills in a GIS environment. The derived DEM will be used to delineate
watershed features using ESRI ArcGIS 10.9 and Arc Hydrology extension (Konadu & Fosu,
2009). The topographic parameters like terrain slope, channel slope, and length will be derived
from the DEM (Kaburi & Odera, 2014).

3.4.7 Filling the DEM

The fill tool will be used to remove imperfections called sinks in a DEM then the filled DEM
will be computed and added to the map, there will be a slight change to the values which is
because the topo to raster tool in Arc GIS is Hydrologically correct DEM (Imran et al., 2019).
O’callaghan and Mark 1984 discovered this extension in the ARC Hydro tool developed by
ESRI and was first used as depression in DEM that are the known sinks generated during data
collection.

Drainage values specify the direction in which water will flow out of a cell and are assigned
when each cell in the grid's flow path is created, therefore the fill stages must be completed
before the flow direction grid is created. The filled DEM will be computed and added to the map
and there is a slight change to the values, which is because the topo to raster tool in Arc GIS
generates a hydrologically correct DEM (Kaburi & Odera, 2014).

3.4.8 Creating a Flow Direction and Accumulation

The function will define the flow direction and destination of water flowing across the earth
surface of the original DEM generated. This is generated from the filled DEM through the GIS
platform using the Flow direction tool in the ARC GIS environment and numbers are generated

44
automatically by the system to represent the flow direction in this case (Kaburi & Odera, 2014).
This is primarily a function in a GIS environment that calculates the flow into each cell by
simply accumulating the cells that flow into each downslope of the cell. The flow accumulation
tool under the hydrology is used and the flow direction raster will be used to input data and the
flow accumulation raster will be computed and added to the DEM (Kaburi & Odera, 2014).

3.4.9 Creating Watershed Pour Points

In addition, a new empty shapefile will be created to contain the pour identified for the upper
Chawia and Fururu watersheds. The Original watershed with all pour points will be stored as the
shapefile name and the recent/active springs as pour points. Then the spatial reference frame for
the new pour points shapefile is set to be the same as those of the existing DEM layer. The pour
points will be added by zooming in until the individual raster cells in the flow accumulation
raster generated previously and edited accordingly (Kaburi & Odera, 2014).

3.5 Watershed Delineation

Introduction

The process of watershed delineation involves sink filling, flow direction determination, flow
accumulation, and spring definition according to Zhang and Huang, (2019). But the parameters
in DEM processing might affect the accuracy of the delineated watershed and spring network
and thus to compare the effectiveness using simulations A and B. Simulation C steps for times
using different threshold integer values to match the original spring network (Z. Li, 2014). The
watershed will be digitally drawn in a GIS context using the Arc toolbox's hydrology extension.
Fill DEM, Create Flow Direction, and Create Flow Accumulation, Create watershed pour sites,
and delineate watersheds will be used to delineate the upper Chawia and around Ngerenyi
campus watersheds using recently active springs.

The flow direction grid will be used as the input flow direction raster when using the watershed
generation tool. Furthermore, the watershed is estimated and constructed by using the raster
version of the pour points as the input raster. The watershed will be polygonized to enable
computation of the area and clipping of flow accumulation to the watershed to generate the
contributing streams. It will be from here that the calculation of the area of the watershed

45
boundary polygon will be done by adding a field to the watershed shapefile and then using the
calculate geometry tool (Kaburi & Odera, 2014).

3.6 Data analysis;

For watershed delineation and to derive features such as springs, stream networks, and catchment
areas, the data will be examined using the Arc GIS Hydrology tool. The size of the watershed
will be assessed after marking off the watersheds boundaries. The scale of the map and
Longitudes and latitudes will be used to calculate the size of the watershed then converted to
acres.

3.6.1 Photography

Photographs of the Mapped springs in the study site and lineages important to the study will be
taken with a digital camera. A notebook and a camera will be used to record visible details.

3.6.2 Observations and Ground truthing

The actions and what is seen in the natural context will be observed and documented using this
method. This will aid in the verification of information provided by respondents and the
understanding of the actual situation on the ground and to verify RS data.

3.7 To determine the categories in land use and land cover LU-LC through spatial Time
series comparison of the LU-LC maps and overall catchment characteristic

Introduction

This objective aims to determine the change in land uses and land cover with the output of
getting Land uses classes and categories in the fragmented Chawia and Fururu forests over the
years.

3.8 Data sources

3.8.1 Satellite Images

To obtain the land unit map and update the springs drainage map of the watersheds, a set of 4
Landsat-7 ETM the images are made up of eight spectral bands, each having a spatial resolution
of 30 meters. Band 8 (panchromatic) has a resolution of 15 meters. In addition, 2 Landsat-5 TM

46
images with Bands 1 to 5 and 7 have a spatial resolution of 30 meters, and Bands 1 to 7 have a
spatial resolution of 30 meters.

Band 6 (thermal infrared) has a spatial resolution of 120 meters, but it is resampled to 30-meter
pixels. will be downloaded from the USGS website and then merged in a mosaic vital in multi-
temporal analysis (Upadhyay & Gupta, 2012). Mapping of land cover changes based on
historical and recent aerial pictures by using Unmanned Aerial Vehicle (UAV) and Geographic
Information Systems (GIS). Land use and land cover change will be assessed using satellite
images with a spectral resolution of 30 meters, which will provide information on the impact of
human demographic pressure on this region over the previous decades.

3.8.2 Image Preparation, Restoration, Rectification, and Enhancement

This refers to the processes of removing distortions and anomalies within the images that might
hinder image classification hence affecting the analysis of the results. In the first stage, satellite
images from Landsat between 1984, 2000, and 2014 will be used to analyze trends of
deforestation in the area. The date and month of data acquisition were all during the dry season
allowing cloud-free images (Kaburi & Odera, 2014). As previously said satellite, data is divided
into bands and covers various sceneries on the earth's surface.

Scenes P169R060, P169R061, P170R060, and P170R061 will be used in this research. The study
area was located at the crossroads of the four sceneries. Each scene was constructed
independently by merging the necessary bands, namely bands 1-Blue band to distinguish soil
from vegetation and deciduous from coniferous vegetation, 2-Green band to emphasize peak
vegetation for assessing plant vigor, 3-Red band to discriminate vegetation slopes, 4-Near
Infrared to emphasize biomass content and shorelines, 5-Short wave band to discriminate
moisture content of soil vegetation; penetrates thick clouds, and 7-Short wave band to show
hydrothermally altered rocks associated with mineral deposits. This was done in a false-color
combination in an ESRI ArcGIS environment via a procedure called layer stacking. The four
scenes that cover the study area were brought or blended to cover the entire region.

This was also done in an ESRI ArcGIS environment using the mosaicking technique, which
combined the four different sceneries into one. Images from the 1990s, 2000s, and 2019 were
subjected to this treatment. The mosaic was re-projected using the ESRI ArcGIS version 10.8

47
environment to comply with the projection of the shapefile of the study region, i.e. UTM zone 36
south (projection), Clarke 1880 (spheroid), and a Datum of Arc 1960 (Kenya), from which the
study area was a subset (Price, 2011).
These processes were used to the image to compensate for data degradation within the image by
removing systematic distortions and altering the shape of the image to fit the research area's base
map. The importance of image restoration is that it improves scene illumination, bringing out the
differences in locations within the study area. It helps to deal with challenges such as shadows,
and atmospheric conditions such as haze (Kaburi & Odera, 2014).
By predicting a replacement value based on the nearby pixel, an image correction approach was
used to the image to approximate the look of the aforementioned faults. Two methods correct for
atmospheric challenges – Physical modeling and empirical approximation. To deal with
distortions coming from differences in altitude, the velocity of the sensor platform, and sensor
characteristics, image rectification was performed by performing geometric correction by geo-
referencing the image.
Geometric adjustments can be made in two ways: using a mathematical transformation to
calculate geographic coordinates, and using a resampling technique to synchronize data from
multiple satellites. Because it is simple, fast, and does not modify the original pixel value, the
mathematical technique was employed in this study with the nearest neighbor as the resampling
method.
The image enhancement function was used to improve the image's visual quality to aid image
interpretation by increasing visual distinctions between the image's features. Algorithms were
used to adjust the contrast brightness, picture sharpness, and color renderings of features in an
image. Two methods of enhancements were applied in this study – brightness, and contrast by
using histogram equalization to increase contrast over the whole image and edge enhancement to
bring in contrast in linear features such as river channels (Kaburi & Odera, 2014).

3.8.3 Image Processing and classification

The surface is covered by vegetation like forests and agriculture trap crops and holds the water in
the root of plants whereas the built-up and rocky land use affects the recharge of groundwater by
increasing runoff during the rain, so it is necessary to check what quiet features are covered the
study area’s land surface. The Landsat -7 satellite image were used for the study to seek out the

48
land use and land cover of the study area and the supervised classification method was used with
the maximum likelihood for discussion purposes (Deribe & Debalke, 2021). The classification
was done using a supervised classification technique, which necessitated the selection of training
sites and the Area of Interest (AOI) tool was used for this.

Forestland, grassland, farmland, and water bodies all received training data. The classification
was carried out using the maximum likelihood approach once enough training sites have been
selected (more than 600 pixels for each class). Iterations were used to examine the product until
pixel hygiene was attained. The attribute table was filled up, and areas were calculated and saved
in a manner that most remote sensing software can understand. To determine whether areas have
changed, change detection was done in the Arc Map 10.8 environment. Between 2000 and 2010,
1990 and 2013, and 2010 to 2019 (Price, 2011).

3.8.4 GIS image processing tool.

Depending on the Satellite imagery information available ESRI ArcGIS program was used to
analyze images and track changes in land use and land cover in Chawia and Fururu forest
fragments over time (Rasuly et al., 2010). To simulate the rate of land cover changes and map
out endangered regions in the study areas, three key phases, pre-processing, processing, and
post-processing techniques, were used. To fulfill the study objective, the digital image
processing software ArcGIS was used for pre-processing to rectify atmospheric, radiometric, and
geometric flaws found in satellite imageries (Rasuly et al., 2010), and sub setting /merging and
filling of images were introduced to prepare satellite data for the next step. In the processing
step, ArcGIS was used to generate the false-color composite, by combining near-infrared, red,
and green which are bands 4,3,2 together for all images and this was done for vegetation
recognition because the chlorophyll in plants reflects very well to near-infrared than the visible
area (Rasuly et al., 2010).

Forest, agricultural, grassland, and water body classifications were defined and classified for all
land uses. Ground control points (GCP) were plotted on photographs utilizing a global
positioning system (GARMIN Colorado GPS) from places with the study area's classes. This
was utilized to verify the training sites (specified classes) in terms of spectral signature, and then
a supervised classification approach was used to create ultimate land use maps using the
maximum likelihood algorithm. (Rasuly et al., 2010). In the post-processing step, Final

49
vectorized classified maps were entered into the Arc GIS setting then a regression model was
applied to model relations between independent variables such as distance to settlements, aspect,
slope, rainfall, and elevation with land cover changes as (dependent variable) were defined based
on observations.

By specifying and calibrating the final model, the unknown value of deforestation was calculated
and predicted based on some related recognized environmental factors and a logistic regression
model was introduced to find the probability of membership in two contradictory classes such as
forest area being destroyed (Rasuly et al., 2010).

50
 CHAPTER FOUR

RESULTS & DISCUSSION

3.9 To map the physical location of springs and current status springs in the fragmented
Chawia and Fururu watersheds

Introduction
In this study, the physical mapping produced Ninety one (91) springs were visualized in Chawia
and Fururu forest ecosystems, , Fifty three (53) springs were in Chawia whereby thirteen (13)
were found to be inactive and thirty (30) were still active in Chawia Compared to Thirty eight
(38) springs in Fururu which were found to be all active respectively, which were mapped
within and around the two forest fragments as shown in (figure 6 &7) below.

51
Figure 6:Showing physical location of the mapped springs in Chawia forest & Figure
7:Showing physical locatikon of the mapped springs in Fururu Forest.

4.0 To identify and delineate the original watersheds in the fragmented Chawia, Fururu
forested watersheds using GIS and RS.

Introduction

The watersheds of the study site were defined using Arc Hydro tools included in Arc GIS, and
the results were mapped in Figure 5. The entire process of the study adhered to the methodology
flow described in 3.3.2 above. Although remote sensing is a potent data source and is frequently
used to monitor environmental hazards, groundwater cannot be directly measured using remote
sensing since it does not exist on the surface. As a result, numerous studies have tried to
minimize the spatial uncertainty of groundwater. The former watersheds were also delineated in
Chawia and Fururu fragments and the Land cover and land uses were determined as well. In this
study the two forested watersheds that form part of the Taita hills forest patches and the study

52
found that Fururu gained at least 10% between 1987 and 2022 as shown on Figure 2 below.
Chawia lost at least 10% between 1987 and 2022 as shown in (figure 3) below . From this study,
eight (8) Sub-watersheds were delineated and thirty seven (37) micro-catchments were also
delineated.

Figure 8:showing delineated watersheds in Chawia & Figure 9: Showing delineated watersheds
in Fururu

4.1 To determine the Land uses and Land cover classes through spatial Time series
comparison of the overall catchment characteristics

4.1.1 Land Use Land Cover Analysis from 1987 to 2022.

Groundwater recharge may be influenced by land use and land cover. For instance, places
covered in forest and waterbodies have a high likelihood of groundwater recharge while areas
with bare land and built-up watersheds have high runoff and low groundwater recharge.
Cropland contributes more to groundwater recharge than rocky, barren, and built-up areas by
permitting a higher rate of infiltration. (Guduru & Jilo, 2022). Forest, agriculture, bare land,
including residential areas, and waterbodies are the main LULC in the Chawia and Fururu

53
watershed. These LULC have been categorized as indicated in Fig. 7 based on the features that
are associated to them when recharged.

The Forested watersheds' majority is made up of cultivated land Changes in land use and cover
patterns play a significant role in determining the state of the water resources. Due to climate
change and land use practices with high water requirements, water resources are severely
strained (Gebre et al., 2015). Land use pattern changes and their estimation describe the
utilization of land resource by manmade activities, particularly agriculture and urbanization.
Hydrological inferences from land use patterns can help to understand the changing scenario of
water demand

In this study, Four different Land cover classes were determined both within and around the two
forested watersheds fragments for the period 1987-2022. In Chawia, moderate forest had the
highest acreage in terms of cover, which showed a negative decrease from 191 (ha) which is
(26.09%) to 125 (ha) (17.08%) between 1987 and 2022 respectively as in table 6. Grasslands
increased within the forest from 30 (ha) which is (9.93%) to 98 (ha) (32.45%). There is evidence
from the LC-LU maps that Chawia land cover decreased from 1987 to 2022 while Fururu land
cover maps showing great increase from 1987 to 2022 as in table 6.The maps show vegetation
change and different forms of vegetation types. LULC is dynamic in nature and its dynamism
affects several parameters. It is hence observed that surface temperatures are unique
characteristics of every LULC classes. So, it was possible to make much more accurate
classification for LULC when the parameter of surface temperature was considered (Sinha et al.,
2015).

Fig. 4 shows the trend of the Surface Temperature (ST) profile for LULC classes. Open forest,
dense forest and forest blank had narrow range of ST that could not be easily distinguished for
classification based on ST profiles. On the other hand, settlement and water body had similar ST
range in Figure 7. This was purely due to the fact that the settlement areas were un-uniformly
randomly distributed in pockets that are near or within the forests.

The changes in Chawia could have been because of Forest encroachment or even human
disturbance (Agriculture, Livestock) as shown in figure 4& 5 below.From this study, Five
different land use types were determined as; Dense Forest, Moderate forest, Grasslands, and

54
Shrubs in Chawia forested watershed and Dense Forest, Moderate forest, Grasslands, and Shrubs
in Fururu forested watershed respectively as shown in Table 6 below.

The Land use types determined were found to be impacting the forested watershed and thus
reducing the number of springs in the study area which are essential for human use and
ecosystem regulation therefore the change in Land use type has caused these springs to dry up
and that is the reason why the Chawia forested watershed is degraded and has lost a lot of cover
whereas Springs found in Fururu forested watershed were found to be active this is due to the
fact that the watershed is conserved and the forested area is still intact.

The use of remotely sensed data has made it feasible to examine the changes in land use and
cover more quickly, cheaply, and accurately. GIS offers an appropriate framework for the study,
updating, and retrieval of geospatial data. GIS and remote sensing technology offer effective
techniques for analyzing changes in land use and cover. In order to properly manage and utilize
resources, relationships and interactions between human and environmental phenomena must be
better understood through timely and accurate land cover change detection. (Sewnet, 2016).

Chawia phenological Land cover From 1987 to 2022

35.00
32.12 32.45

30.00
Area in percentage

26.09
25.00 22.95 23.46
21.72 22.03
21.04
20.11
20.00 19.22 18.72 18.58 18.59 18.44 19.13
17.08

15.00 12.91 12.84

12.58
9.93
10.00

5.00

0.00
Area in %1987 Area in %2009 Area in %2017 Area in %2019 Area in %2022

years
Dense Forest Moderate forest Grasslands Shrubs

Figure 10: Showing the Phenological Land cover changes in Chawia from 1987-2022

55
 Fururu phenological land cover changes since 1987 to 2022
35 33.33
29.82 30.56
30
26.32 25.93
25 24.07
22.66
Area in percentage

22.22
21.88
18.52 19.53 18.75
20 17.19 17.54
15.74
15 14.04
12.28
11.11 11.11
10 7.41
5

0
Area in %1987 Area in %2009 Area in %2017 Area in %2019 Area in %2022
Years

Dense Forest Moderate forest Grasslands Shrubs

Figure 11:Showing the Phenological Lnad cover Changes in Furur Forest from 1987-2022

Table 5:Land cover change (Ha) from 1987 to 2022

Land use type Area in Area in Area in Area in Area in

Chawia (ha) 1987 (ha) 2009 (ha)2017 (ha) 2019 (ha) 2022
Dense Forest 123 139 141 119 118
Moderate 191 168 154 94 125
forest
Grasslands 30 38 39 97 98
Shrubs 134 133 144 168 137
Fururu
Dense Forest 17 10 15 7 8
Moderate 22 28 29 25 24
forest
Grasslands 9 6 3 7 2
Shrubs 17 12 20 26 33

56
4.1.2 Percentage of change detection in Land use Land cover from 1987-2022

90% % Phenological change in Vegetation Index From 1987-2022

80%
Highest NDVI value for each year in %
70%

60%

50%

40%

30%

20%

10%

0%
1987s 2009s 2017s 2019s 2022s
Vegetation Index (NDVI)
Chawia Forest Fururu Forest

Figure 12: Showing difference in vegetation Index (NVDI) in percentages

The figure 6 above shows Highest (NDVI) value for each year in percentages. The value range
of the NDVI is -1to1. Negative values of NDVI (values approaching -1) correspond to water.
Values close to zero (-0.1 to 0.1) generally correspond to barren areas of rock, sand, or snow.
Low, positive values represent shrub and grassland (approximately 0.2 to 0.4), while high values
indicate temperate and tropical rainforests (values approaching 1).

57
 Landuse status Chawia-Fururu Forest Ecosystem in
Percentages
Grasslands Agriculture
Waterbodies 0.04%
0.23% 7.13%

Forest land Barelands

26.29% 0.01%

Built-up
Areas
66.3%

Agriculture Barelands Built-up Areas Forest land Grasslands Waterbodies

Figure 13: Showing Land use status in chawia and Fururu forested ecosystems

58
Generally, the study indicate that the indigenous forest in the Taita Hills decreased by 66.3% due
to the establishment of exotic plantations and more built up areas. Significant land use changes
have taken place in Chawia and Fururu forested watersheds between 1987 and 2022 showing a
relative matrix change within each land use/land cover class for different time scenes as shown
in Table 6 above where the most changes occurred large with increasing built up areas, forest
land and Agricultural land. From this study, the changes are attributed to large scale run off and
erosion events that may be due to the decreased infiltration capacities associated with LU-LC
changes. Chawia forested watershed experienced significant LU-LC changes between 2019 and
2022 while Fururu forested watershed experienced significant changes between 2009 and 2019
with the dominant changes happening within the dense forest and Grasslands as shown in figure
4 and 5 respectively.

Figure 14: Map showing Chawia-Fururu Land use in 2022.

Trends in forest cover change for Dense forest, Moderate forest, grassland, and shrubs forest
areas varied among the forested watersheds (Table 6). The Dense forest area in Chawia

59
decreased between 2009 and 2022 while the moderate forest decreased between the year 2009
and 2019 before increasing in 2022 again. In Fururu, the dense forest area decreased between
1987 and 2009 before increasing in 2017; but decreased again in 2019 and 2022. The moderate
forest area increased between 1987 and 2017 but reduced in 2019 and 2022 as shown in (Tabel 6
and Figure 4 and 5) respectively.

4.1.3 Land Cover Analysis from 1987 to 2022.

Figure 15:Showing Chawia Land cover in 2022 & Figure 16:Showing Land cover in 2017

60
Figure 17:Showing land cover in 2019 & Figure 18:Showing land cover in 2009

Figure 19:Showing Land cover in Fururu in 1987

61
Figure 20:Showing Fururu Land cover in 2009 & Figure 21:Showing Fururu Land cover in
2017

Figure 22:Showing Fururu land cover in 2019 & Figure 23:Showing Fururu land cover in 2022

62
4.2 Accuracy summary

As in table 8, the overall accuracy of the data is 80.90% as shown below

Table 6: Accuracy summary

Accuracy summary
Class name Reference Classified Number producer users
s
Forestland 71 70 55 84.0% 86.3%
Grassland 21 19 17 90.96% 92.29%
Agriculture 16 19 15 96.4% 95.37%
Built-up 2 2 2 100% 100%
Totals 110 110 89

Overall Classification Accuracy = 80.90%

4.2.1 Error matrix

Table 7: Error matrix

Reference
Classified Forestland Grassland Agriculture Built-up Total
Forestland 55 14 1 0 70
Grassland 12 7 0 0 19
Agriculture 4 0 15 0 19
Built-up 0 0 0 2 2
Totals 71 21 16 2 110

4.3 Discussion

63
This study used ArcGIS software, an ASTER DEM, and satellite imagery to determine the
significance of watershed features for managing water resources in the Chawia-Fururu wooded
watershed in Taita Taveta, Kenya. Due to the implications on water supplies and the
environment, the hydrological response to human activity-induced land cover changes in dry
places has received more academic attention in recent years. (Mahmoud & Alazba, 2015).
Environmental issues brought on by the construction of dams and a reduction in forest area,
which led to the drying up of the majority of mountain springs, were taken into consideration
while choosing the research period and location for this study.

In the global scale springs are the major source of water in the mountainous and hilly areas of the
earth surface and are important to terrestrial ecosystems but the springs are drying up at a rapid
rate thus threatening human throughout the world (Magazine, 2020.). The Taita Hills' ability to
supply water has been lessened as a result of the cutting down of indigenous trees, which has
decreased the catchments' ability to retain water. Over the past 50 years, the water levels in
numerous springs, rivers, and streams have plummeted, and some have almost entirely dried up.
Older residents claim that this progressive decrease in water levels peaked in the 1980s and
1990s.

Formerly, the Chawia/Fururu forested watershed, in particular the areas are in the steep upper
zones, used to remain rainy all year long (J. Hohenthal et al., 2015b). Older residents claimed
that while they used to go fishing in the Taita Hills' springs and rivers as children, there are no
fish there now due to the degraded water quality and quantity.

According to the study done by Madhav Dhakal at Nepal argues that in the current situation of
springs in the middle hills and are drying up or have already dried up and the study show that the
problem is becoming more severe and that is why the communities are facing severe water
shortages and thus bringing water stress in our society (Thapa, 2021). In this study, it was found
out that most of the springs have dried up and thus the local communities cannot access the
terrestrial water supply.

Ninety one (91) springs were mapped and visualized in Chawia and Fururu forest ecosystems,
where the watersheds of Chawia and Fururu were delineated and the hydrological, stream
network and flow direction were developed based on the DEM method used. Due to
overexploitation and overuse of natural resources and environmental degradation is putting a lot

64
of pressure on water resources (J. Hohenthal et al., 2015b). According to the study by Hohenthal
et al. (2015) indigenous forest cover has been reducing drastically since the 1960s due to
population increase, expansion of intensive agriculture, plantation of exotic tree species, and
road construction. According to the authors (Madhav Dhakal et al,. 2020), they argue that the
reason for the drying up of the springs can be attributed to three major factors; anthropogenic
factors; Climate change, and seisimic events and Infrastructure development, such as building
highways, hydropower tunnels, and cement irrigation canals, as well as land-use changes like
deforestation and overusing natural resources, are examples of anthropogenic activities. Water
scarcity is also a result of unplanned urbanization and population growth, which increases
consumption and increases reliance on springs and other water sources.

The rural Taita hills mountain watersheds have severe water shortages, particularly during dry
spells and years with low flow. People in all of the chosen watersheds have the same issues.
Existing springs have been losing water, some perennial springs have become seasonal, and the
seasonal springs have totally dried up (Adhikari et al., 2020). Since not all villages have access
to an enhanced water delivery system, springs are the only consistent and trustworthy source of
water supply in the villages.

In the past, springs supplied enough water to meet their basic needs but the locals noticed that
there wasn't as much rain as there used to be two to three decades before. However, there has
been a water deficit in the villages in recent years. Water production from their springs is
insufficient for their daily needs. Population pressure, altered land use, infrastructure expansion,
rising water consumption in tandem with the adoption of modern technology, and the detrimental
effects of climate change are some of the root causes of water shortage (Adhikari et al., 2020).
When hydrologic changes are put on watersheds, they react as an interconnected system. Most
regional, governmental, and international environmental problems have some connection to land
use or land use change. These land use concerns range from small-scale agriculture to
widespread deforestation, and many of them can become highly contentious politically and
emotionally during public discussion. It is crucial that we specify the scientific questions
involved in each issue and then gather the necessary data to answer the questions in order to
provide decisionmakers with unambiguous scientific recommendations. Research on watersheds
offers a clear framework for gathering such information. When it comes to forest management,

65
satellite imagery has almost exclusively been used by GIS specialists or remote sensing teams,
who have the expertise required to manage and manipulate various types of geospatial data. The
emergence of new forest management solutions is creating new use cases for satellite imagery,
addressing previously unanticipated needs, and ultimately driving value outside of GIS
departments (Forestry Management | Satellite Monitoring | Planet | Webinar, 2019.). In an
investigation by (Uddin et al., 2015) to evaluate and keep track of natural resources, particularly
forests, in the small Lorpa watershed in the Jumla district of Nepal. The findings show that tree
cutting, particularly of larger, older trees, is still an issue in the watershed.

The Copernicus SENTINEL-2 mission consists of a constellation of two polar-orbiting satellites

that are phased at 180 degrees from one another and placed in the same sun-synchronous orbit.
Its huge sweep width (290 km) and long revisit time (10 days at the equator with one satellite
and 5 days with two satellites in cloud-free circumstances, resulting in 2-3 days at mid-latitudes)
which enable monitoring of changes to the Earth's surface (Sentinel-2 - Missions - Sentinel
Online - Sentinel Online, 2020.). In this study, the aim was to test Sentinel-2A/B images
suitability for detecting small water bodies (springs) and wells in taita hills, and in this study
medium spatial resolution of Sentinel-2A/B photos were taken on three distinct dates of four
representative Chawia and Fururu watersheds for this purpose. For each date and watershed, 60
points (30 in water regions and 30 in land areas) were randomly placed 20 meters apart from the
edge of each digitized water polygon to verify the findings (600 in total). According to a study
thet was done by (Patle, 2020) in the Nahra nala watershed using SENTINEL-2B image mapping
using this successful test produced an overall accuracy of 95.52% and a Kappa coefficient (k) of
0.92. Due to its superior spectral resolution in the Near Infra Red (NIR) region and precise
spatial resolution, SENTINEL-2B data has found wide-ranging applications in the fields of
study. These applications include categorization, mapping, and monitoring. When examining
small-scale locations like watersheds, villages, and blocks, policymakers and planners rely
heavily on the SENTINEL-2B satellite image (Patle, 2020). The authors Georeferenced and
delineated the Nahra nala watershed as well.Land use in Chawia and Fururu has been swamped
by increased horticulture and brushing as a result of the rapid growth of the population, and this
has forced the neighboring local area to climb the slopes, creating actual erosion, soil
disintegration, and lowering water tables in the slopes. Consequently, the incorporation of GIS
and RS techniques advances watershed depiction and the land cover characterization from DEM

66
information and from Landsat. The use of GIS and RS in concentrating on the Chawia and
Fururu watersheds helped in information securing on the hydrological foundation of the two
watersheds, group the land uses, and land cover changes in the catchments.

In a study that was done by Jasrotia et al., (2016) on the integrated remote sensing and GIS
approach for delineation of groundwater potential zones in Devak and Rui watersheds of Jammu
and Kashmir, India. Remote sensing, GIS, and Fieldwork techniques were combined and used to
delineate the groundwater sources map of Devak and Rui watersheds and the results from the
study revealed that the excellent zone covers 13.5%, good zone 22.7%, moderate zone 15.8%,
low zone 18.5% and run off zone 29.5% of the total study area. The results depicted the
groundwater potential zone found to be helpful in better planning and management of
groundwater resources in the study area(Jasrotia et al., 2016). In this study the Chawia and
Fururu forested watersheds were delineated from data acquired from sentinel_2 and landsat8
satellites.

Globally, the majority of people on earth live in cities. Currently, it is estimated that 5 billion
people live in cities. The urban population is predicted to exceed 6.5 billion within a decade.
Terrestrial ecosystems will be impacted by urban growth, which will result in higher demands
for the services they supply.

The United Nations (UN) highlights the value of ecosystems and environmental services to
people as natural products and services, on which social progress and economic growth directly
depend, and which are currently disappearing globally at an alarming rate (Abad-Segura et al.,
2020).

LU-LC Classification was performed.Although species-level classification was possible, 10

optimum LC classes were mapped and contrasted in this study without affecting the precision of
each class. Rangeland predominates throughout the study region; there are just a few patches of
forest and agricultural land. Water, fuelwood, and other forest resources are in short supply, and
the three nearby communities' rising demand further reduces their availability. The watershed is
progressively being used at unsustainable levels for agricultural production, while overgrazing
and overharvesting of fuelwood and lumber are contributing to the destruction of the
environment(Uddin et al., 2015). Therefore in this study, it was similarly found that the Four LC
classes were determined and Shrubland seemed to dominate the study area of Chawia watershed

67
and there just a few patches of the dense forest in Chawia and Fururu as well but had an increase
of the grassland forest.

As a result, therefore, the utilization of remotely sensed data and the integration of GIS and their
techniques provide accurate and timely information for detecting and monitoring the land cover
and land use. In this study, the classification accuracy was 80.90% where the study areas were
classified into six-land use and Land cover classes namely; Built up, Agriculture, Bare land,
Water bodies, Grassland and forest land as defined in the classification scheme. The results also
showed that Chawia land cover has reduced where for 1987 the dense forest was (123ha)
(19.22%), 2009 (139ha) (21.72%), 2017 (141ha) (22.03%), 2019 (119ha) (18.59%) and 2022
(118ha) (18.44%) whereas Fururu Dense forest decreased from 17ha (29.82%) in 1987, 10ha
(17.54%) in 2009, 15ha (26.32%) in 2017, 7ha (12.28%) in 2019 and 8 ha (14.04%) in 2022.
Several recent studies have evaluated the consequences of such change on hydrology in various
parts of the world, and research has demonstrated that land cover classes can be used to produce
model inputs to analyze the effects of land cover change on watershed responses. (Mahmoud &
Alazba, 2015). For instance, one study measured the effects of major stream flow changes
brought on by LULC modification in the Laohahe Basin in China. The observed stream flow was
shown to have been strongly influenced by LULC changes, dams, irrigation from rivers,
industries, livestock, and human consumption. Using the Soil and Water Assessment Tool, a
similar study examined the impacts of land cover and climate change on the Gilgel Abay River
in Ethiopia's Upper Blue Nile Basin. It found that both had minimal effects on the river's low-
flow conditions.(Mahmoud & Alazba, 2015).
Therefore from the study done in Taita hills, Chawia lost 10% of the forest cover whereby
Globally, over the decades since 2010 the net loss of forest was 4.7 million hactares per year
much higher rates significantly (Ritchie & Roser, 2021). According to Ritchie & Roser, the
earth’s cover has changed over the past 10,000 years shortly after the end of the great ice up to
the present. The loss of 10% cover for Chawia may be as a result of change in the land use
activities where most people are encroaching towards the forest and participating more on
Agricultural activities that has made the forested ecosystem lose its cover wheras Fururu from
this study gained a cover by 10% due to more conservation efforts done in preservation of the
ecosystem unlike Chawia. The findings indicate that during the past 20 years, a significant
amount of the watershed's natural vegetation has been converted into farmland and populated

68
areas. Within the research area, an increase in urbanization is correlated with a decline in bushes
(Ruttoh et al., 2022). In the midst of the Chawia forested watershed, loss of land cover is mostly
caused by overgrazing and intensive cultivation on marginal soils. Some of the watershed's most
productive agricultural land is being converted into residential and commercial space due to the
area's rapid population growth and high demand for land for these uses. Although there are
policy initiatives to prevent this loss of production, their efficacy in the face of economic demand
is frequently constrained. It is yet unclear if these and other governmental initiatives to lessen
LULC's detrimental effects will be successful (Mainuri et al., 2022).

Similarly, according to the study that was done by (Winkler et al., 2021) According to there
estimates, between 1960 and 2019, at least 17% of the Earth's land surface moved between the
six land types. According to their estimates, between 1960 and 2019, at least 17% of the Earth's
land surface moved between the six land types. The total land change extent, which accounts for
all individual change events (as well as areas of multiple change), is 43 million km2, or almost a
third of the world's land area. This indicates that, since 1960, an area of land that is roughly twice
the size of Germany (720,000 km2) has changed annually on average and it has been most
evident in Africa and South America, regions of the tropics and sub tropics (Winkler et al.,
2021). The authors found out that 0.8 million km2 net loss of forest area globally, but an
increase in world agriculture. Therefore in this Study, Land use changes were notice to have
shifted from Forested land to Agriculture and built up areas and the land use change has not been
constant whatsoever.According to Winkler et al., 2021, they argue that in addition to globalized
trade, climate change and its effects, such as extreme occurrences, droughts, and floods, are other
significant drivers of land change dynamics that have increasingly affected the rate of land use
change throughout the deceleration phase. Droughts in West and Eastern Africa have an impact
on the utilization of agricultural land which in most parts of kenya has affected Agriculture and
vegetation due to the rise in demand for land and poverty levels in taita hills has made the local
people to change their land use types which as well brings about vegetation change and index .

69
 CHAPTER FIVE

SUMMARY, CONCLUSION AND RECOMMENDATIONS

4.4 Summary and Conclusions

Effective water management and planning are required given the current worldwide water
deficit. The constant and reliable supply of water might be hampered by careless water resource
management. Accurately predicting and responding to the current status of crucial resources is
the first step in correctly planning the use of water resources. Particularly in areas with limited
water supplies, groundwater is a great source of water. The investigation reveals that the general
environmental setup of the Fururu watershed has improved as a result of good farming practices

70
and the restoration of wastelands. The Forested Chawia Watershed, which must be protected
from further deterioration, may benefit from several actions including planting, contour bunding,
and the installation of check dams and percolation ponds. Land use in Chawia and Fururu has
been swamped by increased horticulture and brushing as a result of the rapid growth of the
population, and this has forced the neighboring local area to climb the slopes, creating actual
erosion, soil disintegration, and lowering water tables in the slopes. Consequently, the
incorporation of GIS and RS techniques advances watershed depiction and the land cover
characterization from DEM information and from Landsat. The use of GIS and RS in
concentrating on the Chawia and Fururu watersheds helped in information securing on the
hydrological foundation of the two watersheds, group the land uses, and land cover changes in
the catchments.

As a result, therefore, the utilization of remotely sensed data and the integration of GIS and their
techniques provide accurate and timely information for detecting and monitoring the land cover
and land use. In this study, the classification accuracy was 80.90% where the study areas were
classified into six-land use and Land cover classes namely; Built up, Agriculture, Bare land,
Water bodies, Grassland and forest land as defined in the classification scheme. We must
continue to monitor and mitigate the harmful effects of LULC in order to maintain the existing
and future human populations in a healthy state. This calls for increased efforts and novel
approaches. The implementation of programs like reforestation of natural forests, required
homestead tree planting, controlled tree harvesting, and restrictions on forest encroachment can
achieve this.

Ninety one (91) springs were mapped and visualized in Chawia and Fururu forest ecosystems,
where the watersheds of Chawia and Fururu were delineated and the hydrological, stream
network and flow direction were developed based on the DEM method that was used. The results
also showed that Chawia land cover has reduced where for 1987 the dense forest was (123ha)
(19.22%), 2009 (139ha) (21.72%), 2017 (141ha) (22.03%), 2019 (119ha) (18.59%) and 2022
(118ha) (18.44%) whereas Fururu Dense forest decreased from 17ha (29.82%) in 1987, 10ha
(17.54%) in 2009, 15ha (26.32%) in 2017, 7ha (12.28%) in 2019 and 8 ha (14.04%) in 2022. The
findings reveal distinct patterns in the intensification of land use in the area: cropland growth is

71
encroaching on formerly undeveloped dense forest and woods, endangering vital ecosystems
noted for their biodiversity and leading to net decreases in woody vegetation.

From this study, there is need for urgent efforts to be done for Chawia forested watershed
conservation to avoid its degradation and drying up of most springs. And the conservation efforts
should be maintained for Fururu forested Watershed in order to prevent drying up of the springs
and reduction of the watershed size. The monitoring and analysis of land cover patterns as well
as the assessment of the effects of land use change for upcoming land development projects by
the inhabitants of the study areas were determined to be successful outcomes of the combination
of RS and GIS. As from the study results it is recommended that the Land Use Act should be
Enacted and enforced to put more conservation strategies to the forested areas of Chawia and
more conservation efforts be done on Fururu forested watershed. The study also recommends
that due to the higher rate of drying up springs the Kenyan government and other stakeholders
should put more plans and startegies for the construction of more dams to act as water reservoirs
to store water for irrigation and domestic use due to the increasing number of population in the
forested areas of Taita hills. Sharing knowledge gained from experiences In order to benefit
watershed management users/new projects from experiences learned and to prevent duplication
of efforts, sharing these results and identifying appropriate mechanisms for disseminating such
information are important issues. Integrated Watershed Management (IWM) approaches and
methodologies have been successfully implemented in various parts of the world. Making use of
the proper participatory procedures. Participatory procedures are valued as primary at all phases
of IWM, and experience with them is crucial. One-sided bottom-up or top-down techniques do
not work, as demonstrated by experience. Different strategies and tactics should be judiciously
applied and modified in accordance with unique situations. Containing repeatable and
sustainable actions. To continue, it is necessary to enlist the assistance of all relevant agencies,
groups, and IWM officials. Strong financial and technical support is necessary. Examining and
establishing institutional, organizational, and legal frameworks, including interagency
cooperation and decentralization of authority.

For the sustainability and upgradation of excellent practices, there is an urgent need for a
watershed-level springs restoration program with a participatory approach focused on
local communities, local government and academia, NGO, and other stakeholders.

72
 Increasing the ability of local governments, communities, and other stakeholders to
comprehend the various benefits of springs and the preservation of watersheds.
Before building any new infrastructure, local government and communities should take
the values of watersheds into account.
The requirement to create and integrate a watershed management plan into local
government development plans.
Traditional lakes, ponds, and wallows provide excellent ecological benefits by revolving
the springs downslope, among other things. Therefore, safeguarding these lakes, ponds,
and marshes is essential.
Harvesting rainwater in homes and communities.
Knowledge dissemination through local media

In conclusion, this study has shown the value of multi-temporal Landsat and Sentinel-2 satellite
images in monitoring land-cover change and watershed delineation processes, pinpointing areas
in need of restoration, and assessing restoration tactics for managing tropical watersheds through
hydrologic modeling. Although the methods used in this study were applied to relatively small
watersheds, they can also be applied to large watersheds and river basins as long as Landsat and
Sentinel images are available to derive the land-cover information needed for detecting and
locating the changes as well as for hydrologic modeling. The techniques used in this work can be
easily implemented for watershed land-cover change monitoring, management, and rehabilitation
given that Landsat photos taken since 1987 are publicly available online. In order to classify
sentinel images, this study also used the provided Classification. The Supervised Classification's
expectations that class signatures will be regularly distributed and their high degree of
separability are both satisfied, which contributes significantly to the accuracy of the land-cover
maps that result from the classifications. The MLC may not always be able to produce the best
results because there may not be enough training samples to generate sufficient class signatures
or because they may not have normal distributions. Other classifiers like DT, NN, and SVM can
be used to alleviate this issue but at the expense of longer calculation times. The approach
utilized in this study is transferable to other contexts where groundwater utilization is important.
In order to optimize the advantages of management and more research on water-provisioning
systems and water availability (surface water), managers and policymakers can effectively
examine groundwater potential modeling results.

73
Data Availability

The data and satellite images and data used in this study is readily available from the
corresponding author upon reasonable request.

Conflicts of Interest

The authors declare no conflicts of interest regarding the publication of this paper.

Acknowledgments

The authors strongly acknowledge the funding of (Deutscher Akademischer Austauschdienst) /

German Academic Exchange Service (DAAD) during this study.

Ethical statement

This study was carried out in accordance with the recommendations of the Pwani university
Ethics Review Committee(PUERC) with written informed consent from all subjects. All subjects
gave written informed consent in accordance with the declaration of Pwani university committee.

4.5 References

Abad-Segura, E., González-Zamar, M.-D., Vázquez-Cano, E., & López-Meneses, E. (2020).

Remote Sensing Applied in Forest Management to Optimize Ecosystem Services:
Advances in Research. Forests, 11(9), 969. https://doi.org/10.3390/f11090969
Adhikari, S., Gurung, A., Chauhan, R., Rijal, D., Dongol, B. S., Aryal, D., & Talchabhadel, R.
(2020). Status of springs in mountain watershed of western Nepal. Water Policy, 23(1),
142–156. https://doi.org/10.2166/wp.2020.187
Alqurashi, A. F., & Kumar, L. (2013). Investigating the Use of Remote Sensing and GIS
Techniques to Detect Land Use and Land Cover Change: A Review. Advances in Remote
Sensing, 02(02), 193–204. https://doi.org/10.4236/ars.2013.22022

74
Asgari, M. (2021). A critical review on scale concept in GIS-based watershed management
studies. Spatial Information Research, 29(3), 417–425. https://doi.org/10.1007/s41324-
020-00361-7
Ayele, G. T., Tebeje, A. K., Demissie, S. S., Belete, M. A., Jemberrie, M. A., Teshome, W. M.,
Mengistu, D. T., & Teshale, E. Z. (2018). Time Series Land Cover Mapping and Change
Detection Analysis Using Geographic Information System and Remote Sensing, Northern
Ethiopia. Air, Soil and Water Research, 11, 117862211775160.
https://doi.org/10.1177/1178622117751603
Bera, K., & Bandyopadhyay, J. (2012). Ground water potential mapping in Dulung watershed
using remote sensing & GIS techniques, West Bengal, India. International Journal of
Scientific and Research Publications, 2(12), 1–7.
Bullock, E. L., Healey, S. P., Yang, Z., Oduor, P., Gorelick, N., Omondi, S., Ouko, E., & Cohen,
W. B. (2021). Three Decades of Land Cover Change in East Africa. Land, 10(2), Article
2. https://doi.org/10.3390/land10020150
Changwony, C., Sichangi, A. W., & Murimi Ngigi, M. (2017). Using GIS and Remote Sensing
in Assessment of Water Scarcity in Nakuru County, Kenya. Advances in Remote Sensing,
06(01), 88–102. https://doi.org/10.4236/ars.2017.61007
Chemuku, W. (n.d.). EFFECTS OF FOREST FRAGMENTATION ON FOREST COVER
CHANGE, TREE SPECIES DIVERSITY AND CARBON STOCK IN TAITA HILLS,
KENYA. 179.
Chepchumba, M. C. (n.d.). EXPLORATION OF GROUNDWATER POTENTIAL USING GIS
AND REMOTE SENSING IN EMBU COUNTY, KENYA. 109.
County, T.-T. (2018). Taita-Taveta County Annual Development Plan (CADP) 2019-2020.
Taita-Taveta County.
Das, S. (2017). Delineation of groundwater potential zone in hard rock terrain in Gangajalghati
block, Bankura district, India using remote sensing and GIS techniques. Modeling Earth
Systems and Environment, 3(4), 1589–1599. https://doi.org/10.1007/s40808-017-0396-7
Deribe, H., & Debalke, D. B. (2021). Identification of Ground Water Potential Zones Using GIS
and Remote Sensing Techniques in the Case of Odiyo Watershed, Ethiopia [Preprint]. In
Review. https://doi.org/10.21203/rs.3.rs-239478/v1

75
Erdogan, H. E., Pellikka, P. K. E., & Clark, B. (2011). Modelling the impact of land-cover
change on potential soil loss in the Taita Hills, Kenya, between 1987 and 2003 using
remote-sensing and geospatial data. International Journal of Remote Sensing, 32(21),
5919–5945. https://doi.org/10.1080/01431161.2010.499379
ERTURK, A., GUREL, M., BALOCH, M. A., DIKERLER, T., VAROL, E., AKBULUT, N., &
TANIK, A. (2006). Application of Watershed Modeling System (WMS) for Integrated
Management of a Watershed in Turkey. Journal of Environmental Science and Health,
Part A, 41(9), 2045–2056. https://doi.org/10.1080/10934520600780693
Fadil, A., Rhinane, H., Kaoukaya, A., Kharchaf, Y., & Bachir, O. A. (2011). Hydrologic
Modeling of the Bouregreg Watershed (Morocco) Using GIS and SWAT Model. Journal
of Geographic Information System, 03(04), 279–289.
https://doi.org/10.4236/jgis.2011.34024
Farjana Birajdar, M. S. (2015). Analysis of Watershed Characteristics Using Remote Sensing
and GIS Techniques. International Journal of Innovative Research in Science,
Engineering and Technology, 04(04), 1971–1976.
https://doi.org/10.15680/IJIRSET.2015.0404023
Forestry Management | Satellite Monitoring | Planet | Webinar. (n.d.). Retrieved March 1, 2023,
from https://swiftgeospatial.solutions/2019/11/21/how-satellite-imagery-is-solving-forest-
management-challenges-joint-webinar-with-planet-swift-geospatial/
Gebre, T., Kibru, T., Tesfaye, S., & Taye, G. (2015). Analysis of Watershed Attributes for Water
Resources Management Using GIS: The Case of Chelekot Micro-Watershed, Tigray,
Ethiopia. Journal of Geographic Information System, 07(02), 177–190.
https://doi.org/10.4236/jgis.2015.72015
Guduru, J. U., & Jilo, N. B. (2022). Groundwater potential zone assessment using integrated
analytical hierarchy process-geospatial driven in a GIS environment in Gobele watershed,
Wabe Shebele river basin, Ethiopia. Journal of Hydrology: Regional Studies, 44, 101218.
https://doi.org/10.1016/j.ejrh.2022.101218
Gurjar, G. N., Meena, N. K., Swami, S., Telkar, S. G., & Lyngdoh, E. A. S. (n.d.). Concept of
Watershed Management and its Components. 4.

76
Hohenthal, J. M., Räsänen, M., & Minoia, P. (2018). Political ecology of asymmetric ecological
knowledges: Diverging views on the eucalyptus-water nexus in the Taita Hills, Kenya.
Journal of Political Ecology.
Hohenthal, J., Owidi, E., Minoia, P., & Pellikka, P. (2015a). Local assessment of changes in
water-related ecosystem services and their management: DPASER conceptual model and
its application in Taita Hills, Kenya. International Journal of Biodiversity Science,
Ecosystem Services & Management, 11(3), 225–238.
Hohenthal, J., Owidi, E., Minoia, P., & Pellikka, P. (2015b). Local assessment of changes in
water-related ecosystem services and their management: DPASER conceptual model and
its application in Taita Hills, Kenya. International Journal of Biodiversity Science,
Ecosystem Services & Management, 11(3), 225–238.
Imran, R., Rehman, A., Khan, M., Jamil, M., Abbas, U., Mahmood, R., Mahmood, S. A., &
Ehsan, R. (2019). Delineation of Drainage Network and Estimation of Total Discharge
using Digital Elevation Model (DEM). International Journal of Innovations in Science
and Technology. https://doi.org/10.33411/IJIST/2019010201
Integrated Watershed Management—NVCA. (n.d.). Retrieved April 19, 2023, from
https://www.nvca.on.ca/planning-permits/integrated-watershed-management
Iván, V., Stevenazzi, S., Pollicino, L. C., Masetti, M., & Mádl-Szőnyi, J. (2020). An Enhanced
Approach to the Spatial and Statistical Analysis of Factors Influencing Spring
Distribution on a Transboundary Karst Aquifer. Water, 12(8), 2133.
https://doi.org/10.3390/w12082133
Jackson, R. B., Carpenter, S. R., Dahm, C. N., Mcknight, D. M., Naiman, R. J., Postel, S. L., &
Running, S. W. (2001). WATER IN A CHANGING WORLD. Ecological Applications,
11(4), 19.
Jaeztold, D. R., Schmidt, H., Hornetz, D. B., & Shisanya, D. C. (n.d.). Farm Management
Handbook Of Kenya. 467.
Jasrotia, A. S., Kumar, A., & Singh, R. (2016). Integrated remote sensing and GIS approach for
delineation of groundwater potential zones using aquifer parameters in Devak and Rui
watershed of Jammu and Kashmir, India. Arabian Journal of Geosciences, 9(4), 304.
https://doi.org/10.1007/s12517-016-2326-9

77
Johanna, H., Marinka, R., Emmah, O., Belinda, A., Minoia, P., & Petri, P. (2015). Community
and institutional perspectives on water management and environmental changes in the
Taita Hills, Kenya.
Kaburi, A. N., & Odera, P. A. (2014). GIS and Remote Sensing Support in Watershed
Conservation and Management: Case Study of the Upper Gucha Watershed–Kenya.
IOSR Journal of Environmental Science, Toxicology and Food Technology, 8(7), 35–43.
Kanoti, J. R., Olago, D., Opiyo, N., & Nyamai, C. (2019). An overview of groundwater and
sanitation challenges in Kisumu City, Kenya. International Journal of Innovative
Research and Development, 8(4). https://doi.org/10.24940/ijird/2019/v8/i4/143592-
347932-
Konadu, D. D., & Fosu, C. (2009). Digital Elevation Models and GIS for Watershed Modelling
and Flood Prediction – A Case Study of Accra Ghana. In E. K. Yanful (Ed.), Appropriate
Technologies for Environmental Protection in the Developing World: Selected Papers
from ERTEP 2007, July 17–19 2007, Ghana, Africa (pp. 325–332). Springer Netherlands.
https://doi.org/10.1007/978-1-4020-9139-1_31
Lee, S., Hyun, Y., Lee, S., & Lee, M.-J. (2020). Groundwater potential mapping using remote
sensing and GIS-based machine learning techniques. Remote Sensing, 12(7), 1200.
Li, J., & Roy, D. (2017). A Global Analysis of Sentinel-2A, Sentinel-2B and Landsat-8 Data
Revisit Intervals and Implications for Terrestrial Monitoring. Remote Sensing, 9(9), 902.
https://doi.org/10.3390/rs9090902
Li, Z. (2014). Watershed modeling using arc hydro based on DEMs: A case study in Jackpine
watershed. Environmental Systems Research, 3(1), 11. https://doi.org/10.1186/2193-
2697-3-11
Maeda, E. (2011). Agricultural expansion and climate change in the Taita Hills, Kenya: An
assessment of potential environmental impacts.
https://helda.helsinki.fi/handle/10138/24585
Magazine, N. S. (n.d.). Drying Springs: A Threat To Human Survival. SpotlightNepal. Retrieved
March 1, 2023, from https://www.spotlightnepal.com/2020/12/10/drying-springs-threat-
human-survival/

78
Mahmoud, S. H., & Alazba, A. A. (2015). Hydrological Response to Land Cover Changes and
Human Activities in Arid Regions Using a Geographic Information System and Remote
Sensing. PLOS ONE, 10(4), e0125805. https://doi.org/10.1371/journal.pone.0125805
Mainuri, Z. G., Mironga, J. M., & Mwonga, S. M. (2022). Land Use/Land Cover Changes in a
Disturbed River Watershed Kenya. European Journal of Formal Sciences and
Engineering, 5(2), 1–12. https://doi.org/10.26417/ejef.v3i2.p29-36
Masud, M. J., & Bastiaanssen, W. G. M. (n.d.). Remote Sensing and GIS Applications in Water
Resources Management. 24.
Mirrah, A. A., & Kusratmoko, E. (2017). Application of GIS for Assessment of Water
Availability in the Cianten Watershed, West Java. IOP Conference Series: Earth and
Environmental Science, 98, 012018. https://doi.org/10.1088/1755-1315/98/1/012018
Mkuzi, H. T. (n.d.). ASSESSMENT OF LAND TENURE, LAND USE AND LAND COVER
CHANGES IN TAITA HILLS FOREST FRAGMENTS: A CASE STUDY OF NGERENYI
FOREST FRAGMENTS IN TAITA TAVETA COUNTY, KENYA. 104.
Muthoni, E. (n.d.). Environmental Engineering. Environmental Engineering, 57.
Mwamuye, E. S., Chongomwa, M. M., Rayment, M., & Nadir, S. (2021). Multi-Stakeholder
Approach in Natural Resources Management: The Case of Mida Creek and Gongoni-
Marereni Mangrove Ecosystems in Kilifi County, Kenya. Asian Journal of Environment
& Ecology, 171–180. https://doi.org/10.9734/ajee/2021/v16i430267
Ndalilo, L. A., Maranga, E. K., & Kirui, B. K. (2021). Land Use and Land Cover Change along
River Lumi Riparian Ecosystem in Kenya: Implications on Local Livelihoods. Open
Journal of Forestry, 11(03), 206–221. https://doi.org/10.4236/ojf.2021.113014
Niraula, R. R., Sharma, S., Pokharel, B. K., & Paudel, U. (2021). Spatial prediction of spring
locations in data poor region of Central Himalayas. Hydrology Research, 52(2), 492–505.
https://doi.org/10.2166/nh.2020.223
Pande, C. B., Moharir, K. N., & Khadri, S. F. R. (2021). Assessment of land-use and land-cover
changes in Pangari watershed area (MS), India, based on the remote sensing and GIS
techniques. Applied Water Science, 11(6), 96. https://doi.org/10.1007/s13201-021-01425-
1

79
Patle, D. (2020). Land Use / Land Cover Mapping of Nahra Nala Watershed Using Sentinel-2B
Imagery. International Journal of Agriculture Environment and Biotechnology, 13(4).
https://doi.org/10.30954/0974-1712.04.2020.8
Price, P. E. (n.d.). Image Analysis with ArcGIS 10. 54.
Rajaei, F., Dahmardeh Behrooz, R., Ahmadisharaf, E., Galalizadeh, S., Dudic, B., Spalevic, V.,
& Novicevic, R. (2021). Application of Integrated Watershed Management Measures to
Minimize the Land Use Change Impacts. Water, 13(15), Article 15.
https://doi.org/10.3390/w13152039
Rasuly, A., Naghdifar, R., & Rasoli, Mehdi. (2010). Detecting of Arasbaran Forest Changes
Applying Image Processing Procedures and GIS Techniques. Procedia Environmental
Sciences, 2, 454–464. https://doi.org/10.1016/j.proenv.2010.10.050
Ritchie, H., & Roser, M. (2021). Forests and Deforestation. Our World in Data.
https://ourworldindata.org/deforestation
Ruttoh, R. C., Obiero, J. P. O., Omuto, C. T., & Tanui, L. (2022). Assessment of Land Cover and
Land Use Change Dynamics in Kibwezi Watershed, Kenya. The Scientific World
Journal, 2022, 1–13. https://doi.org/10.1155/2022/3944810
Sarkar, A. (2018). Accuracy assessment and analysis of land use land cover change using
geoinformatics technique in Raniganj coalfield area, India. International Journal of
Environmental Sciences & Natural Resources, 11(1), 25–34.
Sentinel-2—Missions—Sentinel Online—Sentinel Online. (n.d.). Retrieved March 1, 2023, from
https://sentinel.esa.int/web/sentinel/missions/sentinel-2
Sewnet, A. (2016). Land use/cover change at Infraz watershed by using GIS and remote sensing
techniques, northwestern Ethiopia. International Journal of River Basin Management,
14(2), 133–142. https://doi.org/10.1080/15715124.2015.1095199
Singh, P., Gupta, A., & Singh, M. (2014). Hydrological inferences from watershed analysis for
water resource management using remote sensing and GIS techniques. The Egyptian
Journal of Remote Sensing and Space Science, 17(2), 111–121.
https://doi.org/10.1016/j.ejrs.2014.09.003
Sinha, S., Sharma, L. K., & Nathawat, M. S. (2015). Improved Land-use/Land-cover
classification of semi-arid deciduous forest landscape using thermal remote sensing. The

80
 Egyptian Journal of Remote Sensing and Space Science, 18(2), 217–233.
https://doi.org/10.1016/j.ejrs.2015.09.005
Srinivas, P., Sarala, C., & Chowdary, P. P. (n.d.). INTEGRATED WATERSHED
MANAGEMENT USING REMOTE SENSING AND GIS TECHNIQUES. 8.
Swami, M. V. A. (n.d.). Model Watershed Management Plan for Shivapur Village.
TAITA TAVETA County Integrated Development Plan 2018-2022—Devolution Hub. (n.d.).
Retrieved November 28, 2021, from https://www.devolutionhub.or.ke/resource/taita-
taveta-county-integrated-development-plan-2018-2022
Taloor, A. K., Pir, R. A., Adimalla, N., Ali, S., Manhas, D. S., Roy, S., & Singh, A. K. (2020).
Spring water quality and discharge assessment in the Basantar watershed of Jammu
Himalaya using geographic information system (GIS) and water quality Index(WQI).
Groundwater for Sustainable Development, 10, 100364.
https://doi.org/10.1016/j.gsd.2020.100364
Thapa, P. (2021). The Relationship between Land Use and Climate Change: A Case Study of
Nepal. In The Nature, Causes, Effects and Mitigation of Climate Change on the
Environment. IntechOpen. https://doi.org/10.5772/intechopen.98282
Time-Series Study. (2010). In N. Salkind, Encyclopedia of Research Design. SAGE
Publications, Inc. https://doi.org/10.4135/9781412961288.n465
Uddin, K., Gilani, H., Murthy, M. S. R., Kotru, R., & Qamer, F. M. (2015). Forest Condition
Monitoring Using Very-High-Resolution Satellite Imagery in a Remote Mountain
Watershed in Nepal. Mountain Research and Development, 35(3), 264–277.
https://doi.org/10.1659/MRD-JOURNAL-D-14-00074.1
Umrikar, B. (2015). GIS Technique in Management of Watershed Developed along the Konkan
Coast, Maharashtra, India. Journal of Geographic Information System, 07(03), 280–293.
https://doi.org/10.4236/jgis.2015.73022
United Nations University, & Institute for Water, E. and H. (UNU-I. (2012). The global water
crisis: Addressing an urgent security issue. Institute for Water, Environment and Health.
Upadhyay, P., & Gupta, S. (2012). Introduction To Satellite Imaging Technology And Creating
Images Using Raw Data Obtained From Landsat Satellite. 1, 41–45.
Vora, K. (2011). APPLICATION OF REMOTE SENSING & GIS IN WATERSHED LANDUSE
DEVELOPMENT. 3.

81
Waiyasusri, K., & Chotpantarat, S. (2020). Watershed Prioritization of Kaeng Lawa Sub-
Watershed, Khon Kaen Province Using the Morphometric and Land-Use Analysis: A
Case Study of Heavy Flooding Caused by Tropical Storm Podul. Water, 12(6), 1570.
https://doi.org/10.3390/w12061570
Wang, K., Franklin, S., Guo, X., He, Y., & Mcdermid, G. (2009). Problems in Remote Sensing
of Landscapes and Habitats. Progress in Physical Geography - PROG PHYS GEOG, 33,
747–768. https://doi.org/10.1177/0309133309350121
Wekesa, C., Kirui, B. K., Maranga, E. K., & Muturi, G. M. (2020). The Fate of Taita Hills Forest
Fragments: Evaluation of Forest Cover Change between 1973 and 2016 Using Landsat
Imagery. Open Journal of Forestry, 10(01), 22–38.
https://doi.org/10.4236/ojf.2020.101003
Winkler, K., Fuchs, R., Rounsevell, M., & Herold, M. (2021). Global land use changes are four
times greater than previously estimated. Nature Communications, 12(1), Article 1.
https://doi.org/10.1038/s41467-021-22702-2
Yuksel, A., Gundogan, R., & Akay, A. (2008). Using the Remote Sensing and GIS Technology
for Erosion Risk Mapping of Kartalkaya Dam Watershed in Kahramanmaras, Turkey.
Sensors, 8(8), 4851–4865. https://doi.org/10.3390/s8084851

APPENDICES

This study took approximately 10 months as shown below.

Appendix 1: Work plan

82
 se O No De Ja Fe M A Ma Jun
p ct v c n b ar pr y e

- -
2 - - - 2 - - -
Activities 2 22 22 22 3 23 23 23 -23 -23

Thesis Writing

1st Draft

and
submission

Changes on
comments

2nd Draft
submission

Final Writing

Thesis defense

Final
preparation

Thesis
Submission

83
Appendix 2: Budget

Item Item description Price per Amount

unit
Personnel
Principal investigator The PI will be charged to conduct 561,340.00
objective research that generates
independent, high quality, and
reproducible results
Field Assistant He will provide reports in data collection 103,630.00
for the project
Administrative cost Institutional transcription 9,000.00
Consultant charges This will be done and targeted for the 25,000.00
final analysis of key findings
Transportation
Local public transport This involves the use of public means of 4,800.00
transport Like matatu and Boda to the
study site
Field stationery
Photocopy, Books, Entails the materials needed like printing 58,000.00
Whiteboards and other stationery like pens, books
among others
Field Equipment
Laptop & internet This is to ease the research by providing 60,000.00
a working platform and internet to access
online materials and software’s
GPS 3,000.00

Additional project- This will involve other activities that 15,000.00

84
 related Activities may arise up and be miscellaneous.
Grant total 839,770/=

Appendix 3: Equipment used

2. GPS Device To Map and record springs locations

3. Notebook and pen To collect information about springs’ parameters

(surrounding) physical outlook.

85