Professional Documents
Culture Documents
Draft Report For The Landcover Change Study and GIS Based Land Suitability Analysis 2
Draft Report For The Landcover Change Study and GIS Based Land Suitability Analysis 2
Table of Contents
Acknowledgment...................................................................................................................................9
Executive Summary..............................................................................................................................10
1.0 Introduction..............................................................................................................................14
1.1 Background............................................................................................................................15
2.1 Landcover and Land Use (LCLU) change mapping for the past 10 years...............26
3.1 Landcover and Land Use (LCLU) Change Mapping of West Nile District..................43
4.0 Results for GIS Based Land Suitability Analysis in West Nile Region.............................................110
4.1 Layers used in the GIS Based Crop Land Suitability Analysis................................110
5.0 Conclusions...................................................................................................................................144
6.0 Recommendations........................................................................................................................146
7.0 References....................................................................................................................................147
List of Figures
List of Tables
Acknowledgment
I would like to thank GIZ Country Office for giving me this opportunity to undertake
Landcover and Land Use (LCLU) change mapping for the past 10 years and produce
the Land Suitability Analyses for West Nile Region. Special thanks go to Ms. Heinze,
Claudia, Head of Component I Response to increased demand on Government
Services and creation of economic opportunities in Uganda (RISE) for her guidance
during this assignment and for putting in place the technical team of GIZ and District
Local Government officials who helped us carry out field validation mission in West
Nile District and making our stay in West Nile Region very comfortable.
I also want to thank Mr. Thomas Ujjiga, the Technical Advisor Response to Increased
Demand on Government Service and Creation of Economic Opportunities in Uganda
(RISE) for the guidance and support during the District stakeholders’ consultation
meetings in all the 7 District. I would like to appreciate the support rendered to us by
Ms. Tumuhairwe Diana the communication specialist of RISE project and Ms. Alina
Zalewski, Development Advisor Arua/ Madi-Okollo during the district consultations.
Thanks to the drivers of GIZ Arua Field Office for wonderful and comfortable drive in
the field.
Finally, I would like to take this opportunity to thank my team Mr Alex Tumukunde, Mr
Kelonye Wiseman, and Mr. Ampeire Nobert for the great work done. Your efforts
during the field validations, soil sample collections, and district consultation meetings
are appreciated.
11
Executive Summary
The bjectives of the assignment were to undertake Landcover and Land Use (LCLU)
change mapping for the past 10 years (i.e. 2010; 2015; 2020) in the District of
Adjumani, Arua, Madi-Okollo, Moyo, Obongi, Terego, Yumbe and Arua City; and
produce the Land Suitability Analyses (LSA) to derive potential suitability for different
Land Utilization Types (LUTs). Landcover and land use (LCLU) mapping was
carriedout using a method that combines both expert knowledge and supervised
classification.
The examination of the landcover and landuse change in different ecosystems for three
epochs 2010, 2015 and 2020 from the District of Adjumani, Arua, Madi-Okollo, Moyo,
Obongi, Terego, Yumbe and Arua City was done, while paying attention to particular
phenomena, such as climate change, soil moisture content change, soil fertility and
groundwater change. The analyse was done on landcover and landuse change in
different geographical locations, such as woodlots, forests, annual cropland, wetlands,
bushland, settlements and grassland. Cconducted GIS-based Land Suitability Analyses
using remote sensing software for image processing to derive landcover and landuse
maps.
Adjumani District
From 2010 to 2015, Adjumani District lost 140,959ha of forests and 51,173ha of
bushland. The district experienced a lot of deforestation between 2010-2015 in the
central part, degradation across the district and little regeneration was observed.
From 2010 to 2020, Adjumani District lost 129, 952 hectares of forests and 53,454
hectares of grasslands. The district experienced a lot of deforestation between 2010-
2020 in the central and southern parts, degradation across the district and little
regeneration was observed in south and north. From 2015 to 2020, Adjumani District
lost 103,819 hectares of woodlands and 51,256 hectares of wetlands. The district
experienced a lot of degradation between 2015-2020 across the district. From the
suitability modeling, the central part of adjumani district is highly suitable for Maize,
Soya and Simsim growing while the north and south are moderately suitable.
Arua District
From 2010 to 2015, Arua District lost 3,764 hectares of Bushland and 71 hectares of
wetlands. The district experienced regeneration with little degradation in north east
12
part of the district. From 2010 to 2020, Arua District lost 11,994 hectares of Bushland
and 77 hectares of wetlands. The district experienced regeneration in central and of
lot of degradation in East and North East part. From 2015 to 2020, Arua District lost
8,230 hectares of Bushland and 17,184 hectares of woodlands. The district
experienced degradation and regeneration in all sub-counties. From the suitability
modeling analysis, the Southern and north east parts of Arua district are highly
suitable for beans growing while the North West and south west are moderately
suitable. The North east parts of Arua district are highly suitable for Cassava growing
while the rest of the district is moderately suitable. The whole of Arua district is
moderately suitable for maize growing with few atreas of the district highly suitable.
Arua city
From 2010 to 2015, Arua City lost 4,736 hectares of Forests and 132 hectares of
woodlands. The district experienced deforestation in all sub-counties, degradation in
the central part and regeneration in the East. From 2010 to 2020, Arua City lost
2,935 hectares of Forests and 961ha of Bushlands and 1,130ha of Woodland. The
district experienced deforestation in all sub-counties, degradation in the central and
East. From 2015 to 2020, Arua City lost 1,599 hectares of Bushland and 998
hectares of woodlands. The district experienced degradation in the Eastern parts.
From the suitability modeling analysis, the east parts of Arua City are highly suitable
for beans growing while the west and some parts of central are moderately suitable.
The urbanization is expanding in the central part of the City. The whole of Arua City
is moderately suitable for beans’ growing while the west parts are moderately
suitable and central not suitable. The whole district is moderately suitable for sweat
potatoes growing.
Madi-okollo District
From 2010 to 2015, Madi-Okollo lost 13,633 hectares of Forests and 20,271 hectares
of Wetlands. The district experienced deforestation, degradation and regeneration in
all sub-counties. From 2010 to 2020, Madi-Okollo lost 39,899 hectares of Bushland,
4,011ha of Forests and 14,768 hectares of Wetlands. The district experienced
deforestation and degradation across all sub-counties, regeneration in the central
sub-counties. From 2015 to 2020, Madi-Okollo lost 63,282 hectares of Bushland and
13
Obongi District
From 2010 to 2015, Obongi District lost 14,916 hectares of Forest land, 2,472
hectares of Wetland and 73 ha of Woodland. The district experienced deforestation
and degradation, degradation and regeneration in all sub-counties. The district
experienced deforestation, degradation and regeneration in all sub-counties. From
2010 to 2020, Obongi District lost 8,512 hectares of Forest land, 845 ha of Bushland,
7,435 hectares of Grasslands, 8881ha of Wetlands and 73ha of Woodlands. The
district experienced degradation and regeneration in all sub-counties. From 2015 to
2020, Obongi District lost 7,181ha of Bushland, 6409ha of Wetlands and 10,940ha of
Grasslands. From the suitability modeling analysis, 50% of Obongi district is highly
suitable for cassava growing while the 30% are moderately suitable and 20% not
suitable. 50% of Obongi district is highly suitable for maize growing, 30% are
moderately suitable and 20% not suitable. 70% of Obongi district is highly suitable for
simsim growing while the 5% are moderately suitable and 25%.
Yumbe District
From 2010 to 2015, Yumbe District lost 11,559 hectares of Forest land, 3,010
hectares of Wetland. The district experienced deforestation and degradation in East
and West sub-counties and regeneration in all sub-counties. From 2010 to 2020,
Yumbe District lost 89,629 hectares of Bushland, 7926ha of Forest, 3,214 hectares of
Wetland and 1,072ha of Wetland. The district experienced deforestation and
degradation in East and West sub-counties and regeneration in the West. From 2015
to 2020, Yumbe District lost 98,976 hectares of Bushland, 204 hectares of Wetland
and 2,736ha of Woodland but gained Forest cover of 3,633ha. The district
experienced degradation in all sub-counties and regeneration in west sub-counties.
14
From the suitability modeling analysis, 60% of Yumbe district is highly suitable for
beans growing while the 40% are moderately suitable. 60% of Yumbe district is
highly suitable for cassava growing while the 35% are moderately suitable and 5%
not suitable. 85% of Yumbe district is highly suitable for groundnuts growing while the
10% are moderately suitable and 5% not suitable.
Moyo District
From 2010 to 2015, Moyo District lost 20,652 hectares of Forest land and 2,841
hectares of Wetland. The district experience a lot of deforestation in central and
north, degradation in West and East of the district. There was lot of regeneration in
all sub-counties. From 2010 to 2020, Moyo District lost 12,775 hectares of Forest
land and 2,427 hectares of Wetland. The district experienced deforestation in central
and north east parts, degradation in southern parts and regeneration in all sub-
counties. From 2015 to 2020, Moyo District lost 15,276 hectares of Bushland, 11,139
hectares of Grassland, and experienced degradation and regeneration in all sub-
counties. From the suitability modeling analysis, 85% of Moyo district is highly
suitable for cassava growing while the 10% are moderately suitable and 5% not
suitable. 75% of Moyo district is highly suitable for groundnuts and simsim growing
while the 20% are moderately suitable and 5% not suitable.
Terego District
From 2010 to 2015, Terego District lost 8,677 hectares of Forestland, 2,697 hectares
of Wetland, experienced degradation in the East part of the district, and regeneration
was also observed in all sub-counties. From 2015 to 2020, Terego District lost 5,808
hectares of Forestland, 23,506ha of Bushland, 2,841 hectares of Wetland and
experienced a lot of degradation although regeneration was observed in the sub-
county of Odupi in the East. From 2010 to 2020, Terego District lost 39,419 hectares
of Bushland, 261 hectares of Wetland, and experienced a lot of degradation in the
East and scanty deforestation. From the suitability modeling analysis, 55% of Terego
district is highly suitable for cassava growing while the 40% are moderately suitable
and 5% not suitable. 90% of Terego district is highly suitable for simsim growing
while the 5% are moderately suitable and 5% not suitable. 60% of Terego district is
highly suitable for Sorghum growing while the 40% are moderately suitable.
15
1.0 Introduction
1.1 Background
Human activities, especially the conversion and degradation of habitats, are causing
global biodiversity declines [1]. The notable activities include the cutting down of trees,
charcoal burning and poor farming methods, among others, which undermine the
functionality of ecosystems [2]. The rapid conversion of natural vegetation, for
example, to farmlands, could be attributed to farming techniques and agronomic
approaches that aim at modern agricultural intensification [3]. For instance, in the
Equateur province of the Democratic Republic of Congo, the agricultural expansion
through shifting cultivation is the main proximate cause of deforestation [4]. In the
northern portion of the Brazilian Atlantic Forest, generally 76% of the households use
fuelwood regularly and consume on average 686 kg/person/year of tree biomass;
poorer people, however, consume 961 kg/person/year [5,6].
consist of objects represented by polygons that are assigned class labels indicating
the objects’ land use [17]. Therefore, land use/cover information can be directly
interpreted from appropriate remote sensing images [18].
In response to the global refugee crisis, the United Nations (UN) has adopted the
New York Declaration for Refugees and Migrants (2016) on an international level. The
NYC Declaration has become the basis for the Comprehensive Refugee Response
Framework (CRRF), with the overall objective of the CRRF to support governmental
and administrative bodies affected by refugee crises at national, regional and local
level and thus to strengthen a ‘whole society’ approach to refugee response. In
Uganda, the reception of refugees and the development-oriented design of refugee
policy are regulated in the Refugee Act of 2006 and the corresponding regulations of
2010 (Refugee Regulations). The main implementation strategies of the legal
17
framework are the so-called Self-Reliance Strategy (SRS) of 1998 and the Settlement
Transformation Agenda (STA) of 2016. The STA has been integrated in the National
Development Plan (NDP III) which is regarded as a good practice.
In addition, UNHCR and the World Bank have collaborated with the Ugandan
government to develop the Refugee and Host Population Empowerment Strategy
(ReHoPE) which aims to implement the STA. The strategy includes an urgent
appeal from the UHNCR to international development partners to participate in the
implementation of ReHoPE.
The RISE program has two specific objectives: 1) Strengthen local authorities’
coordination and development planning, as well as local authority-led service delivery
to refugees and the host populations and; 2) Increase economic self-reliance of
refugees and host populations. The RISE program targets 50% host communities and
50% refugees as beneficiaries.
The German Government has commissioned the Water Supply and Sanitation Program
18
for Refugees and Host Communities in Northern Uganda through the Special Initiative
of the German Federal Ministry for Economic Cooperation and Development (BMZ):
Tackling the root causes of displacement, reintegrating refugees. The programme is
jointly implemented with the Ministry of Water and Environment (MWE) during a period
of three years.
Within the refugee settlements in Northern Uganda, universal access to basic drinking
water supply cannot yet be ensured. Despite the pledges of aid organizations for
new wells and boreholes, a significant part of the drinking water in refugee
settlements is still provided by water tankers. Hygiene conditions are poor and only
some of the households have their toilets and hand washing facilities. Many people
use communal latrines and open defecation is common. Because of the refugee
influx, the demand for drinking water and sanitation has also increased in host
communities. Operation and maintenance of existing hand pumps and newly installed
solar- powered wells with stand posts is currently still inadequate. Therefore, measures
to sustain water supply and sanitation services in line with national and international
refugee strategies cannot be achieved yet.
Many studies on LULCC change have been conducted in Uganda at different spatial
and temporal scales for different reasons. For example, Anaba et al. (2017) focused
on the impacts of LULCC on water quality in the Murchison Bay of Lake Victoria.
Wadembere and Kobugabe (2017) investigated urbanization and its effects in western
Uganda. In northern Uganda Gorsevski et al. (2012) assessed the effect of armed
19
conflict on forest ecosystem while Nyeko (2010) looked at land use change and water
resources management in Aswa basin of northern Uganda. In eastern Uganda, Frank
Mugagga et al. (2015) were interested in land use change, carbon stocks, and climate
variability in Mt. Elgon region while F Mugagga et al. (2012) investigated LULCC and
implications to landslides occurrences in Mt. Elgon, eastern Uganda. On the other
hand Li et al. (2016), Zhou et al. (2017), Jagger and Kittner (2017), and Mwanjalolo et
al. (2018) LULCC on a national scale. What emerges from the literature is that LULCC
is complex and driven by both natural and anthropogenic factors that interact in space
and time.
meanwhile increased from 84,695 km2 in 1990 to 95,211 km2 in 2000 then to
100,000 km2 in 2005–2010, and finally to 105,317 km2 in 2015; Forest decreased in
area from 49,334 km2 in 1990 to 36,655 km2 in 2005, and finally to 19,535 km2 in
2015; Wetland change however, has been case specific with local studies reporting a
mixed of increasing and decreasing trend as later discussed. However, on a national
scale, wetland area appears generally stable.
According to the Ministry of Water and Environment (MW&E), 2015 Uganda lost on
average
122,000 ha/year of forest between 1990 and 2015, in favor of agricultural area
expansion. Similarly the Ministry of Water and Environment (MW&E) (2016), indicated
that national area of wetlands declined by 30% between 1994 and 2008, and slightly
increased by 0.03% between 2008 and 2014.
The numbers of urban centers increased from 67 in 1991 to 259 in 2014, while urban
population increased from 0.6 million people in 1969 to 6.4 million people in 2014
(UBOS, 2014). The implication is that the interaction of population growth and other
drivers of LULC change is favoring agriculture and built-up/urban areas expansion at
the expense of forest, woodlands, wetlands, as later discussed in Section.
On local scales, Barasa et al. (2011) reported increase in agriculture and decrease in
forest between 1975 and 1999 in western Uganda. Similar trend was also reported in
the same region by Majaliwa et al. (2010) between 1973 and 2009. In eastern
Uganda, Frank Mugagga et al. (2015) reported increased agriculture at the expense
of forest between 1960 and 2006. Increase in agriculture and urban areas expansion
was also reported between 1987 and 2005 by Mbazira (2014), and between 2007–
2013 by Akello et al. (2016) in eastern Uganda. In central Uganda, a similar trend
was reported between 1995 and 2003 by Anaba et al. (2017), while in northern
Uganda Nyeko (2010) reported a decrease in agriculture over the period 1986–2001.
The reported trends also confirm what is observed at national level, except for the
Acholi subregion (northern Uganda).
Note that the trends of forest cover and agriculture in northern Uganda are different
from the national trend and those in other regions. First, only 9.5% of the reviewed
21
LULCC studies were from northern Uganda. Fewer studies meant less representation
of the actual situation on the ground. Northern Uganda also experienced armed
conflict from 1986 to mid-2000s. During the period of the armed conflict, people were
restricted to live in Internally Displaced Person’s camp (IDP’s). The people had little or
no access to their land, and could therefore not exploit the resources therein. With
return of peace in northern Uganda, IDP’s life ended in the late 2008. There are now
good roads linking all parts of the region to the grater markets within Uganda, and with
South Sudan and DR. Congo. There has also been rapid expansion of commercial
farming (by the so-called ‘investors’) in the northern District of Nwoya and Amuru in
the last ten years. The dynamics and the trajectory of LULCC in northern Uganda
have thus changed. It is not surprising that the study of Gorsevski et al. (2012) found
high rate of deforestation in northern Uganda coinciding with the time after life at the
IDPs camps.
Kobugabe (2017) and Hartter and Southworth (2009) in western, and Gorsevski et al.
(2012) in northern Uganda all emphasized population as the main driver of LULCC.
However, Nakalembe et al. (2017) found that it was lack of policy implementation but
not population driving LULCC in northern Uganda. On the other hand, Nyeko (2010)
reported insecurity as the major driver of LULCC in northern Uganda and not
population or policy implementation. In eastern Uganda, Place and Otsuka (2000)
discussed unemployment to be the key driver of LULCC, whereas according to Frank
Mugagga et al. (2015), it was population and lack of cultivatable land driving LULCC
in eastern region. On a national scale, however, increased demand for forest
products, lack of policy implementation, and agricultural area expansion have been
reported as major drivers of LULCC in addition to population Li et al. (2016), Zhang et
al. (2017), and Jagger and Kittner (2017).
(Source: NFA).
years (i.e. 2010; 2015; 2020) in the District of Adjumani, Arua, Madi-Okollo,
Moyo, Obongi, Terego, Yumbe and Arua City;
ii. Produce the Land Suitability Analyses (LSA) to derive potential suitability for
different Land Utilization Types (LUTs).
The assignment involved the examination of the landcover and landuse change in
different ecosystems for three epochs 2010, 2015 and 2020 from the District of
Adjumani, Arua, Madi-Okollo, Moyo, Obongi, Terego, Yumbe and Arua City, while
paying attention to particular phenomena, such as climate change, soil moisture content
change, soil fertility and groundwater change; analyse landcover and use change in
different geographical locations, such as woodlots, tree plantations, farmlands, grazing
lands, wetlands, hills, slopes and mountains; study the landcover and use change as
resulting from different practices (e.g. bush fires or burnings, forest harvesting);
scrutinise the landcover and use changes resulting from settlements (e.g. locals,
refugees); examine the landcover and use changes resulting from infrastructure
development (e.g. roads, facilities, schools, towns, markets, health centres); conduct
GIS-based Land Suitability Analyses using the ArcGIS software. The consultant used
remote sensing software Erdas Imagine for image processing to derive landcover and
landuse maps.
The consultant used the major determinants of land use such as demographic
indicators (population size and density; technology; level of affluence; political
structures); economic factors (systems of exchange or ownership; slope; altitude, and
attitudes and values). The consultant modelled inputs that included physical, and
economic factors and considered the interaction between different actors (e.g. public
sector, academia, research institutes and think tanks; environmental NGOs and CSOs,
host and refugee communities) and their influences on the environmental context.
25
The study final report provides a set of useful possible recommendations that will help
the DLGs to make informed policy decisions to counteract environmental degradation
throughout the selected refugee-hosting District. Furthermore, the consultant produced
data that will contribute to fill certain data gaps identified by the National Environment
Management Authority (NEMA) as limiting the capacity of local administrators. The
overall analysis, comprising both the landcover/use change study and the LSAs, will
help both programmes, the DLGs and other programme stakeholders to understand
how the factors causing landcover/use change could interface with the planned/desired
District development.
The study produced Landcover and Land Use (LCLU) change mapping for the past 10
years in three epochs of 2010, 2015 and 2020 in the District of Adjumani, Arua, Madi-
Okollo, Moyo, Obongi, Terego, Yumbe and Arua City; and produced Land Suitability
Analyses (LSA) to derive potential suitability for different Land Utilization Types
(LUTs).
2.1 Landcover and Land Use (LCLU) change mapping for the past 10 years
Collection of ancillary data is important as a baseline for the project and was
collected through national teams as reference material. This activity involved
gathering information from previous reports, publications and studies within the
scope of the project, historical reference locations, existing land use land cover
maps, high spatial resolution imagery and other satellite imagery data. The sole
purpose of this activity was to have sufficient spatial and related attribute information
on land use land cover mapping and from relevant key stakeholders involved in the
project.
On identifying and collecting existing ancillary data, stakeholder involvement was key
in assessing and verification of the consistency and accuracy of the data collected.
This also required reworking of the existing methodology and documenting of
metadata from existing land use land cover maps. The specific tasks undertaken
during this activity therefore included gathering existing and or historical land use
maps and previously collected ground reference data, iidentifying through
consultative forums the land use land cover classification scheme to be used within
the country, re-working and documenting the metadata of the existing land use land
27
cover and related vegetation products and eensuring that enough relevant ancillary
data is made available for classification of satellite imagery to the required
classes/categories.
Quality checking and assessment of such datasets involved checking for consistency
and completeness based on elements such as data Completeness that assesses the
presence or absence of data at the various levels of coverage to determine if at these
levels, the datasets are sufficient to adequately represent various spatial
phenomena, lineage which assesses the historical significance of the dataset in
terms of the original source that it was derived from and the various transformational
changes it has undergone to its present format, positional accuracy which assesses
the nominal accuracy of the dataset based on their true locations on the ground to
determine the tolerance of a dataset, attribute accuracy that provided deviation of the
descriptive information associated with a spatial dataset and their conformity to the
true representation and topological consistency that addressed the fidelity of data
structure relationships thus linking the datasets to the model structures used. The
specific tasks under this activity included receiving ancillary data, performing initial
quality checks and assessment and requesting for clarification from all relevant
stakeholders, verification, re-working and validation of the ancillary data, subjecting
the data to independent quality assessments and identified data gaps with relation to
classification of the required land cover categories.
Quality checks are important for remote sensing imagery as it forms a critical
component for achieving high quality results for the outputs intended. These quality
checks were based on the geometric and radiometric aspect of the images using
both visual means and statistical methods to achieve stipulated Root Means Square
Error (RMSE) of 0.5m or better from orthorectification.
For image pre-processing, consideration was made to the various image parameters
such as dates of acquisition, cloud cover and the type of sensor involved. This
consideration is important in ensuring consistency in producing image mosaics and
therefore consistency in the representation of land use and land cover characteristics
of a particular area. For understanding of the various land covers in West Nile
District, considerations were made to perform an initial unsupervised land image
classification on the extracted image mosaics to get an initial picture of the
representation of the various land use land cover characteristics.
The specific tasks within this activity therefore included carrying out image statistics
to determine the suitability of maximum likelihood classification or Support Vector
machine classifiers/ algorithms, acquisition of Landsat satellite imagery datasets for
West Nile District for the epochs 2010, 2015 and 2020, image pre-processing where
the satellite imagery data were orthorectified to match with existing data using ground
control points and DEMs ,and thereafter based on the quality of the images,
calibration, mosaicking and clipping was done to prepare them for classification,
geometric and radiometric image corrections were applied to the 3 epochs including
atmospheric corrections (TOA), initial parameters were established to ensure that the
imagery collected and pre-processed met the needs of the project. Such parameters
included image acquisition dates, amount of cloud cover, and the type of sensors for
each of the imagery scenes selected, carried out image enhancement to improve on
visual interpretation, carried out preliminary unsupervised image classification on the
29
3 epochs to provide for initial understanding of the various land use land cover types
in West Nile District. The following imagery 2010, 2015 and 2020 were purchased
and processed;
Figure 3: West Nile Satellite Imagery for 2010 used for classification
Figure 4: West Nile Satellite Imagery for 2015 used for classification
30
Figure 5: West Nile Satellite Imagery for 2020 used for classification
This therefore called for a detailed study of the existing classification schemes to
guide in choice of an appropriate structure to represent the land cover characteristics
of West Nile District with guidance from policy documents, the objectives of the
mapping exercise, specific stakeholder interests, area definition and set standards
through consultative forums.
Reference was also made to the IPCC guidelines used in developing globally used
standards that also meet country specific classification scheme standards. IPCC land
use land cover categories for schema 1 level of classification contains the following
classes: Forest land (Dense Forest, Moderate Forest, Sparse Forest, Planted
31
Figure 6: Flow chart showing the steps taken for final products
The initial stage of image processing involved extraction of the individual bands
followed by layer staking where the various bands of the multispectral Landsat image
were composited into a single multi-band image. This ensures that the various bands
of a multispectral image are utilized in determining the land use land cover
characteristics of a given area. Considerations were made to the thermal band and
the panchromatic band to ensure consistency in the spatial resolutions of the images
since they differ in spatial resolution from the rest of the bands.
Statistical computation using Eigen values, Eigen vectors and mean spectral values
are useful in assessing the spectral quality of the imagery. Generating such statistical
information from satellite imagery was useful in determining variability and correlation
33
between bands of the multispectral image that is significant in indicating the ability to
discriminate various classes during classification.
Image mosaics were created from the composites as products of layer stacking to
represent wall-to-wall coverage of West Nile District. Subsets representing west Nile
District extracted using the AOI file depicting the country boundary were produced for
the 3 epochs of the dataset.
Supervised classification (object oriented) was done in Arc GIS to produce output
files in vector format. On screen digitization was employed to conduct the
classification. Visual interpretation was essential in determining the classes. This
process involved extraction of information from the satellite images as polygons
which were then coded by assigning them to the correct class.
Thorough knowledge of the different land cover classes is important. The most
current year which is 2020 was the first to be classified. Quality checks were
performed on the first draft 2020 classification using high resolution imagery that was
accessed through the Google Earth platform before the ground truthing exercise. The
first draft 2020 was verified using field validation points. Necessary corrections were
done and the final draft produced followed by the generation of the accuracy report.
The 2020 Land Cover served as the base file for classifying the years 2010 and
2015. The 2020 file was saved as the 2010 and 2015 to avoid confusion or
replacement of the file. The 2010 and 2015 images were then overlaid. The changes
are easily noticed. Editing of the 2020 file based on the 2010 and 2015 images
produced the Land Use Land Cover files for the years 2010, 2015 and 2020.
Change analysis was conducted to generate change maps that show the exact areas
where changes occurred and the transition among classes. The land use and land
cover statistics were generated in ArcGIS and tables created in Excel. The
development of land cover land use maps were done at the scale of 1:50,000 for the
District maps.
Reference information and training data for classifying imagery for earlier dates i.e.
1990-2010 was developed from review and study of existing land use land cover and
vegetation data, reports and publications and through use of Google Earth platform
for the high resolution times-series imagery. Change detection was also used to
distinguish irregular changes and for identification of erroneous classes. Ground
reference data is collected from the field on randomly generated points at selected
zones for image classification accuracy assessment. A number of criteria were
considered when evaluating the suitability of any ground reference data set for land
cover classification such as: Sufficiency of reference samples to achieve required
confidence levels; a random method has to be considered and should be systematic
and representative of the area of study; and the reference data must be of around the
same time as the satellite image.
A plan for selecting locations for collecting ground control points for data verification
using stratified random sampling involved identifying a sample of each land cover
class proportionate to the population size of the class when viewed against the entire
population. Then the number of points per land use category was identified and used
to generate the number of such points within a sampling frame. The stratified random
sampling technique was then be applied to randomly distribute points across the
sampling frame and across each land cover category in relation to their area
coverage in the sampling zone.
Accuracy assessment to analyze and modify the result of the classification was
conducted for the 2020 land cover classification with a recorded accuracy of 85%.
The accuracy assessment of the land use land cover maps was produced using a
confusion matrix to compare the reference points generated from the land cover
classification to the sample points collected from ground reference locations. A report
of accuracy analysis with the methods used and source of reference points being
indicated.
Ground referencing activity was done in all the 7 District including Arua District.
Ground reference data (a total of 1,400 points) were collected from the field on
randomly generated points at selected zones using field verification form (Annex I).
35
The probability sampling design was the preferred approach. It combines random or
stratified sampling to get points to validate the land cover predefined in the first draft
classification and perform accuracy assessment of the same. In view of this, the data
collection method was systematic, and representative of the entire area that had
matrices. The Kappa statistic incorporates the off diagonal elements of the
error matrices and represents agreement obtained after removing the proportion of
agreement that could be expected to occur by chance.
1- Chance agreement
Table 2: Compliance matrix between interpreter and validation based results
FIELD DATA
Bar Croplan Cropl Gra W
2015_LULC Closed Closed Fore Open Open Riverine Urba Wat
e d and ss et TO PA
CLASSIFICA Bushla Woodla st Bushl Woodl Vegetati n er
Are Plantati Small lan lan TAL %
TION nd nd land and and on Area body
a on Scale d d
66.6
Bare Area 2 0 0 0 0 0 1 0 0 0 0 0 0
3 7
Closed 75.0
0 9 1 0 0 0 0 1 1 0 0 0 0
Bushland 12 0
Closed 83.3
0 0 5 0 0 0 0 0 1 0 0 0 0
Woodland 6 3
Cropland 94.1
0 0 0 16 1 0 0 0 0 0 0 0 0
Plantation 17 2
Cropland 100.
0 0 0 0 10 0 0 0 0 0 0 0 0
Small Scale 10 00
75.0
Forestland 0 0 1 0 0 9 1 1 0 0 0 0 0
12 0
94.7
Grassland 0 0 0 0 1 0 18 0 0 0 0 0 0
19 4
Open 87.5
0 0 0 0 0 0 0 7 1 0 0 0 0
Bushland 8 0
Open 70.0
0 1 1 0 0 0 0 1 7 0 0 0 0
Woodland 10 0
Riverine 100.
0 0 0 0 0 0 0 0 0 1 0 0 0
Vegetation 1 00
80.0
Urban Area 0 0 0 0 1 0 0 0 0 0 4 0 0
5 0
100.
Waterbody 0 0 0 0 0 0 0 0 0 0 0 1 0
1 00
100.
Wetland 0 0 0 0 0 0 0 0 0 0 0 0 1
1 00
TOTAL 2 10 8 16 13 9 20 10 10 1 4 1 1
10
CA%
100 90 62.50 100 76 100 90 70 70 100 100 100 0
Kappa=105(90) – ((3*2)+(12*10)+(6*8)+(17*16)+(10*13)+(12*9)+(19*20)+(8*10)+(10*10)+(1*1)+(5*4)+(1*1)+(1*1))
(105)2 – ((3*2)+(12*10)+(6*8)+(17*16)+(10*13)+(12*9)+(19*20)+(8*10)+(10*10)+(1*1)+(5*4)+(1*1)+(1*1))
The overall accuracy of classification imagery dated 2020 was 85.71% and the
Kappa coefficient was 83.86%. for the whole of west Nile Region. The individual
district accurancy assessment are:
iii. Arua city: The overall accuracy of classification Arua City imagery dated
2020 was 80.14% and the Kappa coefficient was 76.67%.
iv. Madi-okollo District: The overall accuracy of classification Madi-Okollo
imagery dated 2020 was 90.70% and the Kappa coefficient was 86.92%.
v. Obongi District: The overall accuracy of classification obongi district
imagery dated 2020 was 80.01% and the Kappa coefficient was 76.67%.
vi. Yumbe District: The overall accuracy of classification Yumbe imagery dated
2020 was 80.56% and the Kappa coefficient was 77.20%.
vii. Moyo District: The overall accuracy of classification Moyo imagery dated
2020 was 85.03% and the Kappa coefficient was 81.48%.
viii. Terego District: The overall accuracy of classification Terego imagery dated
2020 was 84.08% and the Kappa coefficient was 80.57%.
Preparation of the final land cover maps for the 3 epochs resulting from the land use
land cover classification were presented in user appropriate formats and to the
required cartographic standards in both hard and soft copy formats. A proper legend
was developed.
Degraded areas were easily identified with changes such as transition from
wetland/woodland/bushland to grassland/bare area/agricultural land. Degradation
occurs when land lose the capacity to provide important goods and services to
nature. Deforestation was considered as any change from forestland to any other
landcover class.
Changes detection results were represented in form of statistics tables and graphs.
Maps were used to represent areas of potential deforestation, degradation and
regeneration from various spatial analysis techniques that would reveal vulnerable
areas within the ecosystem of West Nile District and the potential drivers of such
changes. The maps were prepared in a uniform coordinate system (projected
coordinate system WGS 84 36S).
The consultant worked together with the technical experts provided by both
programme team and DLG to define the objective of the suitability assessment per
39
district. The objective of the suitability assessment differed per District, as each district
has different strategic objectives in terms of local economic as well as infrastructure
development. The decision criteria suitable for each of the identified objectives in each
of the district were selected. The consultant related the land attributes to the suitability
for a particular objective. For final criteria selection, the consultant was guided by the
reviewed literature, stakeholder consultation meetings with relevant actors and
acquired the most District.
For each of the decision criterion, a suitability map was generated by transforming the
GIS map layer containing the relevant land elements into suitability scores. The
criterion and sub- criterion suitability maps were then weighted and combined to
generate final suitability maps of each one of the District of Adjumani, Arua, Madi-
Okollo, Moyo, Obongi and Yumbe.
The following steps were followed to develop GIS-based Land Suitability Analysis;
i. Identified factors having a significant effect on land suitability for suitable
developments or interventions and based on principles of sustainable
development and the evaluation of criteria suitability.
ii. Applied analytic hierarchy process (AHP) method for the assessment of the
relative importance (weights) of the criteria and sub-criteria.
iii. Applied simple additive weighting (SAW) method for estimation of the land
suitability index (LSI).
40
Only those quantitative factors that are directly and indirectly related to the
considered land suitability for development activity were selected and during the
factor identification process. ArcGIS and Erdas Imagine software were used for data
conversion and analysis. The final district suitability maps produced were grouped
into highly suitable, moderately suitable, and not suitable areas for developments or
interventions.
The stakeholders and project expert teams agreed on factors that to be used and
42
Decisions were made in an organized way to generate priorities for comparisons and
scale of numbers were needed to indicates how many times more important or
dominant one element is over another element with respect to the criterion or
property they were compared. The Analytic Hierarchy Process (AHP) is a theory of
measurement through pair wise comparison. The comparisons were made using a
scale of absolute judgements that represents how much more; one element
dominates another with respect to a given attribute. The derived priority scales were
synthesized by multiplying them by the priority of their parent nodes and adding for
all such nodes. In this study, for the determination of suitability analysis, preference
was given to physical, ecological, climatic and socio-economic parameters by using
correlation analysis module. Once the database of all GIS layers was created, the
consultant will extract the GIS attribute data and writing to ASCII files.
2.2.4 GIS-based Land Suitability maps development
The multi-criteria evaluation was carried out through the following tasks;
(i) converted the data to the BEST CHOICE format;
(ii) performed a multi-criteria evaluation of the alternatives; and
(iii) performed a sensitivity analysis to examine the robustness of the results
of evaluation ranking,
Integrated analysis for land suitability was done where weightage was given physical
parameters, by comparing each parameter with respect to the other parameter. The
weightage were assigned for all the parameters, according Importance matrix for
suitability analysis scale. Finally, all the land suitability map layers were combined in
a GIS environment to prepare the final land suitability model based on weighted
index model. The land use suitability map was grouped into three categories namely
highly suitable, moderately suitable, and not suitable.
43
The suitability results were reviewed in ArcGIS environment through extract, sorting
and writing the ASCII files ranking results. The ASCII files were imported to the existing
Dbase file and joined the ranking in the Dbase file to ArcGIS software coverage
depicting the alternatives and viewing the results.
3.1 Landcover and Land Use (LCLU) Change Mapping of West Nile District
Landcover and Landcover Change mapping was carried out using satellite data from
the years 2010, 2015 and 2020 for District of Adjumani, Arua, Madi-Okollo, Moyo,
Obongi, Terego, Yumbe and Arua City.
In 2010, Adjumani district had 47% forest landcover followed by Bushland and
Annual crops (Table 3).
In 2015, Adjumani district the highest land cover was annual cropland (40%) followed
by grassland (25.7%) and Bushland (20%) (Table 4).
In 2020, Adjumani district highest land cover was annual cropland (60.9%) followed
by Bushland (20.1%) and woodland (6.6) (Table 5).
From 2010 to 2015, Adjumani District lost 140,959 hectares of forests followed by
51,173 of bushland (Table 6). The district experienced a lot of deforestation between
2010-2015 in the central part, degradation across the district and very little
regeneration was observed (Figure 15).
From 2010 to 2020, Adjumani District lost 129, 952 hectares of forests and 53,454
hectares of grasslands. Other landcover classes gained (Table 7). The district
experienced a lot of deforestation between 2010-2020 in the central and southern
parts, degradation across the district and little regeneration was observed in south
and north (Figure 16).
From 2015 to 2020, Adjumani District lost 103,819 hectares of woodlands and
51,256 hectares of wetlands. Other landcover classes gained (Table 8). The district
experienced a lot of degradation between 2015-2020 across the district (Figure 17).
In 2010, Arua district highest land cover was Annual cropland (74.9%) followed by
Bushland (17.1%) and Forest (7%) (Table 9).
In 2015, Arua district the highest land cover was annual cropland (44.9%) followed by
woodland (24.5%) and Bushland (12.5%) (Table 10).
In 2020, Arua district the highest land cover was annual cropland (74.9%) followed by
Forests (16.3%), and Grassland (3.2%) and Woodlands (3.2) (Table 11).
From 2010 to 2015, Arua District lost 3,764 hectares of Bushland and 71 hectares of
wetlands. There was reduction in annual cropland (Table 12). The district
experienced regeneration with little degradation in north east part (Figure 24).
From 2010 to 2020, Arua District lost 11,994 hectares of Bushland and 77 hectares
of wetlands. There was reduction in settlements (Table 13). The district experienced
regeneration in central and of lot of degradation in East and North East part (Figure
25).
From 2015 to 2020, Arua District lost 8,230 hectares of Bushland and 17,184
hectares of woodlands. There was also a reduction in settlements and wetlands
(Table 14). The district experienced degradation and regeneration across the district
(Figure 26).
In 2010, Arua City the highest land cover was annual cropland (75.8%) followed by
Forests (15.3%), and Woodland (3.3%) and Bushlands (2.9%) (Table 15).
In 2015, Arua city the highest land cover was annual cropland (81.3%) followed by
settlements (5.4%), and Bushland (4.7%) and Woodlands (2.9%) (Table 16).
In 2020, Arua City the highest land cover was annual cropland (83.3%) followed by
settlements (7.3%), and Forest (6.9%) and Woodlands (2.3%) (Table 17).
From 2010 to 2015, Arua City lost 4,736 hectares of Forests and 132 hectares of
woodlands (Table 18). The district experienced deforestation in all sub-counties,
degradation in the central part and regeneration in the East (Figure 33).
From 2010 to 2020, Arua City lost 2,935 hectares of Forests and 961ha of Bushlands
and 1,130ha of Woodland (Table 19). The district experienced deforestation in all
sub-counties, degradation in the central and east parts (Figure 34)
From 2015 to 2020, Arua City lost 1,599 hectares of Bushland and 998 hectares of
woodlands. There was also a reduction in Grassland (Table 20). The district
experienced degradation in the Eastern parts (Figure 35)
In 2010, Madi-Okollo highest land cover was annual cropland (46.5%) followed by
Bushland (25.3%), and Wetlands (17.1%) (Table 21).
In 2015, Madi-Okollo highest land cover was annual cropland (36.8%) followed by
Bushland (37%), and Grasslands (14.5%) (Table 22).
In 2020, Madi-Okollo highest land cover was annual cropland (74.6%) followed by
Wetland (9.7%), and Forests (6.9%) (Table 23).
From 2010 to 2015, Madi-Okollo lost 13,633 hectares of Forests and 20,271 hectares
of Wetlands (Table 24). The district experienced deforestation, degradation and
regeneration in all sub-counties (Figure 42).
From 2015 to 2020, Madi-Okollo lost 63,282 hectares of Bushland and 27,825
hectares of Grasslands (Table 26). The district experienced degradation in all sub-
counties and regeneration in the central and north parts of the district (Figure 44)
In 2010, Obongi District highest land cover was Bushland (25.2%) followed by Forest
(21.6%), and Annual cropland (19.6%) (Table 27).
In 2015, Obongi District highest land cover was Bushland (32.7%) followed by
Annual Cropland (29.1%), and Grassland (16.4%) (Table 28).
In 2020, Obongi District highest land cover was annual cropland (42%) followed by
Bushland (24.2%), and Forests (11.5%) (Table 29).
From 2010 to 2015, Obongi District lost 14,916 hectares of Forest land, 2,472
hectares of Wetland and 73 ha of Woodland (Table 30). The district experienced
deforestation and degradation, degradation and regeneration in all sub-counties,
(Figure 51).
In 2010, Yumbe District highest land cover was annual cropland (47.3%) followed by
Bushland (43.1%), and Forests (7.5%) (Table 33).
In 2015, Yumbe District highest land cover was Bushland (47.1%) followed by Annual
Cropland (37.3%), and Grassland (10.2%) (Table 34).
In 2020, Yumbe District highest land cover was Annual Cropland (88.9%) followed by
Bushland (4.6%), and Forest (4.1%) (Table 35).
From 2010 to 2015, Yumbe District lost 11,559 hectares of Forest land, 3,010
hectares of Wetland (Table 36). The district experienced deforestation and
degradation in East and West sub-counties and regeneration in all sub-counties,
(Figure 60).
From 2010 to 2020, Yumbe District lost 89,629 hectares of Bushland, 7926ha of
Forest, 3,214 hectares of Wetland and 1,072ha of Wetland (Table 37). The district
experienced deforestation and degradation in East and West sub-counties and
regeneration in the West (Figure 61).
From 2015 to 2020, Yumbe District lost 98,976 hectares of Bushland, 204 hectares
of Wetland and 2,736ha of Woodland but gained Forest cover of 3,633 ha (Table 38).
The district experienced degradation in all sub-counties and regeneration in west
sub-counties (Figure 62).
In 2010, Moyo District highest land cover was annual cropland (41.7%) followed by
Forests (23.4%) and Bushland (13.2%) (Table 39).
In 2015, Moyo District highest land cover was Bushland (42.2%) followed by Annual
Cropland (34.7%) and Grassland (12.5%) (Table 40).
In 2020, Moyo District highest landcover was Annual Cropland (52.3%) followed by
Bushland (27.5%) and Forest land (11.1%) (Table 41).
From 2010 to 2015, Moyo District lost 20,652 hectares of Forest land and 2,841
hectares of Wetland (Table 42). The district experience a lot of deforestation in
central and north, degradation in West and East of the district. There was lot of
regeneration in all sub-counties (Figure 69).
From 2010 to 2020, Moyo District lost 12,775 hectares of Forest land and 2,427
hectares of Wetland (Table 43). The district experienced deforestation in central and
north east parts, degradation in southern parts and regeneration in all sub-counties.
(Figure 70).
From 2015 to 2020, Moyo District lost 15,276 hectares of Bushland and 11,139
hectares of Grassland (Table 44). The district experienced degradation and
regeneration in all sub-counties. (Figure 71).
In 2010, Terego District highest land cover was annual cropland (66.3%) followed by
Bushland (22.2%) and Forest (8.6%) (Table 45).
In 2015, Terego District highest landcover was annual cropland (52.6%) followed by
Bushland (36%) and Grassland (6.7%) (Table 46).
In 2020, Terego District highest land cover was annual cropland (93%) followed by
Forests (3.5%) and Bushland (1.9%) (Table 47).
From 2010 to 2015, Terego District lost 8,677 hectares of Forestland and 2,697
hectares of Wetland (Table 48). The district experienced scanty deforestation in the
district and degradation in the East part of the district. Regeneration was also
observed in all sub-counties (Figure 78).
From 2015 to 2020, Terego District lost 5,808 hectares of Forestland, 23,506ha of
Bushland and 2,841 hectares of Wetland although there was gain in Woodland
(Table 49). The district experienced a lot of degradation. Regeneration was observed
in the sub-county of Odupi in the East (Figure 79).
From 2010 to 2020, Terego District lost 39,419 hectares of Bushland and 261
hectares of Wetland. There were gain in Woodland and Forestland (Table 50). The
district experienced a lot of degradation in the East and scanty deforestation. (Figure
80).
4.0 Results for GIS Based Land Suitability Analysis in West Nile Region
4.1 Layers used in the GIS Based Crop Land Suitability Analysis
The West Nile District GIS-based Land Suitability Analysis was conducted, whereby
land suitability was determined through a systematic and multi-criteria analysis.
Climatic and ecological factors were used in the model analysis to see how they could
interface with the planned/desired development through assembled ranking of the land
elements within the landscape. The ranks were used to prepare the different layers
used in the multi-criteria analysis used to determine suitability of site for a specific
development or intervention.
For each of the decision criterion, a suitability map was generated by transforming the
GIS map layer containing the relevant land elements into suitability scores. The
criterion and sub- criterion suitability maps were then weighted and combined to
generate final suitability maps of each one of the District of Adjumani, Arua, Madi-
Okollo, Moyo, Obongi and Yumbe.
During the district stakeholders’ consultation, each district identified three priority
crops for production under the Enterprise (Table 51). The crop suitability analysis and
modeling was done on only these priority crops.
Figure 81 above shows the organic matter content of the study area. The soil organic
matter consists of plant and animal detritus at various stages of decomposition, cells
and tissues of soil microbes, and substances that soil microbes synthesize. The
percentage organic matter layer was used in the suitability model by grouping of the
organic matter content which best supports high yields for specific crop. Different
crops require different quantities of soil organic matter content for growth.
113
Figure 82 shows the altitude of the study area. Altitude refers to the height of a
particular location in relation to sea level or ground level. Altitude Suitability is the
categorization of different heights above sea level in which specific crops are best
adapted to have the highest production. Different crops require different altitude for
growth.
114
Figure 83 shows the rainfall of the study area for 2020. Rainfall is the amount of rain
that falls in a place during a particular period usually measured by the depth in
inches. Rainfall Suitability is grouping the amounts of rainfall received in a region by
whether these amounts will be sufficient for supporting crop growth for the entire
growing period or whether it falls into either excess or deficit hence reduces the
production in a way. Different crops require different amount of rainfall for growth.
115
Figure 84 shows the temperature of the study area for 2020. Temperature is a
physical quantity that expresses hot and cold in a location. Temperature suitability
refers to the categorization of optimum temperature conditions for specific crop
production whereby increase or decrease from these thresholds results in the
reduction in growth of shoots, roots or a complete stop of growth and death of plant.
Different crops require different temperature for growth.
116
Figure 85 shows the soil depth of the study area. Soil depth defines the root space
and the volume of soil from where the plants fulfil their water and nutrient demands.
Soil depth suitability refers to grouped soil in terms of how deep or shallow they are
and the effect this has in comparison to the depth in which the specific crop produces
the most based on how water, nutrients and mechanical support are accessible to
the plant. Different crops require different soil depth for growth.
117
Figure 86 shows soil drainage of the study area. Soil drainage is a natural process by
which water moves across, through, and out of the soil as a result of the force of
gravity. Soil drainage suitability refers to grouping the effect of the ease of flow or
retention of water on a specific crop by either increase or decrease in crop health and
productivity. Different crops require different soil drainage for growth.
118
Figure 87 shows the soil ph of the study area. Soil pH is a measure of the acidity and
alkalinity in soils. Soil pH suitability refers to determining and grouping different pH
levels according to the effect they have on improving nutrient uptake by specific
crops thereby affecting the production. Different crops require different soil drainage
for growth.
119
Figure 88 shows the soil texture of the study area. Soil texture refers to the
proportion of sand, silt and clay sized particles that make up the mineral fraction of
the soil. Soil texture suitability is the grouping of different soil characteristics such as
water-holding capacity, permeability, and soil workability which affect plant growth.
These differences act by the promotion of specific plant growth through the actual
condition the crop requires in order to get maximum support and water through which
nutrients are availed to the crop. Different crops require different soil texture for
growth.
120
Figure 89 shows the slope of the study area. Slope is the rise or fall of the land
surface. Percentage slope suitability refers to the categorization of the slope based
on the optimum requirements for the specific crop. Crop stability lessens or increases
based on the adjacent steepness of the land. Different crops require different slope
for growth.
121
The maize require 50-100cm of rainfall, temperature of 21°C- 27°C, good effective
depth, favourable morphological properties, good internal drainage, an optimal
moisture regime, sufficient and balanced quantities of plant nutrients and chemical
properties, soils with a pH from 5 - 8 but 5.5 - 7 is optimal, altitude range of 800 -
3000 m above sea level, large amounts of nitrogen (N), phosphorus (P), and
potassium (K) as well as a smaller amount of sulfur (S) and some micronutrients.
122
From the suitability modeling, the central part of adjumani district is highly suitable for
maize growing while the north and south are moderately suitable (Figure 90).
Sim sim grows well in rain-fed area, requires high heat and light requirements and is
sensitive to low temperatures. Growth and fruiting are favored by temperature of
around 37 degrees centigrade, moderate rains, high soil temperatures for germination.
From the suitability modeling, the central part of adjumani district is highly suitable for
simsim growing while the north and south are moderately suitable (Figure 91).
123
The maize crops require 50-100cm of rainfall, temperature of 21°C- 27°C, good
effective depth, favourable morphological properties, good internal drainage, an
optimal moisture regime, sufficient and balanced quantities of plant nutrients and
chemical properties, soils with a pH from 5 - 8 but 5.5 - 7 is optimal, altitude range of
800 - 3000 m above sea level, large amounts of nitrogen (N), phosphorus (P), and
potassium (K) as well as a smaller amount of sulfur (S) and some micronutrients.
From the suitability modeling, the whole of Arua district is moderately suitable for
maize growing with parts of the district are highly suitable (Figure 95).
127
Sweet potatoes grow best where the air temperature remains very warm, from 75° to
95°F (24-35°C) throughout the growing season, preferably 55 to 65 degrees) with
high humidity, annual rainfall between 850 and 1500 mm, high altitude ranging from
1000 to 2000m and requires sandy loam soils. The whole district is moderately
suitable for sweat potatoes growing. (Figure 98).
130
Millet crop thrives well in areas of warm, dry climate with annual rainfall of 45cm,
Temperature ranging between 25 to 30°C. Millet grows well in sandy loamy soils and
shallow black soil having good drainage and humus. Requires ph range from 6-7.5.
From the suitability modeling analysis, 85% of Madi-Okollo district is highly suitable
for millet growing while the 5% are moderately suitable and 10% not suitable (Figure
100).
132
Sim sim grows well in rain-fed area, requires high heat and light requirements and is
sensitive to low temperatures. Growth and fruiting are favored by temperature of
around 37 degrees centigrade, moderate rains, high soil temperatures for germination.
From the suitability modeling analysis, 35% of Madi-Okollo district is highly suitable
for millet growing while the 45% are moderately suitable and 20% not suitable
(Figure 101).
133
Groundnuts require temperature between 27-30˚C for good germination and growth,
minimum annual rainfall of between 450 to 1250 mm and warm climate, fine textured
soils, high yields are obtained on slightly acid soils with pH 6.0 to 6.4, well drained,
light coloured, loose, friable, sandy loam soil, well supplied with calcium and a
moderate amount of organic matter. A well-drained soil facilitates adequate
exchange of air to meet nitrogen, carbon dioxide and oxygen requirements of the
groundnut crop. From the suitability modeling analysis, 75% of Moyo district is highly
suitable for groundnuts growing while the 20% are moderately suitable and 5% not
suitable (Figure 103).
135
Simsim grows well in rain-fed area, requires high heat and light requirements and is
sensitive to low temperatures. Growth and fruiting are favored by temperature of
around 37 degrees centigrade, moderate rains, high soil temperatures for germination.
From the suitability modeling analysis, 75% of Moyo district is highly suitable for
simsim growing while the 20% are moderately suitable and 5% not suitable (Figure
104).
136
The maize crops require 50-100cm of rainfall, temperature of 21°C- 27°C, good
effective depth, favourable morphological properties, good internal drainage, an
optimal moisture regime, sufficient and balanced quantities of plant nutrients and
chemical properties, soils with a pH from 5 - 8 but 5.5 - 7 is optimal, altitude range of
800 - 3000 m above sea level, large amounts of nitrogen (N), phosphorus (P), and
potassium (K) as well as a smaller amount of sulfur (S) and some micronutrients.
From the suitability modeling analysis, 50% of Obongi district is highly suitable for
maize growing, 30% are moderately suitable and 20% not suitable (Figure 106).
138
Simsim grows well in rain-fed area, requires high heat and light requirements and is
sensitive to low temperatures. Growth and fruiting are favored by temperature of
around 37 degrees centigrade, moderate rains, high soil temperatures for germination.
From the suitability modeling analysis, 70% of Obongi district is highly suitable for
simsim growing while the 5% are moderately suitable and 25% not suitable (Figure
107).
139
Simsim grows well in rain-fed area, requires high heat and light requirements and is
sensitive to low temperatures. Growth and fruiting are favored by temperature of
around 37 degrees centigrade, moderate rains, high soil temperatures for germination.
From the suitability modeling analysis, 90% of Terego district is highly suitable for
simsim growing while the 5% are moderately suitable and 5% not suitable (Figure
109).
141
Groundnuts require temperature between 27-30˚C for good germination and growth,
minimum annual rainfall of between 450 to 1250 mm and warm climate, fine textured
soils, high yields are obtained on slightly acid soils with pH 6.0 to 6.4, well drained,
light coloured, loose, friable, sandy loam soil, well supplied with calcium and a
moderate amount of organic matter. A well-drained soil facilitates adequate
exchange of air to meet nitrogen, carbon dioxide and oxygen requirements of the
groundnut crop. From the suitability modeling analysis, 85% of Yumbe district is
highly suitable for groundnuts growing while the 10% are moderately suitable and 5%
not suitable (Figure 112).
145
5.0 Conclusions
Adjumani District
Arua District
Arua City
Madi-Okollo District
Moyo District
The district experienced deforestation and wetland loss between 2010 to 2015, and
2010 to 2020. Between 2015 to 2020, the district experienced loss of Bushland and
Grassland. 85% of Moyo district is highly suitable for cassava growing while the 10%
are moderately suitable and 5% are not suitable. 75% of Moyo district is highly
suitable for groundnuts and Simsim growing while 20% are moderately suitable and
5% not suitable.
146
Obongi District
The study shows a lot of change in 2010-2015 and 2010-2020 and less change in
2015-2020. Most of the changes occurred in the north and southwest of the district. A
lot of wetlands were degraded into agriculture in 2010-2015. The district experience
regeneration of landcover in 2015-2020. From 2015-2020 there was an expansion of
settlement in the north of the District. The whole Eastern part of the district is not
suitable for cassava, Maize and Simsim as it contains wetlands and restricted areas.
The central area of the district in Gimara Subcounty is a restricted protected area that
has also been degraded.
Terego District
Yumbe District
The study shows a lot change in 2010-2015 and 2010-2020 and more degradation of
environment from 2015-2020. A lot of deforestation and degradation occurred in the
west of the district in 2010-2020 and 2010-2020. The district experienced
regeneration of landcover in 2010-2015 compared to 2010-2020 and 2010-2020. In
2020 there was an expansion of settlement in southeast of the District. The north
area of the district contains the restricted protected area Mt Kei that has been
degraded in 2010-2020 and 2015-2020. The whole district is suitable for cassava and
beans but highly suitable for Groundnuts.
147
6.0 Recommendations
i. There is an urgent need to sustain the productivity of agricultural lands
through matching crop requirements with the resource available through land
suitability analysis.
ii. The District Local Government Planners and land users should use the
produced land-use and landcover change and suitability maps to guide in
identifying alternative land uses.
iv. There is a need to ensure restoration of the lost landcover and protection of
the remaining landcover.
viii. For sustainability, there is a need to build the capacity of DLGs to generate
future landcover maps and carry out land suitability analysis of the remaining
crops using GIS and Remote Sensing applications.
148
7.0 References
1. Newbold, T.; Hudson, L.N.; Hill, S.L.; Contu, S.; Lysenko, I.; Senior, R.A.; Börger,
L.; Bennett, D.J.; Choimes, A.; Collen, B.; et al. Global effects of land use on local
terrestrial biodiversity. Nature 2015, 520, 45. [CrossRef]
[PubMed]
2. Popp, A.; Calvin, K.; Fujimori, S.; Havlik, P.; Humpenöder, F.; Stehfest, E.;
Stehfest, E.; Bodirsky, B.L.; Dietrich, J.P.; Doelmann, J.C.; et al. Land-use futures in
the shared socio-economic pathways. Glob. Environ. Chang. 2017, 42, 331–345.
[CrossRef]
3. Sambou, S.; Lykke, A.M.; Sambou, H.; Guiro, I.; Sambou, B.; Mbow, C. Land use-
land cover change and drivers of deforestation in the Patako protected area (Center-
West of Senegal). Am. J. Environ. Prot. 2015, 4, 306–317. [CrossRef]
4. Samndong, R.A.; Bush, G.; Vatn, A.; Chapman, M. Institutional analysis of causes
of deforestation in REDD+ pilot sites in the Equateur province: Implication for REDD+
in the Democratic Republic of Congo. Land Use Policy 2018, 76, 664–674.
[CrossRef]
6. Specht, M.J.; Pinto, S.R.R.; Albuquerque, U.P.; Tabarelli, M.; Melo, F.P. Burning
biodiversity: Fuelwood harvesting causes forest degradation in human-dominated
tropical landscapes. Glob. Ecol. Conserv. 2015, 3, 200–209. [CrossRef]
7. Smith, P.; Clark, H.; Dong, H.; Elsiddig, E.A.; Haberl, H.; Harper, R.; House, J.;
Jafari, M.; Masera, O.; Mbow, C.; et al. Agriculture, forestry and other land use
(AFOLU). In Climate Change 2014: Mitigation of Climate Change. IPCC Working
Group III Contribution to AR5; Cambridge University Press: Cambridge, UK, 2014.
11. Alkama, R.; Cescatti, A. Biophysical climate impacts of recent changes in global
forest cover. Science 2016, 351, 600–604. [CrossRef] [PubMed]
149
12. Kweka, E.J.; Kimaro, E.E.; Munga, S. Effect of deforestation and land use
changes on mosquito productivity and development in western Kenya highlands:
Implication for malaria risk. Front. Public Health 2016, 4, 238. [CrossRef] [PubMed]
13. Were, K.O.; Singh, B.R.; Dick, Ø.B. Effects of land cover changes on soil organic
carbon and total nitrogen stocks in the Eastern Mau Forest Reserve, Kenya. In
Sustainable Intensification to Advance Food Security and Enhance Climate
Resilience in Africa; Springer: Cham, Switzerland, 2015; pp. 113–133.
14. Natkhin, M.; Dietrich, O.; Schäfer, M.P.; Lischeid, G. The effects of climate and
changing land use on the discharge regime of a small catchment in Tanzania. Reg.
Environ. Chang. 2015, 15, 1269–1280. [CrossRef]
15. Rawat, J.S.; Kumar, M. Monitoring land use/cover change using remote sensing
and GIS techniques: A case study of Hawalbagh block, district Almora, Uttarakhand,
India. Egypt. J. Remote Sens. Space Sci. 2015, 18, 77–84. [CrossRef]
16. Zhang, F.; Zhan, J.; Zhang, Q.; Yao, L.; Liu, W. Impacts of land use/cover change
on terrestrial carbon stocks in Uganda. Phys. Chem. Earth Parts A/B/C 2017, 101,
195–203. [CrossRef]
17. Yang, C.; Rottensteiner, F.; Heipke, C. Classification of land cover and land use
based on convolutional neural networks. ISPRS Ann. Photogramm. Remote Sens.
Spat. Inf. Sci. 2018, 4, 251–258. [CrossRef]
18. Lillesand, T.; Kiefer, R.W.; Chipman, J. Remote Sensing and Image
Interpretation; John Wiley & Sons: Hoboken, NJ, USA, 2014.
19. Masette, M.; Isabirye-Basuta, G.; Baranga, D.; Chemurot, M. Levels of tannins in
fruit diet of grey-cheeked mangabeys (Lophocebus ugandae, Groves) in Lake
Victoria Basin forest reserves. J. Ecol. Nat. Environ. 2015, 7, 146–157.
20. Shaban, K.S.; Okumu, J.O.; Paul, O.; Joseph, O. Assessing community-based
organizations’ influence on trees and grass planting for forest, soil and water
management around Mt. Elgon National Park in Uganda. For. Trees Livelihoods
2016, 25, 161–172. [CrossRef]
21. Arinaitwe, K.; Rose, N.L.; Muir, D.C.; Kiremire, B.T.; Balirwa, J.S.; Teixeira, C.
Historical deposition of persistent organic pollutants in Lake Victoria and two alpine
equatorial lakes from East Africa: Insights into atmospheric deposition from
sedimentation profiles. Chemosphere 2016, 144, 1815–1822. [CrossRef] [PubMed]
24. Musau, J.; Sang, J.; Gathenya, J.; Luedeling, E. Hydrological responses to
climate change in Mt. Elgon watersheds. J. Hydrol. Reg. Stud. 2015, 3, 233–246.
[CrossRef]
26. Nakalembe, C.; Dempewolf, J.; Justice, C. Agricultural land use change in
Karamoja region, Uganda. Land Use Policy 2017, 62, 2–12. [CrossRef]
27. Li, J.; Oyana, T.J.; Mukwaya, P.I. An examination of historical and future land use
changes in Uganda using change detection methods and agent-based modelling. Afr.
Geogr. Rev. 2016, 35, 247–271. [CrossRef]
28. Solberg, S.; May, J.; Bogren, W.; Breidenbach, J.; Torp, T.; Gizachew, B.
Interferometric SAR DEMs for Forest Change in Uganda 2000–2012. Remote Sens.
2018, 10, 228. [CrossRef]